How about the AD869x series.

@You Shengquan: Thanks for Your suggestion. This seems like a good chip for many high frequency biomedical applications. I did make amplifiers for intramuscular EMG amplifiers and similar maybe 10 years ago and then used then contemporary state of the art amplifiers from Linear Technology.

To give some background to or my present requirement, I have not found any solutions to the requirement of low input capacitance combined with wide bandwith fulfilled by any integrated circuits that I have found so far. Some low input capicitance amplifiers are described in older LT app notes, like LT45p2. Bob Pease has a circuit in his book "Troubleshooting analog cicuits" that works well as a active probe. I have saved lots of money over the years using these instead of expensive Tek probes. I need a single ended amplifier, but would be very interested in good ideas on low c input bridge amplifiers as well, since this has been a problem in the past for me. Such a circuit is much more difficult to design than a single ended circuit and a main reason why cpapacitive sensors often are realized as half bridges.

The published circuit that comes closest to my present requirements is a bootstrapped FET amplifier published in "Capacitive Sensors p 162" by Larry K Baxter. I think its on Google books. This circuit can have wide bandwidth and an input capacitance of around 0,005pF can be achieved through positive feedback to the input guard. I think guard driving is a necessary technique for such an amplifier. I have a specific application, but such an amplifier is of much general interest in capacitive sensors, high voltage dividers, biomedical microelectrode signals and lots more. I am surprised that so little is published and that so few components seem to be available for these purposes. I can actually use a narrow bandwidth around my frequency of interest, which could help S/N but it seems all solutions I found so far are broadband amps.

Please contact me [email protected]. Making capacitive sensors is basically all we do. Practically all the guard amps we have designed and use meet this spec.

See also this exotic circuit idea...

http://www.philbrickarchive.org/1964-1_v12_no1_the_lightning_empiricist.htm

... and the explanations in the Wikibooks story below:

http://en.wikibooks.org/wiki/Circuit_Idea/Revealing_the_Mystery_of_Negative_Impedance#Generalizing_the_N-shaped_negative_resistance_idea

@Cyril I must say I enjoy the style of these old documents as the Philbrick ones. Very informative no BS and with an air of optimism and enthusiasm...

Much the same spirit as in the later writings of Bob Pease and Jim Williams. I must highly recommend reading their works to younger folks in EE who actually want to learn design and construction of working circuits, not just simulating them.

The capacitive feedback amplifier in that document is actually the same principle as what presently usually named low-Z amplifiers. It can ideally null out the input capacitance. It is perhaps one of the best principles for building low frequency amplifiers for very low capacitances. But for faster speeds, the capacitive load and drive capability and also stability of the OP comes into play.

For higher frequencies driven guard amplifiers are likely a better solution since the summing point does not load the output of the amp so much.

My post was mainly concerned with the problem of how that technique can be improved by limiting BW to the pass-band, improving guard driving, choosing recent components to increase performance etc. It seems that sinusoidal drive and using the highest possible carrier frequency is what is required if optimum dynamics and measurement signal BW is the goal. I would of course be happy to learn if there are other viable alternate techniques as well.

An also if there are any solutions to stable high CMRR bridge amplifiers with very small input capacitance. (Which has been a recurring design problem for me).

LT1807 has 2pF input but has low impedance input

Unfortunately to circuits that cancellates stray capacitance is its dependency with temperature