Trouble at Step 38 and on...

[Archmart]

Hey Hey!

Please help!

I've almost finished a J99 Rev. 3.3 with 990Cs from John Hardy.

(I previously successully completed 2 A12s and 2 C84s without much trouble... well, one bad SC25, but there's a new one on the way)

All tests out ok up through step 37.

At Step 38, I have trouble.

Initially, I measured U2, between pin 4 and 6, and it responded appropriately. Then I did the same on U1, for good measure. It stayed around 20 millivolts. I thought maybe since it wasn't actually specified, I was ok to move on, thinking perhaps U1 testing wasn't necessary as it wasn't specified.

I proceeded to try the "To adjust the input bias current" instructions in the Sticky. I shorted pin 2 to pin 3 at the xlr input. I switched the jumpers to the ADJ position.

My first reading was -27mV with switch 1 fully counter-clockwise. My second was -350mV at fully clockwise. I turned the R45 until I got near -27mV. Then I remeasured and found it way off. (it becomes a little blurry here) I think it was near zero and I tweaked it to up around -27 again. Actually, I think it wouldn't go that far. I went back to fully clockwise and it was back up in the 350 range. I thought I'd try to tweak it to split the difference and took it to 150ish. When I checked at counter-clockwise I was still down in the low levels and then I kept tweaking back and forth never getting them anywhere closer to each other.

I -think- I then tried the same things on U2. I think it acted the same.

I -think- the next thing I did was power down and swap the two 990Cs. I don't 'think' it was any different.

I started over and got the same kinds of results.

At some point I swapped in another of my 4 990Cs.

At some point the measurements started showing crazy numbers like 17 volts and 20 volts and the two 990Cs were getting really hot. (The output transformer also gets pretty hot)

(I don't know what an appropriate level of heat is on these components, but it was a little too hot to touch and remove without using a napkin.

Right now here are my readings as I go back and test...

I'm still ok up through step 37.

Step 38... (with the jumpers NOT in adjust mode)

With 990C (A) in U1 and 990C (B) in U2:

U2 (990C (B)) Pin 4 to pin 6: 15.86 volts

U1 (990C (A)) Pin 4 to pin 6: 19.28 volts

With 990C (B) in U1 and 990C (A) in U2:

U2 (990C (A)) Pin 4 to pin 6: 16.55 volts

U1 (990C (B)) Pin 4 to pin 6: 20mV, dropping over a couple of minutes past zero to -3mV, then coming back to around -2mV, then proceeding again to -15mV (or so... it keeps moving very slowly).

With 990C (B) in U1 and 990C (C) in U2:

U2 (990C (C)) Pin 4 to pin 6: Slowly rising to 225mV and dancing up and down from there.

U1 (990C (B)) Pin 4 to pin 6: 23mV

I suspect this is enough info for now. Obviously something's wrong, maybe even more than one thing.

Please help me troubleshoot this and get it working. I'm going to hold off on starting the next one until I find out what I've done wrong.

Thanks,
Archmart

[Archmart]

Anyone?

[tpryan]

It sounds like you may have some faulty 990Cs. Set the regulators for +/-18V. Set the jumper at J6 to ADJ. Install an op-amp at U1 only and apply power. If the output voltage (pin 6) goes to some large value (usually close to either the positive or negative supply voltage) and stays there, you have a bad part. Test each 990C and let us know what you find.

[Archmart]

Hey Hey!

Thanks.

I'll check it out tonight.

So... I'm rolling R28 and R32 (or the equivalent on this revision - I'm just going by the online instructions pdf right now) back from +/-24v to +/-18V, putting the J6 jumper at ADJ (the two pins closest to ADJ), installing one 990C at a time in U1, and measuring between U1 pins 4 and 6. This SHOULD give me the "dropping below 5mV" scenario and if it doesn't it'll probably be at something around the +/-18 or 24, which would mean it's a bad 990C. Is that right?

And am I in danger of ruining a good 990C in this process? I've only plugged 3 of my 4 in so far and I'm wondering if I should be careful with my last untouched one. If I've ruined them, I'm out $50.00 each...

Thanks for the help,
Archmart

[Archmart]

Hey Hey!

Ok...

990C (A) gives me 13.5V.

990C (B) gives me -25mV initially and slowly drops a couple of mV to .23mV.

990C (C) gives me -26mV.

990C (D) gives me -22mV.

All in U1, of course.

So this would lead me to think that there's something way wrong with 990C (A) and either something minimally wrong with the other three or something wrong in my circuit.

And as I noted, I also had trouble when using 990C (B) and 990C (C).

So is there possibly also something wrong with my U2 circuit?

Please advise.

Thanks,
Archmart

[tpryan]

990C-A is bad, the other three should be fine. Try proceeding with the bias compensation adjustment with each op-amp separately. That is, install a good 990C at U1, jumper J6 to ADJ, and adjust R45 according to the procedure in the sticky. Then, power down and remove the 990C at U1. Install a second 990C at U2, jumper J8 to ADJ, and adjust R50. Reinstall U1, jumper J6 and J8 for normal operation, and apply power. If the offset voltages settle at 0V, you're done.

[Archmart]

Hey Hey!

Great. Thanks.

So you don't think the -22 - -26mV readings are bad...

I'll follow these directions tonight and let you know.

Thanks,
Archmart

[Archmart]

Hey Hey!

Ok, so I reset the voltages to +/- 24V.

Then I plugged in a good 990C (B) in U1 and set the jumper to ADJ.

My SW2 CCW measurement was -25.6mV.

My SW2 CW measurement was -335mV.

I turned R45 until I measured very close to -25.6mV.

Then I turned SW2 back to CCW: I measured somewhere around -1mV.

Then I turned R45 up to -25.6mV.

SW2 back to CW: I measured -326mV.

Back to CCW: +3mV... Raise it up to -25.6mV (approx)

Back to CW: -316mV. Lower it to around -25.6mV.

Back to CCW: +1mV... Raise it to -25.6mV.

Back to CW: 335mV.

Now frustrated again I power down and take the 990C out of U1.


Then I plugged in a good 990C (C) in U2 and set the jumper to ADJ.

My SW2 CCW measurement was -144mV.

My SW2 CW measurement was -1.29V.

I turn R50 and the measurement jumps to 16.23V

SW2 back to CCW... still 16.23V.

R50 makes no difference.

So now do I have 2 bad 990Cs?

Help?

Thanks,
Archmart

[tpryan]

If you test the suspect 990C at U1 do you get similar behavior? If so, it's probably bad.

Remember that the whole idea of the bias compensation circuit is to make the servo's job easier. If you can't get the voltages any closer than 200mV, or whatever, it's not necessarily a problem. If the DC voltages settle at 0V when you move the jumpers at J6 and J8 to their normal positions, that's all there is to it.

[Archmart]

Hey Hey!

I'll try the putting the 990C (C) in U1 and see what happens first thing tonight. But then it's the weekend so I wonder if you could give tell me what to do in a couple of scenarios that may result.

If it's bad in both U1 and U2, then it's bad.

But it was good in U1 before, at +/- 18v and good in U1 at +/-24, so if it then shows bad again in U1... I'm thinking that each time I put one in U2 it goes bad.

So, if it shows good in U1, then I've got something wrong with U2, so what should I check next?

Or, if it shows bad in U1, then something in U2 is blowing my 990Cs (at 50 bucks a pop), so what should I check next, before destroying more 990Cs?

Thanks,
Archmart

[tpryan]

If you're worried about damaging 990Cs in U2, lift one end of L2. You don't have to remove it from the board, just disconnect one end to break the circuit. That will let the output of U2 swing wildly without drawing excessive current. If an op-amp behaves in U1 but not in U2, check the voltages at U2, pins 1, 2, and 6 and let us know what you find.

[Archmart]

Hey Hey!

Lifting L2... That's with a 990C in U2 so I can do the bias test on U1? I think I'm confused here.

And Checking the voltages at U2, pins 1, 2, and 6. That's without a 990C in either socket, yes?

Thanks,
Archmart

[tpryan]

Disconnecting L2 allows you to screw around with both bias adjustments without worrying about damaging either op-amp.

If an op-amp behaves in U1 but not in U2, measure the voltages at U2 pins 1, 2, and 6 with the op-amp in the circuit and let us know what you find.

[Archmart]

Hey Hey!

Thanks for your continued help, Tim.

I'm about to proceed with this advice.

Let's say I test and still have some trouble, but want to get some more work done this weekend. Based on measurements so far, would you say it's likely that I've done something wrong and should postpone starting my second J99 until we find the problem or is it more likely I just have bad components so I could reasonably safely proceed with the other one?

Thanks,
Archmart'

[Archmart]

Hey Hey!

Both 990C (A) and (C) had the way out (17, 18 volts) measurements in both U1 and U2.

I tested (B) and (D) and both were good in both slots so I went ahead with adjusting the bias. I got within .1mV on one of them and within 1mV on the other.

I un-shorted pins 2 and 3 of connector 1, resoldered L2, and powered it up.

U1, pins 4 to 6 measures -5.5mV.

U2, pins 4 to 6 measures 4.3mV.

(these numbers are moving a little bit, settling up and down a tenth of a mV per second or two)

... Ok, so I went ahead and installed it in the rack.

It sounds great. Really great. Big, full, clean.

I dunno... are my numbers close enough? You've referred to zero, but also "less than 5mV". Is it an issue that one is positive near 5 and the other is negative near 5?

Well, at least as far as I can tell I can go ahead assuming I was doing it right. I guess I've got to have the difficult conversation with John (Hardy) about whether something I did might have blown the two 990Cs or whether they were just bad.

Thanks for you help Tim! I'm sure glad I didn't have to go the whole weekend wondering what the problem was.

Thanks,
Archmart