Precision Peak Detector

Thanks Thomas,

I'm going to give this circuit a try. By any chance, do you think this would work with one of TI's TL062 op amps? I ask this because I happen to have them on hand!

Alex

Thomas,

Thanks for the quick reply. I've put this circuit together in TINA TI. However, simulation indicates that this circuit does not work well for higher frequencies. My end application uses a 300 kHZ sine wave. When I simulate this, the circuit fails to work acceptably.

Do I need a faster switching diode? Perhaps I need to go with some kind of half-wave rectifier? Do you have any ideas on this?

Thank you,

Alex

Thomas,

Thanks for the reply. I searched for and found another circuit that appears to do exactly what I want it to. I was able to get it to simulate "well" with the THS4221 op amp. I'm probably going to sample the THS4222 (dual version), and see if it works well in reality. Let's consider this case closed for now.

Thank you,

Alex

Will the OPA211AIDR work in this circuit? Working on a full-wave rectifier that can operate between 100KHz-1MHz.. Our system uses 100KHz but we want to catch any spikes (peak detection) which we calculate to be about 1MHz.

Hi Tony,

The absolute value circuit can be made to work in the 100 kHz to 1 MHz frequency range, but a wider bandwidth, faster slewing operational amplifier than the OPA211 must be used. And you must be willing to burn some current to get the higher speed operation.

I've redesigned the original absolute value circuit using the fast OPA354 CMOS operational amplifier. See the circuit shown below - the TINA-TI Spice file is attached. I used two OPA354 amplifiers in my simulation, but the dual, OPA2354 could be used instead. I expect that higher bandwidth, faster slewing operational amplifiers will produce a higher accuracy absolute value function.

Regards, Thomas

PA - Linear Applications Engineering

For the most part the circuit works great. There are some things that I have questions about.. I am using the OPA23254AIDGKR. I have it on a surf board with components wired out. When ever I turn on V+ (5V), the output jumps to V+. We want it to stay at 0V. I was thinking that maybe somehow V+ was coming through the diode D1 and putting a cap there to filter the DC would help.  What do you think? 

 6470.OPA354_No_wave_forms.TSC 

Hi Tony,

There isn't any direct path to V+ through the OPA2354 amplifier section having the diode in the feedback path. I expect that adding a capacitor anywhere in the signal path would disturb the ac performance. I tend to think the problem may be related to a start up timing relationship between V+, the mid-supply source (VS1) and possibly VG1. Do you have VS1 bypassed with a capacitor? I didn't include one because the sources in TINA can be set up as ideal.

You may want to put a scope on the supply line, mid-source supply, signal source and outputs during power-up to see if there is a condition that is forcing the output to V+. How do you get it out of the output lock-up condition?

Thomas

PA - Linear Applications Engineering

I built the circuit as per schematic. So no capacitor on the input. I read the V+ as soon as I turn on power. And that is without the input signal. I have been experiment with a cap on the output. Which works great but the output signal is now distorted. The cap that seems to work best is 4700 pF. Again this is happening with or without the input signal connected. Once it is turned on it functions great. I just need to get rid of it so not to get a false error at start up. 

Made any head way on this?

Hi Tony,

The absolute value circuit that I provided is intended to be driven by a low impedance source; the closer to zero Ohms the better, but a 50 Ohm source impedance should be fine. If the source impedance is very high, or infinite like an open circuit, the OPA354 output will scream up to its positive supply rail level. If that is the case, try placing a resistor from the union of R1/R2 to the +2.5 V high side of  VS1. This will place the inputs of U1 near +2.5 V and keep them from following the output to the positive rail. Also, the VS1 connection at U2's  non-inverting input must be properly bypassed to ground with a capacitor. Something around 100 nF should be a good starting value.

If you find you just can't get the circuit I recommended operating as needed, you may want to consider some other designs. TI has an application note, SBOA068, that provides some alternate absolute value circuits. However, you will have more experimenting to tell if any one circuit meets all of your criteria.

The SBOA068 applications note is available via the TI website.

Regards, Thomas

PA - Linear Applications Engineering