500KHz peak detector

[ricko_uk]

Hi,
I have to make a peak detector of up to 500KHz. Using some of the classic configurations require a more expensive (high speed high slew rate op-amps).

I have come across a comparator-type peak detector below but it start introducing error at 100KHz

Any other suggestions for higher frequencies up to 500KHz? Or at the absolute minimum 250KHz?

Many thanks

 

[graybeard]

what is the range of signal amplitudes?

do you need both positive and negative peaks?

what are the source and load impedances?

you need to define your requirements more completely

[vk6zgo]

An old style "peak & hold" detector can work up into UHF, but may have limitations, in that very fast peaks in the midst of a lower frequency signal may be missed.

If you need to detect rarely occurring transients, you are pretty much stuck with doing it digitally, & will need a very high sample rate, plus a lot of memory.

If the transients you want to see are very infrequent, maybe the device could check what it has "seen" for a set previous time, flag that as "no transient found", empty & re-use the memory.

[Circlotron]

Precision AM detector?
I have a circuit at home I'll post later.

[ricko_uk]

Thank you Circlotron and Greybeard,

Circlotron, look forward to it. Thank you!

Greybeard,
- signal amplitues as mentioned from few tens of mV to 4.5V. It depends on the configuration for every specific application. Most cases are between 1.5 and 4.5V.
- the signal is as just a fixed sinewave
- detection of both positive and negative but no need to differentiate (i.e. a full wave rectifier would solve that)
- source and output impedances are buffered with op-amps so generally speaking at this stage no major issues there that I can think of

Thank you both

[Circlotron]

Here it is.
Hope it serves as a good starting point.