DC offset question

[tschanrm]

I was reading a thread over at audio karma about DC offset. Thread can be found here:

http://audiokarma.org/forums/showthr...ghlight=offset

In it, it mentions that a good dc offset is less than 15mv. Above that value, amplifiers can have problems (distortion, etc.). With the tripath amps, many people are reporting dc offset around 80mv. I want to build a tripath based amp, but I am worried about dc offset causing damage to my speakers. Because of this, I have a couple of questions:

1. Does that mean that the tripath amps with this high dc offset (80mv+) are malfunctioning?

2. Is there a difference between the dc offset found in solid state amplifiers compared to the dc offset found in tripath's amplifiers?

3. Does the dc offset in the tripath amps add any distortion to the amp?

I've tried searching about what dc offset is (besides being a aquare wave); all I've found is that it is bad for speaker and headphone voice coils (because it heats them up). Any help would be appreciated.

[Pierre]

What do you mean when you say that "offset is a square wave"?
Offset is only a DC level present at the speaker output with no input. Above 0.5V or so is bad, but to my knowledge levels around 100mV or so are only unacceptable for electrostatic loudspeakers (correct me if I'm wrong).

[classd4sure]

Hi,

I dont' know if above 50mV is "bad", but no greater than 50mV seems to be the accepted "standard".

I've run a homebrew switcher with about 100mV offset with no problems. Can't say what it does for distortion.

Tripath do seem to have unreasonably high levels of offset from what I've read here.

My UCD180's have ~2.5mV max.

Regards,
Chris

[theAnonymous1]

Depends on which Tripath amp(s) your talking about. I have 3 AMP-1B's with not a single mV of dc offset on any of the 6 channels.

[fredos]

Dont stress about DC offset of 100 mv or less in class D amplifier. In linear amplifier, this can introduce distortion because the input differential is not matched. In class D, offset can be caused by offset in error comparator or asymetry in the triangle wave, so it's not critical for overall dynamic performance, like in linear amplifier...So dont stress with small offset!

Have a nice day!

Fredos

www.d-amp.com

[Pano]

Originally posted by theAnonymous1
I have 3 AMP-1B's with not a single mV of dc offset on any of the 6 channels.

That's a bit unusual, or perhaps lucky! I see Tripath amps all over the place on DC offset; from a few millivolts to over 100. Some of it seems to come from the chip; some of it from the external components.

I'll agree with Fredo, don't worry about 100mV or less if you have any sort of "normal" speaker set-up. E.G, 95dB/W or less. In very sensitive drivers, the DC offset can cause a little non-linearity, as it tends to push the cone away form its neutral point. The DC offset can also act as a sort of "brake" on the movement of the cone. Whether that will sound bad or good is hard to guess.

There is a simple circuit published by Tripath to eliminate DC offset on the small chips, you can see it is now included in the latest TA2024C docs

[tschanrm]

Ok, thanks for the replies everyone, makes me feel more comfortable now!

[theAnonymous1]

panomaniac, the AMP-1B auto trims the offset to 0mv at startup. BTW, my AMP-1Bs don't even HAVE an output now do to a silly silly mistake.

[Pano]

Originally posted by theAnonymous1
the AMP-1B auto trims the offset to 0mv at startup.

Oh Yeah, I had forgotten about that. Nice feature of the 2022 chip, the modulator feedback loop is external. Allows for an auto offset trim. Nice! On the lower power chips it has to be done a the input.

What did you do to your amp?

[theAnonymous1]

I was trying some different bypass caps on the 5v supply that daisy chained all 3 amps to see if I could get rid of some of the hiss. I either shorted the supply for a brief moment or the bypass cap caused some kind of surge in the line, but either way all 3 amps are stuck in error mode now. I think the chips are dead.