Negative peak detector with OPA615

user676896

Other Parts Discussed in Thread: BUF602, OPA615

Please refer to Figure 48 of the OPA615 datasheet for a peak detector circuit. http://www.ti.com/lit/ds/symlink/opa615.pdf

If I need only negative peak detection ( between -1 to -5v ), Should I keep the the positive peak detection portion ? If I remove the poitive detecion section ( OTA etc ) , How to provide feedback to pin 11 of SOTA ?

Thanks

over 10 years ago

0 Rohit Bhat over 10 years ago

TI__Genius 15975 points

Hello,

I think if you need only the negative peak detection, you can remove the positive detection section and keep the BUF602 path. You can operate the SOTA stage by itself by having the inverting pin 11 connected to 200ohms to ground or having it connected via the OTA in emitter follower configuration. The SOTA stage is a comparator and does not necessarily need a feedback signal.

Best Regards,

Rohit

0 user676896 over 10 years ago in reply to Rohit Bhat

Prodigy 80 points

Thanks Rohit for the clarification.

Is there any suggested part number for the diode to be used ? Will the forward voltage drop across the diode introduce error in measured peak voltage ? Can I remove the diode as I need to detect only negative peaks? if we provide Logic high at HOLD pin , will the output voltage stay at peak without discharging ?

Thanks .

0 Rohit Bhat over 10 years ago in reply to user676896

TI__Genius 15975 points

Hello,

You can use a fast switching diode for your application, such as 1N4448W. What maximum input frequency do you need to detect the peak for?

I think the forward voltage drop across the diode could introduce an offset error equal to a diode drop of 0.7V in the measured peak voltage. I don't think removing the diodes would work for peak detect, since diode is the reason why the peak detector circuit works. The diode along-with the charging capacitor acts as a negative peak rectifying circuit.

I would think if Logic high is provided to HOLD pin, the output voltage should stay at its peak since the diode prevents the capacitor from discharging and the input impedance of BUF602 is also high. One parameter which you could consider for the selection of fast switching diodes is the low reverse leakage current to prevent the capacitor from discharging.

Best Regards,

Rohit

0 user676896 over 10 years ago in reply to Rohit Bhat

Prodigy 80 points

Hi Rohith

Requirement is to detect the peak of a 10nSec negative voltage pulse occuring at 50mSec intervals .

1N4148W forward voltage drop is 0.7 - 0.85V . So using a schottky diode with 0.3v voltage should provide lower error in peak measred . I will select a part with very low everse current spec too. Is there a PCB layout available for reference ?

Thanks

0 Rohit Bhat over 10 years ago in reply to user676896

TI__Genius 15975 points

Hello,

Although the schottky diode has a lower dropout voltage of 0.3V compared to regular p-n junction diode of 0.7V and has better switching performance, the reverse leakage current for the schottky diode would be worse compared to a normal p-n junction diode. You would have to choose between those two based upon the trade-off that your system can tolerate.

For PCB layout reference, you could refer to the user's guide for the OPA615 and BUF602 evaluation boards: or . Let me know if you need the gerber files as well as I can dig that up.