Help!!! :( Serious DC Offset with Pimeta
Apr 11, 2006 at 3:55 AM Thread Starter Post #1 of 42

FallenAngel

Headphoneus Supremus
Joined
Mar 7, 2006
Posts
9,627
Likes
72
Location
Toronto
Hi guys,

I've just "finished" putting together my Pimeta and I'm getting a serious DC offset!

I plan to do Class-A, but have not installed the JFETS yet.

I wanted a gain of about 4, so here are the components used:

R1R/R1L : 4.42k
R2R/R2L : 1M
R3R/R3L : 3.32k
R4R/R4L : 10k
R5R/R5L : 10k
R6R/R6L : 1M
R8R/R8L : 0 (Jumpered)
R10R/R10L : 1k
R11R/R11L : 210

R1G: 4.42k
R3G: inf
R4G: 4.42k
R11G: 210

OPAG: AD8610AR
OPALR: AD8620AR
BUFFERS: 1 for each BUF634P
D1: STPS1045D (The other one on Tangent's part list)

The DC offset (latest):
OG to OR: 5.72v
OG to OL: 4.99v

Any ideas??? I'm totally confused as this is my first serious build (asside from a working TREAD).

Thanks!!!
 
Apr 11, 2006 at 4:13 AM Post #2 of 42

MisterX

Headphoneus Supremus
Joined
Dec 30, 2002
Posts
3,880
Likes
23
What are you hoping to achieve by not jumpering R8?
confused.gif


You may as well cut the left and right wires off the output jack cause you get the same exact results.
wink.gif



Quote:

The DC offset:
OG to OR: 1.66v
OG to OL: 1.66v


With no R8 and/or no jumper in place of R8 you should not be able to measure any offset at all.

Got this mounted in a case?

Measure the dc volts from inground to outground and tell me what you get.
 
Apr 11, 2006 at 4:47 AM Post #3 of 42

FallenAngel

Headphoneus Supremus
Joined
Mar 7, 2006
Posts
9,627
Likes
72
Location
Toronto
I guess I read something wrong cause I thought you ignore R8 if not using Class-A, and now I can't find it anywhere
confused.gif
. Well, I just put it back and updated the scenario. R8R/R8L = 0 (Jumpered).

No case, just sitting on the mousepad on my desk right now

Quote:

Originally Posted by MisterX
What are you hoping to achieve by not jumpering R8?
confused.gif


You may as well cut the left and right wires off the output jack cause you get the same exact results.
wink.gif





With no R8 and/or no jumper in place of R8 you should not be able to measure any offset at all.

Got this mounted in a case?

Measure the dc volts from inground to outground and tell me what you get.



 
Apr 11, 2006 at 11:26 AM Post #4 of 42

tangent

Top Mall-Fi poster. The T in META42.
Formerly with Tangentsoft Parts Store
Joined
Sep 27, 2001
Posts
5,969
Likes
56
Quote:

Originally Posted by MisterX
With no R8 and/or no jumper in place of R8 you should not be able to measure any offset at all.


Sorry, that's not the case. You still have a path to -IN, and there's still a feedback loop around the op-amp (R5 + R6) so the op-amp should still be happy in this broken setup. It's not forced into comparator mode or anything bad like that.

I haven't thought deeply on it, but it seems to me that what you're measuring with R8 open is the "force" applied at the op-amp's -IN to create the ~300x inner loop gain on the op-amp. This certainly will give you a weird reading, if you're expecting to measure the amp's output DC offset.

FallenAngel, have you remeasured DC offset since jumpering R8? It's quite possible that the amp is working now.
 
Apr 11, 2006 at 2:40 PM Post #5 of 42

FallenAngel

Headphoneus Supremus
Joined
Mar 7, 2006
Posts
9,627
Likes
72
Location
Toronto
I have measured the offset now. It is the values posted, 5v on Right, 6v on left. As for "working", well, I already tried plugging in $2 headphones into it, they actually got hot!... then they got toasty and one of the drivers blew
very_evil_smiley.gif


When I didn't have the R8, DC offset was equal on both channels at 1.66v, since jumpering it, it is now at 6v and 5v for Right and Left respectively.

I also really wanted to know, how important is it for R3 to be 5-10x smaller then R5? I have mine actually larger, hope that's not the problem.
 
Apr 11, 2006 at 2:49 PM Post #6 of 42

n_maher

Resistorous Conflagorous
Joined
Nov 20, 2004
Posts
8,405
Likes
31
As with trying troubleshoot anything it would be most simple if you do the following:

1. Return the amp to a "known good" configuration with regards to the resistors.

2. Post pictures of the board, top and bottom. Make sure they are clear, good quality, high resoluation images.

Nate
 
Apr 11, 2006 at 3:32 PM Post #8 of 42

n_maher

Resistorous Conflagorous
Joined
Nov 20, 2004
Posts
8,405
Likes
31
Quote:

Originally Posted by FallenAngel
I can post pics of the board after I get home today, but as for "known good" config, I built it like this from the start and I'm not sure I ordered too many more resistors but I'll see what I can find.


If you need resistors just shoot me a PM, I'll send you what you need to build it with the "stock" gain of 11. Hopefully someone with more knowledge can chime in about whether or not the values you have chosen could be causing the problems you're seeing. In the mean time I wouldn't power up the amp again, if you're having stability problems it'd be a shame to toast an opamp and compound the problem.

Nate
 
Apr 11, 2006 at 3:38 PM Post #9 of 42

FallenAngel

Headphoneus Supremus
Joined
Mar 7, 2006
Posts
9,627
Likes
72
Location
Toronto
Thanks for the offer buddy, I might take you up on it some time, but I work a few blocks away from a surplus store where I should be able to find everything.

I was working that formula and have come up with this attempt:

R3: 1k (This I'll need to get)
R4: 3.32k (Take from R5 in old config)
R5: 4.42k (Take from R3 in old config - now almost 5x R3)
R6: 1M (Keep same, hopefully no problem that itn's not ~100x inner loop)

The gain will change from 3.2 to 4.2, but I don't think I'll mind much, especially with the new phones coming in soon
biggrin.gif
 
Apr 11, 2006 at 3:50 PM Post #10 of 42

n_maher

Resistorous Conflagorous
Joined
Nov 20, 2004
Posts
8,405
Likes
31
Like I said, open offer, anytime.

And with a decent volume pot (I don't know if you're using the Alps or the Panny or another) you should get plenty of adjustment with a gain of 4, unless your cans are really hard to drive (don't know what you're using here either).

Certainly report back if you do get some swapping done.

Best of luck,

Nate
 
Apr 11, 2006 at 3:52 PM Post #11 of 42

tangent

Top Mall-Fi poster. The T in META42.
Formerly with Tangentsoft Parts Store
Joined
Sep 27, 2001
Posts
5,969
Likes
56
Quote:

Originally Posted by FallenAngel
When I didn't have the R8, DC offset was equal on both channels at 1.66v, since jumpering it, it is now at 6v and 5v for Right and Left respectively.


This is sounding more like a broken chip or a bad connection, then. Post pictures.

Quote:

how important is it for R3 to be 5-10x smaller then R5?


Not very. Lowering R4 is better for changing gain than raising R3, but I don't see how the values you've got now could possibly be causing the offset. Something else is wrong.
 
Apr 11, 2006 at 3:59 PM Post #12 of 42

FallenAngel

Headphoneus Supremus
Joined
Mar 7, 2006
Posts
9,627
Likes
72
Location
Toronto
Thanks for the help guys, I'll have the pictures up tonight.

Wow, I just found out what I was doing before, but I mixed up R8 with R10. I do not have Class-A yet, and I have R10 populated with 1K. Should this be removed until I populate JFETS? I'm guessing though that this should have no effect.

Tangent's Great Document:
R10

If you bias the op-amp into class A with JFET cascodes, you should also add R10. JFETs have an input capacitance, and you don't want a capacitor across the output of the op-amp. R10 is a kind of "insulation" between the two, so they don't affect each other as much.

The exact value of this resistor is not critical, but it does have to fall within a certain range. If it's too low, it doesn't do its job very well; 100 Ω is probably the smallest value that will provide appreciable benefit. If it's too high, it will cause the cascode to fall out of regulation during parts of the swing of the op-amp's output, negating its advantages; 1 KΩ is about the largest you can get away with.

Optional? Yes. Populate it if you bias the op-amp into class A. Leave it empty otherwise.
 
Apr 12, 2006 at 3:11 AM Post #14 of 42

SubRosa

New Head-Fier
Joined
Mar 8, 2006
Posts
47
Likes
0
Trying cleaning the flux off the board, it also looks like you've got some cold joints there.
 
Apr 12, 2006 at 4:04 AM Post #15 of 42

FallenAngel

Headphoneus Supremus
Joined
Mar 7, 2006
Posts
9,627
Likes
72
Location
Toronto
Made some changes to the resistor values, updated top post.

R3: 3.32k
R4: 10k
R5: 10k
R6: 1M

This got the DC offset to about 1.7v on L/R, which is a huge improvement from 5v+. Of course, still not good enough yet.

Also, unfortunately while rearranging the resistors, I managed to snap a pin off the BrownDog adapter on the AD8620AR and had to do some creative fake pin creation (1/2" SPC lead) - therefore, new pictures:




SubRosa: I'm not using flux, as for the cold joints, I'll see what I can do about fixing them, but damn it's a pain in the ass to redo all of them.
 

Users who are viewing this thread

Top