Failed voltage checks on Low-current and B-side high current boards

[Loading]

I recently finished building my crack-a-two-a and I passed all resistance and voltage checks with the original build.
I used it for around 48 hours over the course of the past week without any faults and decided to incorporate the twoquiet upgrade.

After the upgrade, I find that there is 561 ohms on terminals 4 and 12 and the other values are as specified.
I have verified all of the connections appear to be correct, but am unsure where to go next.
Would the twoquiet upgrade alter the original resistance values or have I done something wrong?

 

[Paul Birkeland]

If you change the position of the Coarse attenuator, that value will move depending on the position of the switch.

Terminals 4 and 12 in the C2A built manual specify 0-100K ohms, so you're still within those limits.

 

[Loading]

That's good to know. I thought that since the upper bound was 100k ohms, the value should be closer to that.

Having passed that, the voltage checks now fail on the low-current board and the B-side high-current board.
I find the following voltages are wrong (measured values with expected in parentheses):

• 
  Low-current
  • IA: 107V (150V)
  • KregA: 0.58V (3-6V)
  • bRegA: 107V (150V)
  High-current (B)
  • OA: 107V (150V)
  • bB: 108V (150V)
  
  
  I have double-checked all connections and am unsure of the problem. Please advise on where to start.​

 

[Doc B]

Start by swapping the 6AQ5 tubes side to side and see if the voltages that are off follow the B side 6AQ5 tube to the A side.

 

[Loading]

Swapping the tubes did not fix the problem. The voltages were the same.

 

[Paul Birkeland]

Something that was marginally connected before has been bumped or broken while installing the attenuator. What is OB on that high current board that is giving the bad reading? 

If you pull both 6AQ5 tubes and run the amp with just a 12AU7 and a 6080, what are the voltages on the high current C4S board under those conditions?

Typically what you have posted would have me suspicious that a 431 regulator has been damaged, but typically this kind of things happen when you slip with your meter probe and touch two pads at once on accident and you get a little spark that's enough to damage the sensitive regulator.

 

[Loading]

I just checked the 431 regulator on the A side and I think that's the problem. All of the pins have continuity to each other and I removed it from the board and this is still the case. I must have mistakenly shorted two pads when testing the voltage the second time and missed the spark.

 

[Loading]

Would this be the proper replacement for the damaged component?
The markings appear to match.

https://www.mouser.com/ProductDetail/512-LM431ACZ

 

[Loading]

I looked around more and I have these on hand.
https://www.digikey.com.au/en/products/detail/texas-instruments/LM431ACZ-NOPB/148198
Is this compatible?

 

[Paul Joppa]

Probably, but the circuits were designed around the ON Semiconductor version, which had (at the time) better documentation of stability margins and noise. It's available at Mouser.

Some no-name bargain versions have been noisy, like 20dB noisier than standard. They may have been fakes, of course.  :^)

 

[Loading]

Ok, I will purchase the ON ones from mouser then. I appreciate all the help.

 

[Loading]

By chance, I discovered I received an extra LM431A component when I built my mainline amp previously.
I installed it and now all voltage checks pass perfectly.
Before I use the amp, I have a question on the safe voltages for the headphone jack.
I measured

Tip: 18 mV
Ring: 32 mV

Are these safe values for the DC voltage at these points? I understand ideally the DC voltage should be as near 0 as possible here.

 

[Paul Birkeland]

That is 0.018V and 0.032V, which is plenty close.

 

[Loading]

That's great to hear.
I appreciate all of the help and can't wait to listen.

Is there any way to lower these readings any further on the headphone jack?
Just curious as I know some other amps I've worked with had DC offset adjustments.

 

[Doc B]

Those number are insignificant, not even worth calling a DC offset. If it was something like 8 or 10 volts it would be worth addressing. The 2.2Kohm resistors on the headphone jack serve to drain off any buildup of DC offset.

 

[Paul Birkeland]

DC offset adjustments are typically present in solid state (and some tube OTL amps) with bipolar power supplies and directly coupled outputs.  On a capacitor coupled amplifier, the cap blocks DC and the resistor from the output of each coupling cap to ground ensures that there's no DC present at the output and provides a path for charging on powerup.  The tiny bits of DC that you're seeing at the output could be from the DC voltage on the other side of the cap moving around slightly, and possibly from a tiny bit of capacitor leakage.  Still 0.032V across the 2.2K resistor there is possibly 0.0000145A of leakage current, and with 300 Ohm headphones, the DC appearing at the outputs would drop to 0.0043V, which is even closer to 0 than what you had before.

 

[Loading]

Thank you so much, that makes a lot of sense.

From a theoretical perspective then, could you reduce this even further by using caps that are less "leaky"?

 

[Paul Joppa]

Very probably.

It's always possible there is some other cause, and replacing the 100uF output cap is an expensive way to test this, but on he other hand it's a popular upgrade mod - and the C2A has room for it.

 

[Paul Birkeland]

From a theoretical perspective then, could you reduce this even further by using caps that are less "leaky"?

You could, but this is like driving all over town to buy a car for $19,999.98 rather than $19,999.99.