Question on peak detector circuit

Hi David,

It seems that you have not attached the circuit. Is it possible to re-post the attachment?

Best Regards,
Rohit

Slowly learning how to use the system. Here's the schematic...

U11 doesn't seem to get feedback until after U12. Why not design each stage as a working subsystem. U11 should purely be a peak detector and U12 should purely be a buffer. Right now they are mixed and unless I simulate it can only speculate that this will be unstable.

Try simulating with a similar op amp in LTSpice (I guess TINA-TI if you like it but I find LTSpice much easier to use). Just use a basic op amp in there, no need to use the LMP7715 to validate the schematic above, I don't see why any op amp wouldn't be able to do this.

Would you mitigate pulling them apart by directly attaching R17 to the non-inverting input of the peak detector U11 and providing current limiting resistance so that when the ideal diode function pulls the inputs of U11 apart you prevent a damaging amount of current from flowing?

Hi Bryce,

In looking for an OpAmp that would meet the differential signal requirement, I came across the OPA365. If you go to the bottom of the first sheet of the data sheet, you'll see the circuit that I wanted to use. The data sheet shows a small cap to the negative input. I'm guessing that this speeds up the circuit a little, but it's essentially the same.

I'll need to wait until I get parts in, but the little I was able to test, it appears that the hold-up time for this circuit was much higher than for the case of the simple peak detector followed by a voltage follower. In either case, now that I think the problem is solved, the hold-up time should be more than enough.

Just another comment: I found that the reverse leakage specs on the MBR0520 were not very good. So, I'll be going to a diode with a lower leakage specification.

Thanks for your input,
Dave

The OPA365 looks like it should work.

They are optimizing for higher speeds with the capacitor to the inverting input. Check out this app note (hopefully no one minds non-ti ap notes, couldn't find a TI one quickly, let me know if this is an issue!)
cds.linear.com/.../dn61f.pdf



As for the MBR0520 yeah, that diode is overkill at 0.5A forward current rating. the power doides will have much higher reverse leakage in general.

Hi David,

Glad you found a solution.  The OPA365 is a good choice.  Consider the IN4148 for the diode, it doesn't have the lowest forward drop, but does feature low (~25nA) reverse currents at reverse voltages of 25V. 

Bryce,

To answer your question from earlier, you're correct and series resistances could be added to prevent damage to the amplifier.  However, the amplifier is placed into an undesirable state any time that the absolute maximum ratings are exceeded and the circuit performance would likely be much worse than a circuit with an amplifier that didn't contain the diodes.  Thanks for helping!

Hi David,

As Collin pointed out, the LMV7715 does have diodes across the inputs.

The peak detector circuit in the LMC6482 datasheet uses a second diode that keeps the first amplifier from going fully open loop. This should help.

Also, try using a BAT54 Schottky diodes instead, less voltage drop.

Regards,