Confusion on home made buffer, one channel clear, two channel not?

[Destroyer OS.]

This my power supply.

http://img.villagephotos.com/p/2003-4/144711/DOS317.gif

15 - C is my buffer

http://www.borbelyaudio.com/adobe/ae699bor.pdf

So I thought I had it all plugged in and the right channel sounded correct at good. I could not test to see if it sounded better with it because the left channel made weird noises like it was not plugged in, a small zapish sound and mostly buzzing. Apparently the RCA cord fell off when I put the buffer into my little wood stand thing.

When I plugged the left channel in and fired it up again I get really bad distortion and static? The right channel is no longer clear?

The only other details of this setup are thus... all grounds from RCA's are running to system ground. The two shunted resistors for input are going to earth ground. Earth ground is connected to chassis from plug.

[Destroyer OS.]

I have concluded these means the + and - power do work.

The right channel works by itself. Power is happening, signal is moving. The question is just why the left when powered is causing a problem. Could it be a bad JFET? I still have like 15 so I could test new ones. I tried not to over heat them. However would one bad JEFT make both channels sound distorted?

[Destroyer OS.]

Anyone got any ideas?

[robertor]

hi, my idea is there is something wrong on the power supply.

for the negative(-24) you must to use the LM337.

then the LM317 provide the +24v and the LM337 the -24v.


or the drawing of the power supply schematic is wrong.

[Destroyer OS.]

Hm this shows either, http://www.national.com/images/pf/LM337/00906701.pdf

My understanding was to get -24v I just reverse the output leads, like in the picture. A friend helped me with this, who is extremely good with electronics, but could of been wrong.

The other question would be why did it work on one channel but not with both plugged in?

[Joris]

Two pieces of 317 will also do without trouble, just don't screw them uninsulated on the same heatsink, that will short circuit...

[Destroyer OS.]

I did not mount them to a heatsink considering the amount of power in this circuit is a couple of mvs. There is definitely power going. It worked before with power but due to bad PDF instructions I had the JFET's reversed and they blew up. That is fixed but now the noise issue is baffling.

[Destroyer OS.]

I have been reading up on negative voltage... the way it is setup is correct.

At the moment I am realizing having the grounds on the RCA jacks tied into "system ground" seems unnecessary. From what I can tell the +24 and -24 circuits complete eachother... so long as the JFET's are switching then +24 and -24 rails should make a complete circuit. The problem lies in the fact that they must reference to ground. I wonder if this means that I need to make the shunted inputs (500k I think I have them at) go to "system ground" in order to give +24 and -24 reference points to ground, in affect giving them their voltage.

The grounds from RCA's and the signal should be it's own circuit, especially to keep DC out of everything.

I guess I will have to try this soon.

[Destroyer OS.]

Ok I the power normal now by switching the ground from the resistor shunts to the system ground. However even without the buffer on I get buzzing through my little test amp and speakers. I am going to put capacitors on the inputs, already tried outputs... I was thinking DC is feeding back into it but I am not sure?

It sure sounds like a ground loop hum or some sorta disconnect, but how it is happening is behind me unless my RCA's do need to be touching system ground inside of the buffer, as well.

[Destroyer OS.]

I tried to the RCA's and found no success.