Seeking help with Peak Detector circuit

Thread Starter

boattow

Joined Oct 20, 2008
4
Hi all,
I have been reading through lots of threads and articles as well as breadboarding some circuits without good results so since there are people here with much more knowledge than I, I'd like to get your input.

I have constructed a circuit that is receiving a sinewave at a fixed 76Hz. The whole idea is to measure the signal level. This signal starts out very small in the microvolt range and I amplify and bandpass filter it. I have a nice clean sinewave at this point. Next I capacitive couple the signal through an op amp which has a 5volt single supply so the end result is similar to half wave rectification and I end up with only the positive side of the original sinewave. Next stage I have a op amp configuration that can either attenuate or amplify the signal so it ends up near zero (10 to 20mV) and the peaks can go to 5volts. The final configuration is for a digital pot controlled by a microcontroller to adjust this circuit. On the perfboard I have a multi turn trimmer for testing purposes in this circuit and it works fine. In fact up to this point everything is working as needed. The last circuit is to take the signal and give a DC voltage equal to the peak value of the positive sinewave peaks (~10mV to 5V). I will feed this into an A/D on a PIC microcontroller.

The input to the peak detector is going to be anywhere from approximately 10mV to 5V. The 2 circuits I found and tried so far will not handle this range and the output is non-linear when the signal input is low. I tried the circuit using an LM393 published in many places on the web including on this forum because it states it will accept an input as low as 10mV but this circuit didn't work well as I described. With low signal inputs it was nonlinear (not accurate) and it would only go as high as approximately 2.5V. Someone else in another thread stated that they had issues with it when the signal frequency was below 1000Hz. I'm at 76Hz range.

The end device will not have a continuosly changing signal level such that it can be slow. In other words, it will be turned on, the signal level checked, then the sensor will be moved, then the signal level checked, etc. So it can be relatively slow. I hope I've given enough detail. Any suggestions are greatly appreciated.
Boattow
 

Thread Starter

boattow

Joined Oct 20, 2008
4
Thanks, I'll see how I can get a schematic. In the meantime, in summary I have a 76Hz sinewave centered around zero volts, the bottom half has been eliminated leaving only the top half which can vary from about 10mV to near 5Volts. I need a clean DC voltage representing the peak value. I believe a peak detector is what I need but it needs to be able to handle that voltage range at 76Hz. Thanks.
 

gootee

Joined Apr 24, 2007
447
You could also search for "envelope detector", unless you have a good way to reset a peak detector, since their detected value can only rise, not fall, unless they are reset.

But since you have a sine wave, you could also use an averaging detector, or an RMS-to-DC converter.

Doing it in analog, it would probably be a lot easier if you kept both halves of the original signal, and put it through an ideal full wave rectifier opamp circuit, instead of clipping off the bottom half. Keeping both halves would double the frequency of the rectified signal, which would make it much easier to detect or average.

It sounds like you just need to use an averaging detector on the full-wave-rectified signal, to end up with the equivalent DC value. Low pass filter?
 

Audioguru

Joined Dec 20, 2007
11,248
An opamp is usually used with a dual-polarity supply to make a peak detector. Instead, you used a dual comparator that is not an opamp.
Without seeing your schematic then we don't know what you are doing.
 

Thread Starter

boattow

Joined Oct 20, 2008
4
Thanks for the suggestions all. I've decided to just run the signal into an A/D on the PIC and just sample for the peak value rather than to continue to look for a hardware solution.
 
Top