Radio Boulevard
Western Historic Radio Museum

Rebuilding the R-390A Receivers

Brief History - 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:
(2)1967 EAC R-390A Receivers
Collins R-648/ARR-41
Collins R-389/URR Detailed Restoration Log
Recreations of Arvin R-725/URR, ASA R-390A,
Clark AB Black Panel R-390A, USMC OD Panel R-390A

by: Henry Rogers WA7YBS/WHRM

     PART 1 - History, Assessing your Receiver, Main Frame, RF Module, IF Module, Audio Module, Power Supply, PTO


     PART 2 - Front Panel Restoration, Other Details, Contractors List, Receiver Alignment, Expected Performance, R-390A Diversity Operation, Rebuild-Dynamotor R-648/ARR-41


      PART 3 - Restorations (2) 1967 EAC, Recreations: ASA R390A, Black Panel NSA R-390A, USMC Olive Drab front panel
R-390A, Arvin Industries R-725,


      PART 4 - R-389 Restoration WARNING! Extreme ODC detailed information,  R-392 Info,  Other R-390 Variants, Security Dial Cover, CV-979 Cabinet


1967 Arvin Industries R-725

The R-725 is a R-390A that was modified by Arvin Industries to use a special-build Series 500 IF module that was an updated R-390 IF module (no mechanical filters.) Additionally, the R-725 has a ferrous metal shield installed over the PTO and a "hum bucker" addition that allows the PTO and BFO tube heaters to run on DC voltage. For the full story go to Part 3 of this article under "Creating an Arvin Industries R-725 Receiver."


Front Panel Restoration

Silk-screened Panels - All Collins and Motorola R-390A receivers use front panels that have silk-screened nomenclature. This presents a problem if the panel is in rough condition. About the only solution is to look for a decent condition replacement panel. If originality isn't an issue, a silk-screened panel can be replaced with an engraved panel. Be aware though, that most Collins and all of the other contractors used a short serial number tag with the exception of Motorola. The Motorola contracts used a 3" long tag with different locations for the mounting holes. If you're trying to maintain originality with a Motorola contract R-390A, then you're going to have to find another Motorola front panel. (NOTE: Early contract Collins R-390A receivers also use the long 3" data tag.)

Engraved Panels - Repainting an engraved front panel is very easy. Be sure to mask the back of the panel (if you're going to paint it) where the panel mounts to the Main Frame and also a small area by the upper left mounting screw for the Carrier Level meter. Some panels will have nomenclature on the back that identifies some of the components. These are usually the earlier silk-screened front panels as later engraved ones have nothing on the back but the paint. Be sure to use automotive grade paint that is purchased from an automotive paint dealer. This type of paint will have special hardeners that make the paint really durable. Also, professional paint will dry "ultra-thin, hard and flat" which will help make filling the engraving a lot easier. After painting, let the panel dry at least overnight before doing the engraving fill. 

End-user Panel Repaints - The R-390A specifications state that the front panel is to painted medium gray. The manuals give a specific part number for the paint but it seems the shade of gray did change over the years from contractor to contractor. Also, after fifty+ years on the planet, most of the original paint on the panels will have faded somewhat. When choosing the color for repaint, try to get as close as you can to your receiver's original panel color by having the original paint matched at an professional automotive paint supplier. Use the back of the panel for matching since it's probably the least faded. Only use automotive quality paint for repainting the front panel. Nearly all R-390A panels are found painted gray, however, sometimes the end-users did repaint the front panels totally non-specification colors. Very light gray with black filled nomenclature panels are fairly common and are certainly the product of some rebuild facility that wasn't required to follow specifications. The USAF had banks of R-390A receivers at Clark AFB in the Philippines that had black panels that were actually a black anodized finish. Once and a while, olive drab panels turn up, supposedly painted that way by the USMC. At any rate, there's ample evidence that R-390As were painted colors other than gray when the end-users had some reason to do so. Remember, all R-390A receivers left the contractor's facility with gray panels (and that's original) but it can be considered "acceptable" to paint the R-390A panels colors other than gray if there is believable evidence that the color was actually used on a receiver that was operating in a commercial or military capacity.

Engraving Fill - I use Artist's Acrylic paint to match the engraving fill paint. If you use pure white it will look way too bright. You should mix a color that is close to that found on manila folders - kind of beige color. Apply the fill paint to only one control nomenclature section at a time. Use a small paint brush and dab the paint into the engraving. Don't try to just paint into the engraving - you have to dab the paint into the engraving to have enough there and, of course, you'll have some paint around the engraving - that's normal. Let the fill paint set for about one minute. You'll now have to remove the excess paint around the area. Use a dampened paper towel folded very flat to remove the excess remaining paint on the panel. You'll have to be careful not to "pull" the fill paint out of the engraved area, so keep the paper towel pieces small and only use them once. You'll have to have several damp paper towel pieces ready as you do each area on the front panel. Also, I've found that if you dampen the paper towel pieces using Glass Plus instead of water the Artist's Acrylic comes off much cleaner. These paper towel pieces should be just damp - not wet! Generally, you'll have to do each engraved nomenclature section twice to get a good fill. Let the first coat dry for about five minutes before applying the second coat. When finished let the panel dry overnight. The next day you can apply carnauba wax to protect the panel and the engraving fill which will enhance the overall panel appearance.

IMPORTANT NOTE: Don't use Windex to dampen the paper towel pieces. Windex contains ammonia which might damage the new panel paint. Glass Plus doesn't contain ammonia but works very well to remove the excess paint without damaging the panel paint.

The Back of the Front Panel - The backside of the front panel has several clamps for holding the harness in place. Also, there is a printed circuit board mounted on the back of the front panel just above the frequency readout bezel. There is a lot of mechanical stress on the various wires when the front panel is lowered so check all of the wires to the pots and switches for any breaks or other problems. Since you have removed the front panel for repainting (or replacement,) then you'll be remounting all of the pots and switches along with the phone jack, dial lock and the zero adjuster. Note also that there is a grounding lug on the upper left (as seen from the front) stud of the CARRIER LEVEL meter. This provides a chassis connection for one of the AGC delay capacitors.

Use the Correct Lock Washers - Each nut that secures a potentiometer or switch should have a internal tooth lock washer installed. There are five 6-32 flat head phillips screws that mount into the Main Frame bed and into one of the vertical dividers (into pem-nuts) on the underside of the bed. These screws each have a #6 conical external tooth lock washer installed. If you're missing the conical lock washers, they are available from McMaster-Carr (boxes of 100 - they're cheap.) Notice that the three 6-32 flat head phillips head screws that mount the cable clamps have split ring washers mounted on the back side of the clamps (along with nuts.) The eight 10-32 flat-head screws that mount the panel to the main frame vertically will thread into Nylock pem-nuts, so no lock washers are required.

Mounting the Front Panel to the Main Frame - When mounting the front panel, note that the Dial Lock has to fit over the KC tuning shaft lock-plate. Leave the Dial Lock loosely mounted so it can be rotated to clear the lock-plate. Once the front panel is mounted, you can rotate the Dial Lock into position (the locking grips on each side of the locking plate with the locating tab in the hole one the backside of the front panel) and tighten the mounting nut.   >>>

photo above: 1961 Capehart contract R-390A OD panels were sometimes the product of the USMC
but, in this case, I built-up this recreation of the USMC Capehart R-390A in 2010 (more details in Part 3.)

>>>   If you have the two large shaft bushings (KC and MC Tuning) and the three small shaft bushings mounted to the front panel you'll find it difficult to guide the shafts thru the bushings because of the harness length. Although you could dismount the harness clamps, it's easier to just plan ahead and slide the rear panel shaft bushing onto the shafts and then mount the front panel to the Main Frame. You'll find with the large openings, it's really easy to guide the front panel over the shafts with the harness clamps tightly mounted. Once the front panel is mounted, then slide the rear bushing forward and slide on the washer and thread on the front bushing.

Once everything is mounted to the front panel and the front panel is fully mounted to the Main Frame (and tightened,) then you can go ahead and adjust the panel shaft bushings for the best feel when rotating the controls.

Front Panel Bearing Adjustment - If you want your R-390A to tune "light and easy" then you're going to have to adjust the front panel feed-thru bearings. These are on the KILOCYCLE and MEGACYCLE tuning shafts.

After a thorough cleaning of the RF module gear box, you probably noticed that the KILOCYCLE tuning was very light and easy to manipulate. As you reinstalled the slug racks, the tuning became slightly more difficult to manipulate but was still very light and easy. When the front panel was installed, all of a sudden the tuning seemed to drag and was noticeably more difficult to manipulate. This is caused by the two panel feed-thru bearings. When the RF module is removed and then reinstalled, it's very slightly, differently oriented and the same goes for the front panel. Only a slight misalignment of the panel bearings will cause a "heavy-feel" to the tuning.

Before the front panel is reinstalled, slide the rear bushing onto each shaft. Fit the front panel into position over the shafts and begin installing the mounting screws. With all of the screws tightened that secure the RF module to the Main Frame and all of the screws tightened that secure the front panel, then slide the rear bushings forward and mount the washer and front bushing nut. Note how the bearings can be moved within the feed-thru mounting hole. This is to allow proper placement of the bushing to act as a guide and bearing for each shaft.

Using a 5/8"open-end wrench, lightly tighten the KILOCYCLE bearing nut being careful to not move the position of the bearing itself. Then try the KILOCYCLE tuning. If the tuning is very light then try to just slightly tighten the bearing a bit more - not too much - the bearings don't have to be mounted "super-tight." If the tuning is still light then the adjustment is fine. Usually, no matter how the bearing is adjusted, there will be a slight increase in the "drag" because of the bearing itself. The adjustment is to achieve the lightest "feel" while still providing support for the shafts.

Do the same procedure for the MEGACYCLE tuning although this tuning is much more difficult anyway since you're moving so many of the slug-racks and there's also a detent about every turn of the shaft. Adjust this bearing for the best "feel." You can also apply a drop of machine oil on the shafts to help lubricate the oil-lite bronze bearing that is inside each of the feed-thru bearings.

The end result will be a KILOCYCLE tuning that is very easy to manipulate and feels great when fine tuning is required.

The three small (.25" shaft) bushings are adjusted in the same manner (ANT TRIM, BANDWIDTH, BFO.)

CARRIER LEVEL meter and LINE LEVEL meter - If a R-390A receiver has ever been through a military decommissioning, a surplus house or other "official-type" of handler, both meters will most likely have been removed. There was (is) a concern that the radium-phosphor mix that coated scale and needle represented too much of a long-term radiation source and that was a cause for meter removal and its "proper" disposal. It wasn't necessarily a concern over a single receiver's meter set but, collectively, many receivers represented a fairly large amount of radium in one area that was going to be radioactive for centuries (radium's half-line is 1600 years.)

Whether or not the radiation level from the two meters on a single receiver is anything to be concerned about is up to the individual user/owner to research and decide. Since the meters are about the same size as an old style radium dial wrist watch a comparison might be interesting. Most data available on radium watch dials seems to indicate that a 1 year exposure by wearing that type of watch was equivalent to ONE coast to coast flight on an jet airliner (24 mSv.)

When proper original type meters are found (taken from parts sets apparently) they will likely be in "rough" condition. Sometimes it's difficult to remove the outer metal cover, so rather than damage the meter further, just refinish the metal cover. It's easy to mask the glass and give the body a light coat of flat black paint. If there are heavy scratches or gouges, these will have to be removed with either a file or Al-Ox paper followed by a paint job. I've also "touched up" the meter covers and then used 0000 steel wool applied "lightly" to even out the finish. Be sure to use the gaskets (if you have them) between the meter body and the front panel.

If you want to get inside one of the original type meters - DON'T EVEN TRY! It's a difficult operation that usually ends up ruining the meter. Proper tools are necessary and one should wear protective gloves. If you're contemplating changing the scales to something non-radioactive, this would probably create more of a "radiation problem" since now the scale with the radium is out in the open. How do you dispose of it? It's better to just leave the meter "sealed." Most R-390A owners feel that the meters are safe when installed in the receiver. At a distance of three inches the meter's radiation leakage is not even measurable. It's the phosphor material that "glows" in the phosphor-radium mix. However, the phosphor "glow" is depleted fairly rapidly (it can happen very quickly if the meter scale is exposed to constant sunlight.) So, be aware that a non-glowing meter is still radioactive. As mentioned, radium remains radioactive for several centuries.

A Phosphor-material mix "glows" and Radium provides the energy

Grab Handles - These are made out of stainless steel and can be easily cleaned up with 0000 steel wool and a small brass brush for the washers and the flanges. Wash with Glass Plus before installing.

Dial Cover - This cover, unfortunately, takes a lot of "hits" and as a result is sometimes found dented, scratched or both. Inside the cover is painted with zinc-chromate primer which is bright yellow-green color. Usually the inside is okay but if the cover has dents to be removed, it might need repainting after the body work is finished. Outside the cover is semi-gloss black. Automotive-quality paint should be used for painting the exterior of the cover. NOTE: Minor defects can be touched-up with black lacquer.

Knobs - The knobs also take a beating and many times will need to be restored. First strip the old paint off with a high-quality stripper. Go over the knobs with a wire brush afterwards. Use a high-quality automotive paint in semi-gloss black. Let the paint set overnight. Mix "manila" Artist's Acrylic as described above for front panel nomenclature fill and use the same procedure to add the index line for the knobs. Let this set up for a day and then give the knobs a coat of carnuba wax and install. NOTE: I've found that many times the knobs can be "touched-up" with good results. I use black lacquer paint. I sand the paint chips a little and then dab the lacquer into the chip. As the lacquer dries, it will shrink and become level with the rest of the paint. Inspection of "original" knobs usually will show this method was used in the depots for minor defects.

Tubes - It's very important that ALL of tubes that are installed in a newly rebuilt R-390A are either NOS or are in "used tested as new" condition. For the receiver to operate at its full capabilities ALL of the tubes have to be in "as new" condition. As found, nearly ALL R-390As that I've reconditioned had a bad RF amplifier 6DC6 and all three oscillator tubes 6C4 types always seem to be very weak. The 6AK6 tubes in the AF module also seem to commonly be found in very weak condition. The two 5654/6AK5 tubes also seem to be found weak fairly often. The two rectifier tubes 26Z5 usually test marginal (it's just an emission test for rectifiers.) The 5749/6BA6 tubes seem to last quite a long time (lots of them in the IF module) and the PTO tube is rarely bad. Also, the 5814A/12AU7 tubes (seven used in the receiver) seem to have a long life. BUT, test ALL of the tubes and only use ones that test as "NOS" or "used test as new" and replace any weak or marginal tubes.

Tube Testers - The military TV-7 is a small and easy to use gm-type tester but it's not the only good tube tester out there. Besides, the TV-7 has become very, very expensive these days thanks to audiophiles and tube collectors. There are several other tube testers that work the same way as the TV-7 to measure the tube's mutual conductance. Unfortunately, most dynamic mutual conductance tube testers will be expensive, but maybe just not as expensive as the TV-7. Many of the TV service industry-type of Hickoks do sell for much less than the TV-7 and use exactly the same circuit. Hickok 600, or the 6000. Hickok 534 and many others were produced for TV repairmen and these are all gm-type testers that sell for much less than a TV-7.

Reducing "High Line" AC using a Line Bucking Transformer - The R-390A was designed to run on 115vac input voltage. Tube heater voltages and the non-regulated B+ voltage is dependent on 115vac being the line voltage used. Today, line voltages are nearly always minimum 120vac and often run up to nearly 125vac. Short term operation, such as testing or alignment, isn't usually affected by "high line" operation. Long term operation however will affect tube life and the higher non-regulated B+ can cause an increase in overall heat generated. It's very easy to add an external Line Bucking Transformer to lower the "high line" down closer to the specified input AC voltage. Using a 6.3vac filament transformer (that was designed for 115vac primary voltage) will lower the AC line by about 7.0vac, e.g., lower a 124vac line down to 117vac. Our line is 123.8vac here in Dayton and Line Bucking with a 6.3vac filament transformer lowers the line to 116vac. For receivers I use 6.3vac @ 3A transformers and for transmitters I use 6.3vac @ 8A. It's not critical since the VA is only for the drop, not the entire current the equipment draws. There are lots of easy hook-ups shown on the Internet (search on "line bucking transformer.") Of course, an autotransformer (Variac or Powerstat) could be used just as easily but filament transformers are plentiful and cheap. Besides, the Variac belongs on the test bench.


Other Details

The Contractor Companies, Contract Numbers and Build-Years for R-390 & R-390A Receivers


1951  -  Collins Radio Co.  -  contract 14214-PH-51 (contract 14214-PH-51 was also used for R-389, R-391 and early R-390A receivers)

1952  -  Motorola  -  contract 26579-PH-52


1954, 1955  - Collins Radio Co.  -  contracts 375-PH-54 or 08719-PH-55  (early R-390As built on 14214-PH-51 contract)

1955, 1956, 1958  -  Motorola  -  contracts 63-PH-54, 14-PH-56, 14385-PH-58

1959, 1960  -  Stewart-Warner  -  contracts 42428-PC-59, 20139-PC-60-A1-51

1960  -  Electronic Assistance Corp.  -  contract 23137-PC-60 (may have been for modules only)

1961  -  Capehart Corp.  -  contract 21582-PC-61

1962  -  Amelco  -  contract 35064-PC-62

1963  -  Teledyne-Imperial  -  contract 37856-PC-63

1963  -  Stewart-Warner  -  contract DA-36-039-SC-81547

1966  -  Communications Systems  -  contract FR-11-022-C-4-26418 (may have been for modules only)

1967  -  Electronic Assistance Corp.  -  contract FR-36-039-N-6-00189

NOTES:  The first few hundred R-390A receivers built by Collins will have long data tags with 14214-PH-51 contract number. Remaining Collins receivers have short data plate with either 375-PH-54 or 08719-PH-55 contract numbers. Amelco was supposedly an alternate name used by Teledyne. Electronic Assistance Corporation was owned by the same conglomerate that owned Hammarlund in the sixties. It's often reported that Hammarlund owned EAC but the "Hammarlund Connection" for EAC is vague at best.

Ovens - Crystal Oscillator, PTO, Cal. Crystal - There really isn't a need to have these ovens operating. Maybe it was necessary when the receivers were operated by the military and were set up to copy enemy DFSK signals but today's amateur operations don't require that degree of frequency stability and the operation of the ovens increases the heat within the receiver substantially. To turn off the all ovens, look for the switch on the lower right corner of the rear panel. It's marked "ON" and "OFF" and to switch off just align the screwdriver slot of the switch to be inline with "OFF." It's surprising how many R-390s and R-390As will be found with the ovens still operating. It's not necessary and just creates more heat and consumes power unnecessarily. A Note on All-Matching Modules - This is generally an indicator that the receiver has not been used extensively and has not gone through any sort of echelon rebuild. These types of receivers are desirable in one sense since they usually haven't been brutalized by careless technicians. Most enthusiasts consider the "non-matching modules" equipped R-390A to be inferior since it has obviously been worked on in the past. If you intend to use an R-390A "as delivered" then the all-matching modules type gives you a chance this "out-of -the-box" operation might be possible. However, if you intend to rebuild the R-390A before putting it into operation, then the "non-matching modules" type will be a more reasonably priced option. All of the modules are basically the same regardless of which contactor built them. There are minor differences but they are all interchangeable. You will find that the early Collins and Motorola IF modules don't have adjustments on the inputs and outputs of of the mechanical filters unless they've been upgraded. The ECO was issued in mid-1956. It might seem that the early Motorola RF transformers are of a higher quality than the later EAC units (that used American Transformer units.) However, in early RF transformers it's common to find a stuck rotor and stator on the trimmers. This isn't usually found on the later-manufactured RF transformers. These minor issues are just an evolution of production methods. The contactors had to meet a detailed specification when building each module and all modules will perform to spec after a rebuild. Certainly, if you enjoy the rebuilding process, then a "non-matching modules" R-390A will be your most economical route and you won't have to be concerned about disturbing the unit's originality. However, now at 50 years old, even the all-matching 1967 EAC R-390A receivers should be thoroughly checked over before operating.

The Receiver Alignment

Initial Power-up after Rebuild  You should have the R-390A receiving stations on all bands with power on. You can turn on the CAL and the BFO. Then tune to a 100kc calibration signal. Now rotate the MC knob through each of the bands listening for the calibration signal. Normally, you'll hear the CAL signal on every band although it will be at various tone-frequencies depending on each band's particular alignment at this time. You should hear the signal on every band though. This will assure you that everything is basically working and the receiver is responding to an input signal on each tuning range.

Things to do Before Proceeding to the Fixed-IF, Variable-IF and RF Alignments - you should first do your PTO end-point error correction, if necessary (it will be.) Be sure that the Calibration Oscillator is set correctly and the Veeder-Root counter is in sync with the PTO (you should have checked both before doing the EPE adjustment.) Also, check the Crystal Oscillator output at E-210 and be sure the voltage there is between -3.5vdc and -8.0vdc on the 8mc to 31mc tuning ranges. The voltage should appear when the MEGACYCLE knob is on its detent for each band. You don't have to check the .5mc to 7mc bands because these crystals and associated oscillator circuits are both fundamentally and harmonically operated and also work in either a double or triple conversion scheme. They were actually checked in the higher ranges. Be sure your DIAL ZERO is mechanically centered within its approximately one-sixth turn of the KILOCYCLE tuning when calibrated to XX.000 on the Veeder-Root counter.

Error in Army TM 11-5820-358-35 and more - Field Depot and Maintenance Manual for the R-390A from December 1961. Alignment instructions, page 116, paragraph 76b (2) indicates that URM-25 Signal Generator should be tuned to 18.75mc. Actually, the correct frequency is 18.25mc. This error was very obvious to technicians doing the alignment and is very well known. Interestingly, the earlier TM 11-856A Technical Manual for the R-390A has the correct 18.25mc information, so this later error was probably a typo that wasn't caught in proof-reading. There are many conflicts in the Army TM schematics due to the reluctance of the Army to change the schematic. Instead "NOTE" is placed on the schematic that should guide you to a list of upgrades that changed the schematic or component value. Just be aware that if you're troubleshooting to the component level to double-check if a circuit value is in conflict with what is shown on the schematic.

How to do the Balanced Input Alignment - Make up a test resistance that consists of two 68 ohm 1/2 watt carbon resistors in series. Each separate leg of the resistors will push into to each terminal of the Balanced Input Twin-ax connector. The junction of the resistors will be connected to the signal generator. When performing the alignment of the RF stages first adjust the Balance trimmer on the the RF transformer for the minimum voltage on the DIODE LOAD as read on an (analog) VTVM. The Balance adjustment will not reduce the DIODE LOAD voltage to zero - you're setting the Balance trimmer for the minimum voltage. Be sure the ANT TRIM is set to 0. With the minimum voltage set, now proceed with the RF adjustments for that section of the receiver. Recheck the Balance trimmer adjustment after the particular RF section has been aligned. The Balance should still be close but will probably need just a slight adjustment for minimum voltage at the DIODE LOAD. Recheck the RF alignments - but there should be no significant change and just a slight "tweak" should be all that's required.

Does it Really Matter - Balanced or Unbalanced Input? - If you do the Balanced Input alignment, then the answer is yes. Correctly aligned you might see an improvement if you are using an adapter that grounds one side of the two terminals and connects an unbalanced antenna (that is matched for the received frequency) to the other terminal. This method runs the signal through a set of tuned coils before going to the RF amplifier stage. If you haven't performed the balanced alignment then it's very possible that a reduced signal level may be experienced with this method of connecting the antenna. In this case, connect your antenna to the Unbalanced input (but, eventually, you'll want to do the Balanced Input alignment.) Originally, the Balanced Input was for dipole antennas that used a balanced feed line in the 100 Z ohm range utilizing the "twin-ax" type coaxial cable. Nowadays, hardly anyone runs a balanced antenna directly since most transmitters operate into unbalanced loads. Additionally, the "twin-ax" type cable has a significant db loss per foot. If you're using a tuned antenna with antenna coupler, you might find that the Unbalanced input works better. This may also be the case if you're using a vertical antenna directly fed with coax. Go ahead and do the balanced alignment and then test both Balanced and Unbalanced inputs to see which nets the best results with your particular antenna.

Stagger Tune the IF or Peak Adjustment of the IF? - Stagger tuning will give the IF bandwidth the maximum flat top possible with the mechanical filter selected. Early receivers were "peak" tuned for maximum response and may give a frequency bandwidth somewhat less than the mechanical filter bandwidth. Stagger tuning was used on the later IF modules and does give a better response in the receiver that is generally flat out to the limits of the mechanical filter's bandwidth. You can align early IF modules using the Stagger Tune method. Be sure to check the IF transformers to verify that the Q-spoiler resistors are present. Some IF transformers were modified by clipping out the Q-spoilers to have more IF gain but this also narrows the bandwidth. The Q-spoilers should be installed and connected for the best bandwidth. Updates to TM11-856A have the later procedure for stagger-tuned IF alignment, as does TM11-5820-358-35, or it can also be accessed from many sources on the Internet.

IF Cans without the Alignment Hole - Early production IF modules will have shields over the IF transformers that don't have a hole for alignment. This was to assure that the receiver's IF alignment wasn't tampered with in the field. When the receiver went back for repair or alignment the technicians had a set of covers with holes that were installed for the alignment and when the alignment was finished then the "non-hole" originals were re-installed. If you can't locate an extra set of IF cans then it will be necessary to drill an access hole for alignment. All later IF shields had the hole anyway and many early ones are found nowadays with the hole already drilled since the TM directs the technician to drill a hole for alignment purposes.

Slug and Trimmer Adjustments - One caution on adjusting the slugs and trimmers for the RF alignment. First set the Veeder-Root counter on the R-390A to the specified frequency. Then set the RF Signal Generator to its specified frequency. You'll notice that by "rocking" the Signal Generator frequency that there is a "peak" output on the Diode Load that is very slightly different from the Veeder-Root counter setting. Be sure to use the "peak" Diode Load voltage set by the Signal Generator frequency adjustment. Once you've set the signal generator to have the peak Diode Load voltage, then adjust the proper slugs or trimmers for further maximizing of the Diode Load voltage. Remember, the slug adjustments don't affect the receiver's frequency readout accuracy. That's a function of the PTO accuracy and the Crystal Oscillator. Always set for the "peak" response by "rocking" the RF Signal Generator and then adjust that band's slugs and trimmers for maximum output on the Diode Load.

What to do about early IF modules without the Mechanical Filter trimmer caps? - The trimmers were added with the 1956 contract R-390A receivers. The earlier IF modules will have fixed-value 110pf mica capacitors to tune the input and output of each mechanical filter. When checking these earlier IF modules, it will be necessary to measure the output  and see whether or not each of the Bandwidth positions (16kc, 8kc, 4kc and 2kc) are more or less equal for a constant, known-value input signal. The easiest way is to use the CAL and tune to zero beat and watch the CARRIER LEVEL meter. For instance, if 16kc and 8kc measure 50db on the meter but 4kc measures 40db and 2kc measures 50db, then something is wrong with the 4kc mechanical filter. When doing this test, you'll notice that the fixed tuned mechanical filters are not equal but are usually fairly close - within 5db of each other. If the four filters are not close in their equal response it will be necessary to reselect the tuning capacitors on the filter that is different (hopefully, it's only one filter that has been affected.)  Try to select capacitors so that each Bandwidth position results in a fairly equal output in all four Bandwidths (mechanical filter derived bandwidths, that is.)

Make up an adjustable trimmer that has a range of about 80pf up to around 130pf. Remove the cover from the mechanical filters on top of the IF module and then remove the 110pf fixed capacitor on the particular mechanical filter.  >>>

>>>  Now "tack solder" the trimmer in its place. Power-up the receiver and select the bandwidth for the particular mechanical filter and adjust the trimmer for maximum reading on the CARRIER LEVEL meter. Remove the trimmer and measure the capacitance with a digital capacitance meter. Install a silver mica capacitor of that value to the mechanical filter. This may be enough to get the mechanical filter tuned enough for equal response but it usually isn't.

To do the capacitance selection for the mechanical filter output will require accessing the underside of the IF module. You can loosen the BANDWIDTH and BFO knob-shafts and pull them forward. Then undo the IF OUTPUT coax cable and loosen the three captive screws. The IF module can now be lifted in the front and placed in a vertical position. Use rubber spacers to "prop up" the IF module. Remove the fixed silver mica on the particular mechanical filter and "tack solder" the adjustable trimmer. Power-up the receiver and switch the BANDWIDTH control for the particular mechanical filter and adjust the trimmer for maximum reading on the CARRIER LEVEL meter. Remove the trimmer and measure its capacitance and install that value silver mica on the mechanical filter. Replace the IF module and the knob-shafts and give the receiver a final test to see how the new tuning compares. Hopefully you'll be able to retune the mechanical filter to be within 5db of the other mechanical filters. Needless to say, the 1956 upgrade that added adjustable trimmers to all of the mechanical filters made everything a lot easier.

Most digital capacitance meters aren't even close when measuring pico-farads not to mention at that scale the test meter leads will also add C. It makes this sort of fixed-value tuning difficult to do accurately. Unless the input/output fixed C is way off, it's probably better to just leave the MF tuning "as-is." The spec was within 5db of the other MFs. It would have to be over 10db before I'd try this method of retuning the MF. And,...actually an easier approach is to install the MWO for adjustable trimmers - well, actually just rob the trimmer assemblies out of a junk IF deck, test and install.


Expected Performance

I've used many R-390A receivers in my various ham radio station set-ups over a period of many years. My first R-390A was a 1959 Stewart-Warner version that worked pretty well "as-is" when I got it from a ham swap meet in 1991. I used it with an Eldico SSB-100F transmitter I had and performance was great. A few years later I obtained an excellent EAC version from 1967 but I sold it to buy a 1951 contract Collins R-390 installed in a CY-979 cabinet. I still own and use the R-390 on a regular basis. I sold the 1959 Stewart-Warner after obtaining a 1955 Collins R-390A. This Collins R-390A was given to me as payment for repairing and rebuilding a Motorola R-390 for a fellow ham. Eventually, I "wheeled and dealed" my way into a 1956 Motorola R-390A, a 1961 Capehart R-390A and a "Blue Striper" survivor from St. Julian's Creek Annex. In 2016, a bargain-priced 1967 EAC was purchased. It needed a little TLC and turned into a great receiver. Another '67 EAC showed up at a bargain price in 2017. In 2019, I traded some BC gear to a fellow collector for his entire stock of R-390A parts sets and modules. From that collection I was able to recreate an original Arvin Industries R-725 (R-390A with R-390 IF deck.) I've used all of these receivers except the "Blue Striper" at one time or another, both as an SWL receiver or as a Station receiver.

Here's what I like about the R-390A,...

1. If you absolutely must know exactly where in the electromagnetic spectrum you are listening, the R-390A and its family are the most frequency-accurate readout available in vacuum tube receivers. It's easy to achieve 1/2 kc accuracy or better. The mechanical-digital readout eliminates the vague interpretations of reading analog dials.

2. If you're bothered by QRM, remember the R-390A was designed to intercept radio signals from the USSR, China, East Germany and other Communist countries and be able to successfully copy those signals through any kind of interference whether natural or man-made. The mechanical filters allow the best in steep slope bandwidths. When operating CW, you can also switch in an 800 cycle audio filter. You can literally copy one CW signal with another CW signal almost on top (yes, I've done it,... many times.)

3. When the R-390A is rebuilt and correctly aligned it is very competitive as far as sensitivity is concerned. Are there more sensitive receivers? Of course, but sensitivity isn't all that's required to successfully copy weak signals. Nowadays, the ability of a receiver to function well in a noisy EMI environment is far more important than an "under a microvolt sensitivity" spec. After all, when your local EMI noise floor is S9, what good is .25uv sensitivity? When all the available controls are taken into account along with the mechanical filter selectivity (and the user is very familiar with the operation and capabilities of the receiver) the R-390A is almost unbeatable as a station receiver.

4. Stability is the best in vacuum tube designs. Drift is non-existent.

5. You have two individual audio outputs on an R-390A. The LOCAL AUDIO is normally used to drive a 600Z ohm speaker set-up but you can also use the LINE AUDIO for the same thing - simultaneously! The LINE AUDIO was normally used to drive data devices like RTTY TUs, etc., but there's no reason it can't drive any 600Z load - like another speaker. I've set up a speaker in one room run by the LINE AUDIO and a second speaker in another room run by the LOCAL AUDIO. Independent audio levels in separate rooms. Really neat.

6. Everything about the R-390A's construction is "heavy-duty" and its use metal knobs imparts a massive "feel" to the receiver's operation. The R-390A has a certain impressive presence that attracts the attention of ham shack visitors. This seems to be true whether the visitor is familiar with the R-390A receivers or not. 

The following might be concerns for some users,...

1. On SSB and the Meters - Although there were a couple of military SSB adaptors , the CV-591A (aka MSR-1 or MSR-4) and the CV-157, available and several modifications have been published and other add-on devices for demodulating SSB produced, none of these are necessary for demodulating SSB signals. Unfortunately, many new R-390A owners have only used modern equipment (with SSB Product Detectors) before going to the R-390A which only has a simple Envelope Detector. They expect the R-390A to be adjusted for SSB reception just like their modern receiver - RF GAIN at maximum with the AVC (AGC on the R-390A) on and volume level set by the AF GAIN (LOCAL GAIN on the R-390A.) The R-390A can't be operated like that when receiving CW or SSB. Before product detectors came along it was standard procedure when receiving CW or SSB to reduce the RF GAIN and advance the LOCAL GAIN (AF GAIN) so that the proper ratio of incoming signal to BFO injection could properly demodulate either CW or SSB. AGC was usually turned off but it depended on the receiver. With the R-390A, AGC can be left ON to limit the maximum response, if desired. If the R-390A's BFO is properly set-up, its position allows selecting either upper or lower sideband. Now, you do lose the function of the CARRIER LEVEL meter in this method of reception but who cares? The CARRIER LEVEL meter measures DB over 1uV and its accuracy depends on the RF GAIN setting. If you were planning to use the CARRIER LEVEL meter for CW or SSB signal reports, most stations wouldn't even know what you're talking about when your report was so many "db over 1uV." It's all a relative measurement anyway, dependent on the frequency and conditions. It's better to operate the R-390A as a standard pre-product detector receiver for CW and SSB and just give your contacts an estimated R-S-T report.

2. More on the subject of SSB reception and Modifications - The CV-591A family of SSB adapters were built by The Technical Materiel Corporation. These adapters work from the IF output, therefore you lose the Noise Limiter function, the 800 cycle filter function and the dual audio section of the R-390A receiver if you utilize only the audio output section of the CV-591A. Now, if you happen to have an extra speaker, you can connect one to the CV-591A output (8.0Z or 600Z) and the other one to the R-390A's LOCAL AUDIO (600Z only.) If you want to do SSB or CW you can use the CV-591A and its speaker. If you want to do AM, then use the R-390A's LOCAL AUDIO and speaker. The CV-591 will provide excellent, distortion-free SSB reproduction and they are well-worth using. The only disadvantage is the price that the CV-591A is fetching today, sometimes selling for as much as the R-390A . 

On modifications to enhance SSB reception,...most of the mods that have been published do not improve the receiver's overall performance. Most modifications on any piece of vintage radio equipment will enhance performance for one area at the expense of overall performance. Besides, modifying a vintage receiver to make it operate like a "modern" piece of equipment seems to go against the whole idea of collecting, restoring, operating and preserving these classics in the first place. You're better off to learn how to use the R-390A properly and when you do, you'll find that modifications are not necessary for great performance in all conditions and in all modes.

3. On Audio Quality - Audio reproduction is not as bad as a lot of "BC-Quality AMers" complain it is. The mechanical filters provide a specific, very steep-sided bandwidth but some AM op-listeners are used to the "bell curve" that many early vacuum tube receivers had with only two fairly broad-tuned IF amplifiers. "Ringing" or a "hollow sound" were the usual complaints about the mechanical filter bandwidth. If you change the LOCAL AUDIO coupling capacitors to .022uf and then use a high quality 600Z transformer with a large speaker in a good enclosure, the audio sounds very nice, especially in the 8KC bandwidth (which is really close to 11KC) on AM with marginal signals or 16KC with a really great signal level (AM or SW Broadcast but it's hard to find quality material being broadcast.) You'll have to do the same thing to the LINE AUDIO if you want to run dual audio lines to two separate speakers. However, if you're really into high-fidelity, then you can take the signal from the DIODE LOAD and run it through a shielded cable to a high-fidelity audio amplifier that's connected to a large hi-fi type speaker system. At 16kc bandwidth, AM signals will sound incredible. As with the SSB adapters though, you'll lose the NOISE LIMITER and 800 cycle audio filter functions with this "hook-up" unless you provide a separate speaker on the LOCAL AUDIO line (just in case you want to do CW.) For most users though, especially if you're using typical communications loudspeakers, the stock audio sounds pretty good and the original .01uf coupling caps are fine. I've only substituted the .022uf on a couple of R-390A receivers (and that was done years ago) and it's difficult to tell the difference between those receivers and the ones with stock coupling caps.

Here are some disadvantages to using the R-390A  -  minor stuff, really,...

Break In Operation - The R-390A "Break In" function requires using a T-R relay with Normally Open (NO) contacts that change to Normally Closed (NC) when in transmit. This function operates the R-390A Break In relay and the R-390A Antenna Relays. The Break In function of the R-390A is opposite that of a typical receiver Stand By function that will require the T-R relay to provide a NC state for receive and NO for transmit. Most T-R relays, like the Dow-Key type, will usually provide a set of DPDT auxiliary contacts that allow connecting the R-390A Break In to one of the NO set of contacts. Most vintage military transmitters will have internal T-R relays that provide receiver antenna switching and also NO and NC contacts on the sending relay to control the receiver standby.

The Weight Issue - No doubt, the R-390A is a heavy receiver weighing in at close to 80 lbs. Here's a hint for when you have to move the receiver. Remove the Power Supply module and the AF module. These two modules will reduce the weight by well-over 15 lbs. With the covers off and the two modules out the receiver weighs about 60 lbs - much easier to move. Of course, you do have to get the receiver to the work bench to remove the covers and modules. This hint is really for moving the receiver longer distances, like to another room (when you don't have a roll cart) or ESPECIALLY moving up or down stairs.

Cabinets - If you want the R-390A to be mounted in a cabinet you have two choices. First, is to find the proper CY-979 (or CY-979A) aluminum cabinet. This is a high-quality, military cabinet that is designed for the R-390-family of receivers. They are expensive. Originals were built from the early fifties up well-into the sixties. In the 1990s, an ad in Electric Radio offered CY-979 cabinets for about $150-$200. These cabinets were restored by W5MC and ink-stamped on the interior with an ID. There is some confusion on these restored cabinets as it wasn't clear in the ad if these cabinets were rebuilt old ones or new recreations. It doesn't seem likely that someone could have built a CY-979A complete with shocks and skids and then sell it for so little. But, since any original contractor ID was removed in the restoration process, the only ID is the ink-stamp inside which usually has a date with it (from the 1990s.) I've only seen one of these W5MC CV-979A cabinets and it was exactly like the original, with the screens inside the louvers, proper shocks and skids, etc. I would have to conclude that these W5MC cabinets are "restored" originals. If you are going for the CY-979 cabinet and you're willing to pay a high price be sure that the one you decide on has the shock mounts and the skids. It's fairly common to find CY-979 cabinets with the shocks and skids removed. These cabinets are incomplete and should be priced accordingly. For more details on the differences between the CY-979 and the CY-979A, go to the section on these cabinets at the bottom of the page in Part 4 of this write-up. Other than the CY-979, any other cabinet that is for 10.5" by 19" panels with a depth of 15" will also work. Several sources sell a new Hammond cabinet of this size. Although advertised as a "R-390A Cabinet" it really is just a cabinet in which the R-390A will fit. Price is several times less expensive than the CY-979.

Quick Check for Prospective R-390A Purchases -  The following "Quick Test" assumes you are at the seller's QTH and have AC power available. The test doesn't require anything other than the powered-up R-390A and can be used for any R-390 or R-390A receiver that you are interested in purchasing. Hopefully the seller will allow this easy test since you don't need an antenna (a common excuse from sellers for not providing information is lack of an antenna.) Though a loudspeaker connected to LOCAL AUDIO would help, you don't actually need a loudspeaker but make sure both LOCAL and LINE gain controls are set to "0" if a speaker (600Z ohm load) isn't connected. If you bring your own 600Z ohm speaker, if the seller is agreeable, connect it to the receiver LOCAL AUDIO terminals.

Put on the CAL and the BFO, set the KC to xx.500 kc and rotate the MC change thru all bands to hear if the CAL oscillator is received on all bands. Turn off the BFO, make sure the RF Gain is fully advanced, AGC is selected and then, watching how much the CAL oscillator shows on the Carrier Level meter, run thru the bands again. The meter level should read over 40db. 50db is more likely if everything is aligned. Be sure to check all bands for CAL Carrier Level. With the CAL still on, check the EPE on any band. Be sure to also check the linearity every 10kc. That's an easy test that shows a lot. Recent alignment should have no more than 0.5kc EPE and an unknown PTO might have up to 4kc EPE but should still be linear with just a slight change each 10kc increment. These two tests are easy to perform and tell you a lot of info. If both tests look good, the receiver is probably in good shape. This applies to any R-390A that you want to perform a "quick check" to.

Since the CAL oscillator is practically connected to the Antenna Input, it serves as a onboard signal generator. If you hear the CAL oscillator when tuning through the 100kc markers on all bands (if you have a speaker connected,) then most of the receiver is working well enough to receive a strong signal. The reading on the Carrier Level meter will show generally how sensitive and in alignment the receiver is. Most rebuilt and aligned R-390As will show 50db+ on the meter at around 10mc. Check around 5mc and the meter should read the same or slightly higher. Check around 15mc and the meter will probably read less but still be around 40db. The EPE check just tells you if the receiver has been recently aligned by someone who is thorough and did do the EPE adjustment. EPE < 1kc, recent alignment. EPE > 4kc, typical of "as found" PTO.

You never know how a seller is going to react to this testing, so before doing any testing, be sure to describe what you want to do, what you're going to be looking for and why. Most honest sellers would welcome more detailed information on what they're selling.


Dual Space Diversity Operation with the R-390A Receivers

If you're lucky enough to own two R-390A receivers and have room for widely separated antennas, you can easily set up the pair to operate in Dual Space Diversity. Good separation of the antennas would be at least one wavelength at the frequency of operation but usable diversity effect can usually be obtained with closer spacing if necessary. Space Diversity assumes you will be using two similarly polarized antenna and are relying only on the phase differences of the radio wave based on the spacing of the antennas. You can also try "Polar Diversity" which relies on a vertical antenna for one receiver and a horizontal antenna for the second receiver. Polar Diversity doesn't require that the two antennas be separated by great distances and assumes that there will be a benefit from the reception of two different polarizations of the incoming radio wave. This assumes that some splitting and rotation of the radio wave will occur as it propagates through the ionosphere and is returned to earth. Generally, space diversity helps with fading signals and polar diversity helps with phase distortion due to wave rotation. 

With either method of Dual Diversity reception, the receiver set-up is the same. You will be connecting the DIODE LOAD from each receiver together. The receiver that you plan on operating as the "master" will have to have the DIODE LOAD terminals jumped while the "slave" receiver doesn't have the terminals jumped. The "slave" receiver is only operating to the detector stage and its audio output is not used. You can connect 500 ohm resistors across the LINE AUDIO and the LOCAL AUDIO on the "slave" receiver. You will also have to install the jumps to connect AGC DIV terminals together on each receiver. You will also have a wire connecting the AGC DIV from each receiver together. A speaker on the LOCAL AUDIO is only required on the "master" receiver.  To listen to just the "slave" receiver, turn the RF GAIN on the "master" receiver to 0 and what you hear thru its speaker is the "slave" receiver. Also, if you want to listen to just the "master" receiver, turn the "slave" receiver's RF GAIN to 0 and what you hear thru the speaker is the "master" receiver only. With both receivers operating and connected to their respective antennas, tune in a strong shortwave broadcast signal. Have both receivers' RF GAIN set to about 8. Don't set the RF GAIN on either receiver to "full on" (10) or each receiver will "fight" the other one for control of the AGC line. By alternately reducing the RF GAIN of each receiver to 0 you should be able to end up with both receivers tuned exactly to the signal. Once the SW BC signal is tuned in on each receiver you will need to "balance" the RF GAINs. Slowly increase the RF GAIN on the each receiver alternately to the point where you see the CARRIER LEVEL meter showing some response. Adjust the RF GAIN on each receiver until you have the highest CARRIER LEVEL readings on each receiver without one receiver or the other "overloading" the AGC line. When "overloading" occurs the CARRIER LEVEL meter on one receiver will drop much lower in its reading and with a reduction in the RF GAIN of the other receiver you'll see the meter reading jump back up. By "balancing the receivers" you get the best diversity response and the best sensitivity. You will note that the two receiver's CARRIER LEVEL meters will react differently since each receiver is responding to a phase difference in the radio wave based on the separation of the antennas. You should see deep fades that cause one CL meter dip while the other receiver's meter remains steady. You should also see a reduction in phase distortion if you are using the polar diversity set up.

Remember, you can only use AM reception on this type of Dual Diversity. That's because CW or SSB reception requires the BFO to be in operation and the BFO dominates the detectors and spoils the diversity effect. For RTTY reception special TUs were used like the CV-116 that was designed for diversity RTTY. Diversity CW reception required Tone Keyers.

So, give Dual Diversity reception a try if you can. It's interesting and sometimes beneficial to copy. 

photo above
: A Dual Diversity set-up with the 1956 Motorola on the bottom and the 1955 Collins on top. Both receivers are fully rebuilt (including recapping) and both are fully aligned. Though the speaker looks like a Hallicrafters PM-23, it isn't. It's a homemade wooden cabinet that houses an 8" Jensen speaker and the louvers are made of aluminum. The cabinet is painted gray wrinkle finish.


Rebuilding, Refurbishing and Recreating Projects


Dynamotor Retrofit for the R-648/ARR-41

The R-648/ARR-41 is the so-called "Airborne R-390A." Certainly the nickname comes from the intended use of this Collins-design/build along with its many similarities of construction and operation to this receiver's "bigger brother," the R-390A. However, Collins' engineers borrowed a lot of the circuitry from the 51J Series of receivers. Since the R-648 was destined to be used in aircraft, it had to be light-weight. Rather than the 80 pound, hefty-weight of a R-390A, the R-648 weighs-in at about 30 pounds. To achieve this lighter weight package, the receiver is much smaller, with reduced-size mechanics and components. Of course, most of the R-390A features and circuits are not even present. Still, with 17 tubes that provide two RF amplifiers, three 500kc fixed-frequency IF amplifiers, frequency coverage from 190kc to 550kc and from 2mc to 25mc, two mechanical filters (9.4kc and 1.4kc - the 9.4kc MF was later changed to 6.0kc) and a mechanical digital frequency dial, the R-648 does have a few "R-390A" features. The circuit conversion methods were borrowed from the 51J/R-388 by using a dual variable IF and fewer crystals in the Crystal Oscillator along with the 500kc IF. Double conversion is used on all frequency ranges except for the 2-3mc band and the 3-4mc band where single conversion is used. Audio output has three stages and was designed for a headset of 300Z ohms or greater. Typical antenna input Z is 50 ohms.

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.   

photo above: Top of the chassis showing the various plug-in modules.

The R-648 operated on +28vdc battery-charger buss available on the aircraft. The +28vdc buss supplied the series-parallel connected tube heaters, the dial lamps and also operated the "on-board" dynamotor. The dynamotor supplied +250vdc at 100mA. A regulator tube type 0A2 was used to also provide +150vdc from the +250vdc. A divider network also provided +31vdc for the AVC bias. Unfortunately, like a lot of the dynamotor-operated military gear, many R-648 receivers have been converted to operate on an AC power supply. Some modifications were well-designed and their incorporation into the receiver was accomplished with minimal modification to the receiver while still providing the necessary voltages. Most AC power supplies utilize the original dynamotor module chassis since this then has all of the circuitry for the 0A2 regulator and the divider network for the +31vdc. Usually two power supplies and two transformers are necessary, one for the +250vdc supply and one for +24vdc for the tube heaters. Some better conversions utilize a DC-DC converter to provide the +250vdc and then the receiver is operated on +24vdc input. In essence, providing a solid-state substitute for the dynamotor only. If you don't have the original dynamotor, this is probably the best solution to keeping the R-648 as original as possible.

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.

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

As to the R-648 operation, the GAIN control operates RF gain when in CW and AF gain when in AM. The Sensitivity is a screwdriver adjusted pot accessible on the front panel behind a "toilet seat" cover. The overall Audio Gain is a screwdriver adjusted pot accessible on the rear panel. The audio output impedance works well with 600Z ohm loads and a loudspeaker with a 600Z ohm matching transformer will have plenty of volume. The manual doesn't specifically give an output impedance only stating that the load should not be less than 300Z ohms.

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 "Inspect, Repair As Needed."

Dynamotor Retrofit - In 2017, I purchased an original R-648 dynamotor. It was just the dynamotor, not the entire power supply chassis. When looking at the AC power supply that had been built and installed in the R-648, it was apparent that it was built on the original dynamotor/power supply chassis. Even the regulator tube circuitry and many other original parts were still within the chassis. I purchased the dynamotor with the idea that I could rebuild the AC power supply back into the original dynamotor configuration.

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

Servicing the Dynamotor - Most dynamotors found today haven't had anything done to them in several decades. The R-648 dynamotor only needed a servicing. This consists of removing the end-bells to access the brushes, the commutators and the bearings. I mark the brushes for location and orientation so when reinstalled their fit to the commutator is correct. With the brushes removed, I use a small piece of 400 grit Al-Ox paper to clean the commutator surfaces. I wash the commutators afterward with denatured alcohol. I run some light-weight oil through the bearing to clean out the old grease. I then work new grease into the bearings. These bearing are somewhat "sealed" but it's still possible to push new grease into the bearings. I then reassemble and test. My test was to run the dynamotor for 15 minutes while monitoring its output voltage (+260vdc) and the running current. No problems were experienced so the dynamotor was ready to mount to the stripped chassis.

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. 

Building the Dynamotor Circuit - Without a good photo of an original R-648 dynamotor chassis, I was pretty much left with fitting everything into the small front area. I used ground lugs and tie points as necessary. I followed the silk-screened positioning if it was possible (most of the time, it was.) Of course, the components aren't anything like the originals in appearance but they are the correct values. The +250vdc chokes were easy but the +27vdc chokes had to carry 2 amps so the wire for those chokes had to be at least 18ga or so. Once the circuit was built, I had to test the operation and voltages. With +27vdc input on pin 10 (+) and pin 11 (chassis and -) I had +260vdc, +150vdc and +33vdc on the proper pins of the connector. This completed the test and the dynamotor chassis was then installed into the R-648. Building the Power Cable - I gave KDWC some measurements and pin orientations and he was able to find the correct seven pin power connector identification number. From this, he located a used connector on eBay. Upon receiving the connector, I removed the wires from the pins. Only four pins are used for the R-648. Audio output is on pin A, +27vdc is on pin E, Chassis Ground is on two pins, B and D. The wires need to be 16 gauge on the power pins and 22 gauge is large enough for the Audio Output. The proper size wires were taped together and then a braided shield (harvested from old RG-8 coax) was installed and the entire cable wrapped with electrician's tape. I made the cable six feet long. The connector was installed on one end with the shield connected to a chassis-ground pin. On the other end of the cable spade lugs were soldered to the wire to allow a good connection to a +27vdc power supply. I used an old linear supply good for 10 amps - overkill, no doubt, but solidly stable.
Routine Problems - Before going any further, I needed to test all of the tubes. I was surprised to find that three of the 5749 tubes in the IF module were bad and both 5726 tubes were bad. All other tubes in the receiver tested good. Odd that most of the tubes in one module were defective. Anyway, a quick test of the receiver with ALL good tubes installed and the performance improvement was very good but there was a problem that involved all of the bands below 4mc.

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.   >>>

>>>   Further testing revealed that 2-3mc and 3-4mc bands seemed to be "tuning backwards." This had to be some type of mechanical problem. I checked over the description and the drawing of the gear box. The description said that below 4mc, the receiver tuned counter-clockwise to increase frequency. Hmmm. Mechanically, the tuning was the same on all bands - clockwise to increase. Finally I read the description in the section before the low frequency mechanical tuning. I read that the gearbox always tunes the same direction and that the "appearance" of tuning counter-clockwise was due to a sliding mask over the dual, opposing dial scaling. The mask switching was cam-driven and changed scales at 4mc. In my case, the dial mask wasn't moving at all. The problem was caused when I had replaced the front panel with an original condition panel from Fair Radio. I hadn't noticed that inside the dial cover the spacing between the dial mask and the left side dial lamp was really, really small. I had the dial cover too close and too low and that was blocking the dial mask from moving. I installed a second dial cover spacer taken from the old "hacked up" front panel to give a slight bit more clearance. Upon reassembly, the dial mask now operated correctly and the tuning followed exactly what the manual instructed it should.
Test and Alignment - I noticed that the slug rack for the 190kc to 550kc was way out of mechanical synchronization. This particular slug rack is very easy to realign only requiring pushing the spring-loaded gear rack back and sliding the slug rack into the correct position. All other mechanical synchronizations were correct. The IF alignment requires accessing the switch arm S402D to input the 500kc signal, otherwise it's straight forward. There's also a PTO output transformer that needs to be matched to the IF frequency. The Variable IF consists of two bands, 2-3mc and 3-4mc and the alignment is by slugs for the low-end and trimmers for the high end. The various 1mc wide bands are set-up by the various crystals in the Crystal Oscillator that are operating at either the fundamental, 2nd and 3rd harmonics and then mixing with the PTO and the Variable IF. The exception is the 2-3mc band and the 3-4mc band which both are single conversion and bypass the Crystal Oscillator. Alignment of 190kc to 550kc and the bands from 4mc to 25mc are via slugs and trimmers in four sets of RF transformers. Once all of the mechanical issues were resolved the actual alignment was easy,...well,...sort of. Alignment Using Nav Manual - To say that the Nav manual is difficult to use is an understatement. First, errors abound. The alignment procedure has so many misidentified components that a significant amount of time is spent looking for the non-existent parts and then trying to figure out what the actual part ID is. Even when the procedure is correct, the user is jumping from the alignment page to tables in other sections of the manual. The most serious error is the synchronization of S402 as the instructions list the wrong switch tab for mechanical alignment. The test points locations are shown in another section of the manual rather than with the alignment section. I read thru the alignment instructions first (a couple of times) and then added (in pencil) the actual locations of the specified test points before doing the alignment so I wouldn't have to keep jumping back and forth between sections. Penciled corrections were also added "as found." Certainly, after one has gone thru their first alignment of a R-648 and all of the test points and errors are then known, any subsequent alignments, of course, would be much easier. Be prepared doing your first R-648 and read the instructions first, make notes and that should ease the frustration somewhat. Probably the Navy figured that only their experienced techs that had been sent "to school" on the R-648 were going to be working on the receivers so the convoluted and inaccurate instructions wouldn't be too serious of a problem.
Performance - The R-648 does have a lot of gain and is a very sensitive receiver. The automatic switching for the GAIN control in CW or VOICE is actually very practical and makes the switch from CW (or SSB) to VOICE (AM) easily accomplished. Since in CW the GAIN controls the RF gain, SSB is easy to demodulate. Selectivity with the mechanical filters is excellent. The IF bandwidth has such steep sides that strong AM signals suddenly are heard and suddenly aren't heard as the receiver is tuned through the entire 9.4kc bandwidth (late-build receivers have a 6kc Voice MF.) The action of the AVC time-constant does sometimes block AM signals if they are tuned through their passband rapidly. Slow tuning in AM allows the AVC time to control the receiver sensitivity. Even though the CW bandwidth is only 1.4kc SSB signals sound just fine. Audio quality is pretty much communications-grade and it's easy to drive a 600Z ohm speaker to loud volumes. There's no "break-in" provided. If used as a station receiver you'll have to provide good isolation during transmit and reduce the GAIN as necessary (this would be for voice only, in CW the receiver can be used to provide a sidetone.) It's small size and light-weight makes the R-648 easy to locate within the station landscape. I'm using my R-648 set up with my DY-12 operated ART-13. The receiver operates on a +26.5vdc 6A power supply and transmitter dynamotor operates from a PP-1104-C DC power supply (+28vdc at 50+amps.) Pseudo Shock Mount - The mount that can be seen in the top photograph came with the receiver. However, it didn't really look this way when I got it. It had rubber feet and wasn't painted. I replaced the rubber feet with the proper type, metal and cushion type shock feet. I painted the shock mount black to sort of match the receiver. On the whole, while certainly not as "cool" as an original would be, it is better than having the receiver cabinet set directly on the table. Or worst, having rubber feet installed on the bottom of cabinet. Original R-648 shock mounts are rare but maybe one will turn up someday. Until then, this one provides the necessities well enough.   NOTE: An original R-648 shock mount did show up on eBay a couple of years later. The seller also had the receiver but that was on a separate auction. Needless to say, the shock mount sold for double the price of the receiver. The shock mount went for $300 and the receiver went for $150.



R-390A PART 1                     R-390A PART 4                     Return to Home Index



Donations to Radio Boulevard - Western Historic Radio Museum's Website

If you enjoy using Radio Boulevard - Western Historic Radio Museum's website as an information resource and have found our photos, our hard to find information or our restoration articles helpful, then please consider a donation to the WHRM website. A small donation will help with the expenses of website operation, which includes website hosting fees, data transfer fees, research, photographing and composition. WHRM was a real museum that was "Open-to-the-Public" from 1994 to 2012 - eighteen years of operation. WHRM will continue to provide its on-line information source with this website, which has been in operation since 1997.

Please use PayPal for sending a donation by clicking on the "Donate" Button below



Radio Boulevard
Western Historic Radio Museum

 Vintage Radio Communication Equipment Rebuilding & Restoration Articles,

 Vintage Radio History and WHRM Radio Photo Galleries

1909 - 1969

- 60 years of Radio Technology -



This website created and maintained by: Henry Rogers - Radio Boulevard, Western Historic Radio Museum 1997/2022