Collins R388/URR Receiver Resource Page
I own a Collins R388/URR, 51J-3. I acquired it via one of the newsgroups by placing a want ad approximately one year ago. I do not recommend looking for radios on Ebay, as I have been burned twice by sellers who refuse to pack their radios correctly. I had an HT-37 Hallicrafters transmitter and a beautiful R388 beat to death by the good folks at Unmitigated Package Smashers (UPS). My second R388/51J-3 is in fair shape and I have done some minor modifications to it (product detector and AGC). This particular model receiver has a special place in my heart, as a borrowed one almost stopped mine in mid-beat in 1986: the chassis was, unknown to me, hot with AC due to a leaky bypass capacitor. I had my hand on the shell of a PL259 and layed my other hand on the chassis. WHAM!!! My hands shook for three days!!>
The purpose of this webpage is to accumulate all the notes I have obtained via email and other means as a way to possibly assist others in their quest to learn more about this fine receiver. I wonít be including the various manuals in PDF format as this information is readily available on the various Collins websites. If you have any data to pass on, I would be glad to include it here on this page. Where possible, I have credited the source of my information. My receiver has been named "Fred" by my kids...gotta love 'em!
Collins NObsr-49132 (AN/URR-23A): 91 units
Collins NObsr-52527 (AN/URR-23A): 280 units
Collins 19624-PHILA-50: 1,000 units
Collins 1908-PHILA-51: 845 units
Collins 3096-PHILA-51: 8 units
Collins 3131-PHILA-51: 919 units
Collins 3135-PH-51: 1,175 units
Collins 3155-PH-51: 1,379 units
Collins 3164-PH-51: 153 units
Collins 3167-PHILA-51: 498 units
Collins 3357-PHILA-52: 1,672 units
Collins 3362-PHILA-52: 109 units
Collins 3469-PHILA-52: 20 units
Collins 3470-PHILA-52: 1,449 units
Collins 25635-PHILA-53-36: 2,319 units ---- contract included "Break-in On-Off" switch
Collins 25067-PH-54-55: 436 units
Collins 21318-PH-56: 88 units
Collins 30951-PHI-57: 121 units------ contract included R187
Collins 37003-PC-62: 171 units------ contract included R187 and "Break-in On-Off" switch
Collins TCG-38466 (for USCG): 28 units
Total known production: 12,735 units
Many of us choose to modify our receivers to make single-sideband reception enjoyable without having to back way down on the RF gain and utilize the full sensitivity of the receiverís IF stages. There are several different modifications available, depending on your source. Without doubt, if one does the SSB modification, one must also do the AGC modification as well, or the receiver wonít perform properly and you may get distorted audio. For reference, the particular modifications known to me are as follows:
Lee, CQ Magazine, April 1961
Sherer, CQ Magazine, December 1968
Orr, Ham Radio Magazine, February 1978
See also articles from Orr (Ham Radio 2/78), Scherer (CQ 12/68), Lee (CQ 4/61) Lankford (Hollow State News #28 and #36) and Odell (CMJ 9805-06)
Component references are for R388/51J-3/J-4. (the 51J/J-2 are different.)
This is Dallas Lankford's version of the product detector:
-Replace R-161 with a 47K 1/2 watt resistor.
-Lift pin 2, V-114 from ground
-Locate the wire connecting pin 5, T-106 to pin 7, V114. Move it from pin 7 to pin 2, V-114.
-Add components labeled "(new)" on the schematic.
-Note the photo and location of the terminal strip for the output network.
-Note that unlabeled components on the schematic are unchanged from the original circuit.
I have found that all of the switching wiring can be done on the front panel harness. Only one shielded wire needs to go from the front panel to the chassis. The remaining connections can be done at the BFO switch, at the Limiter switch and at the Audio Gain control. Simply disconnect and tape off the wire that connects R-154 ("high side") to S-116. These 2 terminals will be used to connect the switched audio. (see diagram)
PRODUCT DETECTOR PROCEDURE:
These first notes are about Dallas Lankford's Product Detector. I have implemented this one three times with excellent results. I used it because I liked what Dallas addressed with his version:
1. Strong BFO harmonics through 7 mHz were fixed with an LC instead of an RC pi network for the output. For the network he used the Orr configuration (blocking cap between plate and pi net, keeps B+ off the network). Component values used were: 0.01uF, 560pF, 1mH, 560pF.
2. "Motor-boating" was fixed by adding a 10uF 450 volt electrolytic in parallel with C-219 (Frisco Roberts, 10/78 Ham Radio). "Motor-boating" is a low grade oscillation when the AF gain was turned up. Sounds kind of like the noise kids make to imitate a motorboat. Morris Odell described it as a raspberry sound.
3. There is maximum simplification of the mod, while maintaining required signal levels. This also addresses poor access to components.
-R-160, R-162, C-218 and C219 are unchanged from original BFO circuit.
-R-161 changes from 33K to 47K. I accidentally did one radio, first pass, without changing R-161. It worked pretty well! Subjectively, the minor value change does seem to improve performance.
-Dallas did not use the Orr cathode resistor and cap (220 ohm/0.01uF). In this respect his circuit looks like Lee's and Scherer's.
4. For the product detector input network (at pin 7, new V-114) he used a 2pF cap and 8200 ohm resistor. With 2 pF, I had to go to a 15K resistor to get balanced audio levels between AM and SSB. I have had good results (last 2 radios) using 5 pF and an 8200 ohm resistor. I recommend this starting point. I have tried values from 2.2 to 5 pF and 4700 to 15k ohms.
AUTOMATIC GAIN CONTROL PROCEDURE:
The following notes are about the AGC solutions I have tried with the product detector. Between three radios now, I have tried all of the AGC mods in CQ and HR, but not HSN. AGC is the simple part, but very important!
1. Lee's mod is a simple, changing R-144 to 5 meg. It works very well (on my R-388) if "bias" is in the -60 volt range (notes follow). In about 4 years of regular use I have not had the predicted AGC amp tube failure. Best of the lot for AM.
2. Orr's mod is a little more complex and works well (I have it on two 51J-4's). The Orr mod uses a diode-resistor network for R-144 and more time constant capacitance added at C-205B. Use a 1N4148 diode. With the recommended 1N270, the decay is way too fast (even with the added 1.5uF cap!). Again, biases must be right to prevent AGC pumping on AM and distortion on stronger SSB signals (more on this later). (see schematic)
3. Scherer's mod did not work, although it looks like it should have (maybe my mistake).
4. I have not tried Lankford's mod. Others have reported good results with it.
I would start with Lee's, then try Orr's if Lee's does not work. By "does not work", I mean that SSB signals will still sound distorted with the RF gain all the way up. A trick that I used to make this installation easier was to attach a small (2 or 3 lug) terminal strip to one mounting screw of C-205. Remove R-144. Bring a wire from pin 3 of V-110 to the terminal strip. Now either mod can be done between this terminal strip, ground and the approriate lug of C-205, all where you can get at it! For the Orr mod, I soldered the two resistors and diode into one package, covered them with heat shrink tubing and installed them as one component.
BIAS FIX PROCEDURE: YOU MUST DO THIS FOR AGC TO WORK RIGHT!
There is no doubt, bias levels drive the AGC performance differences. Morris Odell wrote a short article on this for the Collins Monthly Journal (CMJ). For this discussion bias is as measured at pin 1 or 2, V110, no antenna. Basically lower (closer to zero) voltages here work better (Odell's conclusion also).
1. -Specificed level is -56 volts.
2. -My early (1950 contract) R-388 reads -60 V. It performs very well with the simple Lee AGC mod.
3. -My 51J-4's (sn's 6093/3419) Both started out reading near -70 volts. Neither worked with the Lee mod. The Orr mod worked, but I got some AGC pumping. In both cases, I was able to bring the Bias back to around -65 and the Orr mod worked very well (no pumping and excellent AGC action).
The radio gets bias voltages through a resistor string on the transformer HV center tap: R-164,165,166 (& R-184 on late models). That string sees full HV current, so the bias is porportional to the total current load in the radio (E=IR, R=550 to 562 ohms). (Note that the parts list in the J-4 manual has the wrong value for R-164,166. The Schematic is correct.) Based on my experience and discussions with other users, excessively negative bias seems to be a problem for late models. It's probably a result of higher line voltages today (the numbers work out about right). There are three solutions. All involve bringing the radio back to the early circuit configuration. (There is also a fourth. Run a bucking transformer!)
1. -Removing R-186 will return the radio to the early (51J-3/R-388) AF PA curcuit and will unload the power supply 10 mA or so (changes the operating point of the AF PA). This will move negative bias about 4 to 5 volts closer to zero. I found no degradation in audio quality, making me wonder why Collins even bothered to add this resistor. Morris Odell made the same observation in his article for CMJ.
2. -Bridging R-184 (12 ohm 3 watt, only there on J-4s). This is good for a volt or two.
3. -Returning the dropping resistor scheme for the voltage regulator tube to the "early" configuration, a single 4000 ohm resistor (R-181). Late units have 2 resistors totalling 3200 ohms (R-181 & R-185). Just get the total closer to 4000. It works, This explanation maybe an oversimplification, but:
4. -The voltage at one end (V-116, 0A2/VR150) is always 150 volts, and the other is always high B+ (about 230 volts), so voltage drop is always 80 volts.
5. -Again E=IR, so do the math. You can save 5 mA with this change. That's another 3 volts of bias impact.
To make a long story short, you should measure the bias first to determine the starting point. If it is around -60V, try the Lee mod first. If it is nearer -65V, go straight to the Orr mod. If it is near -70V, make the changes above until it is near -65V.
Don't be intimidated by all of the AGC notes: this is a real simple and reversible modification that makes a big improvement in the versatility of the radio.
Post Item: After the modification, is the AGC popping on Strong signals? Is so, then note that there are three resistors which determine the range of adjustment for the RF gain control. They also convert the bias string (the 3 power resistors on the side wall near the power transformer) to usable voltages for the AGC amp.
So check the following:
1. -R146 which is wired in at the AGC switch, and should be accessible on the back of the front panel. It should be 47K . Check it with AGC in the On position, otherwise the switch shorts the resistor.
2. -R147 which is wired to one of the three power resistors. It should be 27K. Changes in this resistor's value can be the culprit. A new resistor may fix this problem.
3. -R149 which is wired to some terminal strips near the power resistors. It should be 820 ohms.
4. -V111 is the AGC amplifier tube. Just sub another 12AU7 tube and see if the symptoms change.
Additional Note: Early R388s without the mechanical filters don't have R187 in the IF chain, which adjusts the gain. Adding the variable pot and its associated bypass capacitor (in later 51J-3's and 51J-4's), and adjusting the gain down may also help (I am trying this one myself next--just an idea on my part). Compare the two schematics for and add the two components right after the crystal filtering stage.
Aligning the R-388 PTO is tedious, but actually easier to do than to describe! The same technique will work with nearly all Collins PTOs, although others have different frequency ranges.
First, assuming the 70E-15 PTO is already out of the radio, remove the 2 1/2' metal mounting disk in front of the tubes, V001, and the hex cap over the spread coil adjustment. Make sure that the lock collar on the little shaft inside is loose, and that the little shaft can be turned (not always easy)-be gentle- you definitely don't want to damage the shaft! Replace V001 and its shield, but not the metal disk. Of course you must do the alignment with dubious capacitors/resistors replaced, high emission 6BA6s and tube shields installed, lubrication completed, and with the PTO coil cover screwed firmly in place. Remove, carefully clean, and reinstall the little tensioned grounding spring riding on the PTO shaft: it must make good electrical contact with the shaft!
Now make a jig so that you can determine when you have rotated the shaft exactly 1 turn and exact multiples of 1 turn. There are lots of ways to do it, but I solder a 1" piece of wire to a solder lug and fasten the lug to the standoff supporting the 2 1/2" disc on the front of the PTO (between the tubes), so that the wire stands up vertically. Glue a 3" dia. cardboard disc with a 1/4" hole in the middle to a little knob with a setscrew that you can fasten to the PTO shaft. Make an index mark on the edge of the disc, and set the vertical wire index just behind the disc, with the end bent over the edge of the disc to sit in front of the index mark.
Hook up the PTO to +150VDC (regulated) and 6VAC, let it warm up at least 30 min, then check that the PTO output waveform with a scope-it should be a very clean sine wave with no erratic jumps or spikes in the trace as you turn the PTO through its range. Don't try to align the PTO if it isn't working properly!
Attach a frequency counter to the PTO coax and set the PTO frequency to 2.5 MHz, the middle of the PTO's 3-2 MHz range. Lock the knob and attached cardboard disk to the shaft with the setscrew so that the disk index mark lines up with the index wire, and the counter reads exactly 2500.0 kHz. Now turn the knob and shaft exactly 5 turns clockwise, record the readings at each turn (100 kHz interval), then 10 turns counterclockwise and record the reading at each turn. If the PTO alignment is perfect you'll get 2,000.0 and 3000.0 as your end readings and each full turn will change the frequency by exactly 100 kHz.
Since your PTO is probably not perfect, you will see your tracking error at each of the 10 points as a reading slightly greater or less than exact multiples of 100 kHz.
Reset to 2500.0, remove the knob and disc, and turn the spread coil (L002) shaft an eighth turn clockwise, reinstall the knob and shaft to align with the index wire at exactly 2500.0 kHz. Re-check your endpoints. If they are closer to 2000.0 and 3000.0 than before, you're going in the right direction, if they're farther off from exactly 10 turns equaling exactly 1000 kHz, turn the spread coil shaft a quarter turn counterclockwise, reset to 2500, and repeat the endpoint measurements. Seeing how much (what percentage) of the PTO tracking error an eighth turn in the spread coil shaft cancels will tell you how much the shaft needs to be turned to completely cancel the spread error. The response to adjustment of the spread coil is fairly linear. After each spread coil adjustment the knob and disk MUST be reset so that the wire aligns at exactly 2500.0 kHz. Repeat this sequence until exactly 10 turns gives exactly 1000 kHz frequency change.
It's very important to appreciate that turning the spread coil shaft changes BOTH the spread (kHz/turn) and the absolute frequency. This is why you must loosen the knob on the shaft and re-align the index mark on the disk with the wire at 2500.0 after each spread coil adjustment.
You should be able to get the ends (+/- exactly 5.00 turns) to track within +/- 2 kHz of the center of 2500.0. A good PTO will have range ends within +/-1 kHz of the centre. Starting from 2500, one endpoint may be slightly farther out than the other: you need to go for the best fit. After setting the 2000, 2500, and 3000 kHz points you may notice a small remaining error that varies in an uneven fashion as you move across the range. This can only be corrected by readjusting the stack of metal shims that the coil guide rollers run on in the PTO can: you really don't want to go there! In most PTOs this error is less than 1 kHz/100 kHz, and isn't worth worrying about.
There is one situation in which you might want to touch those shims: if the range is very asymmetric, for example 2000 and 2500 are dead on (5.00 turns), while 5.00 turns above 2500 gives you 3007 instead of 3000. In that case you would adjust the shims toward the rear of the PTO-replacing the can after each adjustment, and probably spend another 4 hours in tedious trial and error for best fit!
Instead of struggling with re-tightening the locking collar on the spread coil shaft, I put a tiny dab of RTV silicone sealant from the shaft to the wall: it's much easier to remove if you ever need to do this again. Leave the hex plug off until the sealant has cured-you don't want to seal acetic acid vapors into the PTO!
Reinstall the VFO set to 2500 kHz, leave the setscrews loose on the coupling, and set the MHz dial to 7.0 MHz, kHz dial on the radio to 0, and the fiducial line to the center of the window. Put the crystal calibrator on, set the (hopefully calibrated) BFO to center frequency, and carefully adjust the PTO shaft to give the zero beat with the fiducial line on 0. Tighten the coupling, check the kHz knob for smooth operation, and tune across the dial. The 100kHz cal points should now fall on the kHz dial 0, +/- the small remaining error you couldn't remove.After 50 years these PTOs are pretty well stabilized and seem hold their re-alignments very well. I have the measurements on an R-388 I did in 1983 and it's not changed at all since!
Courage! 73, Fred VE1FA
MORE INFO FROM FRED, VE1FA
There are three 0.01uF capacitors in the R-388's 70E-15 VFO, namely C-005, C-006, and C-008. Probably because of the very limited space, Collins used quite small, long, thin cylindrical brown plastic caps -with dimensions like the old wooden matches, and which frequently short internally! C005+6 are plate bypasses under the tube sockets, and C008, the osc feedback cap, is in the VFO can. Failure of any will make the VFO work erratically or usually not at all. If you have the VFO out, I'd replace all 3 as a matter of course. Incidentally, I've seen some 70E-15s where C005+6 are short tubular ceramics, and they're fine as is.
Don't be afraid to open the VFO can on the bench -if you're into such repairs. Just remove the three small screws and carefully pull the can off its O-ring. While you're there, carefully re-lube the slug lag screw and O-ring with silicone or moly grease-very sparingly. Since they will be pink and dead, you can also discard the dessicant packages you find. Healthy ones are blue, and the ones you find don't dehydrate (regenerate) properly in an oven (in my experience), so you're better off without them. If you're really keen, you can replace them with some sort of little porous cylinder of fresh silica gel taped to one of the rear bearing support posts. Don't use Drierite or other salts-based dessicants!!
Be sure to replace C008 with a high-quality cylindrical cap (0.01uF at 200V, polyester, polystyrene, or Mylar) and clamp it exactly where the original cap was. Try to get a replacement no larger than the original. Don't disturb any other wires or components in the can if you want to keep your VFO linearity.
Note the little coil with its adjustment behind the threaded cap: that's the "spread" adjustment to make each 360 deg.turn of the VFO shaft and dial exactly 100 kHz. Let me know if you need a procedure for setting the spread. Collins spec is that your VFO scale should not be out more than 3 kHz in the 1000 kHz range of the VFO. Using a good frequency meter and some patience to set the spread, you can usually do better than that. When you re-seal the VFO can, do it after letting the VFO equilibrate in a warm dry place with as dry air as possible-definitely not a damp basement! Placing the VFO over a 40W trouble light until it's very warm (but not too hot to touch) prior to closing the can is one way.
Another common 51J VFO problem is that the output coax center conductor fatigues and breaks where it's soldered to the tube V-002 or mixer socket pin. An easy fix, but impossible to see in the radio if the break is under V002! Also, confirm that the VR tube (0A2) is glowing violet.
Finally, there is a simple little modification you may wish to add. Looking at the schematic you will see that there is a 4kohm dropping resistor on the plate of the OA2 regulator tube, just where the VFO picks up its B+. VR tubes are noisy. If you put a 0.1 uF (at >250V) cap from ground to the plate of the 0A2, with the radio on, RF gain at 10, and no signal in, you will hear a distinct decrease in background noise when the cap is in the circuit. You might think that C005+6 in the VFO would clean up the B+ sufficiently, (Collins obviously did) but that doesn't seem to be the case! The 0.1 uF cap is easily mounted beside the 4kohm dropping resistor (R181). Some individual VR tubes are worse than others, but you should hear a reduction in background noise in all cases. While you're there, put a voltmeter, or a 'scope from the VR plate to ground, and be sure that; (A) the DC voltage is 150 +/- a couple of volts; and (B) that the voltage is steady for several minutes. If it's not, check R-181 and replace the VR tube.
73 Fred VE1FA
In my opinion, correcting tracking errors using the shims is not all that difficult if you keep the following in mind:
1. Place 4 index marks on the disk so you can check tracking every quarter turn (25 kHz) Take care to place these accurately.
2. You don't need to reinstall the can completely each time you make an adjustment. You have to put it in place, but you don't need to fasten it. Watch the counter to confirm this. It's probably best to remove the O-ring, which you'll probably want to replace anyway.
3. By all means, don,t start this if you think you might be interrupted.
4. Don't loosen the clamping screw more than necessary to let you move the shims.
5. Start at one end. With the can in place, read the counter. Remove the can, adjust only the shims which are under the roller, replace the can to check. Repeat until you're within 200 Hz or so, then turn thr shaft 1/4 turn and repeat, remembering that the frequency should now be 25 kHz higher (or lower). Continue until you've done all 41 points. Don't forget to tighten the clamp screw when you're done.
Once you've done a few points, it goes surprisingly fast. It took me less than an hour, and the total tracking error has been 300 Hz or less ever since.
If you're replacing the O-ring, use Neoprene or other material which is compatible with petroleum products. These rings are available at bearing shops for a few bucks. Grease it lightly before assembly.
73 Bob, VE3TOU
Return to Philip's Incredible Universe
This page restored March 3, 2007. I apologize for the downtime. Collection by Philip Neidlinger