Fault with MOSFET based headphone amp

[zebra100]

I have recently built a headphone amp which has developed a fault.

http://www.diyaudio.com/forums/showthread.php?s=&threadid=12848

I have tidied up the cables within the case and bolted the boards down since first testing it and now the left channel is noticeably lower than the right. It also seems lighter in the bass department (although it's dificult to tell with the lower level).

I have done all the usual checks, visual and gounding checks with no mismatch between the channels. I have checked the pot and also switched the connections on the CDP. Both channels measured exactly the same give or take the odd .1 of an ohm.

Am I looking at component failure? Could it be the CCS network(LM317T)?

 

[h_a]

Maybe you haven't notived, but your link points to a phono-preamp thread - not to a headphone-amp.

Have fun, Hannes

 

[zebra100]

Oop's, no wonder I haven't had many replies. Thanks for pointing that out. This is the correct link (I hope), post 16 is the CCT digram.

http://www.diyaudio.com/forums/showthread.php?s=&threadid=128487&perpage=25&pagenumber=1

 

[h_a]

Well, if both channels measure the same (at DC) I would replace the mosfet.

You already checked that R1 and the 100k resistor have the correct values?

Have fun, Hannes

 

[Gordy]

So it measured equal before 'bolting' it down?

So un-bolt it and measure it again...

 

[zebra100]

Done that, no difference. I have also changed the LM317 on the affected side, which did lift the level slightly, but they are still not matched. The original LM317 did show signs of overheating. While the boards were out, I reheated all the solder joints just to be on the safe side, as it could have been a dry joint causing the problem. I will try swapping the MOSFET later. Is there any way to check the MOSFET's with a meter before fitting them (i.e matching them)?

 

[zebra100]

Changed the MOSFET and now the inbalence has gone the other way. The right channel is now much lower than the left
(The right channel is much louder than before)
I'll change the right channel just to be on the safe side.

 

[peranders]

Can you report the working points you have? Voltage across each resistor and across the mosfet?

 

[zebra100]

I'm starting to think this project is doomed. I am now getting +ve rail voltage at the junction of R2/R3/RG on the right channel (which has a lower output), which would lead me to suspect O/C R3 or short CCT R2. Maybe I have a faulty component in the bias CCT which has caused the MOSFET breakdown. I do hope this hasn't damaged my last MOSFET. I fear the worst............