DC offset problem

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Jump to Latest
li_gangyi

li_gangyi

Member
Joined 2003
iexitsolutions.com
a servo..it's in the app note for the LM3875...basically it's something like this
 

Attachments

  • servo.png
    servo.png
    5.6 KB · Views: 806
C

carlosfm

Banned
Joined 2003
Vikash said:
If it bothers you than you can adjust the resistance from signal in to ground, or perhaps try the Ci cap as shown on the National schematic.
Click to expand...

Yes, this can be done and lowers DC to very low values.

Max, I suppose you are testing the amp without signal, no source connected.
You have on the input a 22k resistor to ground.
Is this an integrated amp you are doing or a power amp?

In case of an integrated amp, did you try connecting there a pot and measure DC again?
In case it's a power amp, have you tried other values at the input to ground instead of 22k?;)

Oh, :cool: heatsink you have there.:up:
 
maxw

maxw

Member
Joined 2004
Franz G said:
In this application note, you can find another good example of a servo drive.

http://www.national.com/an/AN/AN-1192.pdf

And what about the first sound test of your amp, Maxw? :)

Franz
Click to expand...
Haven't done a sound test yet, i was waiting until I got this sorted first. That looks kind of complicated, I'll try it as a last resort:)

carlosfm said:

Yes, this can be done and lowers DC to very low values.

Max, I suppose you are testing the amp without signal, no source connected.
You have on the input a 22k resistor to ground.
Is this an integrated amp you are doing or a power amp?
Click to expand...

Yes, no signal, no source.
I have made a temp external volume control for testing out of a 10k Taiwan alpha log pot but it the end it will be powered by a buffered preamp so it wont be integrated:)

carlosfm said:

In case of an integrated amp, did you try connecting there a pot and measure DC again?
In case it's a power amp, have you tried other values at the input to ground instead of 22k?;)

Oh, :cool: heatsink you have there.:up:
Click to expand...

You just mean a pot between Input ground and input +?
No I haven't tried any other values yet, will do.

Vikash said:
If it bothers you than you can adjust the resistance from signal in to ground, or perhaps try the Ci cap as shown on the National schematic.
Click to expand...

See my attachment, is that the right way to implement Ci? It what I figured from looking at the diagrams and Brians userguide.

Thanks for your help guys;)
 

Attachments

  • dsc03672.jpg
    dsc03672.jpg
    47.3 KB · Views: 773
C

carlosfm

Banned
Joined 2003
maxw said:
See my attachment, is that the right way to implement Ci?
Click to expand...

See LM3875 datasheet, page 1, figure 1.

I don't know if on Brian's schematic it's R3.
If you post the schematic I'll tell you if it's ok.
Sorry, too busy to search for it now, I'm finishing my new pre.:smash:

Oh, forget the resistors across the output for now, measure DC without anything connected.
What DC values do you have on the output now, with Ci?
 
maxw

maxw

Member
Joined 2004
carlosfm said:


See LM3875 datasheet, page 1, figure 1.

I don't know if on Brian's schematic it's R3.
If you post the schematic I'll tell you if it's ok.
Sorry, too busy to search for it now, I'm finishing my new pre.:smash:

Oh, forget the resistors across the output for now, measure DC without anything connected.
What DC values do you have on the output now, with Ci?
Click to expand...

I have attached a pic with both schematics.
I dropped the resistors on the output and its the same.
My resistor burning up must have been because of a bad connection with my test board because the IC is just sitting in the wholes, not soldered.

So I retried with 22uF electrolytic as Ci and got a DC offset of 2mV!! That must be good right providing I've stuck it in the right place?
 

Attachments

  • untitled2.jpg
    untitled2.jpg
    25.8 KB · Views: 702
maxw

maxw

Member
Joined 2004
I did a listening test today with a Sony CDP, 10k pot and my DIY 2 way speakers. I added the Ci cap as in the above attachment to the unit that had 80mV DC and then was 1-3mV. The other unit I used was the one with no Ci cap and it had an offset 44mV and 20mV when all connected up to pot etc.

So I have tested the Ci cap on two units now and in both cases it reduced DC offset to 1-4mV. So why doesn't everyone use it? does it affect the sound? (I couldn't tell with one channel using it one not);)
 
Nuuk

Nuuk

Member
Joined 2003
decdun.me.uk
So I have tested the Ci cap on two units now and in both cases it reduced DC offset to 1-4mV. So why doesn't everyone use it? does it affect the sound? (I couldn't tell with one channel using it one not)
Click to expand...

Because, as some of us have found, lower DC offset does not necessarily mean better sound, in fact I found the opposite was true. ;)

Try both and hear for yourself. :att'n:
 
C

carlosfm

Banned
Joined 2003
maxw said:
So I retried with 22uF electrolytic as Ci and got a DC offset of 2mV!!
Click to expand...

:D

maxw said:
I did a listening test today with a Sony CDP, 10k pot and my DIY 2 way speakers. I added the Ci cap as in the above attachment to the unit that had 80mV DC and then was 1-3mV.
Click to expand...

:D

maxw said:
So why doesn't everyone use it? does it affect the sound? (I couldn't tell with one channel using it one not);)
Click to expand...

You can bypass the 22uf cap with a 100~150nf poly cap.
Let it :hot:-in for some days.

If you don't want do make an inverting amp (with buffer, preferably), with NI this is the way to go IMHO.
What's debatable is what has more effect to the sound?
- 80mv DC
or
- Ci caps?

Your speakers may not work so well with 80mv DC.
But the effect of the Ci caps, if they are decent quality, is minimal if not undetectable.
 
maxw

maxw

Member
Joined 2004
carlosfm said:
But the effect of the Ci caps, if they are decent quality, is minimal if not undetectable.
Click to expand...

Well I cant hear the difference and I like things that are technically good so I will use Ci on my GC. Cool.

One last Q though, I have used a cheap generic cap that is 35V, cost 50c. The only place I can get better ones in NZ is RS so do you think Elna Starget (for example) caps would be worth paying $10-$15 for? or shall I just use generic low ESR ones?
 
C

carlmart

Member
Joined 2001
preciseaudio.com
maxw said:


Well I cant hear the difference and I like things that are technically good so I will use Ci on my GC. Cool.

One last Q though, I have used a cheap generic cap that is 35V, cost 50c. The only place I can get better ones in NZ is RS so do you think Elna Starget (for example) caps would be worth paying $10-$15 for? or shall I just use generic low ESR ones?
Click to expand...


Like I said on another thread, according to an article published in Electronics World and Wireless World, try a bipolar capacitor as feedback DC block. Something reasonable good, not any speaker cap you may find around.

If the feedback resistors you are using are large, particularly the small one, you can even get to use film caps. Something like 2 x 4.7uF should not be expensive to arrange.


Carlos
 
P

phugger

Member
Joined 2004
S

skyraider

Member
Joined 2004
maxw,

Did you take offset measurements with the input conected? Because if input is connected to a source, your readings might change again. Same goes for heat, aging effect.. etc.
Dont worry too much about offset, especially if its low like in your case. It will change again after sometime.
I (and many others?) personally disgust of having Ci there in the circuit. What I would suggest is to try as many chips as you could to find a low-offset pair.(like what gaincard guys did?)
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Top Bottom