measuring DC offset of opamps

Set the meter to DC volts, 200 mv setting. Connect the black probe to ground. Place the red probe on the output pin of the opamp, pin's 1 and 7 for a dual, pin 6 for a single. Observe voltage and polarity.

If below 2~3 mv, usually it's safe to remove the el cap. If higher, observe polarity. Be sure the + pin on the cap is facing the positive DC and vice-versa.

thanks jim and svart.

this gives my a big help as now I can start to take some reading around the circuit rather than just through darts at a board.

what is your thoughts on coupling caps before pots. even if I measure low dc offset would you still be inclined to leave the caps in place to protect the pot.

IME opamps handle DC just fine, many are designed for precision DC operation. They don't like exceeding the maximum specs. I come from the era of analog computing. Those used opamps to shove DC around. They were tube opamps back then, changed to SS in the mid/late 1960's. The IC's came out in '68 with the LM101.

If pots had a hard time with DC then all those dbx compressors would have failed many years ago. They still operate just fine today moving DC up and down. If you have more than a few millivolts of DC going across a pot, you will hear that as scratch. Using low bias current precision opamps will eliminate the bias current noise when turning a pot, fet opamps are the lowest.

Small amounts of DC can be delt with by a cap, a trimpot or a servo. Modern low offset, precision opamps eliminate the need for those older generation fixes. Yes, the sonics of an electrolytic cap are determental to quality audio. Those that have experienced a direct coupled, capacitor-free audio system already know this. A bipolar el cap is also no fix, it may block DC from both directions but you pay by running the signal through two back connected el caps, doubling the losses. Bipolar caps are not available in low impedance 105 degree formulations either. If you need to use one you are better off using two higher quality el caps wired back to back. 105 degree el caps will last longer as well.

Here are some opamps that will allow direct coupling in many circuits:
LME49720/10
OPA1611/12
OPA1641/42/44
AD8597/99

Besides power rail decoupling, any feedback resistor above about 3k ohms will generate stray capacitance that will lead to ringing or oscillations. A small cap across the feedback resistor usually fixes that.

If any pin on the opamp goes to ground, then yes, you could use that as a ground point. Not every opamp has a pin that might go to ground though, as different circuits use the opamp differently. The 220uf go from each of the + and - rails to ground just like the 100nf. The 100nf typically are for supplying fast transient currents to the opamp and also "shorting" small signal rail noise to ground by forming a LPF. The larger 220uf caps are for bulk capacitance, which can help smooth out larger rail fluctuations and keep reserve current handy more closely to the parts that use it.

Yes, newer opamps can handle DC just fine, didn't mean to insinuate that they didn't, but they don't handle DC and audio at the same time very well, unless designed to do so. Also, thinking back to the craze a few years ago where people were doing strange things in attempt to bias class A/B opamps into class A operation (via active current sinks or simply resistors on the output) I remember reading much about overheating, burned up traces and power supplies due to the overall power consumption increase. Anyway, the general theory might work but in my own study, the increases in perceived audio detail were simply increased distortion, crossover and otherwise from the poor attempt at the opamp output driver to account for mismatched push-pull behavior.

It may very well be a matter of taste anyway. I tried a number of opamps over the years, but always felt the academically pure but clinically sterile super-fast opamps weren't putting wind in my sails. I fell back to the old standards and felt at home. Haven't bothered to spend money on that stuff again, now I just spend it on mics.

Hey Gouge.

As has been said, right tool for the right job. You want to avoid DC offset, you use an Op Amp with good DC precision. Putting DC on pots can be very scratchy, particularly on cheap pots where the brush contact isn't consistent. Nasty stuff - do not want.

The idea on the power filtering caps you mentioned is to remove as much noise as possible just before it hits the V+ and V- input of the Op Amp. Track resistance accumulates, and you want the AC impedance to favor your cap substantially more than the path to your V+ and V-.

Are we going to be continuing your thread on GS or here? That is, should I reply to the GS thread at all?

Well it is best to measure twice, cut once.

If you wanna remove the caps, go with Jim's recommendations. I've a few others I'll dump in the other thread. Might be a little outside the scope of what you're aiming to achieve.

And thanks!

with regards the decoupling caps. which polarity are they installed for each rail?

is DC offset a product of the BIAS resistors differences, the Interals of the op amp, or a comgination or none of the above?

The reason I ask, is I have a small DC offset at the output of a INA217 and its about 10mv and I can get rid of it with a OPA137, but if its a case of just better bias resistor matching I will go down that path first I think

cheers

Wiz