Diode Clamp, Has Anyone Ever Used This?

Has anyone ever used this circuit?
What are it's drawbacks?
What style diode have you used?
What type of sensitivity have you been able to achieve?
Is something like this temperature stable?

I have an AC coupled input that must accept pulses that vary in amplitude from a few hundred uV to a few hundred mV. and varying duty cycle.

That circuit only works for signals that are greater than the forward drop of the diode (≈0.4V for silicon Schottky diodes which have the lowest forward drop).

If the signal maximum frequency is no more than a few tens of kHz, then you could use an op-amp full-wave precision rectifier to detect the signal.

For high frequency signals you could use a tunnel/back diode.

To your questions: all depends on your specifications. Frequency and power will define diode type, also source and load impedance.
I used two anti-parallel diodes like 1N4001 across loudspeaker terminals, to get rid of annoying high-level interference It works very well with 8 Ohm speakers.

Diodes are inherently temperature-dependent. See above recommendations for a good design.

- - - Updated - - -

For the diodes: consider the typical knee voltage of o.7 V for Si diodes, 0.35 V for Ge and GaAs diodes. Zero-bias Schottky may have <0.2 V. Tunnel diodes even lower but tend to oscillate.

crutschow said:

That circuit only works for signals that are greater than the forward drop of the diode (≈0.4V for silicon Schottky diodes which have the lowest forward drop).

If the signal maximum frequency is no more than a few tens of kHz, then you could use an op-amp full-wave precision rectifier to detect the signal.

For high frequency signals you could use a tunnel/back diode.

Click to expand...

+1 go with the full wave precision rectifier

if your input is a series of pulses, just use a wide bandwidth OPAMP to give you pulses of Vcc amplitude. DC clamps (and rectifiers) suffer from a shift in DC as the storage capacitor charges. A big pulse will charge the cap, which will then stay charges and miss the next pulse.
Frank

chuckey said:

if your input is a series of pulses, just use a wide bandwidth OPAMP to give you pulses of Vcc amplitude. DC clamps (and rectifiers) suffer from a shift in DC as the storage capacitor charges. A big pulse will charge the cap, which will then stay charges and miss the next pulse.
Frank

Click to expand...

Why would there be a storage capacitor ? As I understand it the OP only needs to rectify the signal. not smooth it ?

jiripolivka said:

...I used two anti-parallel diodes like 1N4001 across loudspeaker terminals, to get rid of annoying high-level interference It works very well with 8 Ohm speakers......

Click to expand...

Interesting concept. How exactly does this work ? Seems like this would clamp your signal to +/- 0.7v or so, and that too in a non-linear way. So would actually ADD to a distorted output from your speakers if your output was anywhere above 30mW.

Or did you mean headphones ?

Rmorris, you are correct, there need not be any explicit storage cap, but the OP said he is feeding these small variable pulses into an AC coupled amplifier (why???), which I took to mean that he has an input coupling cap. The DC on the other side of this will bounce about depending on the pulse amplitude and duration. its best that the DC on the input side of the coupling cap remains constant.
It would be nice to build a log amp so that very big pulses do not cause any latch up in the opamp, so to reduce the 10:1 pulse amplitudes to ,say, 2 :1.
Frank