That's a horrible circuit, but it would be improved with better sampling diodes. There are 0.25 pF diodes around, like SMS7621 or BAT15.
Classic samplers had blowby compensation to cancel signals leaking through the diode off capacitance.
I never has much luck with full-bridge samplers. Half-bridge worked for me, whit short impulses as the sampling gate. I did a theroretically 40 ps 1bit sampler/TDR which I never finished testing, but that's an easy way to go fast.
I am sorry, my brain is slow today, that should have been the first thing I looked at. For once I have an application that allows for more expensive components :-)
I litterature they state that the 2 diode sampler is not optimal, the 4 diode is better. Do you have an alternative 2 diode sampler circuit?
Cheers
Klaus
In classic times, only slow (1 GHz range) samplers were full-bridge, and faster stuff was 2-diode. The bridge samplers were usually pulsed by an avalanche transistor or a step-recovery diode, usually through a balun.
A 4-diode sampler would be OK, but the sampler time constant, using a rectangular gate, limits bandwidth. A short-impulse sampler is fast, but gain ("Sampling efficiency") is low so usually has an integrator and bias feedback loop.
Gotta get going now. I'll see if I have stuff later.
The old Tek sampler manuals are online, and one of the Tek concept series books is about sampling.
I have a deconvolution program that takes an ugly sampler and makes it into a pretty one. It could increase the bandwidth of a slow-gate sampler but wouldn't fix blowby.
Writing as an idle idiot bystander that has never made a sampler but am interested. Just wondering if it could be a balance issue - like do you need to trim one side of the bridge drive resistors 150 bzw 1k to null imbalances?
piglet
No one in their right mind could ever accuse you of all people of being any such thing, Erich. Don't be so self-deprecatory!
-- This message may be freely reproduced without limit or charge only via the Usenet protocol. Reproduction in whole or part through other protocols, whether for profit or not, is conditional upon a charge of GBP10.00 per reproduction. Publication in this manner via non-Usenet protocols constitutes acceptance of this condition.
Very kind! Just regretted possible waste of bandwidth on a silly idea.
piglet
I compiled the two 4 diodes sampler circuits I tried out (turned out the low capacitance diode was not much better than I had first off)
First circuit is identical to the one from the link I presented first
It does not match the input voltage (could possible just be calibrated out, allthough calibration from get go is maybe asking for trouble)
Second circuit is an ac coupled version. It does not sample to the correct value at first sample, but since I am doing repetitive sampling (periodic), that is tolerable. It matches the input very well after 3 samples
Cheers
Klaus
Have something to add? Share your thoughts — no account required.
Ask the community — no account required