DC offset problem

[whatsnew]

I use this schematic
http://www.geocities.com/rjm003.geo/...gc.html?200629

The DC offset about -200mV.
I have measured the power source were DC -36 and +36.
I decreased R2 value to 6K.
Then i got the lower DC offset.
Could it influnce the sound ?

[halo0925]

By decreasing the value of R2 the low frequency rolloff will be increased depending on the value.
Do you get the DC offset with no input connected?

[whatsnew]

I connected the CD player
but I still get the dc offset

[Leolabs]

My DD3875 read -50mV without input connected,with Rin=Rfb=22k,Rg=1k.Power source are +/-25V.Drop to -9mV with input grounded.Not a single cap is used at the input nor feedback.

Seem that your got -200mV,try to lower the voltage gain.

[AndrewT]

Hi,
I'm surprised that reducing R2 to 6k reduces the output offset.
How are you measuring this offset?

200mV is a bit high, most aim for <50mV and these amps when properly set up can achieve <10mV.

It probably won't affect offset, but is Vrail=+-36Vdc OK for these chips?

The schematic shows AC coupled. Adding a source should not change the DC offset.
Or
have you got the version that is mixed AC coupled on the input and DC coupled on the NFB loop. Reducing R2 in this version will reduce offset. It's due to the complete disregard for the resistances seen by the two inputs brought about by mixing the AC & DC coupling.

Note that Leo is using DC coupling throughout. But, this only works if the source is also DC coupled, otherwise the DC route to source is blocked and you end up with mixed coupling again.

[whatsnew]

How are you measuring this offset?

connect red pin to the pin 3 and black pin to ground

It probably won't affect offset, but is Vrail=+-36Vdc OK for these chips?

I heard about that supply voltage under +-40 would be ok.

The schematic shows AC coupled. Adding a source should not change the DC offset.
Or
have you got the version that is mixed AC coupled on the input and DC coupled on the NFB loop. Reducing R2 in this version will reduce offset. It's due to the complete disregard for the resistances seen by the two inputs brought about by mixing the AC & DC coupling.

I dont know what the AC/DC coupled is.
Can you explain more in detail?
thanks.

[AndrewT]

Hi,
your link shows two schematics.

One is AC coupled i.e. it has a DC blocking capacitor at the input AND it has a DC blocking capacitor in the Negative Feedback Loop (NFB). The Input cap stops DC from your source causing output offset.
The cap in the NFB reduces the DC gain of the amp to one (+0db) and this limits the output offset to match the chipamps input offset which varies a bit with temperature and with the input pin resistances.

The second schematic shows the DC block on the NFB omitted and this raises the DC gain to match the AC gain. Now the input offset (and it's variation) is multiplied by the DC gain and appears on the output as DC offset.

Which have you installed?

[gootee]

Something, somewhere, is way off. We can hope that it is due to some "trivial" construction or measurement error.

Check your wiring and component values again, and try to make sure they are the same as the schematic.

And what sort of LOAD do you have connected to the amp's output (i.e. speaker, resistor, nothing)?

One other (somewhat remote) possibility is that radio-frequency interference (RFI) is entering your chipamp and is being rectified and showing up at the output as a DC offset. That might be less unlikely if you have your circuit only breadboarded, for example, possibly with long wire leads, etc, or have a poor layout, and/or, if there is a relatively strong source of RF nearby. But I'm guessing that it would probably take something on the order of 10mV of 1MHz or greater RF at your + input, to get 200 mV out, with part of it probably being rectified into DC, depending on the frequencies involved.

You should also measure the voltage between the grounded "signal input" (and other "ground" points) and the ground terminal at the power supply, just to be sure that your grounds layout isn't causing severe problems.

Have you measured the DC offset directly at the + input pin of the chipamp (i.e. the voltage across R2), while you're getting the 200 mV output offset? It would probably take something like 6 mV to 8mV DC, there, to get 200 mV DC out (or a similar negative offset at the chipamp's - input pin), depending somewhat on how long your ground leads are, and whether or not you've separated the signal and power ground returns.

Also, are you sure that C2 is OK? Can you measure a DC voltage across C2, when the input is grounded as it was during your testing?

And, did you measure the power supply voltages directly AT the chip's power input pins, i.e. between the actual power supply pins on the chip and your power supply's ground? If the negative power rail's voltage, AT THE CHIP, was between -4 and -5 volts instead of -36v, it could possibly cause such an output offset, with that circuit (assuming you're using the second schematic, at the link you gave).

A photo of your setup might be helpful.

Good luck.

- Tom Gootee

[sangram]

The circuit in figure 2 would give quite a high DC-offset with the values shown (but not 200mV definitely). I used a R2=4k7, R3=220ohm, R4=4k7 and get about 1.5mV. I also separate signal and speaker output ground by a 2.2 ohm power resistor. This seems to keep the offset down.

This is an LM4766. With full AC coupling it tends to raise the output offset to about 30-40 mV on the channels. By decreasing the input resistor R2 the offset does indeed change, and quite drastically at that. IDK the exact reason why, but peranders had explained to me why it has something to do with iinut bias currents. Also the input F3 changes and the cap must be increased to compensate.

Also the power supplies may have something to with it. A difference between the rails (uneven loading, layout issues) will show up directly as DC offset.

On a sceond note I have an amp that has no (<2mV) offset but is fully DC-coupled (no input cap) and on one of my sources gives about 300 mV output on one channel (the source has a bit of offset on that channel which is multiplied by the amp). It operates fine, though I don't use it with that source too often, if at all. No damage to speakers, except that the on-off thump is a little louder than normal - however YMMV.

[Peter Daniel]

Originally posted by sangram
Also the power supplies may have something to with it. A difference between the rails (uneven loading, layout issues) will show up directly as DC offset.

The difference in power supply voltage will not affect the offset, to the point that one rail can dissapear and it will still not produce any excessive DC offset.