75A-4 AVC

[KB5MD]

I have been pulling my hair out trying to find the problem with the AVC in the 75a-4.  It works fine when the set is first turned on but after a few minutes it quits along with the s-meter operation.  I think it must be something getting hot and opening up, but for the life of me, I can't seem to find it.  Anyone else had this problem?  I have checked all resistors and pots associated with the AVC circuit and everything checks okay.....GGGRRRRRRRRRRRRRRR!!!

[Todd, KA1KAQ]

Have you dealt with the condensers? This is a know issue with the A-4, tubulars to silver micas. Could be something as simple as a bad solder joint too, but I'd start with the condensers and go from there. Check the ones common to the AVC line and S-meter circuit, that should narrow it down.

If you google Collins 75A-4 AVC you'll find a few hits that address it.

[WA5VGO]

Almost certainly a bad silver mica is pulling it down. If they are "Sangamo" capacitors, you can bet on it. Try removing the AVC from the stages one at a time until you find the stage that's dragging it down.

Good Luck,
Darrell, WA5VGO

[w3jn]

The 75A-4 has amplified AVC and if the 455 KC AVC amp stage goes south, so does the AVC.

Check voltages associated with V21.  I'm betting you have a leaky coupling cap to the grid of that tube, or a resistor is getting hot and drifting way up in value.

[KM1H]

If a resistor is heating up something else is causing it. Id put my money on any silver mica that has a voltage across it and not just a RF signal.

Carl

[w3jn]

The afore-mentioned couplking cap does indeed have B+ on one end.

[k4kyv]

If you have a good VTVM or DVM, check voltages on each stage per the voltage chart in the book before going to the trouble of ripping out caps to measure leakage.  This will quickly lead to the bad component, probably one of those little  red silver mica coupling caps.  Those things are notorious for becoming leaky with age.  I think I have replaced over a half dozen in my two A4s combined.  They tend to work ok when the set is first turned on, then gradually the receiver becomes insensitive, the S-meter indication goes off scale, the audio becomes distorted and other performance degrades.  The give-away is the bias voltage at the following stage.  It will be way off, sometimes even positive as the leaky coupling cap conducts more and more B+ voltage to the grid.  I specifically recall problems occurring in the i.f. stages and in the Q multiplier (notch filter) stage. I also had one to go flaky in the crystal oscillator stage; it was the resonating cap across the 20m oscillator coil. Ordering a replacement crystal didn't help, but after replacing the cap both xtals worked. I kept the new one in, since the frequency was closer to nominal than with the older one and I now have a spare.

The cap may not check leaky with a normal ohmmeter or capacitor tester. Apparently the insulation gradually breaks down after high voltage is applied. I have an old Heathkit capacitor checker that applies up to +150v DC to the cap for the leakage test.  Leaving the cap in the test mode for a half hour or so will allow the breakdown to show up.  I just replaced the ones that were actually causing the bias to shift and left the others alone. I must have eventually found all the bad ones since I haven't had a recurrence of the problem in years.  The way the A4 is constructed, with the deep chassis and crowded layers of components wired over other components, makes it very difficult to unsolder and re-solder a resistor or cap without melting or burning the thermo-plastic insulation on the wiring to adjacent components.

[WA5VGO]

The cap may not check leaky with a normal ohmmeter or capacitor tester. Apparently the insulation gradually breaks down after high voltage is applied. I have an old Heathkit capacitor checker that applies up to +150v DC to the cap for the leakage test.  Leaving the cap in the test mode for a half hour or so will allow the breakdown to show up.

This is interesting. I have always been puzzled by the fact that I could never detect any leakage. I've even checked them on meggers, and they always checked good. It seemed that the only way to really find them was by substitution. Next time I'll apply voltage and leave it on for a while.

I specifically recall problems occurring in the i.f. stages and in the Q multiplier (notch filter) stage. I also had one to go flaky in the crystal oscillator stage; it was the resonating cap across the 20m oscillator coil. Ordering a replacement crystal didn't help, but after replacing the cap both xtals worked.

These seem to be major trouble spots. I have two 75A4's and both of them have also had these exact failures.

73,
Darrell, WA5VGO

[Todd, KA1KAQ]

I need to go through my front line A-4, it has similar issues. I've had to go in and adjust the crystal oscillator trimmer a few times to get the thing to receive. Wrote it off to an aging crystal the first few times, but it started to happen on multiple bands and 'PJP, 'JN and a few others pointed to the same nasty little caps.

[W1UJR]

First, nice receiver!
Second, wise move to ask rather than dive right in, the experience pool here is amazing.
I've so many sets "improved upon" that it makes me sick.
http://www.w1ujr.net/collins_75a4.htm

Right off the venerable AM Window site, here is the list of the "7 Deadly Caps", which, as the fellows sagely suggested, is an excellent starting point.
http://www.amwindow.org/tech/htm/deadly.htm

Enjoy, and Happy New Year all!

-Bruce

[w3jn]

The cap may not check leaky with a normal ohmmeter or capacitor tester. Apparently the insulation gradually breaks down after high voltage is applied. I have an old Heathkit capacitor checker that applies up to +150v DC to the cap for the leakage test.  Leaving the cap in the test mode for a half hour or so will allow the breakdown to show up.

This is interesting. I have always been puzzled by the fact that I could never detect any leakage. I've even checked them on meggers, and they always checked good. It seemed that the only way to really find them was by substitution. Next time I'll apply voltage and leave it on for a while.

I specifically recall problems occurring in the i.f. stages and in the Q multiplier (notch filter) stage. I also had one to go flaky in the crystal oscillator stage; it was the resonating cap across the 20m oscillator coil. Ordering a replacement crystal didn't help, but after replacing the cap both xtals worked.

These seem to be major trouble spots. I have two 75A4's and both of them have also had these exact failures.

73,
Darrell, WA5VGO

Better is using a DVOM or VTVM to check voltages while the circuit is misbehaving.  In a receiver, if a grid is EVER positive with respect to the cathode there's a problem.

However even a VTVM or DVM can load down a leaky cap, so when you're checking and the problem mysteriously changes radically or goes away when you attach your leads, you still need to suspect that cap.

I've even come across mica caps that acted as small batteries - they'd show a couple volts on a DVM even when completely removed from the radio.

In any event it's not a bad idea to prophylactically replace the 7 deadly caps.

[KA2DZT]

W3JN wrote,  "prophylactically replace the 7 deadly caps"

I thought this was a "G" rated family show.

Fred

[W3SLK]

John said

However even a VTVM or DVM can load down a leaky cap, so when you're checking and the problem mysteriously changes radically or goes away when you attach your leads, you still need to suspect that cap.

John brings up a good point with regards to using meters in different circuits. One the oldest but somewhat forgotten is the use of heat and cold to weed out suspect components. Not nearly as good as 'Blue Shower', some contact cleaners provide enough thermal differential to make them useful. And shot with a heat gun works just as well also.

[KB5MD]

I found several mica caps that were leaking along with a couple of bad disc ceramic caps, replaced them and now it's down to finding the loose connection somewhere in the AVC circuit.  It works great for a while then quits, you can bump the receiver and it comes back up.  So, I get to chase the intermittents some more.

[WD8BIL]

Don't forget to check tubes and sockets!

[Todd, KA1KAQ]

Sounds like you're well on your way. Intermittents are, without a doubt, the most annoying part of troubleshooting. After you've checked solder joints and other obvious points, pull the tubes, clean the pins with some DeOxit, and reinstall/remove them from the sockets a few times to clear any crud off the socket connections.

Then you can start going through the grounds, checking terminals and screw connections. I've found that sometimes it takes going through and loosening then re-tightening any through-chassis connections with the little star washers to get them to bite through the years of oxidation and make a good connection. Tedious for sure, but when all else fails it might be the ticket.

W3JN wrote,  "prophylactically replace the 7 deadly caps"

I thought this was a "G" rated family show.

Fred

It is, Fred. That's why he didn't say "RIP THOSE %$&*@!! THINGS OUT!". 

[KM1H]

When playing with BA's it helps to get a real capacitor tester and not some cheap kit not up to the job.

Sprague made some real nice ones from the 1939 Model 16 to the 60's TC-6A. I have both and they all serve a purpose of testing at rated voltage; the Model 16 goes up to 1000V.

The military ZM-11 is another good performer and is modern enough to not need caps replaced before using.

The Model 16 required all caps and many resistors to be replaced, the good thing is now its easy to get well under 1% tolerance of the design value and the measurement results are right on the money. The 1941 and immediate postwar version was renamed the TC-3.

Carl

[W3SLK]

I swear by my Eico Capacitance-Resistance bridge.

[KM1H]

Most people I know swear AT Eico 

[w3jn]

HP 4261A digital RLC bridge, best $60 I spent on test equipment

Those hammy hambone digital C meters are crap, they give wrong readings if there's the slightest amount of leakage or ESR.  The old sk00l RLC bridges are great though, just a little harder to use than the HP.

[Todd, KA1KAQ]

Can't remember the model number of that massive gray GenRad bridge we had in the lab in co-lage,  but it sure was impressive (and intimidating). Looked like it belonged in a 707 cockpit. It gave me my first exposure and appreciation for GenRad equipment.

That HP looks like the bee's knees.  

[w8khk]

Can't remember the model number of that massive gray GenRad bridge we had in the lab in co-lage,  but it sure was impressive (and intimidating). Looked like it belonged in a 707 cockpit. It gave me my first exposure and appreciation for GenRad equipment.

Maybe it was the GR 1650-B?  This one is great for measuring most of the 75A caps in-circuit...