Maker Pro
Maker Pro

Heathkit AR-15 receiver - problem with multiplex circuit

haskellbob

Apr 13, 2020
21
Joined
Apr 13, 2020
Messages
21
Hello.
I have a Heathkit AR-15 receiver which has a procedure for aligning the multiplex circuit, written in the manual. I follow this procedure, the last step of which is to adjust a choke until the sound goes from garbled to clear. I have done this several times. Each time the FM works fine for a few days, and then it gets garbled again. So I adjust the "garble removing" choke until the sound is clear and... after a few days, the same thing happens. I think there is some kind of a drift?
I thought maybe the core of the choke was physically moving (it seems a far-fetched idea but it was all I could think of) and, since I had another choke to replace the original one if things went badly, I inserted a drop of glue in the choke (having adjusted it first). Well, guess what?
It's a good thing I had another choke to swap in. And I adjusted that one, and after a few days... you guessed it!
I have to think the problem is not with the coil at all, but with some other component that is drifting or changing its value over time.

I have the schematic, although I'm not finding any way to attach it. I will if I can.

Does anyone have any ideas about what could be causing this? I sure would appreciate being pointed in the direction of the source of the problem. Thanks in advance.

Bob
 

Ylli

Jun 19, 2018
401
Joined
Jun 19, 2018
Messages
401
I don't have a full manual, what 'choke' are you referring to?
Annotation 2020-04-12 210946.png
 

PETERDECO

Dec 19, 2019
239
Joined
Dec 19, 2019
Messages
239
Replace the capacitors associated with it, especially the 2700pF. One of the diodes following it might also be the culprit. I went through this years ago with a TV tuner, the old clicker type, and it was a disc cap that was drifting.
 

haskellbob

Apr 13, 2020
21
Joined
Apr 13, 2020
Messages
21
Thanks for the help, and the suggestions. I will let you know how it turns out.
 

Ylli

Jun 19, 2018
401
Joined
Jun 19, 2018
Messages
401
This is a somewhat unusual MPX scheme. More common is a scheme where the 19 KHz pilot is filtered out, amplifier, frequency double, and the used to dem-mux the audio, in this case that regenerated 19 KHz is doubled and used to injection lock a 38 KHz oscillator.

Composite audio comes in to the board on terminal 10 and is fed through C444 into the resonant circuit of L409/C445 to Q410 . The 19 KHz component is amplified by Q410 and is fed through a network that allows some phase adjustment to Q411, where it is further amplified and can be seen across resonant transformer T402. The secondary of T402 has a diode doubler connected to it that doubles the 19 Khz to 38 Khz. The resulting 38Khz (DC component available at TP2) is fed to the stereo detect circuitry (Q417 et al) and to the base of a 38 KHz oscillator Q405. The frequency determining tank for Q405 is C419 & T401.

The oscillator lock range appears to be seriously limited.

Tuned to a strong FM Stereo station, use the built in or an external meter to make sure the voltage at TP2 is within spec. Again, I don't have a full manual, just the schematic so I can not tell you what 'spec' is. Is the 'Stereo' indicator light responding properly? If so, the voltage at TP2 is probably OK, if not, then we need to look prior to this point. If an oscilloscope is available, it would be diagnostic to look at the signal at TP2.

If the signal at TP2 is good, then for some reason the oscillator itself is not locking. C419 could indeed be the culprit, or unlikely but possible a bad diode in the ring demodulator.
 

haskellbob

Apr 13, 2020
21
Joined
Apr 13, 2020
Messages
21
Thanks for all of that analysis. I can't say I understand it all, but I get the basic gist of it. The stereo indicator light works fine. I replaced C419 today and adjusted T401. It worked for a while, but it just now went garbled again. I will check the diodes in the ring demodulator next, but I will wait a bit. It's a chore to open that unit up and get at the components...

I'll see what I can find out about the voltage at TP2 and whether I can check it.

Just to be sure, the diodes in the ring demodulator are D 423, D424, D425 and 426, right?

Since your own experience with this came down to a disc capacitor, I wonder if my problem is one of those too. Which capacitors are the most likely culprits? Are they the ones between each pair of diodes?

I'm curious what it means when an oscillator "locks". For that matter, I don't know what an oscillator is and does, or what exactly oscillates. But I don't want to tax your patience. You can go into that as much or as little as you want. I'm actually very green and just desperately trying to get a solid AR-15 to work for me; my dad had one and I have always wanted one myself.

OK. Thanks for the help. When my reservoir of patience fills back up I'll go back inside the patient to see what I can find.

- Bob
 
Last edited:

Ylli

Jun 19, 2018
401
Joined
Jun 19, 2018
Messages
401
I would think measuring the voltage at TP2 would be part of the alignment instructions. Look for the section that talks about adjusting L409 and/or T402.

Correct on those diodes.

The 2700 pF cap failure was something that Peterdeco saw, not me. (I've never seen the insides of an AR-15, although I do have, and have worked on, an AR-1500A and numerous other Heathkits)

Can't imagine any of those 470 pF caps in the diode ring would go bad, and even if they did should not cause your problem. In general, ceramic discs do not have a very high failure rate. The 2700 pF is the only cap in the oscillator circuit where a small drift in value would make any difference. That's not to say that your problem might not be caused by a bad disk.

In electronics, an oscillator is a device or circuit that takes a DC input and creates an AC output. In this case, Q405 and associated components form an oscillator that outputs a 38 KHz AC signal. That signal is developed across T401 and fed to the ring demodulator. This application requires that the oscillator output be extremely accurate in both frequency and phase. A simple circuit like this is no where near accurate enough. So an external 38 KHz, at a relatively low level, derived from the received signal, is injected into the oscillator and forces the oscillator to 'lock' to that reference signal.

See if you can determine if the TP2 voltage is within spec. Again, this will be a section that uses the internal meter to adjust L409 and/or T402.

When you need to readjust T401 to get rid of the 'garbles', do you always need to adjust it in the same direction, or do you sometimes need to adjust it CW and other times CCW?
 

haskellbob

Apr 13, 2020
21
Joined
Apr 13, 2020
Messages
21
There is a section about tests to make before finishing the kit, which use the signal meter to measure voltages at various points, but none of those points is TP2. I do think performing those tests might help locate the problem. Could it not be one of the transistors?

TP2 is used in the multiplex adjustment sequence. They have you put a jumper between TP2 and the 12V pin on coil L409, and then check to see if the sound is garbled. But I can't find a voltage specified for TP2.

I replaced the .001 capacitor C413, just prior to the oscillator transistor, Q405, since the Troubleshooting section mentioned that having that capacitor "open" might be the cause of "garbling" (it's their word). But no dice. The same thing happened; it worked fine for a short while and then re-garbled. Very discouraging!

I'm not sure if I always have to adjust in the same direction. I will pay attention to that detail. I THINK it's always counter-clockwise, but I'm not completely sure.

The four diodes all check out fine, so I don't know where to go next.

I'd welcome any more suggestions.

Thanks!
 

Attachments

  • multiplex voltage check.pdf
    785.1 KB · Views: 2
  • multiplex adjustments 1.pdf
    895.3 KB · Views: 1
  • multiplex adjustments 2.pdf
    901.1 KB · Views: 1

duke37

Jan 9, 2011
5,364
Joined
Jan 9, 2011
Messages
5,364
I cannot pretend to understand the circuit but a locked oscittator will need a good signal to lock it and if the signal has insufficient amplitude or wrong impedance then lock will be lost. Obviously any drift should be investigated.
 

Ylli

Jun 19, 2018
401
Joined
Jun 19, 2018
Messages
401
The section of the adjustments you posted in ..1.pdf and ..2.pdf are the right adjustments. Have you gone through them in full? I don't quite understand why they have you detune L409, except perhaps to reduce the signal level in the following stages to allow better adjustments - but then I would expect them to have you readjust L409 back to peak, which I do not see in the instructions.

Try this: Put a voltmeter between board terminal 3 (black meter lead) and terminal 4 (red meter lead). Tune to a strong FM stereo station and note (and post) the voltage at that point. Adjust T402 for a maximum voltage indicated. Post results. I'm hesitant to have you adjust L409 at this point as I don't know why they are having you detune it in the adjustment instructions, so leave it alone for now. Let me know the results.
Annotation 2020-04-14 113523.png
 
Last edited:

haskellbob

Apr 13, 2020
21
Joined
Apr 13, 2020
Messages
21
Update:
I noticed this morning that the receiver had moved back into the "range" where the multiplex(ing) was working correctly. This had not happened before. So I went back and pointedly went through all of the alignment steps - more carefully than I did yesterday.

I'm hoping that the bit of drift that garbled it yesterday, but which was undone today, has been compensated for by centering the alignment in more precisely. Wishful thinking? I'm afraid that might be what it is, but time will tell.

The voltage I'm getting across points 3 and 4 centers at around .65; it ranges on different strong stations between 0.65 and 0.7 VDC,

Does that seem about right?
 

Attachments

  • p. 132.pdf
    910.5 KB · Views: 2

Ylli

Jun 19, 2018
401
Joined
Jun 19, 2018
Messages
401
The voltage I'm getting across points 3 and 4 centers at around .65; it ranges on different strong stations between 0.65 and 0.7 VDC,

Does that seem about right?
No, that seems quite low. Is that with all the switches back in their 'normal' positions?
I assume you had no problem with the adjustment instructions - that things responded as expected?

For thoroughness, let's measure the voltages on the base, emitter and collector of Q410 and Q411 and be sure they are close to what is listed on the schematic.
 

haskellbob

Apr 13, 2020
21
Joined
Apr 13, 2020
Messages
21
I'll do that as soon as:
a) the sound gets garbled again, or
b) a couple of days have gone by with no garble.

I'll let you know the values I get when I do it.
Thanks.
 

haskellbob

Apr 13, 2020
21
Joined
Apr 13, 2020
Messages
21
Hi. The receiver is still ungarbled... I never answered your question: yes, everything went as expected in the alignment.

Also, I re-measured the voltages at Points 3 (11.97) and 4 (12.72). I don't know what changed or what mistake I may have made in measuring before. What are their ideal values? I don't know how to determine that.

Here are the voltages of the two transistors:

Q410
E - 4.5
B - 4.99
C - 8.43

Q411
E - 1.25
B - 1.9
C - 6.96

Those seem to be fairly close to spec, don't they? I don't know what range of variation to use in evaluating them, but they're at least in the ballpark.

I've got my fingers crossed.

Thanks!

Bob
 

Ylli

Jun 19, 2018
401
Joined
Jun 19, 2018
Messages
401
Also, I re-measured the voltages at Points 3 (11.97) and 4 (12.72). I don't know what changed or what mistake I may have made in measuring before. What are their ideal values? I don't know how to determine that.
I had asked you to measure between them not each to ground. Your new numbers makes 0.75 volts between them is is basically the same as you measured earlier. I would have expect3d at least a couple of volts, but that was only a guess. what you see may be normal.
Here are the voltages of the two transistors:

Q410
E - 4.5
B - 4.99
C - 8.43

Q411
E - 1.25
B - 1.9
C - 6.96

Those seem to be fairly close to spec, don't they?
They are reasonable. The only one that grabs my attention is the B-E voltage of Q410 is only 0.49 volts, that is really not enough to get it into it's linear operating range. But the collector voltage indicate near normal current draw.

So no good answers for you at this point. Would be interesting to trace the 19 KHz through with a scope.
 

haskellbob

Apr 13, 2020
21
Joined
Apr 13, 2020
Messages
21
It's still producing clear sound. Sorry I didn't notice that you'd asked for the voltage between them, but it got measured anyway!
Point of order: The base of Q410 is 4.99 volts. So 5 volts. The schematic says 4.5.

I still have my fingers crossed. I'm hoping the replacement of capacitor C413 has corrected the problem. By the way, can you explain what an "open" capacitor would mean? The problem listed in the Troubleshooting section was "Capacitor C413 open". I think that means nothing gets across, that it's like open air. The opposite of a short. Right? So if the old one was indeed open, and the new one is working, then something is reaching the base of the 38 khz oscillator Q405 that wasn't getting across before. But what would the purpose of that signal or voltage be?

Thanks again. I wish I had a scope. And a signal injector. I'd like to learn how to inject a signal and trace it with a scope, but that will have to be post-coronavirus, if ever.
 

Ylli

Jun 19, 2018
401
Joined
Jun 19, 2018
Messages
401
Point of order: The base of Q410 is 4.99 volts. So 5 volts. The schematic says 4.5.
Issue is the difference between the base and emitter voltage of the transistor. The B-E junction looks a bit like a diode, so there will be no base current until the B-E voltage reaches the 'threshold' of a silicon junction. That is typically 0.6 volts. So theoretically with a B-E voltage of only 0.49 volts, the transistor should be turned completely off. If it is off, it can't amplify or even pass a signal. But the knee is 'soft', and that turn on voltage can vary between about 0.5 volts and 0.7 volts depending on the purpose of the circuit.

If this transistor is indeed 'off' and not properly passing the 19 KHz signal, it could cause low signal levels in the following stages and reduce the stability of the oscillators lock. The collector current draw - that yields a near spec'd collector voltage - could be transistor leakage. But since your meter readings during the alignment procedure were as expected, I don't think Q410 is an issue. But it still would be nice to be able to look at it with a 'scope.
I still have my fingers crossed. I'm hoping the replacement of capacitor C413 has corrected the problem. By the way, can you explain what an "open" capacitor would mean? The problem listed in the Troubleshooting section was "Capacitor C413 open". I think that means nothing gets across, that it's like open air. The opposite of a short. Right?
Right. A capacitor in general passes an AC signal and blocks DC. If it is open, that means it passes nothing. Take a new capacitor and cut off one of the leads - now that capacitor is 'open'.
So if the old one was indeed open, and the new one is working, then something is reaching the base of the 38 khz oscillator Q405 that wasn't getting across before. But what would the purpose of that signal or voltage be?
The signal flowing though C413 is a 38KHz signal that 'locks' the oscillator. An oscillator is like a pendulum, and the signal coming through C413 gives that pendulum a push each cycle and forces it to remain in sync with the pushing. I think C413 is a disc ceramic capacitor, and those do not normally fail open - they can, but it is not common.
 

haskellbob

Apr 13, 2020
21
Joined
Apr 13, 2020
Messages
21
Hi. I have an update. The culprit turned out to be a .001 mfd disc cap. Interesting coincidence with your experience. The manual, in the "Troubleshooting" section, mentioned this as a possible cause. I have heard it's rare for those capacitors to fail, so I wonder what caused it? Maybe its being over 50 years old, sitting in somebody's garage? I don't know its history; I got it on Ebay. But thanks for your help!
 

Attachments

  • culprit.png
    culprit.png
    245.5 KB · Views: 6
Top