My PIMETA has a 200mV voltage offset!
Jan 28, 2004 at 5:53 AM Thread Starter Post #1 of 19

donovansmith

Headphoneus Supremus
Joined
Aug 16, 2002
Posts
2,957
Likes
11
Location
San Antonio, TX
I was wondering why I would hear clicking noises in my headphones anytime the headphone jack was moved and today I discovered that there is a 200mV voltage offset from ground to the L/R outputs. I have checked rechecked everything and cannot figure why there would be an offset like this. All the resistor values are correct and there is no R3G, so no possibility that I accidentally configured for gain. I did install C6, too. Anybody have any ideas?

Edit: This is with no source connected.
 
Jan 28, 2004 at 7:07 AM Post #2 of 19

donovansmith

Headphoneus Supremus
Joined
Aug 16, 2002
Posts
2,957
Likes
11
Location
San Antonio, TX
I spent the last 2 hours trying to figure this out to no avail. The OPA627 in the ground channel seems to be boosting a 0v signal to 200mV somehow but I can't figure out how or why. Funny thing is, despite this the amp sounds perfectly fine. Should I really be worrying about this, is this maybe just normal for this topology?
 
Jan 28, 2004 at 12:59 PM Post #3 of 19

ECM

100+ Head-Fier
Joined
Jan 9, 2004
Posts
175
Likes
19
what kind of headphones are you using? 200 mV can damage them over time. i use a set of grados and when i built my CMOY i had about 30 mV offset. i called the company up and asked them what was the MAX offset the headphones could handle and they said .1V or 100mV.
 
Jan 28, 2004 at 5:22 PM Post #4 of 19

donovansmith

Headphoneus Supremus
Joined
Aug 16, 2002
Posts
2,957
Likes
11
Location
San Antonio, TX
Right now I am using HD497s, but soon it will be HD600s. What bothers me about the dc offset is that the ground on this thing is perfect, but the opamp in the ground channel is somehow putting out 200mV DC yet everything is to spec. The OPA627 is a unity gain stable chip and others have reported success with it used in a PIMETA here. I have already isolated this to the opamp. Does anybody know if this is oscillation and/or a bad chip?
 
Jan 28, 2004 at 5:27 PM Post #5 of 19

robbneu

New Head-Fier
Joined
Jan 22, 2004
Posts
36
Likes
0
I'm curious about this, as well. I'm slowly getting ready to build a PIMETA myself and I'd hate to accidentally damage my Grados.

In a couple of weeks, I hope to have a PIMETA put together, so I might be able to provide some feedback then... until that point, Donovan, I'm hoping someone will post some good information.

Robb
 
Jan 28, 2004 at 6:31 PM Post #6 of 19

tangent

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

- What is the current draw of the amp, and what do you expect its quiescent current to be? If there's a significant mismatch, it's oscillating or you aren't accounting for something.

- Are any of the chips getting warm? (Not counting buffers, if set for higher bandwidth.)
 
Jan 28, 2004 at 6:50 PM Post #7 of 19

donovansmith

Headphoneus Supremus
Joined
Aug 16, 2002
Posts
2,957
Likes
11
Location
San Antonio, TX
I can't measure the current draw (my multimeter doesn't seem to work for this) but I am getting 6 hours average out of 2 9V 150mAH batteries in series, which is what I expect given the high power draw of the OPA627s. All of the BUF634s are in low-bandwidth mode. None of the opamps get warm that I could tell. That's what makes this so bizarre, is that everything seems to functioning okay except for the ground channel opamp putting out a 200mV dc signal. The virtual ground is also perfect.
 
Jan 28, 2004 at 8:00 PM Post #8 of 19

blipblop

100+ Head-Fier
Joined
Dec 6, 2003
Posts
378
Likes
11
maybe the measurement is off?
 
Jan 29, 2004 at 1:37 AM Post #9 of 19

donovansmith

Headphoneus Supremus
Joined
Aug 16, 2002
Posts
2,957
Likes
11
Location
San Antonio, TX
It's possible, but given the loud popping and crackling in the headphones when the plug is wiggled or pulled in or out it seems that a dc offset truly exists. The only things I can think of is that somehow the opamp is configured for gain, a very high gain (200 or so) which I doubt is the case since the unit sounds fine, or the opamp is oscillating like tangent is suggesting. Problem is that I don't have another single channel opamp to test with. The OPA627 is a fast opamp and it might be prone to oscillation more so than something like an OPA132. Strangely it isn't heating up or anything out of the ordinary either. I do have C6 installed, which I guess to help prevent oscillation (I'm just guessing though). R1G and R4G are both 4.7K, and hand-matched to be within 0.1% of each other. There is no R3G installed, and nothing jumpered there. Is there anything else I can check for?
 
Jan 29, 2004 at 3:53 AM Post #10 of 19

donovansmith

Headphoneus Supremus
Joined
Aug 16, 2002
Posts
2,957
Likes
11
Location
San Antonio, TX
I am just curious about what C6 does. I am guessing it may be a DC filter, but am not sure. Tangent's PIMETA parts selection guide mentions this as being essential and that it is suppossed to prevent problems of some sort, but doesn't get more specific. If I remove this, what are the possible downsides?
 
Jan 29, 2004 at 5:29 AM Post #11 of 19

MisterX

Headphoneus Supremus
Joined
Dec 30, 2002
Posts
3,880
Likes
23
Check the PPA parts selection guide for the role of C6.

Quote:

This is a bandwidth-limiting cap for the ground channel. It is necessary to maintain the stability of the ground channel.


Do you think removing it is going to help matters?
 
Jan 29, 2004 at 6:56 AM Post #12 of 19

donovansmith

Headphoneus Supremus
Joined
Aug 16, 2002
Posts
2,957
Likes
11
Location
San Antonio, TX
I didn't realize that's what it was for. I was at work while posting and I didn't think to look up the PPA pages. I think blipblop may be right, my multimeter could be wrong. I notice that after I take the leads off the chip the reading stays the same on the multimeter which is rather strange. There are no chips heating up, no excessive current draws. It sounds fine. In fact, when I get a mA reading between the left or right channels and the ground, it reads at 0.01mA which isn't much at all. The amp is also dead silent at all volume levels even with low-impedence phones. Maybe I'll get a different multimeter and see what kind of readings I get from it. But I guess for now I might as well not worry about this. If there truly is a voltage offset that high I guess my HD497s would have been toasted a while back, but they still work good as ever.
 
Jan 29, 2004 at 3:24 PM Post #13 of 19

tangent

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

I can't measure the current draw (my multimeter doesn't seem to work for this)


Are you inserting it inline with the circuit you're trying to measure? This is easy with 9V batteries: remove one battery, attach the battery clip to just one of the two terminals, and use the meter to jump the other battery terminal to the battery clip.

Quote:

I am getting 6 hours average out of 2 9V 150mAH batteries in series


That implies an average 25mA draw, which is about right for three low-bandwidth BUF634s and three OPA627s, plus miscellaneous draws like the LED.

Quote:

I am just curious about what C6 does.


It's a short high-frequency feedback path for the op-amp. It means that high frequency signals won't have to go clear through the buffer before being fedback to the op-amp's -IN. Any HF signal on the ground channel must be noise or the beginnings of oscillation, so we want to short it out as quickly as possible. Do not remove it.

Quote:

I notice that after I take the leads off the chip the reading stays the same on the multimeter which is rather strange


Hold the leads together. If the meter doesn't zero out, it's time to get a new meter.

Are you on a mV scale?

If everything with the meter checks out, measure +IN on OPAG to OUT on OPAG, and then OUT on OPAG to OUT on BUFG.
 
Jan 29, 2004 at 6:15 PM Post #14 of 19

donovansmith

Headphoneus Supremus
Joined
Aug 16, 2002
Posts
2,957
Likes
11
Location
San Antonio, TX
I tried doing that with the battery, but it doesn't seem to actually bridge the connection so the meter doesn't read anything. My meter does zero out when measuring voltage and I hold the leads together, though. I didn't know if that was normal or not.

I was measuring the voltage from the in and out leads on the opamp and buffer. My 200mV reading came from the OPAG in to the OPAG out leads.

I did get a suggestion by email from a forum member that I should lift up any pins on the opamp that are are not the power or opamp in/out pins (e.g. the offset trim pins). I will try that when I get back home this evening and see if it might help. I never even thought about the fact that a single opamp has some extra connections that may affect the operation of the opamp. I do remember somebody mentioning that they had to lift a few pins on the OPA627 when they stuck it in a Meier Corda amp since some of the pins had different functions than the original opamp in that amp in a thread here a while back.

Thanks for your help!
 
Jan 29, 2004 at 6:30 PM Post #15 of 19

tangent

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

it doesn't seem to actually bridge the connection so the meter doesn't read anything.


Do you mean that the amp won't power up when you try this? No matter whether you get a reading or not, the amp should power up. If it powers up and you can't get a reading, either you need to reverse the leads, or the meter isn't sensitive enough to give you a meaningful value. You need about 1mA resolution or better to make useful readings here.

Also, on some meters there are different jacks for the probe leads for amperage readings.

Quote:

I was measuring the voltage from the in and out leads on the opamp and buffer. My 200mV reading came from the OPAG in to the OPAG out leads.


Do the two measurements I suggested, then. And, make sure you're measuring against +IN on the op-amp, not -IN. While +IN and -IN should be identical in theory, there are factors that can cause small differences.

Quote:

I should lift up any pins on the opamp that are are not the power or opamp in/out pins


The only pin likely to be causing problems is pin 8. Let me know if lifting that one fixes the problem.
 

Users who are viewing this thread

Top