AU-919 DC offset transistor advice!

[Hyperion]

2SK117 Seems to have just the right spec' - that 'input diff pair' - did you replace a dual encapsulated device or were they discreet to begin with - and if so, how did you mount the replacements to ensure close thermal coupling?

 

[dr*audio]

2SK117 Seems to have just the right spec' - that 'input diff pair' - did you replace a dual encapsulated device or were they discreet to begin with - and if so, how did you mount the replacements to ensure close thermal coupling?

I used it to replace an encapsulated pair. I coupled them together with heat shrinkable tubing.

 

[noveltone]

Thanks dr*audio and Hyperion for the 2sk117 info. That looks like a great part, and it's a lot more available than some of the more famous others. This is great info!

This weekend I had a chance to observe some more measurements. Those I will share here. Again it seems that changing the nature of the input of the power amp really changes the DC at the output. You guys are a much better judge of this stuff, however. Please look at these items and let me know if they give you any ideas about where to look in the power amp.

I have some JFETs on the way (hopefully), but I am concerned that they won't solve the problem.

Power Amp Rail voltages:
steady at -56.0V DC, +54.7V DC both channels
<1.0 mV AC

Amp on 2 hours, cover on, power amp separated from preamp, adjusted freshly for the sake of the below comparisons
L:0 mV DC offset, 20mV Bias
R:same

3-ft RCA cable connected to power amp input, other end disconnected
L:+50mV DC offset at output
R:+480mV (!!!!!)

Same cable connecting amp to preamp, all preamp sections adjusted for 0mV, conn/sep switch in back set for "direct coupled"
L:+14.1 mV DC offset at ouput
R: -110 mV

Connected as above, with Jump switch only (bypassing flat preamp)
L:+16.9 mV at output
R:-127 mV

Connected as above, with Subsonic switch only
L:-22 mV
R:+263 mV

Connected as above, but with conn/sep switch set for "cap coupled"
L:-46 mV
R:+546 mV

Cap coupled switch with Jump and Subsonic engaged
same result

Amp allowed to cool, then restarted separated.
L:-36 mV
R:-82 mV

It alarmed me to see the offset moving all over the place with seemingly commonplace switching changes. Especially the one-end-only RCA cable connected to the power amp input setting the Right channel going almost to 0.5 V!

Thanks again for feasting your eyes upon this information. Please share any insights you may have.

Sincerely,
Collin

 

[noveltone]

see diagram showing switches in context

 

[xoaphexox]

 

[EchoWars]

New caps in the amp and preamp? All centering adjustments properly made?

 

[noveltone]

Yep, all new caps(electrolytic and "black flag") in amp, flat preamp, and their power supplies with one exception. I did not replace the large oval-cased power supply caps.

Regarding the centering adjustments, they were done according to the SM. No inputs, no speakers connected, volume at low, shorting plugs on either the power amp input or on aux input for the flat preamp. I even kept the cover on, then just removed it to make the adjustment, then repeating once or twice to get it to be zero for a while. Sound good?:scratch2:

 

[noveltone]

When the preamp is separated, the power amp is subject to very high offset with an antenna of one-ended RCA cable attached to its input.

Now, the problem I identified initially was that despite the hot-amp DC offset centering that would occur, the hot-amp DC offset measurements would tend to drift and settle at pretty high values, 60 mV, and 110 mV. Since I just redid the centering I haven't observed the same thing happen again.

The measurements on page three, with all the switching and odd things connected to the power amp input, were just ways to gather more data about things affecting the power amp input. Maybe this is normal.

(what happened to the post I responded to?) Subscriber thing?

 

[EchoWars]

Deleted the post after I re-read your thread. But I'd set the offset using shorting plugs and see if that yielded better results.

I don't see how the offset can continue to creep in the same direction. You wait for it to warm up (1 hour, 2 hours), and make an adjustment. Yet, you power down, let it cool off, and when you power back up the offset still needs adjustment in the same direction. Something is wrong with this scenario, as after a few repetitions, you'd be maxed-out on the adjustment pot.

I still feel that you don't have a good grasp of the problem definition. That, or I missed something somewhere.

 

[tarior]

Did you ever check the voltages from the regulated power supply(ies)?

 

[noveltone]

Hm. EchoWars I think I see your point about the offset not being able to keep going in the direction it's going. I don't really know where to go from here, however, except to send it to one of you guys. I took measurements with and without shorting plugs, it didn't change the result at all.

When DC offset fluctuates for people, does it fluctuate over long periods, or it is pretty unstable right when you're looking at it? As tarior asked about, the rail voltages I observed on the driver board were stable at the values I mentioned earlier today, not much AC ripple. Same with flat preamp, which is a different supply. It's more involved to get output supply voltages, so I haven't done that yet.

Would it be valuable for me to take the driver board schematic and overlay measured voltages against those listed?

 

[EchoWars]

No, because your error is much less than 1V.

I'd match a pair of 2SK117GR's and replace one of the input FET's with them just to observe the change, if any.

 

[tarior]

As tarior asked about, the rail voltages I observed on the driver board were stable at the values I mentioned earlier today, not much AC ripple. Same with flat preamp, which is a different supply. It's more involved to get output supply voltages, so I haven't done that yet.

I was more curious about the regulated power supply or supplies, not so much the unregulated rail voltages.

 

[noveltone]

So Tarior there is a simple power supply with rectifier, caps, and transistor that supplies the driver board only. Is that 56V supply regulated on the board further? If so, are those the voltages that one sees in the part of schematic I posted early in this thread?

EchoWars, it looks like knowing how that JFET is doing is the next step to narrowing it down. I am ordering some 2SK117 (hopefully GR)'s today. Thanks to Dr Audio and all who've recommended this part.

We shall see. Thanks everyone for your continued interest.

 

[tarior]

OK, the two unregulated supplies on the 2845 board are the collector supplies for the outputs.
The two supplies on the lower half to the 2845 board are regulated. Are their voltages correct say, within +/- 10%?
Also, are the voltages on the F-2844 board correct?

 

[noveltone]

I can say that the regulated supplies on F-2845 are correct, and that F-2844 is correct. Can't say regarding the unregulated output supplies. F-2844 supplies the preamp sections, including the flat amp.

 

[dr*audio]

Connect a DMM to the preamp output and tell me what you read, with nothing bypassed, and no filters. In other words, I want a DC path all the way through the preamp so I can see if there is any offset there.

 

[noveltone]

Thanks for your interest.

The story on the preamp output is a little strange. I measured a persistent 4.5 mV DC on the right channel, while at the flat preamp, I was getting 1 mV. I had the input selector on AUX, so AFAIK no phono stage signal was in the path. There were shorting plugs on the AUX signal, and no bypasses or filters were used.

The 4.5 mV was persistent because it always re-settled there after I adjusted the trimmer in the flat preamp. When I blew air on the flat amp, that number fluctuated by 10 mV or more, but re-settled. I didn't measure PS voltages, but I'll try to do that tonight. The right side phono preamp, which is fed by the same PS as feeds the right side flat preamp, is pretty unstable, too.

Edit: I checked and found that what I said about PS is incorrect. The same +/- rails feel both left and right sides of all preamp sections. The driver boards and output stages have separate L/R power supplies, however.

 

[dr*audio]

Thanks for your interest.

The story on the preamp output is a little strange. I measured a persistent 4.5 mV DC on the right channel, while at the flat preamp, I was getting 1 mV. I had the input selector on AUX, so AFAIK no phono stage signal was in the path. There were shorting plugs on the AUX signal, and no bypasses or filters were used.

The 4.5 mV was persistent because it always re-settled there after I adjusted the trimmer in the flat preamp. When I blew air on the flat amp, that number fluctuated by 10 mV or more, but re-settled. I didn't measure PS voltages, but I'll try to do that tonight. The right side phono preamp, which is fed by the same PS as feeds the right side flat preamp, is pretty unstable, too.

Edit: I checked and found that what I said about PS is incorrect. The same +/- rails feel both left and right sides of all preamp sections. The driver boards and output stages have separate L/R power supplies, however.

Leave the power supplies alone for now until we localize the problem. When you measure the preamp output, you should have the poweramp disconnected. Please verify this, then retest and tell me what you get on both channels for DC. Make 2 measurements; once with the volume control at minimum, then at maximum. Report the readings for both channels in both conditions.

 

[noveltone]

I tested it again tonight, making absolutely sure the power amp and preamp were disconnected by continuity check. Volume min, volume max, flat amp bypassed and not, I got the same results as above. Does that help?