3.3V to 10V PWM - Op Amp Circuit Is Making An Audible Noise - How To Troubleshoot To Make Quiet?

Thread Starter

Mahonroy

Joined Oct 21, 2014
406
Hello,
I created this 10V PWM circuit (from a 3.3V GPIO), using a LMC7101 op amp, and the PWM frequency is 350 Hz. The circuit is making a pretty loud audible noise, and I am wondering how I can attempt to troubleshoot it and make it quieter?

Circuit:
op_amp_noise_1.jpg
All components are 0603 components (except for the 33 ohm resistor is 1206).

I have a limitation where I cannot increase the frequency past 350 Hz, this is as fast as it can go, otherwise I would try increasing the frequency until the sound is gone.

I also am having difficulty telling exactly where the sound is coming from, but its definitely coming from one of these components on the schematic I shared.

Any help or advice is greatly appreciated, thanks!
 

joeyd999

Joined Jun 6, 2011
5,237
You're kinda doing it wrong. That amp likely can't keep up with the sharp edges. Put a high impedance RC filter in front of the amp. DC performance will be much better and the noise will probably disappear.
 

philba

Joined Aug 17, 2017
959
I don't know why your circuit is making audible noise but it doesn't make sense to me. Are you trying to simply amplify the PWM to 10V? or are you trying to convert PWM into a voltage control? I believe it's the latter.

If it's the former, why not just use a transistor? If it's the latter, you need to use a low pass filter on the input like joey said. That may cut your noise.
 

Thread Starter

Mahonroy

Joined Oct 21, 2014
406
I don't know why your circuit is making audible noise but it doesn't make sense to me. Are you trying to simply amplify the PWM to 10V? or are you trying to convert PWM into a voltage control? I believe it's the latter.

If it's the former, why not just use a transistor? If it's the latter, you need to use a low pass filter on the input like joey said. That may cut your noise.
Yes I am trying to amplify a 3.3V PWM up to 10V PWM. Not sure what you mean by a voltage control? (Are you referring to if I am trying to convert PWM into a 0 to 10 analog signal?).
 

Alec_t

Joined Sep 17, 2013
14,280
My guess is C28 is the noise source if it is a ceramic capacitor. If so, replacing it with a plastic film type might solve the problem.
 

philba

Joined Aug 17, 2017
959
As noted, why not use a single transistor?
I made that comment but saw that the OP (er, TS) wants to solve the noise problem, not change his circuit. Though, I bet a PNP transistor would not have the noise problem.
 
Last edited:

Thread Starter

Mahonroy

Joined Oct 21, 2014
406
When I was researching what the best circuit would be to accomplish this, I saw a bunch of pros/cons to each method of either using an op-amp, or a transistor. I saw people complaining about the interference and switching noise (non-audible noise) that people got with a transistor, and it seemed the better rout was to use an op-amp, so I picked that rout. I'm no expert, so I had to research as much as I could and just go with one. It needed to be able to drive up to 300mA, be robust and not failure prone, needed to be able to be short circuited, and able to withstand transients coming in on the signal output line.

I will try removing C28 and see what that does.
 

philba

Joined Aug 17, 2017
959
Perhaps, if we're talking about the signal and not power to the load itself. We don't have all the pieces to the puzzle.
Not sure what you mean - 1% duty cycle would become 99%. But I do agree we don't have all the info. If the TS has access to the thing denerating the PWM, he can invert it there and use aforementioned NPN.
 

philba

Joined Aug 17, 2017
959
OK, so I did a quick little spice simulation to illustrate my point. This is a simple NPN circuit that shows the inversion I was talking about. The resistor values aren't critical. As you can see, the input waveform gets inverted upon output and the duty-cycle (DC) becomes 1-DC. In this example, the DC goes from 25% to 75%.

To get non-inverted output with NPNs, simply add a second NPN stage and use a voltage divider to the output voltage to 0-10V (use a 20K pullup and a 100K pulldown from the collector). Though, a simple MOSFET voltage translator solution also exists.
ckt.pngVin.png Vout.png
 
Last edited:
Top