[tarking]

 
Quote:

How is the grounding on your amp?  The hand waving is a dead giveaway of improper ground.
 
I never saw a big difference with R10 - that's simply the Class-A bias on the opamp, but it's good that it worked for you.
 
I would still take a look at grounding and before deciding on the OPA637 (only for Left/Right - you can't use 637 in ground, 627 can be used), I would highly recommend trying the AD825.
 
Photos would also be great.

Oops... I meant I changed R11L/R/G from 1 kohm to 100 ohm, sorry &

. Grounding is OK, input ground and potentiometer connected via aluminum case, output jack and DC input jack isolated. This has worked great for me in the past with Pimetas. I'll post some photos when I'll get them resized.
 
I have OPA627AUs on BrownDogs already, didn't have time to solder OPA637AUs yet . Well, I got OPA637AUs for free (well, almost and I happen to like them. I admit I have not listened to AD825. The problem is that this part is not available in Finland.

 

[FallenAngel]

R11 separates the buffer from the opamp, definitely something that should be OK with a lower value than 1K.
 
The amp picking up radio interference (your hand close to the chassis) certainly suggest that your input is floating and it's picking up anything around.
 
Photos are easily hosted on imageshack.us and they resize automatically.
 
Good luck

 

[tarking]

Photos added per request..
 

 

 

[FallenAngel]

Looks great!
 
I would also suggest grounding your POT (screw on back of pot to IG of board).  I'm not sure where you have the input ground connected to chassis either...

 

[tarking]

Quote:

Looks great!
 
I would also suggest grounding your POT (screw on back of pot to IG of board).  I'm not sure where you have the input ground connected to chassis either...

 
Thanks, I am pretty satisfied with the end result myself too.

There's no need to ground pot separately because input ground and pot body are connected via aluminum enclosure. Input ground is connected to chassis because input jack body goes through back panel and the hole in the back panel is precision machined i.e. tight . The plastic washers on the inside were there just to make connecting cables easier for me... they do not isolate body from the chassis. All verified with my meter's continuity mode of course.

 

[bikeboy24]

When setting the buffer bias, does it matter whether the source and/or the headphones are plugged in? It may be my imagination, but the reading across the output resistors change.

 

[tangent]

You can have the headphones plugged in while you do this, but you don't want to do that if you're also measuring at the same time.  Leave them plugged in if you're trying to play with bias by ear, rather than by meter.

 

[bikeboy24]

What about the source? It seems like the readings across the output resistors go down when the source is plugged in.

 

[tangent]

That's DC offset from the source.  It's just confusing things.  Unplug it.

 

[bikeboy24]

The weird thing is that if I unplug everything, all the measurements seem way off. The amp is working and everything. I get sound (sounds pretty good!).
 
Wouldn't plugging in a load (headphones) affect the readings across the output resistors though?

 

[tangent]

Yes, which is why I say leave them unplugged if you want accurate readings. Plug them in if you want accurate hearings.

 

[bikeboy24]

After playing around with the amp some more, I noticed a couple things. I'm using BD139/BD140
 
-The 100ohm R11 fixed my hiss issue with lower impedance / high sensitivity headphones (I'm using Grados)
-I'm still having issues with the buffer bias, but I figured out what is causing the strange readings. When the amp is not grounded to the case, all the readings seem normal. As soon as I connect input ground to the case (such as RCAs, grounding strap), the buffer bias voltage will drop. Anyone know why that is? Is this normal? Output ground is isolated from the case. I don't have any shorts.

 

[tangent]

I think you're just measuring the buffer bias voltage plus the source's output DC offset times the amp gain.

 

[NixNada]

Hey my first post here

 
I'm having issues with amp. It's a v1.1 from 2004 i think.
 
It has worked wonderfully the last 6 years. But last year the solderpoints on the input wires started to fail. Perhaps due to my inexperience.
 
I tried to resolder, but now there is something wrong. I got a very loud static hiss in earphone. So i started to investigate.
 
I found that it draws a lot of current about 500 mA from a 9 V battery. Also there is a large voltage drop about 3 volts, so something is drawing a lot of power.
 
The buffer chips (those are the intersil ha3-5033-5) are getting hot all 3 of them. After about 10 seconds too hot to touch.
 
Also i'm reading a 50 mV from the output to ground, which might explain the unpleasing sound i got in my earphones

 
What could be the problem? And what should i do next?
 
I really have no clue. Is it possibly an easy fix or is there somthing serious wrong with my beloved ppa

 
 

 

[NixNada]

Ok something is wrong. I tried tangents guide for tracing a 1khz input signal.
And it fails when i+ should equal i- on the opamp (both channels)
The i+ equals that of the input (source).

According to guide something is wrong with the feedback loop?

I suspect oscillation because the IN (leg 4) on the buffer transistor is building up voltage as it gets hotter and hotter.

I wonder what could be the problem here?