Hello,

I've got a circuit that is essentially a voltmeter, and it receives very short pulses (on the order of microseconds) that need to be measured. I'd like to add a simple circuit between the input and the voltmeter that will hold the maximum input voltage for at least a few seconds. So, even if the input jumps to 10V for just a moment, the meter will display that value long enough to see. Is this as simple as an RC circuit?

Thanks!

 

This is called a "sample and hold" circuit.

 

A sample and hold circuit requires a timing signal to perform the sample. You want a version of that called a peak hold circuit which doesn't require a separate timing signal.

 

OK, the peak detector seems perfect, thanks for that.

I found two designs: one with an op-amp (http://en.wikipedia.org/wiki/Precision_rectifier#Peak_detector) and one without (http://www.afrotechmods.com/videos/peak_detector_circuit.htm).

Am I right in understanding that the op-amp design will give me a true peak, while the simpler design will give me a slightly reduced peak due to the forward voltage drop on the diode?

 

Many decent quality meters will have hold function and also min/max record functions.

Prices are so cheap now you can get a meter with a ton of cool functions for $50 to $60.

 

Many decent quality meters will have hold function and also min/max record functions.

Prices are so cheap now you can get a meter with a ton of cool functions for $50 to $60.

Did you miss the part about the very short pulses?

 

I found two designs: one with an op-amp (http://en.wikipedia.org/wiki/Precision_rectifier#Peak_detector) and one without (http://www.afrotechmods.com/videos/peak_detector_circuit.htm).

Am I right in understanding that the op-amp design will give me a true peak, while the simpler design will give me a slightly reduced peak due to the forward voltage drop on the diode?

Yes, the op amp design is slower, but more accurate. The diode peak detector will have an offset due to the diode forward voltage of 0.6-0.7V.

The problem with the op amp design is the slow response time and slew rate of an op amp which will make it difficult to capture microsecond pulses.

One better approach for a high speed peak detector is to use a fast S&H circuit modified to perform the peak detect function. Look at Application No. 9 in this ap note.

 

Did you miss the part about the very short pulses?

Nope. My Fluke meters have "min/max recording" (which I mentioned above) and it's a very handy feature as it records the miniumum and maximum voltages, even if they are short pulses. I use it all the time for checking PSU rails, it will detect short term sags or overvoltage faults etc.

 

Nope. My Fluke meters have "min/max recording" (which I mentioned above) and it's a very handy feature as it records the miniumum and maximum voltages, even if they are short pulses. I use it all the time for checking PSU rails, it will detect short term sags or overvoltage faults etc.

That's pretty cool. Maybe my Fluke has the same feature (duh).
It must be an analog circuit, unless the Flukes (and others) have fast A/Ds.

EDIT: I have a Fluke 89. It does have FAST MN MAX. Events can be as short as 1mS, with decreased accuracy (3.5 display digits). Our OP wants to capture events on the order of microseconds.

 

Thanks for the info Ron, my Flukes are model 12 and I just went and had a look at the user manual.

As you said in MIN/MAX record mode it won't accurately register the voltage of very fast transients as it is much too slow. The "capture" function for capturing open or short circuits is much faster but not of much use.

It's beginning to look like a job for a 'scope or digital storage 'scope.

Maybe a fast diode into a small cap would charge up the cap to something near the peak voltage of the transients? There will be the diode voltage drop and some leakage issues but it could be workable if he just needs a ball-park value and not a highly accurate value.

 

One better approach for a high speed peak detector is to use a fast S&H circuit modified to perform the peak detect function. Look at Application No. 9 in this ap note.

That looks perfect, crutschow, thanks very much. Is there anything I should keep in mind when choosing an op-amp for the comparator?

 

That looks perfect, crutschow, thanks very much. Is there anything I should keep in mind when choosing an op-amp for the comparator?

The first thing to keep in mind is "don't use an op amp as a comparator". The built-in compensation of standard op amps makes them too slow when used as a comparator for this application.

You want a fast comparator IC such as one of these.

 

Ha. OK, great. Thanks for the recommendations!