OPA2277: Set point noise at the ends of the Pot.

Thomas,

Thanks for the info. I will check those voltages. A little more info for you, essentially what I am doing is setting the span of adjustment with U2. U2A is setting the high limit and U2B is setting the low limit. These are generally set to 0-5V of 1-5V. I tried an experiment where I put a 20 ohm resistor between the wiper of the pot and U1. This almost completely eliminated the noise on the high end (U2A), but the reduction on the low end was still not adequate. This seems odd to me as U2 is powered by +/-15V being at/near 0 shouldn't be a problem. I also tried it 1-5V and the low side was still noisy.

Hi Joe,

when the wiper reaches the end of resistance element, contact problems can occur. The resistance between wiper and resistance element can heavily increase due to oxid layers. The oxid layer often forms a parasitic semiconductor which can zener like a zener diode. This can result in heavily increased noise. Even an instability of OPAmp can occur, because this parasitic zenering effect can oscillate by itself. You can observe this generation of excessive noise, especially if the trimmer is old, oxidized or sulphated. The insertion of a current limiting resistor can help, either at the output of driving OPAmp or between the wiper and the receiving OPAmp.

Another issue is noise injection. Such a trimmer is usually unshielded and when you touch it with a screw driver you can directly inject HF interference into the circuit. This can destabilize the trimming circuit and cause oscillation.

Instability due to unwished load capacitance is another issue which can mostly be observed, when the wiper reaches the end of resistance element, because then there's no isolation resistance between the output of driving OPAmp and the trimmer terminal which looks capacitive. The insertion of a 22...100R isolation resistance can help then.

Capacitive feedback is another issue which can make trouble with trimming circuits. This can often be seen when the ground routing and/or the supply voltage decoupling of OPAmp is improper. In such cases there's usually a stray capacitance from the output of U1A back to the trimmer terminals or even the input pins of U2A. A simple resistor between the wiper and the receiving OPAmp forming a low pass filter in combination with the input capacitance of receiving OPAmp can help. A solid ground plane and proper supply voltage decoupling is the best cure against capacitive feedback, though.

In my circuits I have always an isolation resistor between the output of driving OPAmp and the according trimmer terminal. And I have always a passive low pass filter between the wiper and the receiving OPAmp. In big circuits it's extremely helpful to have passive input low pass filtering at each OPAmp. And, analog circuits do also heavily profit from a solid ground plane...

Kai

Hi Thomas,

This is not a totally new application of the OPA2277. The output structure, which I will post below, is at least close to things we have done in the past. Using it to set the span, U2, is new at least in this configuration. U2A is setting high point of the span, approximately 5Vdc. U2B sets the low point of the span, 0 to 1Vdc.  I checked the output of U1A and it adjusts from 3-8 Vdc, as it should. We are not getting anywhere near the +/- 13 Vdc rails. We are running the circuit with a 250 ohm load connected between C and D.

I am also adding some o'scope captures. The dark blue trace is at TP6. The light blue is at the output of U1A. The green is at the drain of Q1. And the pink is at TP7. This is with a 20 ohm resistor added between the wiper of the pot and U1A-3. As you can see, the input, TP6, remains clean throughout the adjustment range. The other signals start showing noise around the mid point of adjustment and get worse as we get to zero.

Thanks for the help.

Hmmm...it doesn't seem to want to let me insert more files.

set800 scope captures.pdfI guess it doesn't like pdf's.  

Hi Joe,

this isn't noise. This is oscillation! Does is come from the 250R load? Where is it connected to?

Looks like the U1A circuit is unstable.

Kai