Sunn Amplifier Power Amp Puzzle

[crutschow]

i found the IC1 1/2 implemented rising/falling signal/slope detector curious as the second time i see such ... as is **broken link removed** is also and this one here frequency band oriented but here one only seems to work at signal going away from SG e.g. ± 1-st & 3-rd quarter-cycle (i don't want to get down to philosophy of such) - looks like something implemented fast* without much thought to it . . . ?? is such* and the rest of the amp . . .

Sorry but I have no idea what you are trying to say.
Please use complete sentences with punctuation.

 

[ci139]

complete sentences with punctuation

(prev. msg.) is not a definition of the new mathematical hypothesis (where it is more likely to matter) -- such should be auto detectable , customizable by the readers like.

in post #20 one can omit the lower part - while using simulator-X "you" can "flow" a variety of (wave forms from the virtual) function gen. to the amp. inp. and study it's response/ possible shortcomings

 

[crutschow]

(prev. msg.) is not a definition of the new mathematical hypothesis (where it is more likely to matter) -- such should be auto detectable , customizable by the readers like.

That is completely unintelligible to me.
What language are you speaking?

 

[ci139]

(don't trash this thread) estonian(mother),finnish,russian,english(/minor lat.(verb roots) i manage to understand the written texts - so far)

 

[Bud_J]

Then it seems we await the arrival of your Variac.

Well, to my pleasant surprise, a large, heavy package arrived at my door this a.m. well ahead of schedule. I tested it out and, while the dial and meter readouts seem a little off at the far ends of the spectrum, they're accurate in the 115vac range and don't go too far out of sync at the 50-60vac range. Perhaps under load the number will be right on. No hum or buzzing, no screeches when I turn the dial, not loose parts rolling around when I tilt the device. All-in-all I seem to have done pretty well for an off-brand device. I'll keep my fingers crossed.

I got the boards back together in the amp chassis, all parts back in place. Bad transistor replaced (again). Before charging ahead, which has thus far not served me too well, I figure I'd post my plan here to see if I'm on the right track.

Based on our conversation so far, I'm thinking I should place my meter's probe at some safe place along the output of pin 7 of IC1 -- perhaps where it connects with R22 as there seems to be a good "grabbing spot" there. Then, with no speaker connected, slowly bring the variac up as I watch for any deviation from 0vdc.

Yes?

If so, how I should I bring up the line voltage? I imagine if it's too low, then things won't "turn on" enough for a decent test. If I can bring it up to 60vac without seeing any deviation from 0vdc at pin7 of IC1, then would that be a good time to power down and setup to check other voltages? Or would you say that I should plow ahead as long as I see ovdc there? Or might it suddenly "swing" up to rail at any moment?

On a related note, IC1 wants to see +/- 15vdc. So does it make sense to make the initial ramp-up of the line voltage pretty quick? Say, bring it up quickly to 50vac?

Anything else I can do to protect the circuit as I bring up the voltage?

Thanks for all your help so far. Hopefully I'm not over-thinking it now, but I don't want to blow anything else up!

Peace,
Bud

 

[crutschow]

To begin with you could wire a small (start with perhaps 10-25W) incandescent bulb in series (not parallel) with the variac output.
That will limit the current in case of a short and its brightness will give an indication of any excess current draw.
The lamp should only light dimly or not at all at the lower voltages.

I would look at the output (TP6) for starters.
Then I would slowly bring up the AC.
The output may not be 0V initially but should get to near zero once the op amp start working

You might periodically look at the ±15V to see when it comes up.
Once you have a few volts of power to the op amps, then the output should start going towards zero.

When the op amp is working its output voltage will be of opposite polarity to the input (audio amp output).
If the condition persists with the op amp input and output of opposite polarity as you increase the AC then you might want to stop there and look at some of the other voltages to see if anything look unusual.
There are voltages posted on the schematic so most voltages, once the op amps starts to work, should be a rough percentage of the variac voltage compared to 120V, e.g at 60Vac they should be about 1/2.

If you get stuck, measure the DC voltages at all the TP's, the power supply (and any other interesting node), and post them and the variac voltage here.

 

[Bud_J]

If you get stuck, measure the DC voltages at all the TP's, the power supply (and any other interesting node), and post them and the variac voltage here.

Thank you, Crutschow,

I certainly felt a lot better and safer with the light bulb limiter and the variac. Leaving the speaker disconnected (and no dummy load or input either) here are my findings:

1.)
With a 40W bulb in the limiter I had to bring the variac up to 50 or 60 Volts before I registered -1.0 vdc on the -45 supply. I looked for this right at the 3300uF caps. The bulb never glowed at all.

2.)
With a 7.5 W bulb in the limiter, bringing the variac up to 50 or 60 Volts only yielded about -0.3 vdc on the negative supply. The bulb glowed an orange color, somewhat like the harvest moon low on the horizon.

Seeing that this was consistent with my previous findings a while back—being that minimal voltages registered with no load connected—I stopped testing there. I suspect that I could connect the speaker and the voltages would come back up, which is how I had proceeded in the past. However, since both you and Nigel agree that this, too, is a problem with the amp, I figured I'd post back here first. I guess it makes sense to figure out why the voltages barely come up with no load connected. Both you and Nigel agree that the presence of a load should have no major impact on the testing voltages.

In the past when I looked at this, I only scratched my head. It seemed that the speaker was somehow completing a circuit that should be complete without it. I guess this should be a clue, but I remember being stymied by it. I have a bit more experience with the schematic, so I'm going to study it again now that I know for sure this situation is not normal. Maybe something will pop out at me this time.

What are your thoughts? Do you agree that I should trace this problem first? Or do you think I should connect a dummy load and keep working on the newer (Q19 shorting out) problem?

Thanks for all your help thus far --- you too Nigel! --- I really appreciate this. I'm learning a lot!

Peace,
Bud

 

[KeepItSimpleStupid]

Keep the speaker out of it.

With a 40W bulb in the limiter I had to bring the variac up to 50 or 60 Volts before I registered -1.0 vdc on the -45 supply. I looked for this right at the 3300uF caps. The bulb never glowed at all.

What is the positive supply doing relative to the negative supply.?

I would suspect that they should come up somewhat symetrical. You are, having issues with the negative supply of th OP amp.

Check the voltage between the AC terminals of CR11. Get something decent like 10 VAC and measure the DC voltage between the +, - terminals of the bridge. Then look at the voltages across C16 and C17. The C16 and C17 voltages should be symetrical. Also look at the AC voltages across C16 and C17 which should hopefully be small.

For reference, is your meter True RMS?

STOP here.

Save for later:
Look at the voltage across the CE junction of Q13 or wherever it's easy to get too. Turning it down to as low as it goes will be fine for initial testing. Make sure the voltage moves with R42. Hopefully full CCW would be the lowest voltage.

 

[Bud_J]

Check the voltage between the AC terminals of CR11. Get something decent like 10 VAC and measure the DC voltage between the +, - terminals of the bridge. Then look at the voltages across C16 and C17. The C16 and C17 voltages should be symetrical. Also look at the AC voltages across C16 and C17 which should hopefully be small.

Thank you, KeepItSimple!

I believe my meter is true RMS, but how does one tell?

Using my meter, I brought up the variac until I had 10vac across the ac side of the rectifier. I measured the DC its output and had 12.3vdc

Then I measured the DC at the caps (taking the measurements between F2 / F3 and ground) and got:

+1.14vdc & 2.0vac
-0.57vdc & 2.0vac

I do remember testing these voltages in the past at full line voltage and they were rock solid at +/- 45vdc. Not sure why the asymmetry now, but it's clearly there.

For the heck of it I checked the +/- 15vdc supplies and found: -0.34 and +.30, but they start high and wind down to these values (while still being kinda jumpy). There was also 2vac at those points, too, according to my multi-meter.

I have a digital oscilloscope, but the Internet has put the fear of God into me about running it without the device under test being connected to an isolation transformer. I plan on purchasing one of these soon. Probably will be my next electronics-related purchase. I don't know if there's a safe way to check the AC using my scope before I get said transformer. Even after I get it, there's conflicting info about how to use them and whether they should be modified for tech usage, but all that's another thread.

Thanks for your ideas. What are your thoughts on my findings?

Peace,
Bud

 

[crutschow]

I have a digital oscilloscope, but the Internet has put the fear of God into me about running it without the device under test being connected to an isolation transformer.

It's save to measure all the voltages after the amplifier transformer with your scope.

Just don't try to measure the Variac AC output voltage as it's not isolated.

 

[Bud_J]

It's save to measure all the voltages after the amplifier transformer with your scope.
Just don't try to measure the Variac AC output voltage as it's not isolated.

Thank you, Crutschow. That's pretty much what I suspected, but it's good to hear it from a pro. I figured as long as I keep my 'scope grounding clip connected to the chassis, then I can safely measure any voltage after (and including) the rectifier's output.

 

[KeepItSimpleStupid]

I believe my meter is true RMS, but how does one tell?

Post model or link to instructions. Usually they have TRMS posted prominently.

sing my meter, I brought up the variac until I had 10vac across the ac side of the rectifier. I measured the DC its output and had 12.3vdc

Very good. You should have +-5.8 +-10% and your pretty close. 5.8*2. Look at the AC voltage here too. I know, why AC when it's supposed to be DC? You don;t have scope, I presume. The AC voltage measures ripple. TRMS does a better job than a run of the mill meter. We can address this topic later.

hen I measured the DC at the caps (taking the measurements between F2 / F3 and ground) and got:

+1.14vdc & 2.0vac
-0.57vdc & 2.0vac

You have to do better in your explanations. F? has two ends. Something is suspicious because one is about 1 diode drop and the other is about 2 diode drops.

Your used to measuring from ground. I'm asking for voltages across things, and not relative to ground.

Again, do a weird measurement. The DC voltage ACROSS F1 and across F2, same conditions. 10 Vac. Put the probes on the wires or PCB pads going to the holder, not the fuse body. I'm assuming an open frame fuseholder. If you can't include the contact resistances, that's OK too.

And the DC and AC voltages across C18 and C19. You should see symetrical voltages about +-6 V and very little AC voltage.

REMIDER: remote trouble-shooting sometimes has many difficulties.

 

[KeepItSimpleStupid]

Thank you, Crutschow. That's pretty much what I suspected, but it's good to hear it from a pro. I figured as long as I keep my 'scope grounding clip connected to the chassis, then I can safely measure any voltage after (and including) the rectifier's output.

Oops. Just found out you had a scope. That's great.

Then you can compare the AC coupled waveforms across C16 and C18 and C17 and C19 which are essentially relative to ground. Yippee! So, C6 to Gnd, C18 to Gnd, C17 to Gnd and C19 to Gnd

I don't really want to tell my story about the intermittent fuse right now, but suffice to say, it's possible. Likely - No.

At some point, you can check the voltage rating and current of the fuse. A rating of 32 V should not be in there. I'm also not saying it is. Do this AFTER taking the measurements.

 

[KeepItSimpleStupid]

Bud:

Curiosity... Do you have a way to inject a sine wave signal? It's another thing that helps.
Helpful things:
a) The "light bulb" "fuse"
b) An sine wave oscillator. I had a simple battery operated one.
c) A current meter on your Variac. Voltage and current metered is better.
d) A typical Variac fuses the WIPE, not AC in. Sencore did both. The wiper MUST be fused. The AC fuse can be set to the rating of the Variac. The wiper fuse can be changed. There's a BIG differeence with say a 1 A primary fuse and a 1A wiper fuse. At 240 Vac, the 10A primary fuse limits the power to 2400 W. Simple numbers) BUT, at 10 VAC out, you still have to use a 10 Amp fuse or less. At 10 VAC, the output gets limited to 100 W. At, 20 VAC, 200 W etc. The wire ultimately determines the fuse size. A 2400 W Variac can't deliver 2400 W at 10 VAC,
e) Monitor power up waveform (speaker terminals) with out a load. Watch for bridged amplifiers.
f) That DC voltage that I told you about, The bias regulator.

 

[Bud_J]

Post model or link to instructions. Usually they have TRMS posted prominently.
. . .
Your used to measuring from ground. I'm asking for voltages across things, and not relative to ground.

Thank you, KeepItSimple. You're right. I was measuring to ground (or so I thought--more on this in a minute).

As far as the meter, I had another tech ask me about it and there is nothing on it, not even a picture on the web. Very old "NRI Digital Multimeter" but it works great and has a diode, transistor, and cap checker. However, it does not say TRMS, so I'm guessing it's probably not?

Now for the new measurements, and a discovery I made while doing them. With 10vac at the AC side of the rectifier, I measure 0.0vac on the DC side.

Across C16:
6.47vdc
0.0vac

Across C17:
6.47vdc
0.0vac

Across C18:
6.45vdc
0.0vac

Across C19:
6.42vdc
0.0vac

Across F2:
0.028vdc

Across F3:
0.024vdc

While trying to grab an easier contact point to measure C18, which was difficult to get to, I figured, I'd just use the chassis ground. Then I got a very low voltage, which was nevertheless enough to set the light bulb in my head glowing brightly. I started searching for how the two semi-detached PCBs (pre-amp and power amp boards) ground themselves. Long story short, I found there was only one contact point at the "BOTH" input on the pre-amp board that served to connect all the grounds to the chassis.

I connected a lead from there to the chassis and -- voila! -- the "speaker needing to be connected" mystery was solved! It also explained why the "problem" was intermittent for me in the past. It all depended on how I chose to ground the boards to the chassis. Now I know. There's only one way to do it while things are disassembled.

So, after that finding, I went back to Crutschow's email about monitoring TP6 to verify whether the DC-feedback circuit at IC1 was holding things to 0.0vdc. Very carefully, and while constantly testing back and forth between various test points, I felt confident enough to bring up the input to 20vac across the rectifier as measured with my meter. With 20vac there, here are my measurements:

Rectifier Output: 25.6vdc

Across C16: 12.8vdc, 0.0vac
Across C17: 12.8vdc, 0.0vac
Across F2: 3.6mVdc, 0.0vac
Across F3: 7.7mVdc, 0.0vac

Of course I was monitoring TP6 the whole time. By this time I got wise enough to take measurements in the millivolt range. TP6 seemed to settle at about 3.2mVdc but it was quite jumpy, jumping around from 3.2 to 3.6mV then took a dive down into the 2mV range and up to the 4mV range and back again. Not sure if that was just IC1 doing its thing or if that's not good.

Then I checked the +/- 15vdc supplies:
The +15 supply read +5.98vdc after settling (starts high then settles downward).
The -15 supply read -9.1vdc and didn't seem to require as much settling.

You'll note that this is the opposite problem I was originally tracing where (at full voltage before I removed IC1 and blew things up) I got +15 and -13.2vdc. I don't know if this is just because we're not at full voltage yet or what. I had a 40W bulb in the limiter while doing these tests and it did not glow at all. Before these, when I had the voltage lower, I had a 7.5W bulb in there and it glowed a little as I turn up the voltage past 10vac at the supply. I'm guessing that's not a runaway current situation?

I stopped when I did because I didn't know what to make of the jumpy 3.2mV at TP6, nor what to make of the disparity in the +/-15vdc supplies.

Thoughts? Think I should continue cranking the supply voltage? Or is it time to figure out what's going on with the regulator circuits?

Thank you for all your help. It certainly feels good to be doing things in an orderly fashion. I'm learning a lot, which is what I had hoped for.

Peace,
Bud

 

[KeepItSimpleStupid]

Nice! The high current and signal grounds should be connected at only one point.

Now, I'd match the gains of Q9 and Q10 withing about 10-15%. Same is true for Q17 and Q18, but let's not worry about that now.

Do a quick check with an ohmmeter of the low value resistors. Most are symmetrical and could be affected by what's nearby., so look for symetry and not the actual value. You might have to reverse your probes.

150: R13, R38
560: R16 and R46
470: R19, R49
470: R20 and R51
15: R47 and R17

0.33: Check R18, R41, R18 and R50; These should fully open when bad.

Solder a wire from Q13-E and Q13-C; or across C13.
This will disable the bias regulator.

Put your meter across R21. It should stay near 0 V. It cannot exceed 90 mV.

You should be able to bring the rails up to +-20 to +-25 V and check the +-15 V supplies.

Take a look at the DC output at the speaker terminals.

Bring the amp up with the 40 W light bulb to full voltage if you can if the above checks out.

 

[crutschow]

TP6 seemed to settle at about 3.2mVdc but it was quite jumpy, jumping around from 3.2 to 3.6mV then took a dive down into the 2mV range and up to the 4mV range and back again. Not sure if that was just IC1 doing its thing or if that's not good.

That small amount of jumpiness is probably okay.
What's the output of the bottom op amp doing at that time?

 

[Bud_J]

Oops. Just found out you had a scope. That's great.
Then you can compare the AC coupled waveforms across C16 and C18 and C17 and C19 which are essentially relative to ground. Yippee! So, C6 to Gnd, C18 to Gnd, C17 to Gnd and C19 to Gnd

Getting caught up on my assignments ;-) I see I have more to do after this, too, so I have something to look forward to in the morning. Okay, so I finally got to have some fun with my o-scope. Here are my findings. I'm guessing these are all fine, but posting here anyway because I'm going for a perfect 4.0 this semester ;-)

With 20vac across the input of the rectifier, measuring the voltage across the C-E of Q13, I get 99.5 mVdc after it settles down (kind of jumpy at first--goes high then low then climbs back up). R42 has RTV holding it in place, so I was hesitant to mess with it.

o-scope cap measurements:
C16 ~ 160mV p-p sawtooth ripple ~120Hz
C17 ~ 130mV p-p sawtooth ripple ~120Hz
C18 & C19 are hard to get to and I think the previous tests showed there are no bad connections to them.

Out of curiosity I looked at TP6 to see how fast it might be cycling. It showed a somewhat sinusoidal wave of about 100mVp-p at about 10mHz, but these values were jumping all over the place.

Alright, onto the next . . . thanks for all the help (and homework)!

 

[Bud_J]

That small amount of jumpiness is probably okay.

Thank you, Crutschow. I was hoping you'd say that

What's the output of the bottom op amp doing at that time?

I was wondering if I should look there next. IC1, pin7 yes? I will check it in the morning and post back
. . .

 

[KeepItSimpleStupid]

Check the ground connection at C8, C7 before you do anything. It's hard to see how the negative voltage is higher than the supply? The + voltage looks right.