[Avro_Arrow]

I find it's pretty easy to make mistakes at three in the morning...
 
You really should tie the ground side of the regulators together, not
the output side. It would probably work in your last example, but
I would not guarantee it would work as expected.
 
Your voltages are what would be expected.

 

[shake]

Quote:

Originally Posted by Avro_Arrow /img/forum/go_quote.gif
You really should tie the ground side of the regulators together, not
the output side. It would probably work in your last example, but
I would not guarantee it would work as expected.

The thing is, when I had it wired like this:

It wouldn't put out any voltage anywhere.

It's when I have it tied in the middle that the LED lights up, showing that it's working, and I can get voltage readings.  Resistor is on the anode.

 
EDIT:
layout from above.
7915 on top, 7815 on bottom.

 

[Avro_Arrow]

If it works then carry on...
 
"If it ain't broke...Don't fix it"

 

[shake]

Haha, I just want to make sure I won't make something 'splode.  I already had a near death experience playing with AC today...

 

[shake]

My test pair of headphones began heating up without playing any sound when I tested the amp.  What should I check in the circuit?
Regulators, op amps, and buffers also get warm within a minute.
 
EDIT: I'm measuring ~-11.5v offset for left and right channels.
Right channel is still working, left channel is rattling lol.

 

[Avro_Arrow]

That sounds like a major malfunction.
Check over wiring carefully.

 

[shake]

It looks right to my eyes.
Maybe my schematic is wrong?

 

[Avro_Arrow]

The schematic look fine.
Do you have the class A bias installed yet?
If yes, disconnect it.
It sounds like the signal is being dragged to
one rail. Check your feedback loop.

 

[jcx]

the "inner feedback loop" is pointless - I do know in detail where it comes from intellectually - the theory has been thoroughly discredited, Walt Jung soon came to the correct conclusion that Otala"s "flat loop gain" prescription was flawed - none of his later multiloop audio line driver circuits show it - check out his "Op Amp Applications" book (free online)

 

[shake]

How do I know if ground is actually ground?
 
Even with the class A biasing removed, it's still showing a large offset. -14V from L/R to ground.
 
With all negative wires disconnected from power, I'm still reading an offset of -1V

 

[Fred_fred2004]

Can you disconnect the power supply and power the amp with two 9volt batteries, use the center tap of the two batteries as ground, at least that will tell you which part of the circuit has the problem

 

[shake]

Quote:

Can you disconnect the power supply and power the amp with two 9volt batteries, use the center tap of the two batteries as ground, at least that will tell you which part of the circuit has the problem

I'll get right on it.

 

[shake]

Using the batteries, I'm measuring 6-7V.
Maybe I should have gone with another cmoy :/
Does the LMH6321 need to be connected to ground?  It is right now.

 

[jcx]

for debugging I'd move the global feedback to just the input op amp - the buffer should be fine on its own, removes one more concern
 
I'd also reduce the series 1K at the buffer input - it is too high, adds phase shift to the loop ~ 50-100 Ohms is better range
 
you can just unsolder, lift one end to "remove" a R, C, 2 legged component for debugging, just tack solder new part or shorting wire on

 

[shake]

Quote:

for debugging I'd move the global feedback to just the input op amp - the buffer should be fine on its own, removes one more concern
 
I'd also reduce the series 1K at the buffer input - it is too high, adds phase shift to the loop ~ 50-100 Ohms is better range
 
you can just unsolder, lift one end to "remove" a R, C, 2 legged component for debugging, just tack solder new part or shorting wire on

So I shift everything to before the buffer?  What would I have to change to run this without buffers?
 
Could I completely remove the 1K resistor? I don't think I have any on hand below 1K.
 
And yes, there's always the possibility that I'm dumb.