What contributes to DC-offset? - diyAudio
Go Back   Home > Forums > Amplifiers > Solid State

Solid State Talk all about solid state amplification.

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 27th December 2004, 05:05 AM   #1
sam9 is offline sam9  United States
diyAudio Member
 
sam9's Avatar
 
Join Date: Jun 2002
Location: Left Coast
Default What contributes to DC-offset?

In a typical Lin topology amp or a symetical (2 LTPs, 2VAS's) variation, what are the factors or topology features that contribute to DC offset. I'm not talking about a serious offset, just within the normal acceptable range of 40mV or less. For example I have completed two working amps that appear identical. One shows an offset of ~10mV anf the other ~20mV. Even though this is "OK", I'm a little obsessive and think it would be nice if the offsets were the same AND as low as possible. (It's useless to tell me there is no sonic significance - I know that and still obsess.)

Are there component values that can be adjusted? Is there something that should be matched more closely? For example, the BJTs in each differential pair are matched for beta and their RE resistors are much closer than even 1% -- but the "top" and "bottom"
pairs are not to each other pair, etc. Or is matching the PNP/NPN pairs of output devices where one could look?
  Reply With Quote
Old 27th December 2004, 05:19 AM   #2
Duo is offline Duo  Canada
diyAudio Member
 
Duo's Avatar
 
Join Date: Jun 2002
Location: Victoria, BC, Canada
Send a message via ICQ to Duo
The first thing that causes bad offsets in this kind of amplifier is the LTP input stage. Being a differential amplifier, it is a servo incorporated to maintain the nullification of the amplifier output. If this stage is not balanced, the rest of the amplifier will obviously not be balance either.

The important part of the LTP is the two active devices, the transistors. If these have different gain, which they always do to some extent, they will reflect this difference by balancing differently. One way to achieve better offset is to incorporate a current mirror. This is done with the addition of two more transistors in the LTP. Also, matching of the transistors is very useful before deciding to add the mirrror if you have the ability to do so. There are ways to pad the balance of the circuit using a pot as well. This isn't generally accepted by the stringent audiophile circle, however, it is an easy way out. The real problem with pots being their inherent problems with age, heat, and dirt.

I have found that by adding a current mirror and by matching the transistors, output offsets under 5mV are not difficult to obtain. Many of my amplifiers are down to 1 or 2 mV if I spend the time getting all the kinks of the transistors worked out.

As far as the other stages in the amplifier, in most cases, they don't really have a great deal of effect on the balancing of the output offset. One more consideration I should note, is the matching of the feedback and input network resistors. It seems that by keeping the feedback resistor (one from output to LTP, not groung) matched with the bias resistor on the input side can help with balance. This because the LTP must reference through both of these resistors, one being tied to ground, and the other to the output, they must both give the same potential bias to the LTP or it will balance to the difference in the resistors, resulting in offset. In some amplifiers, I have seen 1% resistors used in these areas in order to maintain equal gain between channels for parallel and bridge operation, and also to achieve good offeset nullification.

I'm sure the real engineers here can offer much more detail into this situation, however, this is a chunk of the important theory I use in designing my own amplifiers, and it gives good results.
__________________
-- Duo, W1ngs, VA7MON, and lesser known handles. --
-- http://www.w1ngselectronics.com -- My Work and Projects --
  Reply With Quote
Old 27th December 2004, 05:59 AM   #3
sam9 is offline sam9  United States
diyAudio Member
 
sam9's Avatar
 
Join Date: Jun 2002
Location: Left Coast
In the specific case, current mirrors are not an option. There are other areas where better matching may be possible.

What about top vs bottom matcing in the case of a 2-LTP input stage? Although each pair was pretty closely matched, it was not possible to match one pair to the other at all well with the stock of small sig BJTs I had on hand. They are not terribly expensive, so buying another batch or two until I could get a 4-way match do-able.

I was dubious about using a pot in place of the two RE resistors since with a 2-LTP topology, it would seem that you would need two pots which would be more likely to just confuse things.
  Reply With Quote
Old 27th December 2004, 10:55 AM   #4
diyAudio Member
 
destroyer X's Avatar
 
Join Date: Feb 2004
Location: Recife - Brasil Northeast
Default When supply positive and negative voltage are different, that difference will appear.

The difference will appear at the output too.

This happens here, as i used to make my own transformer, and sometimes they are so well made that i have, as result, 400 mV differences between positive and negative rails.

regards,

Carlos
__________________
DxS, the best from the best...the cream from the cream, the best wide world ever made audio amplifier.
  Reply With Quote
Old 27th December 2004, 11:10 AM   #5
quasi is offline quasi  Australia
diyAudio Member
 
quasi's Avatar
 
Join Date: Sep 2004
Location: Adelaide, Australia
Default One way;

Here's one way Elektor did it with a double LTP.

Quite clever they adjust a small DC voltage right onto the input transistors of the DLTP. They use a trimpot and 2 x 10Mohm resistors. I have coloured* the relevant components in blue.

Cheers

* Australian / British spelling
Attached Images
File Type: jpg dc offset idea.jpg (39.7 KB, 449 views)
  Reply With Quote
Old 27th December 2004, 03:14 PM   #6
sam9 is offline sam9  United States
diyAudio Member
 
sam9's Avatar
 
Join Date: Jun 2002
Location: Left Coast
I like the looks of the Elector scheme. Normally, I would be relectant to "monkey" with the input signal more than absolutlely necessary, but the high (M-ohm) values plus the cap (maybe parralled with larger cap -220uF-) ought to safely isolate the adjustment mechanism.

Since the two identicle amps share a common PS, the difference in offset between them can't be the whole story. Noetheless, I will check the rails more closely with a DMM having more digits. As it stands the look identical to 0.1V. If there is a difference, that provides an excuse to use one of ESP's circuits that force PS DC-offset almost perfectly to zero..
  Reply With Quote
Old 27th December 2004, 05:24 PM   #7
Duo is offline Duo  Canada
diyAudio Member
 
Duo's Avatar
 
Join Date: Jun 2002
Location: Victoria, BC, Canada
Send a message via ICQ to Duo
This use of a high value pot to null the amplifier is one that I've seen in Harman/Kardon equipment. Namely, the citation sixteen.

Being such a high value resistance, I can't really see it having any relevant effect on the signal information. And the system works very well for cutting out that last few mV of offset
__________________
-- Duo, W1ngs, VA7MON, and lesser known handles. --
-- http://www.w1ngselectronics.com -- My Work and Projects --
  Reply With Quote
Old 27th December 2004, 05:51 PM   #8
anatech is offline anatech  Canada
diyAudio Moderator
 
anatech's Avatar
 
Join Date: Jun 2004
Location: Georgetown, On
Most amplifiers from the seventies used a DC balance adjustment pot before op amp servos came to be. Very safe to do and will not colour (Candian spelling) the sound if done correctly. Matching the diff pairs will really help with DC offset and thermal drift. I recommend this first and do it myself when repairing or building an amp.
-Chris
  Reply With Quote
Old 29th December 2004, 08:00 AM   #9
boholm is offline boholm  Denmark
diyAudio Member
 
boholm's Avatar
 
Join Date: Apr 2004
Location: Copenhagen
Then there is the question of stability and drift due to temperature shifts. Even thou' you try to match components (and I do that too) you can never match them 100%. Therefore there will be a slight chance of voltages and currents drifting resulting in a DC output also drifting. And here the DC servo comes in to take care of that.
__________________
Best regards
Bo
  Reply With Quote
Old 29th December 2004, 09:43 AM   #10
Steven is offline Steven  Netherlands
diyAudio Member
 
Steven's Avatar
 
Join Date: Oct 2003
Location: The Netherlands
Hi sam9,

Depending on topology of your amplifier, you may also have to look at the second stage of your amplifier, probably the VAS stage. Suppose the VAS is loaded with a current source (CCS), which is very common, then the VAS transistor should carry the same current as the current source (if the load of the output stage can be neglected or is symmetrical as well). Now, it is important that the VAS stage is designed in such a way that its base current is available from the input LTP when the LTP is in balance. The delivery of VAS base current should not cause an unbalance of the LTP, otherwise the overall feedback will offset the LTP just to get the right VAS current to match the CCS. This can be tuned (adjusted, tweaked) by changing the collector resistor(s) of the LTP.
In your double LTP topology, the positive VAS may have a different open loop current than the negative VAS stage. Overall feedback will then offset the LTP stages to get the two VAS currents equal. The solution is the same is above: adjust the collector resistor(s) of the LTP(s).

Steven
__________________
The Analog Art shows no sign of yielding to the Dodo's fate. The emergence and maturation of monolithic processing finesse has perhaps lagged a bit behind the growth of the Binary Business. But whereas digital precision is forever bounded by bits, there is no limit excepting Universal Hiss to the ultimate accuracy and functional variety of simple analog circuits. - Barry Gilbert, 1973
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
What is DC offset? spooney Car Audio 16 9th September 2013 12:32 AM
DC Offset JannaCassandra Chip Amps 3 26th July 2008 03:20 PM
2,5v dc offset!! nicoch58 Class D 2 4th March 2007 11:02 PM
Help With DC Offset Sybex23 Chip Amps 8 15th November 2004 07:11 PM
GC offset jan.didden Chip Amps 4 3rd July 2004 08:10 AM


New To Site? Need Help?

All times are GMT. The time now is 07:14 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2