Radio Boulevard
Western Historic Radio Museum

Rebuilding the R-390/R-390A
 Family of Receivers
Includes R-390, R391, R-389,
R-390A, R-392, R-648, R-725

Receiver Types - Assessing your Receiver - Disassembly
The Main Frame - The RF Module - The IF Module
The AF Module - The Power Supply Module - The PTO
Front Panel Restoration - Alignment - Performance
Miscellaneous Info on Variants and Accessories,
Restoration logs for Eleven Receivers:
1. R-390A built from Spare Parts, 2. 1967 EAC R-390A, 3. 1967 EAC R-390A
4. Collins R-648/ARR-41, 5. Collins R-389/URR Detailed Restoration Log
6. Collins R-390/URR, 7. Collins R-391/URR Refurbishing with Detailed Photos,
Four Famous Receiver Recreations - 8. Arvin R-725/URR, 9. ASA R-390A,
10. Clark AB Black Anodized Panel R-390A, 11. USMC OD Panel R-390A

by: Henry Rogers WA7YBS Radio Boulevard

BACK TO PART 1 - Receiver Types, Assessing your Receiver, Main Frame, RF Module, IF Module, Audio Module, Power Supply, PTO, Front Panel Restoration, Meters

BACK TO PART 2 - Other Details, Contractors List, Receiver Alignment, Expected Performance, R-390A Diversity Operation, Building a R-390A "Frankensteinesque" using only Spare "Bone-Yard" Parts, (2) 1967 EAC restorations

BACK TO PART 3 - Recreations of Four Famous R-390A Variants: Arvin Industries R-725, Army Security Agency/NSA R-390A, Clark AB/NSA Black Anodized Panel R-390A, USMC Olive Drab panel R-390A

PART 4 - R-389 Restoration - WARNING! This write-up contains "OCD-detailed" information and it's way too long, R-391 Autotune refurbishment with lots of detailed photos, R-390 SN:3113 Repairs, Dynamotor R-648/ARR-41, R-392/GRC-19 Info, Other R-390 Variants, Security Dial Cover, CV-979 and CV-979A Cabinets

 
1963 Western Electric R-392/URR

The R-392 was used in the GRC-19, a portable 100 watt transmitter-receiver combo. The circuit is essentially a scaled-down R-390 with +28vdc the only voltage required.

PART 4

Restorations of Other Receivers in the R-390 Family, Variants, Accessories

R-389/URR - VLF, LF & MW Receiver

I really thought that meticulous attention to detail was a "good thing." The R-389 is a mechanically complex receiver and everything seemed to be going along just fine until I discovered the VFO oven had been turned on and had probably been on for decades,...well,...every time the receiver was powered-up anyway. I thought turning off the VFO oven and thereby reducing the heat would be a "good thing." I found out that was only the beginning of problems that have been on-going for several years now. Well,...read on,...

Basic Description - Electronics - Built along some of the same lines as the famous R-390 receiver, the Collins R-389 is essentially the LF companion receiver of the R-390. The receiver tunes from 15kc to 500kc in one tuning range and 500kc to 1500kc in the second tuning range. The R-389 uses very complex methods, both electronic and mechanical, to achieve its complete MW, LF and VLF coverage while still utilizing a 455kc IF. The receiver uses 36 tubes within five modules that interconnect and are mounted within the main frame. The 15kc to 500kc tuning range utilizes five permeability-tuned RF bands. The 500kc to 1500kc tuning range utilizes two permeability-tuned RF bands. The motor-driven band switching occurs seamlessly as the receiver is tuned from the lowest to the highest frequency within the two tuning ranges.

Two RF amplifiers are used and the first conversion mixes the incoming RF signal frequency with the VFO (470kc to 1955kc output f) plus the 10.455mc Crystal Oscillator (8.5mc to 9.985mc resulting f) to achieve a 10mc IF. The second conversion mixes the 10mc IF with the same 10.455mc Crystal Oscillator to achieve the 455kc IF. This double conversion scheme was to allow complete coverage from 15kc to 1500kc with no gaps in the frequency coverage. Additionally, since the two mixer stages are 180 degrees out of phase, any drift within the conversion mixers is cancelled leaving only the VFO drift. This is similar to how the "drift-cancelling" Wadley Loop operates. >>>

R-389/URR SN: 268 installed in a CY-979A mobile table cabinet

Mechanical Details - The manual tuning of the receiver RF front end uses clutch-coupled gears to rotate the main RF tuning shaft that has worm gears that perpendicularly engage and rotate the gear-driven front and rear line shafts that have worm gears that in turn engage gear-driven vertical screw-shafts (cut with forward and reverse threads) that raise and lower the various slug racks. This gear-driven system seems like it would be fairly heavy to manipulate but with proper (lightly oiled) lubrication it operates with no more effort than a good condition, clean, well-adjusted R-390A gear box. Since the frequency ranges span a great deal of the spectrum (and this requires a lot of knob-turning) a clutch-coupled, motor-drive tuning system is provided. A separate motor-drive system is employed to operate the bandswitch. There are specific frequencies, that as the receiver is tuned past that frequency, the motorized bandswitch operates and automatically changes to the next higher or lower band as required. 

The Veeder-Root counter is somewhat different than that used in the R-390 and provides two sets of digits, one for 15kc to 500kc (lower set) and the other for 500kc to 1500kc (upper set.) The resolution of the digits (tuned f) is to the tenth of a kilocycle (which are the red background digit wheels.) Neither a calibration oscillator or an antenna trimmer are provided (or needed.)  >>>

>>>  Most of the controls are the same as those found on the R-390. The BFO controls, the Noise Limiter, the Local Gain, Line Gain, Line meter range switch, RF Gain, AGC switch, Break-in switch, Audio Response switch and Function switch. The controls that are unique to the R-389 are Motor Drive, IF Bandwidth (five ranges instead of six,) RF bandwidth KC indicator and the single tuning knob. The two meters perform the same functions as the R-390 meters, that is, Carrier Level and Line Level.

More Details - Physically, the R-389 is the same dimensions as the R-390 and will fit into the CY-917 or CY-979 table cabinets. If installed into a table cabinet, the top and bottom covers should be removed. The receiver weighs 82 pounds but, for easier moving (e.g., up or down stairs,) the power supply and AF module can easily be removed and then the receiver weighs around 65 pounds.

Two antenna connectors are available. Balanced input for 125 ohms input impedance from dipoles or other balanced antennae. Balanced is connected to the primary winding of each antenna coil. Unbalanced input is for random length wire antennae. This input is capacitively-coupled through a .01uf capacitor to the RF amplifier coils. The Unbalanced input impedance is not specified but is probably fairly high assuming that end-fed wires were probably the design target Z. The Balanced input utilizes a "Twin-ax" two-pin coaxial connector and the Unbalanced input utilizes a "C-type" coaxial connector. As mentioned, no antenna trimmer is provided so the antenna impedance should be somewhat matched to the particular antenna input used.

Both audio outputs, Local Audio and Line Audio, are 600 Z ohm outputs and can provide about 500mW on Local and about 10mW on Line. The phone jack doesn't disconnect the audio output (LOCAL) from its respective load. There is a series resistor and a load resistor to the PHONES jack to keep the audio level (5mW) from over-driving the headset if the proper 600 Z phones are used.  

The AC power connector is a four-pin military connector that is keyed and held in place with a central screw that has a fold-down, wing-type handle. There are at least two different types that fit,...sort of. The original (CX-1358/U cable + connector PN) connector has a small round cylinder-shaped housing with a cable exit tube on the side. This type will fit in almost any orientation and can be used if the receiver is installed into a table cabinet.  

There is also a large square housing with the triangular top type that will only fit in one orientation that won't interfere with the terminal strip or the fuse housing. Although this later and larger connector will fit and can be used, it isn't the original type.   >>>

Top of the R-389 showing the RF module and IF module mounted in the receiver. The RF module has the slug racks located under the cover. The two tubes showing thru the opening are the two RF amplifiers. Located under the shield cover are the nine slug racks that comprise the seven tuning ranges and tuned mixer and VFO buffer stages. The Crystal Oscillator and first mixer is the to the left and the second mixer is to the right. The IF module is identical to the type used in the R-390 receiver and is located on the left side of the top frame.

Only one contract for R-389 receivers, Order 14214-PH-51-93, was issued in 1951. The build rate was slow so most R-389 receivers were actually built from 1952 up to about 1955. Total build was 856 receivers.

Rebuild is Necessary - After using the R-389 for a few weeks it's become apparent that this receiver has not been "gone through" in decades. There apparently was some minor work performed about ten years ago that involved the meters and the dial bezel. However, no thorough inspection or any rework or alignments have been performed for quite a long time (well-over ten years.) Sensitivity is poor, not even close to spec (2uv.) The motor-drive sometimes "bogs-down" indicating either poor mechanical alignment or lubrication problems (too much grease, as it turned out.) Most of the worm gears and shafts that require lubrication are located under the RF module which has to be removed to perform the lube job. Per the manual, any lubrication should be very light coatings applied with a paint brush with the excess removed afterward. So, as this project gets started I will insert additions regarding the progress here in this section of the R-390A webpage.

Rework started March 11, 2018.

photo left: Bottom of the R-389 showing the Power Supply module on top, the VFO in the center and the Audio and Electronic Voltage Regulator module on the bottom. Motor drive system is directly behind the front panel. The large rectifier (green with fins) is part of the motor-drive power supply. The long metal arm directly behind the tuning knob is the motor-drive clutch actuator that is cam-driven from the MOTOR TUNE control on the front panel. The three coaxial cables behind the VFO are the inputs from the antenna box (two for the Balanced Antenna input and one for the Unbalanced Antenna input.)

R-389 Restoration Log (Started March 11, 2018)

This probably isn't really a "restoration" but more of a full and complete "servicing" of the R-389. Includes the removal of any "non-military" modifications in order to return the circuit to the original design. Before starting it's worth noting that the receiver does function - not to spec, but it does receive signals. This indicates that most circuits are working but probably not aligned where needed. The R-389 is a very "mechanical" type of receiver and this servicing will probably include not only electronic alignments but also mechanical alignments.

Under the RF module after cleaning off all of the grease. Note the two line shafts that drive the gears that turn the threaded shafts that raise and lower the nine slug racks. The motor-drive bandswitch is in the center of the chassis with the gear drive at the very rear of the module. The motor for the motor-drive tuning is in the lower right of the photo. Note the date on the motor - Sep 10 53.

March 18, 2018 - Dismounted Antenna Relay Box and removed bottom plate. I could see that the Balanced Input wiring had been changed from original. The mod has the Bal. Ant. connector wired to go to the output BNC connectors IF the receiver is turned off. When the receiver is turned on, then the Bal. Ant. input is connected to chassis ground. The reason for this mod seems to be prevention of using the Bal. Ant. input. Check April 4th for correction.

March 19, 2018 - Mod to B+ fuse and HV winding CT found. This mod removed the wires from the B+ fuse holder and soldered them together to eliminate the fuse in that circuit. Then the power transformer CT (pin  6) was disconnected from chassis ground in the power supply module and a wire routed from the CT to pin 15 (unused in ps) of J-118. This pin 15 P-118 connection was originally to the DC 20A fuse but was now routed to the B+ fuse holder and then to chassis-ground (on the Local Audio ground terminal.) This mod may have been installed because of the power supply module conversion to solid state rectifiers. It's likely that the "instant on" HV to the electronic regulator circuit tended to blow the 3/8 Amp B+ fuse. With the mod, fuse blowing would be at the power transformer HV winding current draw rather than at the output of the HV rectifiers to the input to the voltage regulator circuit. The mod also changed the fuse to 3/4 Amp.

March 20 - April 3, 2018 - I had to set aside the R-389 project temporarily. A Collins 32V-3 was acquired that needed testing, servicing and clean-up. I'll get back to the R-389 when the V-3 is completed (a few days, hopefully.)

April 4, 2018 - Well, two weeks (two other projects in addition to the V-3) and back to the R-389. De-mod'd the Power Supply by removing wire CT to pin 15 J-118. Reinstalled the correct bare 14 gauge, "tinned-copper" (TC) wire from chassis to both CT pins on power transformer as original. This completed the Power Supply.

Correction - Checked the Antenna Relay Box again and found that it is correct and original. The relay contacts are actually for grounding the antenna inputs when Break-In is actuated.

The gearbox after cleaning out all of the dried grease. The tuning knob has a built-in slip-clutch that's adjusted by the four recessed screws in the flutes of the knob (two screws are visible in the photo.) Behind the knob is the motor-drive clutch that is engaged when the motor drive switch is turned. This pushes down the lever which pivots against the adjustable standoff to the right and presses the clutch in which then drives the gears for motor tuning. Manual tuning is accomplished with the tuning knob and runs the gear train if the motor drive clutch is disengaged (that is, receiver not in "motor drive.") The tuning knob does "spin" while motor drive tuning.

April 6, 2018 - Removed all wires from non-original connections. Cleaned, straightened and re-tinned all lead ends. Installed DC 20A fuse holder and connected and soldered original wires. Re-wired B+ fuse to now have the original connection to the HV rectifier output. Cleaned all "mod writing" off of the back panel using WD-40 and "Goof-Off." This completed mod removal on the Main Frame.

April 15, 2018 - Back to the R-389. I double-checked some of the RF module mechanical alignments just to verify that everything was correct. Cleaned up some residual grease and splatter that I'd missed earlier. Cleaned the Main Frame bedplate. Cleaned the Veeder-Root counter and repainted the decimal point. Reinstalled the RF module into the Main Frame. This has to be carefully done because there are four cables that have to be routed through holes in the bedplate as the module is lowered into place. The position of the module has to be slightly moved so that the three bedplate captive screws thread in along with the two side screws and the one rear screw. When all six screws have been partially threaded into their pem-nut receptacles then all of the screws can be tightened. Don't over-tighten. Just snug is enough.

Mounted the front panel. The only unusual item to remember to install is the spring-loaded shaft that couples the Frequency Range switch to the mechanism that operates the dial mask. There is a steel brace that screws between the bottom of the RF module at the gearbox to the Main Frame that has to be installed. There are eight 10-32 FH screws that mount the front panel to the Main Frame that have to be installed. The motor drive power cable has to be plugged in. The two harness connectors have to be plugged into the RF module. The three antenna coax cables that come from the RF module and are routed through the Main Frame have to be plugged into the Antenna Relay Box. The remaining coaxial cable connects to the VFO.

I noticed that several of the knobs were the incorrect size for where they were installed based on the panel nomenclature size. I referenced the manual artwork and the manual front panel drawing to install the correct size knobs in their proper locations. Went to my R-390A spare knobs box to find nice condition replacements as needed.   

I found a correct style, good condition Line Level meter on eBay. This was cleaned up and installed. Apparently, when the original R-389 meters were removed the meters were "chopped" out with a cold chisel. On both meters, three of the mounting screw holes had paint scraped off however one hole on each meter mounting was severely "gouged" with a huge aluminum burr that prevented the meter from setting flush against the panel. I used masking tape to protect the panel and then used a file to remove the gouge's deformed metal to allow the meters to mount correctly. Once the Line Level meter was installed, the wires were soldered to the terminals. NOTE: It's odd that the original meters would have been removed for radiation since the original R-389 meters didn't have radium coated needles or scales. I've found out that this might not be true. All R-389, R-390, R391, R390A, etc. meters should have had the radium-phosphorus mix applied to the pointer and scale. Although the phosphorus in the coating can become depleted and not "glow" (age and exposure to sunlight depletes the phosphorus,) the radium in the mix remains radioactive for centuries (radium's half-life is 1600 years.) Although I've never found any R-390, R-391 or R-389 meters that actually "glowed in the dark" it looks like there's a coating on the needle and scale that can be seen. Full details in Part 2 "Carrier Level and Line Level Meters."

More grease! This time on the VFO front gear driven switch. Also the oldham coupler was all greased up. The only reason for the grease on the oldham coupler is to hold it in place while installing the VFO. Only a very small dab is necessary on one side only. Once the greasy mess was cleaned up, the VFO was installed. I had set the frequency to 780kc (no particular reason but it's easy to remember - KOH's freq) so, since the VFO position wasn't moved and the RF module was set to 780kc, then the oldham coupler was in the correct position to allow the VFO to mount easily. The VFO is mounted with five captive screws. (I later had to actually synchronize the VFO for the correct output frequency of 470kc at 15.0kc.)

Installed the Audio Module and installed the Power Supply Module.

There was a brass blade screw used on the front panel to mount one of the resistor boards. Replaced this screw with the correct stainless steel flat head Philips screw.

Cleaned IF module as it seemed to have a light coating of oily dirt on it. Installed IF module and connected the three coaxial cables and the power input plug.

The R-389 was now ready for power-up, testing and electronic alignment.

photo left: The RF module with the cover removed. The two tubes between the slug racks are the RF amplifiers. Note the motor-driven bandswitch behind the tubes. The two left-most slug racks tune the VFO and Mixer Output Stages and Output Coupler. The remaining seven slug racks tune the bands 15kc-27kc, 27kc-56kc, 56kc-128kc, 128kc-255kc, 255kc-500kc, 500kc-855kc, 855kc-1500kc. Below the slug racks are the 34 RF transformers. Four for each band (28,) four for the output coupler and two for the VFO-Mixer output. Note the threaded rods that lift the various slug racks and the associated return springs (visible in the foreground openings.) The threaded rods are gear-driven by the two line shafts that are below the RF module chassis.

HINT: There are two 6BH6 tubes located under the left-most slug rack. It is impossible to remove these tubes for testing unless the receiver is tuned to the highest frequency on either tuning range - 500kc or 1500kc. That lifts this slug rack to its maximum height and allows easy access to these two tubes.

April 19, 2018 - Aligned the IF module. 455kc IF transformers slightly off. Gained about 1 vdc on Diode Load after 455kc IF transformer adjustment. Crystal Filter off quite a bit, especially the core adjust that equalizes the output on the .1kc and 1kc positions. AGC adjustment off quite a bit. About one full turn of the adjustment to attain maximum AGC voltage. Ended up with increase of +10db on CL meter. Performed the alignment on the VFO output stage, the Injection Mixer stage and the Output Coupler stage. This requires a RF probe and VTVM to monitor the test jack level and adjust for maximum output. The RF signal generator is not required because the circuit's crystal oscillator is used as the signal source.   >>>

April 20, 2018 - Aligned First Mixer Output Transformer, Second Mixer Output Transformer and the 10,455kc Crystal Oscillator. This entire section of the receiver was out of alignment. The Crystal Oscillator requires using a VTVM to read maximum RF voltage. One adjustment is accessed from under the receiver and the RF Module brace has to be removed (just two screws.) This adjustment was quite a bit out from peak. The secondary (top side adjustment) was severely out from peak. These adjustments resulted in another +10db increase in signal level on the CL meter. The next step is to align all of the slugs and trimmers for the seven bands that cover from 15kc up to 1500kc.

April 23, 2018 - Adjusted R-260. This adjusts the balance on the First Mixer with one signal from the 2nd RF amplifier and the other signal from the Mixer Driver (VFO plus other stages.) This adjustment was off causing about a 2 volt increase over the minimum setting. Reduces Mixer noise.

The next test was to check the remotely tuned loop antenna operating with the R-389. When first acquired this R-389 didn't receive any signals using the loop antenna. I now connected the loop to the Balanced Antenna input and tuned to 314kc which is the frequency of a nearby DGPS node (read - "very strong signal.") The DGPS signal was received and could be tuned with the variable bias control on the loop. I noticed a significant improvement when the shield of the remote tuner was grounded to the receiver chassis. I also used the HP 606B RF signal generator (connected to a small antenna) as a signal source that could be tuned. This verified that the remote loop could be "peaked" or "tuned" to various received frequencies and that the Q of the antenna was fairly sharp. The higher Q with a sharp peak will give good signal strength with low noise since the antenna is only responding to a very narrow tuned frequency.

Next tests will be with the remotely tuned loop on the Balanced Input and listening at night.  - April 24, 2018

Further testing will be required to determine why the Loop Antenna performance is not as expected. I did modify the Loop to perform best with the Hammarlund SP-600VLF. I may have to do further modifications to the Loop to attain best performance with the R-389.

Another possible explanation is the time of year. Late-April isn't the best time for MW or LF signals. Most of the time, I quit listening to this part of the spectrum by late-February because of much higher noise levels and low signal levels, even at night. -  April 24, 2018

So, what I can say right now is that the R-389 seems to be performing well. It does better on a wire antenna on LW. On the AM-BC band signals are very strong and push the CL meter to nearly +80db. The USN MSK VLF stations also will move the CL meter to about +30db. A comparison test with the SP-600VLF might be an interesting experiment.  -  April 25, 2018

Comparison Test - I tuned in a few signals using the SP-600VLF using the wire antenna (during mid-day.) I could just barely hear MOG 404kc. DGPS 314kc was strong. WWVB 60kc strong and NLK 24.8kc and NPM 21.4kc were strong. I tuned in the same signals with the same antenna using the R-389. The results were nearly identical. The exception was I could easily hear the carrier of MOG but I could only hear the MCW signal a few times. My conclusion is that the R-389 performs just about like the SP-600VLF. See September Performance Update below.

After all of the repairs and the alignment, I used the receiver for awhile. When installing a new Twin-ax to BNC adapter on the Balanced Antenna input, I happened to notice that the OVEN was turned ON. I checked the manual and read that the oven was only on the VFO and was designed to raise its internal temperature up to 167ºF - Wow! Anyway, I started testing the R-389 with the VFO oven off. The frequency "jump" never happened after I turned the OVEN off. I left the receiver turned on for hours and always the tuned frequency remained 20kc low.  >>>

Is this "Oven Induced Frequency Change" a component that fails under heat? Or, is it a component the suddenly "gets better" with heat? With the instant nature of the change, I'd think the component fails with excessive heat. I'm just going to leave the VFO oven turned OFF and use the receiver. In hours of operation the VFO never gets as hot as when its oven is ON (167ºF - that's hot!)   May 10, 2018  -  See Jan 2019 Update on VFO below

With the Autumnal Equinox approaching, I began by listening in the early evening on 9/21/18, which was the day before the Equinox. Only two NDBs were heard, MOG and ULS (392kc, Ulysses, KS) and the noise was still pretty severe. I decided that in the next couple of days I would have to try early morning to see if the noise was down and maybe the signals would be up.

At 0530 on 9/24/18 I started listening using the 100'x135' "T" antenna. Right off, I tuned in ZZP 248kc from Queen Charlotte Islands, BC. Multiple NBDs were heard on many frequencies. I tuned around 390kc and heard what sounded like a voice transmission. I widened the bandwidth to 4kc (from 2kc) and at 394kc I easily copied voice weather being transmitted. With BFO turned off, I could hear in the background RWO being sent in MCW. This was the NDB on Kodiak Island that transmits TWEB or Voice Weather. In about 30 minutes of listening, I had tuned in about 30 NDBs, three of which were newly heard NDBs, ZZP 248kc, RWO 394kc and POY 344kc (#328, #329 and #330, respectively.)

I guess this illustrates that besides a quiet location, a large antenna and a superb receiver, good receiving conditions are absolutely necessary for successful LW DX copy and that my concerns about the R-389's abilities to cope with the modern LW reception issues were unfounded. As conditions continue to improve, I'm looking forward to the "peak LW reception" which will be from mid-November thru mid-January.   

Pulling the VFO, removing the outer cover and oven, then the inner cover provided access to the VFO circuit. At first everything looked normal except that it was obvious that the VFO had been opened before and the desiccant packets removed. Closer examination with magnification showed that the rotor of the the trimmer capacitor was covered with white oxidation. Most of the metal parts inside the VFO had some contamination.   >>>

I suspect that corrosion was "growing" on the plates of the trimmer capacitor and intermittently contacting an adjacent plate. The semiconductive nature of the oxidation may have caused a slight capacitance change rather than a short. The heat would cause expansion to increase the clearance enough to eliminate the "contacting" condition. The receiver was originally calibrated for this "non-contacting" condition. When I turned the oven off, the expansion took much longer to occur and I thought that the problem was solved. So, I had calibrated the VFO for the "contacting" condition. For nine months the receiver operated fine and I didn't experience any change in calibration until the IF tube failure and subsequent repair effort.

Cleaning the oxidation in the VFO seems to have cleared up the problem. Since the "contacting" condition was hopefully gone for good, I had to recalibrate the VFO to the new, hopefully permanent, "non-contacting" condition.  

I ended up adding 260pf of capacitance to the VFO LC circuit in order to have the trimmer at "half mesh" and to have the upper limit of the frequency be around 977.5kc. Low end is close to 470kc. One thing to remember is in the low tuning range the upper limit of VFO frequency is 955kc (at 500kc on the receiver) which actually occurs at turn 49. The 977.5kc is the VFO upper frequency at 52 turns from 470kc (turn 0) which should be the "start" frequency at the low end of the 500-1500kc range. In testing, the upper limit VFO frequency is 968kc instead of 955kc which results in about a 10kc error at the upper end of the range. The overall accuracy of the R-389 tuning is dependent on the VFO tuning being exactly 470kc to 977.5kc in 52 turns. The f-change per turn is 9.76kc upon which the linearity of the tuning range depends.

So, why was 260pf of additional capacitance needed? I doubt any of the components in the LC of the VFO changed value. Another possibility is the ferrite core itself. This VFO had its oven turned on as a method to instigate a change in the LC. Heat seems to be one of the factors that can permanently change the characteristics of ferrite over time. This would change the permeability and therefore the L of the coil. R-388 PTOs are notorious for problems with the ferrite core causing end point error problems that require compensation L trimming to correct. I think long term heat has changed the ferrite core in this 70H-1 VFO and the 260pf padding was necessary to compensate for the decreased permeability of the ferrite core.

Comparison to the R-390A PTO - In comparison to the R-390A PTO, the R-389 VFO is significantly different. Although it's labeled a VFO, it is a PTO, that is, a ferrite core does travel through a coil to change the frequency of the oscillator. But, since the output frequency of the R-390A PTO is 2.455mc to 3.455mc and the output frequency of the R-389 VFO is 470kc to 977.5kc (or 977.5kc to 1955.0kc in the high tuning range,) the method of adjusting the end point error is different. A small trimmer inductor is used on the R-390A PTO because the amount of frequency change is very small compared to the operating frequency. With the R-389 VFO, a fairly large, 95pf max C, air trimmer capacitor is used. This is because the lower frequency requires a greater change in either L or C to adjust the end point and C is the easier to use to affect that greater change. Linearity is adjusted with an adjustable stack of shims on the R-389 VFO.

The procedure for EPE adjusting is basically the same for either VFO/PTO. Trim the high end for an accurate frequency and adjust the low end mechanically for accuracy. Each adjustment affects the other so you have to work "back and forth" until the EPE is as low as possible. The mechanical low end adjustment is similar to the R-390A but the coupler between the VFO gear box and the VFO tuning shaft is loosened to adjust the VFO shaft to the correct lower frequency and then the coupler tightened to test the upper end frequency.

R-389 VFO Test and Adjusting Problems - It's useless to run the VFO outside of the R-389 receiver. There are two output networks that are external to the VFO. One for the low range and one that is 2X output for the high range. Although you can power up the VFO removed from the receiver, the frequency output will be very high and not close to any calibration frequencies. The networks are necessary for correct operation (even though for testing they aren't tuned.) It might be possible to introduce some isolation in the form of a resistor on the B+ connection to the Buffer plate but I haven't tried that. I think the easiest method is to first disassemble the VFO by removing the oven shield, the oven and the VFO shield. By having the receiver on its side and using a small cardboard box (6"x6"x6") to set the VFO on, the oven can be placed inside the oven shield and then that placed where the VFO normally is placed in the receiver. The open VFO then sets on the small cardboard box next to the receiver and is connected via the coax and the power cable. This allows the VFO to function correctly with the output network connected. The large "O" ring seal for the VFO shield should be removed. It was needed when the VFO was pressurized with nitrogen but all of that leaked out the first time the trimmer plug was removed. All the "O" ring does now is make shield removal difficult. The shield has to be in place for test-measurements as its removal will lower the VFO frequency about 10kc.

The initial starting point is to set the VFO mechanical stop on the low end to 470kc output as measured on the plate of the buffer tube, V2, with a digital frequency counter. I use a seven-pin test extension socket to access pin 5 of V2 easily. Be sure to have the frequency counter capacitively coupled since there's +180vdc on the buffer plate. Next, count 52 turns of rotation using the oldham coupler mechanical projection as a reference. At 52 turns, the output of the VFO should be at 977.5kc and the upper end mechanical stop will engage. Adjust the trimmer to the correct frequency and then recheck at zero turns again. It should be 470kc but if it isn't, the mechanical zero relationship needs to be changed for 470kc at zero turns on the mechanical stop. Rotate to 52 turns to the upper end mechanical stop and check, frequency should be 977.5kc. Adjust trimmer if it isn't. This back and forth must be repeated until the zero turn mechanical stop is 470kc exactly and the 52 turn mechanical stop is 977.5kc exactly. This sets the correct span of frequencies for both the lower range and the upper range. I've only been able to get close,...never exact. Remember to always slip the VFO shield on when making measurements.

With the VFO back together and installed in the receiver, I ended up with an EPE of about 8kc at the top end. Accuracy is great from 15kc to about 70kc and then the error begins to show. About 300kc the error is just about 6kc and by 470kc the error is 8kc. The plan is, since I've had so much trouble with this VFO maintaining a dependable calibration, I'm going to run the R-389 through this LW season. If the calibration holds reliably then, after LW season, I'll go back into the VFO and carefully get the EPE and linearity to within 1kc.

UPDATE: Dec 6, 2019 - Another Frequency Jump - I was getting ready to test a newly acquired (but used) Pixel Loop antenna's performance with the R-389. Tuning around I didn't hear anything, hmmmm. I already knew that the Pixel Loop worked great on a couple of other receivers. I finally heard a station at 356kc. Expecting to hear NY 350kc, I was disappointed to hear "DC" which transmits on 326kc. A 24kc error. There's no doubt that the VFO is at fault. When the VFO output is correct the R-389 performs great but I've had nothing but trouble with the VFO. It looks like this last VFO rework (padding the C by 262pf to get the EPE somewhat close with the C trimmer set at mid-range) lasted a little over two months. It's been suggested that the ferrite core might be cracked which results in the sudden change. I'm almost sure the problem is long-term heat damage to the ferrite core but why the sudden frequency change happens might be either cracks or maybe the mechanical mounting of the core itself.    More info to come,...
 

UPDATE: June 29, 2022 - Well, here it is two and a half years later and I guess I've gotten over my frustration with the R-389 VFO problem since I put the R-389 back on the workbench. For the past couple of years I've been watching for a spare 70H-1 PTO for sale - hah!,...none, of course! Also, watching the sales of "parts set" R-389 receivers. I probably saw three or four come up for sale in that time period. I've come to the conclusion that there are no R-389 "parts sets" at least as far as sellers dealing in Collins equipment think,...maybe Collins gear buyers too,...every piece of junk is restorable. All R-389 receivers are being sold with the caveat "I got this from an estate and don't know anything about it,..." or, "For parts or restoration,..." or some other obvious obfuscation of the actual condition of the receiver. However, even the worst condition, most incomplete and obviously non-functional R-389 still seems to sell for over $1500. Of course this is on eBay. Although, I do know of a local ham that has two consecutive serial number R-389 receivers but their actual electronic condition is unknown since the ham is more of a collector than anything else and he's never powered-up either receiver (or almost anything else in his collection.) At the present time these two R-389s aren't for sale at any price. So, rather than essentially pay a minimum of $1500 for a 70H-1 PTO, I've put my R-389 back on the bench to see what else I can come up with for a long-lasting fix to the 70H-1 problem. Since it's summertime, I decided to just skip LF and go to the AM-BC band right off. I was surprised that after 2.5 years the R-389 powered-up without any issues other than some random noise that seemed to disappear after about five minutes of operation. I connected an end-fed wire antenna of about 162ft length to the Unbalanced Antenna Input. As I tuned around, it seemed like the AM-BC stations were pretty close to their correct frequency. I checked the Carson City AM station on 1300kc and it was on 1310kc. KOH 780kc was on 785kc. KPLY 630kc was on 632.5kc. All three of these AM-BC stations were pushing the CL meter up to 80db+ with the clear BC frequencies reading about 10db on the CL meter. So, on the 500-1500kc band, the R-389 seemed pretty normal. Next, the 15-500kc band.  I tuned in WWVB 60kc on 61kc. NPM 21.5kc was on 21.5kc and NML 25.2kc was on 25.2kc. Since the USN MSK stations are so close to 15kc and that is the mechanical calibration point of 470kc from the 70H-1VFO it's normal for these stations to be very close to "on frequency." As the tuning approaches past the middle of the VFO span and higher then the error becomes more apparent. One can extrapolate that the 10kc error on 1300kc would probably be a 5kc error at 400kc in the 15-500kc band, which was very close to what I was experiencing a couple of years ago when the VFO was "on frequency." More accurate testing using a RF Signal Generator showed that maximum error was about 8.5kc at the top of 15-500kc and about 17kc at the top of 500-1500kc. This was very close to what I was experiencing a couple of years ago when the VFO was "on frequency" or as close as I could get it. Next testing will be operation of the R-389 until the VFO again goes "off frequency." I'm really surprised that it's back "on frequency" now because when I stored it away it was reading 25kc high at 326kc (DC 326kc was tuned in at 352kc) in the "failure mode" and the receiver's current condition implies that the VFO problem "corrected itself" just setting around for two years. That's pretty hard to believe, unless the problem is entirely thermally induced and possibly mechanical in nature. This could be a cracked ferrite core or it could be something as simple as a bad solder joint. More testing,... July 1, 2022 - I've been running the R-389 several times during each day trying to induce the VFO frequency "jump." So far, there's been no change and the receiver seems to operate very well. However, today I noticed something different that might be VFO-related. The R-389 had been operating for about a half-hour and was tuned to WWVB. I was in the adjacent ham shack room but could still hear the R-389. I thought I heard the audio tone caused by the heterodyne BFO changing randomly. The tone change was very slight, maybe a few hertz, so I thought it might be just a "hearing anomaly" caused by my location in the other room. But, carefully listening at the R-389 and the heterodyne tone was definitely changing slightly. Only a few hertz random change would probably go unnoticed but since the VFO is certainly a likely candidate for the cause, I will look into this indication and possibly use the DFC to verify that the VFO is what is changing frequency and if it might be related to the f-jump problem. July 2, 2022 - R-389 on for 45 minutes before the random few hertz changes were noted. Since the initial problem of VFO frequency jumping always had occurred when the receiver was installed in the CY-979A cabinet, I wondered if heat might be involved. I've been running these tests with the receiver out of the cabinet and with the top and bottom covers off. To further reduce the heat, I installed a line bucking transformer into the hook-up to reduce the AC line voltage from 124vac down to the specified 115vac. Although the B+ is electronically regulated at +180vdc, the series-parallel tube heaters aren't. It might help but probably not. Incidentally, the CY-979A is now home for the Arvin Industries R-725 receiver, so I won't be putting the R-389 into a cabinet in the future (that changed July 3, 2024 with a third CV-979 showing up.) It's probably not a good idea anyway since the tremendous heat from the electronic voltage regulator seems sort of "trapped" when the R-389 was housed in the CY-979A. Same goes for installing the top and bottom covers in that it seems to really trap the heat on the left side of the receiver. Also, I placed spacers under the main frame of the R-389 to elevate it about 1.5" to allow better air flow underneath the receiver. July 5, 2022 - Continued daily operation for about 30 to 60 minutes with no changes. The few hertz frequency instability was traced to the BFO. A signal carrier was found on LF at around 130kc. Using the BFO, the heterodyne produced was very unstable varying every few seconds by a few hertz but always returning to the set frequency. I then coupled the HP 606B RF signal generator to the receiver to heterodyne with the 130kc carrier. Of course, the difference is that the BFO heterodynes at 455kc while this setup heterodynes at 130kc. The R-389 BFO was turned off. The carrier plus heterodyne producing signal didn't vary at all. Both were solid. This isn't the VFO frequency jump problem but it's something else that has turned up. First replace the BFO tube but that probably won't help - and it didn't. I'll have to use a tube extender for the BFO tube to access the pins to see what is really happening. July 12, 2022 - Slight screen voltage changes on BFO seem to indicate that the bypass capacitor is probably leaking (also, screen voltage is about 11vdc lower than expected at +74vdc instead of +85vdc.) Minor problem that can wait. Since the R-389 won't go into VFO failure I decided to go ahead and move it to a listening position where I can power it up from time to time. It seems typical that once I had it moved to its new location - it would fail! This was different though in that no signals could be tuned in, just lots of noise. Turning the receiver on its side I just touched the VFO tube and all signals returned. I pulled the VFO and BFO tubes and swapped their positions. Also swapped the 3TF7 with another ballast tube. The receiver seemed to be back to normal. I'll keep an eye on the VFO tube socket as a possible contact problem. But, for now, I'm just going to listen to R-389 probably about once a week or so and see how long the "self-healing" VFO lasts.  July 22, 2022 - Tests okay.  Aug 1, 2022 - Tests okay.  Aug 17, 2022 - Tests okay. Aug 31, 2022 - Tests okay. With this test, I checked operation on the 500kc to 1500kc band tuning in several AM-BC stations - all close to "on frequency." On the 15kc to 500kc band, I tuned in NPM 21kc, NAA 24kc and NLK 24.8kc along with the normal reception test signal from WWVB on 60kc. All stations were close to frequency on both bands. Incidentally, KKOH on 780kc drove the CL meter up to +95db using 160' wire antenna. Sept 24, 2022 - The R-389 has been setting, unused due to some station reorganization. So, prior to this test, the R-389 had not been powered up for about three and a half weeks. It was then moved to a different desk and moved from side to side and back and forth until I was satisfied with its position on the desk. At this time, it was connected to the wire antenna and powered up. The tuning was where I had left it on 780kc and, within a short time, KKOH came in with a +90db signal. I tuned to the lower range (15kc to 500kc) and tuned in WWVB at 60kc, coming in with a strong signal. I left the R-389 powered up for about 30 minutes with no problems encountered. Oct 12, 2022 - I used the R-389 a few times now in its new location. The receiver isn't in the CV-979A cabinet and both top and bottom covers are removed. Also, the receiver is slightly elevated to induce more cooling. So far, no problems.  Other Tests in Oct - okay. Failure Nov 2, 2022 - VFO f jump about 20kc high on low band. Interestingly, I tried a heat gun to warm up the VFO but only for a minute or two. Then I turned on the oven. After a minute or two, the VFO instantly changed back to the correct frequency. It wasn't a slow drift, it was an instant change as soon as the VFO had gotten warm enough. All through the summer, the VFO had been working fine. Within the last couple of weeks, with Fall, the temperature has significantly dropped. Last check in the ham shack was 62F. Not cold but a lot cooler than in the summer when it's usually minimum 75F in the ham shack. There are only a few parts in the VFO that could be affected by temperature, unfortunately the ferrite core could be one of them. More testing required to see if the oven temp instigates the frequency jump every time. Test Nov 4, 2022 - Cold start off frequency 20kc. Turned oven on. 4 minutes later, the frequency jumped down 20kc to the correct frequency. Next test will see what happens if the oven is left on for 30 minutes which is the normal listening time period. Test July 2, 2024 - It's been a while. Since it's summer and the room temp upstairs is running about 78F, I thought I'd try the R-389. The VFO oven was turned ON. I had KKOH 780kc tuned at 781kc. The receiver operated correctly for about 15 minutes and then jumped frequency about 40kc (20kc on 15-500kc Range.) Later in the afternoon, I turned the oven OFF and powered up the R-389. It came up on frequency at 781kc for KKOH. I let the receiver run for about 15 minutes and no f change occurred. This frequency jump certainly appears to be temperature related. Incidentally, I was given another CV-979 cabinet. It had been left outside under a carport in Yerington, Nevada for years. The paint was faded and a little bit oxidized, but otherwise, it was in nice condition. I cleaned it up and polished the stainless steel inside slides. I removed two wooden skids that were bolted to the bottom of the regular steel skids. The wood was somewhat deteriorated and its removal improved the overall appearance of the cabinet. So, I've installed the R-389 into this CV-979 (contractor Barker & Williamson) since I don't really think it makes very much difference in the VFO versus heat operation. I did leave the top and bottom covers off of the receiver.

R-391/URR  SN:39
Just Like the R-390 Receiver but with Eight Autotuned Channels

Autotune is one of those features that might seem totally useless to someone that wants a band-cruiser for SW listening or someone that enjoys tuning through the HF ham bands looking for interesting stations. However, if you like the mechanics of an automatic tuning system dating from 1950 that uses a 5800 RPM motor running on +28vdc, rotating a line shaft to actuate two tuning modules with clutches, pawls and gears to drive the receiver's complex gear system of tuning to preset frequencies that are selected by an eight channel switch,...that pretty much describes what is different about the R-391/URR if it's compared to its nearest relative, the R-390/URR. The fact that the R-391 Autotune can be easily be "deselected" by unscrewing the Autotune locks on the MC and KC tuning knobs and that allows the receiver to be tuned manually, just like a R-390, will certainly make the R-391 more appealing to the SWL or ham. But, if you can think of eight favorite frequencies you always listen to or you operate on eight different, fixed-frequency vintage mil-rad nets, the Autotune feature is "very cool" and watching the KC and MC knobs turn by themselves is always impressive along with the cacophony of gear-driven racket coming from the receiver is great for demonstrations.

I waited a month or so before going to Ham & Hi-Fi to see exactly what they had left from Bob's estate. Some of the gear had already been sold on eBay but I was pretty sure the R-391 should be at H&HF. The PP-629 had already been sold (on eBay.) Although there was a lot of recognized gear that had belonged to K6OSM, I was mainly interested in the R-391 (the one receiver type from the "R-390 family" that I'd never owned.) Since the R-391 had never been on eBay, a deal was made for some cash plus a trade of a near-mint Nye-Viking MBV-A tuner and I was able to take R-391 SN:39 home.   Feb 2, 2026

Collins R-391/URR SN: 39
As Found,...looks nice, doesn't work, the CX-1358/U power cable wasn't included,...but I had a spare one here.

R-391 showing IF module far left, Variable-IF and tuned RF coils center to right, Crystal Oscillator right rear has covers installed. The cover over the slug racks was removed for the photograph.

1. No power cable - I have an original CX-1358/U power cable that I will use,...it appears to be NOS.
2. No REMOTE cable for +24vdc input for Autotune. I had two of the proper Amphenol 164-44FS(500) connectors so I made up a cable for REMOTE operation. Cable is six feet long and is 16-2 connected to pin D for +24vdc and pin E for chassis-ground.  Feb 6, 2025
3. Solid State diodes and load resistors mounted to plug-in tube bases to allow easy removal. Will replace with 26Z5 tubes, as original.
4. Tested all of the tubes. 32 tubes can be tested on the TV-7. The 3TF7 ballast tube can only be checked for continuity. All tubes on the Audio Module tested better than 100% above minimum acceptable with the exception of V602 which had a 12AU7 tube installed, V602 should be a 12AT7. The difference between the 12AU7 and the 12AT7 is transconductance with the "AU" being a medium gm and the "AT" being the highest gm of the AT, AU and AX tubes. I doubt that very much difference would be noted, however V602 is the first AF stage for both LINE and LOCAL audio and the lower gain AU might have required a slight increase of the audio gain controls. Replaced V602 with NOS-tested 12AT7.
5. Further tube testing - PTO tube tested 75% above minimum acceptable. Calibration Oscillator had one 12AU7 that had one triode below minimum acceptable. The second 12AU7 tested with one triode 50% less than the second triode. I replaced both Cal Osc tubes with 12AU7s that tested all of the triodes at 100% above minimum acceptable. RF module tubes all tested good except I replaced V204 6C4 with NOS tube. Although V204 wasn't "bad" it was only about 30% above minimum. The three 6C4 tubes must all test as NOS for the receiver to perform at its best. Crystal Oscillator tubes tested good. IF module tubes tested good except V510 12AU7 had one triode below minimum and V507 12AU7 also had one triode weak. Both 12AU7 tubes were replaced with NOS tubes. The 3TF7 was tested and measured 14.4Ω so it's okay.   Feb 7, 2025
6. Gearbox has quite a bit of green grease residue everywhere.
7. Carrier Level meter has a very small crack in the glass center-bottom. There's a scratch and dent in the meter metal cover that probably is from whatever hit the meter hard enough to crack the glass. Luckily, it's a very small crack and not noticeable,...unless you know about it.

Since the R-391 was very similar to the R-390 receiver, the PS, IF and AF modules were interchangeable with the R-390. The PTO, Crystal Oscillator and Calibration Oscillator were interchangeable with the R-390 also. The RF deck was exactly the same as the R-390. There was a "locking gear" that's painted green that's mounted to the front of the cam plate (in front of the Vari-IF.) The front and rear panels and the main frame are unique to the R-391. For most of the receiver, a manual for the R-390 can be used, however the proper R-391 manual will have details on the motor-drive, the auto-tune modules, the Remote and Autotune rear connectors and how to input the +24vdc to operate the Autotune.

R-391 bottom showing the power supply module left, PTO and Calibration Oscillator center and the Audio Module right. Note that the Audio Module has two large tubes, type 6082, combined with the three tubes to the lower left of the 6082s comprise the elaborate electronic +180vdc voltage regulator circuit. The four remaining tubes and components make up the LINE and LOCAL audio sections, the SQUELCH relay and the SHARP, MEDIUM and WIDE audio filters.

Autotune Modules - There are two tuning modules, one for KC and one for MC tuning. These are very similar to the modules used in the ART-13 transmitter. A motor-driven line shaft with three worm gears drives the two modules and the one channel indicating wheel. The set-up operation requires motor-driving to a specific channel, then unscrewing the Autotune-locks on both the KC and MC knobs, then tuning to the desired frequency, then screwing-in the Autotune-locks and actuating the motor drive to a different channel followed by returning to the set-up channel to verify that the autotune goes to the correct frequency. This procedure is then repeated for each channel. As mentioned, if the channel selected isn't changed, then the Autotune-locks can be unscrewed and the receiver tuned manually. As long as the channel isn't changed and the Autotune isn't actuated, only that specific channel's set-up tuning will be affected which will require resetting that channel's desired frequency before screwing in the Autotune-locks and going back to complete Autotune functioning.

R-391 showing the Autotune modules. Note the Autotune-locks, the wing-screws that are on the front of the KC and MC tuning knobs.

R-391 showing the Autotune motor from the bottom side of the receiver. The motor is located under the oil-filled audio screen bypass capacitor.

Solid State Rectifiers and Bucking Transformers - It's very common to find solid state diodes installed to replace the two 26Z5 vacuum tubes that were the original rectifiers in all of the R-390 family of receivers. Some SS replacements are military in origin and were installed because the receivers were generally powered up 24/7 and the SS rectifiers reduced the current load and heat somewhat (although the voltage regulator circuit with the two 6082 tubes runs much, much hotter than a couple of 26Z5 tubes.) Also, since the receivers were on 24/7, there wasn't any "cycling" of the B+ that can cause problems in the other vacuum tubes in the receiver. But, in amateur service, the receiver is turned on and off frequently. The 26Z5 tubes will "soft start" the B+ and also have a greater voltage drop across their diodes which is what was expected when the receiver was designed. Note that the SS rectifiers shown in the photo to right have dropping resistors incorporated into these rectifiers to compensate for the lower voltage drop and higher B+ expected when using SS diodes.

If there's a disadvantage to using the 26Z5 rectifiers it would be the cost factor. Expect tubes to be priced from $15 up to $35 each, depending on the seller and where the transaction would be happening. Ham swap meets will be cheap (although I've never seen 26Z5 tubes at any ham swap around here,...) eBay will be expensive although rapid delivery is certainly the main advantage with an eBay purchase (be sure the purchased tubes have been tested and the results are part of the posting.) The 3TF7 Ballast Tube is also fairly expensive at about $35.

Today's very high AC line voltage will slightly elevate the tube filament voltage and the level of rough DC voltage going to the +180vdc regulator circuit. Also, the +12vdc that runs the Break-in and Antenna Relays will be slightly elevated. The AC line voltage here in Dayton Valley typically runs 122vac but sometimes it will be as high as 124vac. I always use a "bucking transformer" on the AC line going to vintage electronic equipment. The R-390 receivers were designed for 115vac input. If a 6.3vac filament transformer is used as the bucking transformer will lower the AC line voltage about 7vac or 115vac to 117vac for my location.

The plug-in solid state diode and dropping resistors

Power ON Testing - Before applying AC power I checked the fuses to make sure they were the correct type and current rating. I used a bucking transformer to assure that the AC in was 115vac. I turned on the AC power to the R-391. Dial lights came on and after about 30 seconds I started to hear noise in the loudspeaker. I had a 10 foot wire test antenna connected. I had the receiver tuned to 10.000mc but only background noise was heard. I turned on the CAL and still nothing was heard. I switched through the bands and found that 8mc, 19mc, 24mc and 30mc were working and in calibration. In fact, I picked up a USCG weather station on 8.775mc. WWV on 20mc was tuned in by using the 19mc band and tuning to 19.999+. The 14mc band seemed to work somewhat but careful checking using a RF signal generator input as a reference signal revealed that 14mc was tuning backwards and was actually tuning somewhere around 13.7mc. All bands below 8mc didn't even receive the CAL signal. The bands from 8mc to 32mc only functioned correctly on 8mc, 19mc, 24mc and 30mc with the remaining bands just picking up noise. Between some of the bands, increased noise was experienced. I double-checked that all of the slug-racks were operating correctly although for the four bands that work to actually function the slug rack would have to be tuning as it should. The "tuning backwards" on 14mc sounds like one of the conversions isn't working and the signal response is just a harmonic response.   Feb 9, 2025

The Problem - Since 8mc, 19mc, 24mc and 30mc operate correctly, I can eliminate the IF module, the PTO, the Power Supply module and the AF module as having any major problems. The non-operation on most of the bands seems to be located in the RF module or the Crystal Oscillator module. The 8mc to 32mc problem could be poor contact in the switching to the various crystals in the Crystal Oscillator or poor contact in the crystal sockets that could be causing the 8mc to 32mc bands to not all function correctly. Of course, the synchronization of the Crystal Oscillator to the RF module might be a problem that also has to be checked. No signals below 8mc could indicate that the triple conversion isn't working. However, the triple conversion problem might actually be related to the apparent Crystal Oscillator problems above 8mc that are all double conversion and are probably due to crystal oscillator issues.

R-391 SN:39  PTO 70H-2 SN:2630CR
The module to the right is the 100kc Calibration Oscillator

Luckily, there are many types of "E" clips that are easily available at almost any hardware store. The clips have to be for a .250" diameter shaft and have a thickness of about .025" to .030" and be spring material. Fortunately, there's lots of open space around where the clips are mounted so the clip shape isn't too critical (all "E" clips are shaped more or less the same) just the shaft diameter the clip is designed for and the material thickness are important. It might also be possible to harvest some "E" clips from "parts sets."

NOTE: This is the second R-390 receiver that I've found where a missing retaining clip was causing a major problem. The first was a R-390 receiver repair for a ham friend. The receiver had been dropped during shipping. It was of course tested and at first it seemed to function fine but as soon as the MC band was changed and more KC tuning done, the fewer signals the receiver picked-up. Eventually, all that was heard was noise. I was asked to repair the receiver by its owner. I found that a "C" clip on an idler gear in the RF module had "popped off" from the shipping drop. This allowed the idler gear to move out of engagement as the receiver was tuned. Once the idler gear was disengaged, the more the receiver was tuned, the more and more it became desynchronized. Engaging the idler gear and replacing the missing "C" clip was easy since I found the "C" clip setting on the main frame bed when I removed the RF module. Synchronizing the RF module was pretty difficult but I got it working and then performed a full IF/RF alignment on the receiver. The owner was very happy with the outcome and gave me a very nice and complete 1955 Collins R-390A as payment (which I still have.)

I suspect the same sort shipping and dropping happened to this R-391. I've found lots of small bits of Styrofoam peanut-type packing material inside the receiver. That indicates the receiver was probably packed and shipped to K6OSM. I don't know about the actual events but I'd guess the receiver was functioning when initially turned on but then as Bob tuned around it quit working as the Crystal Oscillator became desynchronized. Apparently no troubleshooting was attempted and the receiver became a "shelf queen" that was installed in the rack in his shack.

 *"E" clips are like "C" clips except they can be installed from the side of a shaft rather than from the ends like a "C" clip.

Clips Found - Although "E" clips are available at almost every hardware or automotive outlet and are ultra-cheap to buy, I checked the "junk boxes" first. I thought maybe the R-390A Crystal Oscillator might use a similar band switch but it's entirely different. I checked around and found a couple of National right-angle drive assemblies. These used "E" clips to retain the .250" shaft in the mounting bushing, so one drive assembly had two "E" clips. The clips were easy to remove and I was surprised that they just "snapped" into the recesses in the R-391 Crystal Oscillator band switch shaft. I ran the switch through all of its positions to make sure everything was working. I remounted the Crystal Oscillator module and worked with the Oldham coupler until it meshed together. Then I double-checked that the MC setting of 7.xxx agreed with the Crystal Oscillator indicator wheel showing the index line between 6 and 8. Then the Oldham coupler clamp was "snugged-up" (not over-tightened.) Again, I ran through the MC bands to make sure there wasn't any binding as the RF module was driving the Crystal Oscillator band switch. Next, AC power was applied. I had the R-391 set to 10.000mc and, after about 30 seconds, 10mc WWV was coming in. I checked several other bands from the AM-BC band up to 15mc WWV and all bands seemed fairly close in tracking and generally signals were heard. Now, the R-391 isn't finished,...but it's up and running. It still needs a lot more attention to cleaning, lowering the front panel to clean and lubricate the Autotune system, checking out the Autotune functioning and then a full alignment.

Performance - Later in the afternoon, I connected up the Collinear Array antenna to the R-391. WWV 15mc pushed the CL meter up to an incredible +85db with a no signal (background noise) showing +10db. WWV 20mc read +45db on the CL meter. Frequency readout wasn't bad considering the receiver hasn't been calibrated. WWV 15mc was 2kc off,...not bad,...and I hadn't used the ZERO ADJ. yet, so that was "as found calibration." I tuned up to 16.852mc and heard XSQ CW Marine Beacon from Guangzhou, China. Went up to 21.367mc and heard KH6BB calling CQ from the U.S.S. Missouri in Pearl Harbor. I tuned up to 10M and found a few hams up around 28.350mc. I tuned in the USCG Maritime Weather on 8.775mc coming in very strong. There were several hams on 40M. I had tuned the AM-BC earlier in the day. It appears that the R-391 is working pretty well on all bands. It's not completely finished, that is, not fully aligned. But, this testing relieves any anxiety over any possible major problems and now the refurbishment can continue. Although, I think I'll "tune around" on the R-391 for a few days before I get back to the refurb.   Feb 11, 2025

More Checking   Feb 12, 2025

32 Band CAL Reception - I set up the CAL and the BFO for center band and then went through all 32 bands. All bands received the CAL signal indicating that all bands were essentially functional. The CAL signal is from a 100kc oscillator that is connected to the antenna input. The harmonics allow reception of the signal up to 31mc.

End-Point Error - Since I had the CAL and BFO set up, I calibrated the frequency readout at 15.000mc and performed at EPE test and a general linearity check. EPE was 2.2kc end-to-end which isn't too bad for a PTO "as found" and the linearity was good. I noted that the PTO has the "CR" suffix to its serial number. The "CR" suffix wasn't used until about late-1953 (at least on 70E-15 PTOs.) It might be that the original PTO was replaced sometime in the past with a newer PTO. Also, the PTO SN:2630CR might be the total PTOs for R-390 and R-391 receivers or it might be a new roster of serial numbers starting when the "CR" suffix was added. Since the R-391 manual is from 1953 and Order 14214-PH-51 was used on R-390, R-389, R-391 and R-390A receivers (up into 1954) anything is possible.

LINE AUDIO - I connected a second 600Z loudspeaker to the LINE AUDIO output to test that the LINE AUDIO did function correctly. This also tests the operation of the LINE LEVEL meter, the LINE GAIN control and the LINE METER switch that selects the scaling for the meter of -10db, 0, +10db or OFF. Running both LOCAL and LINE audio outputs to separate loudspeakers will produce a dual-monaural effect that sounds better than running just one audio line.

Dismounting Front Panel Reveals Stuck Screws - The front panel needed to be dismounted to allow cleaning the gearbox, cleaning and lubing the Autotune and general inspection. The KC and MC knobs were removed (Autotune-locks have to be unscrewed and removed,) the BFO and Band Width control shafts were dismounted from the IF module, the Ant Trim knob was removed and the ZERO knob was removed. The bushings on the Ant Trim, KC and MC shafts were loosened. The grab handles were removed. All of this part was easy,...not so with the next part. There are four 8-32 FH Phillips screws that mount the front panel to the main frame. These thread into Ny-lock types of mounted nuts behind the main frame. The four screws were tightened beyond belief! I couldn't move them at all with any type of Phillips screwdriver. I was able to remove one of the four when I used the right-angle ratchet Phillips screwdriver but the remaining three screws just became more and more gnarled until nothing would grip in the Phillips head. To remove "stuck" screws usually heat and thin oil will allow breaking loose the screw but that didn't even phase how tight these screws were. Finally, I gave up and used a special ultra-hard drill bit to drill a center hole down the "cross" of the Phillips head. Then I used a drill bit that was the same size as the 8-32 screw shank. When the drill reached the point where the crown of the countersink meets the screw threads, the head "popped off." Easy and no damage done to the front panel. Once I had the front panel dropped down on the bench, I used a pair of vise-grips to remove the studs that were the remains of the stuck screws. The studs backed out easily. I think the screws were so tight that the conical head was gripping the panel's countersink and with the additional grip of the Ny-locks made it impossible to remove these screws any other way. The new mounting screws won't be torqued using an impact wrench but will just be hand-tightened,...reasonably "snugged-up." In fact, I think I'll put a thin film of grease on the screw threads.   Feb 14, 2026
 

Showing CR801 under the chassis of PP-621

NOTE: As mentioned above, since the R-390 receiver also uses the PP-621 Power Supply module, as does the R-389 receiver, the same CR801 rectifier is used and typically has the same problems in those receivers. What is shown to replicate CR801 can also be used for the R-390 or R-389. Interestingly, my R-389's CR801 seems to be functional since the Antenna and Break-in relays work correctly,...at least, it's functioning for the time being.

R-391 Gearbox after cleaning

Anti-Backlash NOT set on KC Shaft - While cleaning I noticed that the KC Split-Gear didn't have any backlash set in. The springs were completely relaxed. Normally, the anti-backlash is set into the MC and KC split-gears when reinstalling the RF module. Although the MC split-gear was set correctly, the KC split-gear wasn't set. The MC and KC split-gears are mounted to their respective shafts using clamps. By loosening the KC gear's clamp, the gear could be moved forward on the shaft enough to clear the mating gear. Then, using fingers the backlash is set and held as the gear is moved back and meshes with the mating gear. About two teeth is the correct backlash setting. Then the gear clamp is tightened.   Feb 15, 2026

Lubricating the Autotune - The line shaft is difficult to access with the RF module installed. However, the worm-gears and the bearings can be reached using a long and very thin plastic rod no more than .060" diameter to apply a small dab of grease to the worm gears. The same plastic rod can be used to apply a small drop of machine oil to the line shaft bearings. The Autotune modules are fairly accessible if the receiver is turned upside-down and the Power Supply module removed. Most of the gears in the two Autotune modules can be seen through the various holes in the module frame if viewed from the bottom. Also, with the PP-621 removed, the rotating spring-loaded detent can be easily seen and lubricated. It's best to not slather grease on the Autotune gears because the grease will just be scattered as the Autotune operates and the grease will probably end up where it shouldn't be. The manual says to use a small paint brush to apply grease and warns to only apply a thin amount. The manual continues to warn not to lubricate unless the Autotune is noisy in operation,...kind of a relative statement,...the 5800 RPM motor whirling straight-tooth spur gears, line shaft worm gears tuning gears in the modules that are rotating all the receiver tuning gearbox gears and the slug racks going up and down,...that's going to be a pretty noisy 15 seconds. It's a noisy operation when the receiver is on the bench, directly in front of you with the covers off. When installed in a relay rack with the covers on, the Autotune function was a much quieter experience.

Remounting Front Panel - Unbelievably, I couldn't find four good condition 8-32 x 5/8" Stainless Steel FH machine screws. None of the R-390A part sets had any mounting screws that weren't bent or had been severely gnarled. For right now, I used the one original that I was able to remove and then one of the fairly straight "parts set" screws. The grab handles actually bolt-through the main frame, so they also mount the front panel. I'll replace the crummy 8-32 FH screws with new stainless steel screws when I get to the hardware store and go through their stainless steel assortment of fasteners.

Front of the Autotune Modules and the Channel Indicator Wheel

Close-up of the Autotune Motor

To Autotune,...or not to Autotune - The R-391 was designed for a specific type of use,...as a communication receiver that could automatically rapid-tune to a desired pre-set frequency and do so accurately. That use is generally not going to be how the R-391 will be operated nowadays by military radio collectors and amateurs. Luckily, it's very easy to set-up the R-391 to go from an Autotune Receiver to a manually-tuned receiver. What I did was to set up the Autotune to use seven of the eight channels and those seven channels were all frequencies that I use or listen to often. Channel 8 was just set to a random frequency, like 10.500mc, the frequency doesn't really matter. For manual-tuning use, the R-391 is Autotuned to Channel 8 and the cycle left to complete itself (15 sec.,) then both Autotune-locks are unscrewed and the +26.8vdc power supply is turned off. From that point, the R-391 can be operated manually exactly like a R-390 receiver. Do not turn on the Autotune power supply and do not switch from Channel 8 while the Autotune-locks are unscrewed and the receiver is being used "manually." If I wanted to sometime demonstrate the Autotune function I first screw in the Autotune-locks, then I turn on the +26.8vdc power supply and then select any channel from 7 down to 1. The receiver will Autotune to that channel in about 15 seconds. Another channel can be selected and the R-391 will automatically tune to that frequency. When the demo is over, I just return to Channel 8 and let the Autotune finish its cycle, then unscrew the Autotune-locks and turn off the +26.8vdc power supply. At that point, the R-391 is back to being a manually-tuned receiver. Channel 8 will always end up being a different frequency in Autotune,...but it really doesn't matter because the Autotune will cycle Channel 8 to the last frequency tuned anyway. Channels 1 through 7 will always stay preset as long as the Autotune-locks are always screwed in when in the Autotune mode.

How CR801 is built and connects in PP-621

Unsolder all of the wires going to the original CR801. Note the wire color codes and make sure the wires agree with the drawing I've provided. Unscrew the nut and dismount the original CR801 rectifier. The 10-32 screw on the replica rectifier now is placed though the hole in the tab and then a split-ring locking washer and a 10-32 nut are installed. Orient the terminals so they are in the same position that the original rectifier terminals were and then snug-up the 10-32 nut. This nut tightens against the nut on the rectifier. The wires are then connected to the correct terminals and soldered. Be sure to recondition the wire leads for proper lead-wrapping and soldering. This is shown in photo 3.

Install PP-621 and Test - The Power Supply module is then reinstalled into the receiver and, with all of the proper connections in place, the receiver can be powered-up. The Antenna Relay should be heard when going from OFF to STAND BY. When then going from STAND BY to AGC, the Antenna Relay should again be heard. With the receiver in AGC and the front panel BREAK IN switch ON, use a test lead to connect the rear panel BRK IN to GND. The Antenna Relay should operate, the BREAK IN relay should operate and mute the receiver. This test then indicates that the replica CR801 is providing the proper DC voltage to operate the relays correctly. In my R-391, the Antenna Relay is much quieter than a R-390A Antenna Relay and I've noticed the same thing on both the R-390 SN:3113 and the R-389 SN:265. However, I can hear it operating. Interestingly, the Break In didn't seem to work at first. A couple of pulses with the test lead and the Antenna disconnect started to work and then a couple more pulses and the muting started to work. I think that both relays hadn't operated in such a long, long time the contacts probably needed to be exercised several times to clean their surfaces.    Feb 26, 2026   

Performance - The R-391 SN:39 compares very closely to the R-390 SN:3113 and to the R-725 SN:74 when it comes to overall gain and sensitivity. Of course, since the R-391 is just a R-390 with Autotune, the similarity in performance should be expected. The R-725 is a R-390A with the Series 500 IF module (R-390 IF module,) so once again, the similarity in R-391 selectivity isn't unexpected. Like the R-390 receiver, the R-391 audio has noticeably more bass response than the "A" versions.

Signals - As far as signals, WWV 15mc easily drives the CL meter to +80db with the "off signal" being +10db. As can be seen in the photo to the right, the last station I was listening to was Trenton Military VOLMET Aviation Weather out of Trenton, ON, Canada,...they were transmitting an USB signal at that time. On Feb 28, 2025, Trenton Military was sending out the VOLMET Weather in the AM mode. I've never heard Trenton on AM before. With an AM signal, measuring the strength is possible and Trenton pushed the CL meter up to +30db, which isn't too bad for the type of signal and their location.

Autotune - Although I have the Autotune working and set up, I would probably only use it for demonstrations. It's so easy to switch back and forth from Autotune to Manual Tuning, it's really nothing to worry about.

Dual Monaural Audio - I've been running dual 600Z loudspeakers from the LINE and the LOCAL audio circuits and that gives the audio a different sound,...dual monaural. Some of the SW-BC stations broadcast an audio that sounds like high-fidelity with lots of bass response. Running both LINE and LOCAL audio circuits really seems to enhance these station's audio quite a bit. Also, the better bass response of the R-391 certainly can be experienced listening to these stations,...lots of QSB though. One could probably get the same sort of audio experience just listening to the AM-BC band. The R-391 is an excellent receiver that, not unexpectedly, performs just about like the R-390 receiver.

R-391/URR SN:39

Collins R-390/URR  SN:3113

A fugitive from the Reno Salvation Army, this R-390/URR seems to have survived its short period of "donationitis" to end up being saved and then well-cared for by W7ZCA for twenty years. Now, in 2026, it has done another twenty year stint with me, WA7YBS. This is a short write-up about W7ZCA-Paul Eisenbarth's Salvation Army adventures and a subsequent repair log (as best I can remember) that runs from 2006 to Present.

Paul offered me this R-390 about twenty years after he had found it. He wanted $300 for it and that included the CV-979 cabinet. Unfortunately, the CV-979 was missing the shock mounts and the skids. I knew I could recoup the money I would be spending by selling a mint EAC R-390A I had, so I bought the R-390. Of course, the receiver worked fine and was in excellent condition, almost mint inside. The problem was the missing shock mounts and skids on the otherwise nice CV-979.

R-390 "as obtained" in 2006

2013 - Broken Slug on Antenna Transformer for 2-4mc - I noticed that the CL meter wasn't responding to signals on the 75M band. I could copy the signals but they were obviously weaker than normal. A few checks and I was sure the problem was in the RF module. A visual inspection revealed that the slug on the Antenna Transformer for 2-4mc had broken the flex wire and it was at the bottom of the transformer. This was an easy repair that just required removing the broken slug and the adjusting screw. Then the flex wire from the slug was soldered back into the adjusting screw socket. Then the slug was reinstalled and the Antenna Transformer aligned to complete the repair.

2014 - Break-in Slow to Operate - This problem kept getting worse and worse. At first the Break-in was just slow to operate but eventually all it did was buzz. The MgCuS2- rectifier CR801 needed to be replaced. It mounts under the chassis of PP-621. I made a replica using 1N4007 silicon diodes. It looked great and functioned fine but since I had used silicon diodes the voltage drop was a lot lower than it had been with MgCuS2- diodes. The Break-in now switched very fast and with quite a positive "click" from the two relays.

Various Station Set-ups from 2006 to Present - I used the R-390 as the station receiver with a Johnson Desk KW when I was up in Virginia City. From 2012 up to 2018, this R-390 was the station receiver out in the shop in Dayton Valley, operating with the T-368. In 2018, I installed the Arvin R-725 as the station receiver and the R-390 became the station's standby receiver and is now used for general listening.

R-390/URR SN:3113
The R-390 has been in the Dayton shop station with the T-368 transmitter since 2012. Since 2018, it's been the standby receiver for the R-725/URR and T-368 station. I can tell the last station tuned was XSQ Maritime CW beacon from Guangzhou, China.

Dynamotor Retrofit for the R-648/ARR-41

photo right:  R-648/ARR-41 SN: 816 COL from 1957 contract NOas57-438. The shock mount is not original but is homemade. It does have the correct type metal and cushion shock feet installed although they can't be seen because of the front piece of the mount.   

Top of the chassis showing the various plug-in modules.

Shown to the left is the top chassis of the R-648. Directly behind the front panel is the gearbox and the RF module containing all of the RF transformers, VIF transformers, slugs and slug racks. To the right of the RF module is the PTO. To the right of the PTO is the dynamotor module that provides +250vdc, +150vdc (via the 0A2 regulator tube) and +31vdc. Behind the Dynamotor is the Audio and RF Spectrum Oscillator modules. To their left is the Crystal Oscillator module and to its left is the IF module with the two mechanical filters. The front panel also "plugs in" and has enough wiring to be considered another module. The front panel hinges down for easy access.

The NARF NORVA sticker on the front panel. There is also another NARF NORVA sticker on the chassis.

Selectivity is controlled by two mechanical filters. On early versions, the AM filter is 9.4kc. This was later changed to 6.0kc to improve selectivity in the Voice mode. CW selectivity used a 1.4kc mechanical filter. The selectivity is automatically switched between the two mechanical filters when either VOICE or CW is selected by the EMISSION switch. The "SHP" positions are "sharp" selectivity and add an audio filter to reduce bandwidth to improve copy in noisy conditions.

Although the R-648 is called "The Airborne R-390A," don't expect the receiver to have the robust construction or all of the capabilities of its namesake. The R-648 has ample sensitivity, adequate selectivity in most circumstances, a minimal amount of knobs to adjust and lots of volume available if the proper load impedance is supplied. Its light-weight and small size make it easy to find a place for on the bench. 

NARF NORVA - NARF stands for "Naval Air Rework Facility" and NORVA stands for "Norfolk, Virginia." These stickers showed the receiver rework dates. The first quarter of 1975 is marked on this tag. Also note "IRAN." This isn't the country Iran, it's an acronym for the facility's process of "Inspected, Repaired As Needed."

This particular R-648 had been severely "hamstered" in that the original front panel power box connector was removed and a Jones plug installed to allow a 120vac power cord to be plugged-in the front panel. Much of the "behind the panel" wiring had been "hacked" to incorporate this modification. Luckily, I was able to obtain a good condition, original R-648 front panel from Fair Radio. This panel was complete but had a slight bend that needed to be straightened before it fit correctly. 

Unfortunately, the AC power supply couldn't be wired into the front panel unless I wanted to incorporate some modifications and that was what I was trying to avoid. The only solution was to rebuild the power supply into the "original dynamotor" unit and run the R-648 on +27vdc, as originally done.

I removed the power supply module from the R-648 and removed all of the parts that were not original. This left the chassis, the 0A2 regulator, the 11 pin connector, a choke, a couple of resistors and a couple of ground lugs.

photo right shows the R-648 when it had the homebrew AC power supply installed.

Finding Components - One thing I found out was that none of the R-648 manuals, that I had copies of, provide a parts list. Only one photo of one side of the dynamotor chassis is in the manuals. The schematic for the module is called "simplified" but it's the complete schematic. Luckily, the sheet metal had silk-screened component locations that helped in getting the correct locations for remounting the new (replacements for the missing originals) components. Fortunately, the 1.0H choke was not removed but the other three chokes were missing. The original 4K regulator load resistor, about a 5 watt resistor, had been replaced with a 1/2 watt carbon resistor that had become severely burned. Once I had found all of the replacement parts I was ready to rebuild the dynamotor chassis. 

Weird Problem - I had run the R-648 a little when it operated on 120vac so I thought that it was performing correctly. Now, with the dynamotor installation complete and good tubes in the set, I tested performance more thoroughly. Reception from 4.0mc on up to 25mc seemed normal with plenty of sensitivity and dial accuracy. Below 4.0mc, the sensitivity dropped dramatically to where no 75M ham signals could be heard. On 2.0 to 3.0mc no signals were received. Oddly, on 190kc to 500kc an AM BC station was tuned in at around 350kc. With the receiver on the bench, a signal generator was used as a signal source to confirm that below 4.0mc, the R-648 was not operating correctly.   >>>

R-392/URR - Component Receiver for the AN/GRC-19

AN/GRC-19 - The AN/GRC-19 was a portable transmitter-receiver that utilized the Collins-designed T-195 transmitter, a 100 watt output, completely auto-tuned antenna matching marvel that was combined with the Collins designed R-392/URR receiver, also a marvel of packaging most of an R-390 receiver into a cabinet about half the size of the standard R-390. Although separate units, the T-195 and the R-392 were interconnected via a Power Input-Trans Cont cable that allowed the T-195 to power the R-392 receiver. The T-195 operated on +28.5vdc at 40amps and internally had two dynamotors (original configuration.) Later, the dynamotors were replaced with DC to DC power units. Both units mounted to a large shock mount that was typically bolted to the back-seat area of a Jeep or other type of vehicle. The GRC-19 had to be virtually weather-proof and the R-392 is completely a sealed unit (it was claimed that a R-392 would actually float in water - for a while at least.) Most R-392 receivers are in great condition inside because of the weather-proof housing. The entire receiver operates only on +28vdc.

R-392 Circuit - A stout, small and fairly lightweight receiver, the R-392 still has a lot of the features found on it's big brother, the R-390. Frequency coverage is .5mc to 32mc in 32 tuning ranges each with 1mc of coverage. Permeability tuning using slug racks driven by a complex gear train with a PTO, variable tuned IF and fixed Crystal Oscillator providing double and triple conversion is very similar to the R-390 receiver's front end as is the frequency read out provided by a Veeder-Root digital counter. 25 tubes are used in the double and triple conversion circuit that also provides 2 RF amplifiers and 6 IF amplifiers. Also, the IF stages are similar to the R-390 in that mechanical filters are not used for the selectable 8kc, 4kc and 2kc bandwidths. Data modes, e.g., portable RTTY, could be received via the IF output connector (the T-195 was capable of FSK transmission.) The Audio Output is 600 Z ohms and accessed from either of two twist-lock type connectors marked AUDIO or it can also be accessed from the POWER INPUT-TRANS CONT (PI-TC) connector. There is no phone jack on the R-392 because in the GRC-19 configuration the audio was routed to the T-195 (via the PI-TC connector) where typically a telephone handset, the H-33, was used for both transmit (microphone) and receive (earpiece.) The typical field speaker, if used, was the weather-proof (and terrible sounding) LS-166 (shown in photo.) A Noise Limiter circuit is activated with the Function switch and a Squelch function is also available. When operated as the GRC-19 there is a short interconnecting cable between the T-195 transmitter and the R-392 receiver using the PI-TC connector that allows the two units to function together with the T-195 providing Break-in or Stand-by functions along with receiver to transmitter Signal Relay capabilities.

AN/GRC-19 Portable 100W Transmitter & Receiver shown on its Jeep mount
 

photo left: This is the upper deck of the R-392. The modules are the RF Module that has all of the RF transformers, the tuning slug racks and slugs along with the variable IF transformers and its slug racks and slugs. The module the has all of the trimmers is the Crystal Oscillator module. Although reduced in size, these modules are very similar to the R-390/URR receiver.

photo right: This is the lower deck of the R-392. The tuning gear box is visible behind the front panel. Center left is the Crystal Calibrator module, then the PTO and center right is the Audio module. At the rear top is the IF module. Like the R-390/URR, the R-392 modules interconnect using cables and plugs. All modules are secured using captive screws that have their heads painted green. The blue dots on the tube tips are my indication that the tubes have passed a tube test.

photo below-right: This R-392 photo was taken about 2012 using a 3.1mp Olympus digital camera. I had to flash the photos then to get enough detail. Where I took the photos generally had several tungsten-filament lamps on at the same time. The tungsten lamps tended to "yellow" the photos and I was using a very crude photo editor at the time that couldn't correct for it. The olive drab wasn't represented correctly in that version of this photo. I now have a much better photo editor, so I've taken the old photo and "color-corrected" it to better represent how the R-392 receivers generally look.

Operation of the R-392 as a Stand Alone Receiver - To operate the R-392 as a "stand alone" receiver, a separate +25vdc to +28vdc power supply will be required and it should be capable of at least 3 or 4 amps. The typical +24vdc computer-type power supply with the voltage adjusted up to +26.5vdc will work fine.  BE SURE TO USE AT LEAST +25VDC - - - +26.5VDC IS BETTER!  Receiver performance will begin to drop off as the supply voltage is reduced below +25vdc and the R-392 will barely function below +24vdc. The GRC-19 system was designed to operate with the vehicle running or with some sort of charging system used with a 24 volt battery set-up. The typical battery-charging system voltage would have been around +28vdc (although this depends on the engine RPM and the condition of the batteries.) The GRC-19 will only operate marginally at +24vdc since the T-195 is spec'd at +28.5vdc input. When operating the R-392 as a "stand alone receiver" the operating voltage is applied directly to the receiver power input rather than through the GRC-19 system (which usually had some IR drop in the cabling to the receiver power input, thus the +28.5vdc was somewhat less at the receiver power input.) IF a switching computer PS is used be sure to provide shielded cable for the voltage input wires and ground the shield to the minus of the PS which should also be ground-chassis to the R-392. This will reduce the RFI created by some types (especially older types) of switching power supplies.

More "Stand Alone" Info - When the receiver was in active use with the military in a GRC-19, it wasn't really a problem that the supply voltage to the R-392 was a bit high. After all, better gain was then available at the receiver and, at the time, the tubes were easily available. Today, the tubes are still fairly cheap but why abuse them unnecessarily? Although the apparent gain of the receiver can be increased by running the supply voltage at +28vdc, most of the tubes utilized in the R-392 are 26.5 volt filaments and a properly operating R-392 will function great at +26.5vdc supply voltage. This assumes that only one power supply is going to be used and LINE and PLATE are connected together. Increased performance is possible by operating the LINE at +26.5vdc and operating the PLATE at a slightly higher voltage, up to maybe +30vdc. This requires two power supplies and separate voltage wires in the power cable going to the PI-TC connector.

IMPORTANT NOTE ON TESTING 26.5 VOLT TUBES: When testing the 26.5 volt tubes, be suspicious of readings using a typical mutual-conductance tube tester, e.g., the TV-7, etc. Some tubes will show very little gain, perhaps as much as 60% lower than minimum specified useable test level, and yet these tubes will function fine in the R-392. This is probably due to the tube tester's method of powering the tube versus the R-392 circuit's application of DC voltage on the heaters with +28vdc plate voltage. The best indicator of the tube's usability is by substitution in an operative receiver. Naturally, tubes that read high on a tube tester are going to work best but don't necessarily discard the 26.5 volt tubes just because they show "bad" in a tube tester. Try them in the receiver, you might be surprised.

1963 Western Electric R-392/URR

1. Test all tubes on a quality tube tester. Be aware of the caveat on the 26.5 volt tubes and how they will test on mutual conductance tube testers. Clean all tube sockets with De-Oxit. This is probably not necessary on most R-392 receivers but is a routine problem preventative measure.

2. Disconnect all cables and clean sockets and plugs with De-Oxit. I usually pull all of the modules and check everything over. I clean the bed plate although usually it's not all that dirty. If the interior of your R-392 is dirty or the receiver has been out of the cabinet for a very long time, you may want to pull some of the RF transformers to inspect their contacts. If cleaning is necessary, the procedure is very close to that used on the R-390A RF module and is covered extensively in this web-article in the "RF module" section in Part 1 of this web-article.

3. Check operation of the gear box. The feel of tuning the receiver should be fairly light but probably not quite as light as a recently lubed R-390A gear box. It's pretty much the same box and the Veed-Root counter is exactly the same as the one used in the R-390A receivers. If the R-392 gear box seems tight then it might benefit from a cleaning and light lube. They are usually pretty well preserved since the receiver is in a sealed cabinet. Again, the gear box cleaning and lube procedure is above in the section "RF Module" in Part 1 of this web-article.

4. Check operation of all switches and pots. Lube switches with De-Oxit. Pots are usually sealed (Allen-Bradley types) but if they aren't then give them a shot of De-Oxit.

5. Reinstall the modules. Connect all of the cables. Reinstall the tubes. Power up the receiver.

1. Check mechanical alignment. Make sure the PTO and the gear box are close +/-15kc between maximum and minimum span.

2. Adjust Carrier Level Meter first, then proceed to IF alignment. This is like the R-390 (non-A) in that the IF is peaked rather than stagger-tuned. This is because there are no mechanical filters in the R-392.

3. Synchronize the PTO to dial readout. Check that the Crystal Oscillator is adjusted for peak when going from band to band.

4. Variable IF alignment, then the RF alignment.

5. Adjust Calibration Oscillator.

Variations in the R-392 Receivers - The initial contract in 1951 was from Collins Radio Co. but soon, just like the R390A, several other contractors built the R-392 receivers. There are some variations from early production to the later receivers. Early receivers will use 26A6 tubes for the RF amplifiers while later production used an improved version of that tube, the 26FZ6. The change to the 26ZF6 was to help with cross-modulation problems when using the receiver near operating transmitters. Most of the later manuals specify that either the 26A6 or the 26ZF6 can be used as RF amplifiers. Early panels have silk-screened nomenclature while later panels are engraved. The 2kc-4kc-8kc BANDWIDTH nomenclature layout is closer together on early panels but spaced at 90º on later panels. Cabinets on early models have large flutes that run front to back while later cabinets have five "ribs" that entirely encircle the cabinet running parallel with the front panel. These "ribs" strengthened the cabinet significantly. Like many contractor-built items, the color tint of the olive drab paint used varies from contract to contract with some receivers appearing very light brownish-OD while others appear dark greenish-OD. R-392 production ended in the mid-1960s. Ample sensitivity, super-accurate frequency readout and decent audio (from a good speaker - not the LS-166) not to mention the "extreme" military looks along with a small and lightweight package (well,...52 lbs) have made the R-392 a popular receiver with many military collectors and even some BA collectors.

Other R-390 and R390A Versions and Variations

R-391A/URR - Autotune version installed into a R-390A receiver. May have been a conversion and not an official model produced.

R-390A/URR - Squelch Addition - If you've wondered why there is a blank position when turning the knob fully CW on the FUNCTION switch, it was to add a Squelch circuit like the R-390 had. There was also a blanking plate on the Audio module that was intended for the location of the Squelch circuit if it was ever added. I've never seen this option or mod installed on any R-390A. Usually, there's a knob stop installed under the FUNCTION switch mounting nut that prevents rotating to the "blank" position but many times the knob stop is left off during reassembly by those who don't know its function.

R-390A/URR - Navy Power Input/Signal Input-Output Connector Kit - This was a metal chassis with three box connectors that mounted to the rear panel and covered the two terminal strips. When installed it allowed for quick-disconnect of the receiver from input/output connection harness.

USN "DIODE LOAD" Pin Jack - This is an US Navy modification that's encountered once in a while. It added a front panel mounted green colored pin jack to the front panel, usually between the PHONE jack and the LOCAL GAIN control. The front panel was black ink-stamp ID'd "DIODE LOAD" by the pin jack. This jack was added by the Navy to make a routine check on the receiver easier. The Diode Load voltage, essentially the detector voltage resulting from a specific signal level, could be measured at the front panel without removing the receiver from the rack mounting (or the necessity of somehow gaining access to the rear panel.) Apparently this was a part of a monthly requirement to access the tested receiver's operating condition.

U.S. Army Security Agency (ASA) and other Military Security Groups R-390A - Some R-390A receivers will be found with a turns-counting dial with a 10:1 reduction called a "Microdial" installed on the BFO shaft. Sometimes this is referred to as the "NSA" conversion since collection of worldwide signal intelligence is the primary duty of that agency. The reason for the Microdial was for precision adjustment of the receiver to allow the proper bandwidth and placement within that selected bandwidth for either FSK or Dual FSK signals. The Microdial allowed adjustment of the BFO to +/- 13kc rather than the normal +/- 3kc. Since the Dual FSK signals had two Mark and two Space frequencies, the total of four separate frequencies had to be properly placed within the narrowest bandwidth that allowed for no attenuation of the signal - and this was at a time when 850hz shift was the normal military FSK used. The AN/FRA-86 was one such FSK/DFSK demodulator used. These receivers were used in listening and DF facilities where banks of special recorders would monitor and record various types of signals that were of interest to the agencies. Certainly there were many other uses for these types of R-390A receivers since the variety of enemy clandestine signals seemed endless. Likely source of the variation would be the military intelligence and security agencies. Interestingly, the "Microdial" is also shown and described in the USN publication "Cryptologic Collection Equipments" (Ch. 3 "Introduction to Receivers, Fig 3-3) for Naval Security Group personnel. Photo shown to the left is from that publication which is viewable as a PDF, Chapter 3 only, from www.navy-radio.com   Also, I have a "parts sets" Motorola R-390A that has the remains (the base) of the Microdial still present. It has two large stickers on the rear panel titled "United States Army Security Agency" with places for noting installed MWO numbers and the date of installation. This surely indicates that the ASA (1947-77) used the Microdial R-390A receivers. The ASA was part of the NSA.

R-1247/GRC-129 - An entire upgraded SSB comm system for the USAF. Used Manson Labs modified R-390A receivers. Apparently, not all receivers were modified exactly alike. Some had external oscillator inputs. Some receivers had Manson Labs synthesizers installed along with Manson Labs Product Detector modules. Some receivers ended up being used by NASA but these receivers weren't built for NASA. USAF was the source and primary user. The SBM-1102 Stabilizer Kit is the synthesizer that was installed in some of the Manson Labs installations. Apparently these synthesizers weren't reliable and use was limited. Unknown if the SBM-1102 could be purchased separately. Manson Labs was a subsidiary of Hallicrafters.

R-9xx - No details other than possibly a R-390A with LED dial readout installation.

R-1981/TSC-25 - Modified to operate with external frequency synthesizer built by TMC.

R-5076/GRR - This may have been a Canadian version assembled by Zenith Corporation. No specific details known.

R-390, R-390A  Accessories

Security Dial Cover

What were these used for? The most-often-heard purpose is for preventing unauthorized (or unintentional) observation of a listening frequency by radio surveillance room visitors that didn't have the necessary security clearance. This didn't necessarily mean "spies" or "moles," it was just that in certain radio surveillance rooms, the frequencies monitored were classified and not everyone that might have a valid reason to come into that radio room had the necessary security clearance for seeing or knowing the frequencies monitored.

The usual procedure for the radio operator was to raise the cover, tune to the desired frequency and then lower the cover back down. That way the actual tuned frequency was only visible for a short time. Sometimes the covers were left up but immediately lowered if anyone entered the room.

It was also possible to use these covers in a "lights out" situation where the receiver had to be kept in operation. The lowered cover would block the light from the illuminated dial and keep the area darkened.

LS-206A/U Rack Mount Dual Speaker

If you have an LS-206, one use for it is to connect one speaker to the LOCAL AUDIO output and then connect the other speaker to the LINE AUDIO output. That way you can run both audio outputs from your R-390A and also operate both speakers in the LS-206 simultaneously.

The LS-206 was a component part of the AN/GRC-26 Mobile RTTY Communications Hut. The receiving portion of the GRC-26 had two R-390 receivers that were rack mounted with a CV-116 Diversity RTTY Converter and the LS-206. The dual speakers allowed the operator to monitor the signal while tuning in each receiver. Once the RTTY signal was tuned in on each receiver the speakers could be turned off.

The Oneida Electronics version of the LS-206A/U. The contract number on this LS-206 is 15488-PP-63 implying that the contact date is 1963. Serial number on this unit is 19.

Navy Department Bureau of Ships/Hugh H. Eby Company - CU-168/FRR Antenna Coupler

CY-979/URR  Cabinet, Electrical Equipment, CY-979A/URR

The CY-979/URR cabinet was specifically designed for the R-389, R-390, R-391 and the R-390A receivers. All versions of these receivers will fit into the enclosure perfectly with a "wrap around" fit to the back of the receiver that provides a good seal along with screened ventilation holes and louvers to keep out most live insect invaders. The cabinet is made out of aluminum painted smooth gray. The original CY-979 included "skids" that were part of the mounting system. Each skid had a slight up-bend on each side that would slide into the receiving part of the mounting system and thus secure the cabinet in position. The skids were made out of heavy gauge steel painted gray. The CY-979 was specifically referred to as a "mobile table cabinet" and was primarily used in towed huts or vehicles. There was also a CY-917 "light duty table cabinet."

The CY-979 shock mounts were identified as MT-1179(A)/U and were made by several different manufacturers. Barry, Lord Mfg, and Sani-Cap were some of the mount contractors. The mount's flange is bolted to the skids with four flat-head screws the are threaded into tapped holes in the mount's flange. In some versions, each the screws also had nylock nuts installed to keep the screws from loosening. The cabinet mounts to the top of each shock mount with a single bolt. Some versions of the CY-979 have a metal data plate that is mounted top-center of the cabinet. Other versions will have a decal data tag in the same location. The decal data tag is shown in the photo above (note the 1955 contract date.)

 By the late fifties, the CY-979A/URR was being produced. This type of cabinet changed the way the rear opening was made. Apparently, the rounded corners of the earlier versions were expensive to produce so the newer "A" versions of the CY-979 cabinets have a rear opening with square corners. The CY-979A cabinet shown was built by Taffet Electronics, Inc. and the ID is silk-screened on the inside-bottom of the cabinet. This silk-screened ID is quite large and very easy to find and see. Taffet was originally "Taffet Radio and Television, Inc." but changed their name to "Taffet Electronics, Inc." around 1962 because of a law suit.

In the 1990s, W5MC offered CY-979 cabinets in an ad that ran in Electric Radio magazine. The price was between $150 and $200. These cabinets apparently were from either a large collection or some storage depot. The cabinets were "restored" by W5MC and this generally included a new powder coat paint job. Then they were offered for sale. Since the "restoration" removed the original manufacturer's ID, W5MC added a small ink-stamp ID placed on the inside bottom of the cabinet with a "1990s date." Sometimes these are referred to a "reproductions" but even in the 1990s the cost to manufacture an exact reproduction of the CY-979 with shocks and skids along with screened vent-holes would have cost much more than the $200 selling price. These are more accurately called "restored" versions.

Many times you'll run across CY-979 cabinets that are missing the shock mounts and the skids. This seems to have been commonly done to make the cabinet look less "military." If the CY-979 is missing the shocks and skids, it is INCOMPLETE and should be priced substantially less than what the complete CY-979s sell for (which is a lot, these days!) Condition is ALL IMPORTANT!

NOTE 1: However, when you're not really looking for something, sometimes it just turns up. In 2018, I was out in Yerington, Nevada picking up a homebrew transmitter from WA1ICI. Under John's very large carport I saw what looked like a CY-979A cabinet in amidst a large pile of other electronic equipment parts that were all stored outside. Kind of sun-faded, full of sand with wooden slats bolted to the skids but still, it was complete. I asked John WA1ICI, "Isn't that a CV-979?" John's response was, "Oh, I don't know,...if you want it, you can just take it." So, I did. It cleaned up fairly well, still a little sun-faded looking but with the wooden slats removed, it looks much better. It's now housing the R-389 receiver. This CV-979A was a Barker and Williamson contract version. 

NOTE 2: Prices paid for items always changes. I thought that an incomplete CY-979A cabinet shouldn't be too expensive but,...just recently (8/2022) a nice condition CY-979A cabinet that did have the shocks but did not have the skids sold for $675 (plus shipping) on eBay. I guess, draw your own conclusions on the price of a complete CY-979. I have seen quite a few nice CV-979s sell in the $400 range.

photo right: The silk-screened ID of the Taffet Electronics, Inc. CY-979A/URR cabinet. I was given this CY-979 cabinet but it didn't look like these "after" photos. It was severely bent on top where something very heavy had been setting for years. A lot of "body work" was necessary to get the cabinet straight and to try not to damage the paint anymore than it already was damaged. Eventually though, I had to repaint this "Taffet" cabinet. Since the ID was on the inside, the repainting didn't damage this important information. This cabinet is now housing the R-725/URR receiver (which sets on top of the T-368 transmitter.)

Operating a properly rebuilt and aligned R-390 or R-390A today is a pleasure. Pride of ownership is derived from knowing that your receiver was originally built to exacting specifications by some of the foremost American electronics companies and that its performance is the best that could be achieved using vacuum tube technology. Also, in knowing that it would be literally impossible to produce such a receiver today. The cost would be staggering and complexity something most companies would either be incapable of building or something they would definitely want to avoid building due to the excessive cost versus limited profit factor.

Admire your R-390A or other member of the R-390/R-390A Family in knowing that your receiver could only have been built during one specific era - when vacuum tube technology and vacuum tube receiver designs were at their zenith. At a time when this country produced the most advanced electronics apparatus in the world and our government was willing to provide the military with the very best equipment, regardless of cost.
 

Hard Copy Material:

1. Military TMs, Navy Manuals, etc. - the originals provide detailed information on all aspects of receiver design, operation and repair. There are several manuals available and the proper one for your receiver depends on when the receiver was manufactured. Check the publication date of the manual versus your receiver build-date. Army manuals for procedures,...Navy manuals for schematics

On Line Material:

1. www.r-390a.net - the most complete instructions for gear box teardown and reassembly, lots of other information, too.

2. www.navy-radio.com has lots of information on R-390A receivers from a USN perspective.

3. Literally, there is so much data on the R-390A on the Internet it's impossible to list it all. Just do a search on "R-390A" and you'll have pages and pages to look through not to mention photographs galore.

Eyeball, On the Air, E-mail:

1. Thanks to Craig McCartney W6DRZ for all of the data on CD on the R-390A variations.

2. Thanks to Moe Sellali CN8HD/W9 for the detailed R-725/URR information along with the data plate and the proper IF output connector for the R-725/URR.

3. Thanks to Mike Student W7MS for providing the information on the USMC R-390A with OD front panel and many other contributions

4. Thanks to John Svboda W6MIT for the PTO test fixture along with lots of other R-390A info and for many other contributions.

5. Thanks to Les Locklear for sending the Chuck Teeters R-725 article published in Electric Radio January 2006

6. Thanks to Paolo Viappiani for information on many of the R-390A modifications-variations and details on the Microdial.

7. Thanks to all of the many R-390A enthusiasts that have provided information over the years either by e-mail, eye-ball or over the air conversations.
 

R-390A PART 1                    R-390A PART 2                    R-390A PART 3                    Return to Home Index