The marvelous disappearing capacitor
lcamtuf.substack.comHave used that JFET trick on large area photodiode amplifiers in a satellite based radiation detector. Works quite well.
Also quite similar is actively driving the shield used in many high impedance sensor circuits (most notably capacitive touch sensing) to reduce the parasitic capacitance you would get when ground is used for the E-field shield:
Can a similar technique be used to increase the bandwidth of an amplifier driving an LED?
I think that if what you're looking for is to increase the speed at which you can drive something which has some parasitic capacitance, you would do that by using the techniques that are employed by e.g. gate drivers.
This is called a bootstrap circuit. It is basically positive feedback that changes the input impedance. You can see this sort of jfet circuit not only in trans-impedance amps, but also in a lot of audio circuits.
Digikey has a useful article as well. Note how in Figure 5 the PD is reverse-biased to reduce the capacitance without any extra components.
https://www.digikey.co.uk/en/articles/how-to-use-photodiodes...
I'd like to see an I-V plot for various levels of light, for the photo diode.
AoE and AoE X has a bit on TIAs as well, their "final form" circuit idea enhances the basic JFET follower idea with an additional emitter follower to reduce source impedance and by AC-coupling the compound follower's output they get to set favorable DC working points.
The tricky bit is going to find a suitable replacement for the BF862.
Phil Hobbs has also written a lot about photodiode bootstrapping, and he was also a very unhappy camper when the BF862 went EOL. Anyone working in this area should check out his book (see https://electrooptical.net/), which is very much in the same spirit as AoE. I used to read a lot of posts by Phil on Usenet via Google Groups, before Google decided they had better things to do.
For that matter, I have a feeling Michal (lcamtuf) has a book or two in him that could stand up next to Hobbs and Horowitz/Hill on the shelf. Super impressed with his work over the years. (Many may recognize him as the 'Guerilla Guide to CNC' guy, which has been linked on HN several times.)
CPH3910 may be worth considering as a BF862 replacement. I don't recall the part number but there's also a dual CPH3910 in a 5- or 6-pin SOT package that could be used with the transistors in parallel.
Depends on the critical parameters, but 2SK2394?
You can try to source old JFETs but all of the ones advertised online are fake.
It has been a couple years, but iirc the BF862 back then had a pretty unique combination of low capacitance, low e_n and a low 1/f knee. The 2SK2394 is also earmarked for use as an AM radio amplifier, same as the BF862, which probably also explains why these went from jellybean to rarity some time ago...
This reminds me of a somewhat similar circuit, the capacitance amplifier:
https://en.wikipedia.org/wiki/Capacitance_multiplier
Which leads me to think about planned obsolescence in electronic circuits. In many electronic circuits the electrolytic capacitors are the first to degrade, limiting electronics lifetime to 10+ years. What if those capacitors were replaced with amplified capacitors where the capacitor was not electrolytic, say a film capacitor. Maybe the circuit would work the same and last 20-30 years, eliminating a lot of electronic waste and reducing environmental carbon due to manufacturing? Part of the solution to climate change has to be manufacturing products that last longer. Electrolytic capacitors, built in batteries, fans, and parts made of rubber instead of silicone are some of the prime components that cause failures. I have a suspicion car companies have tuned steel formulations and paint to cause rust after a specific number of years as well.
Very cool! Can this trick be used for piezo sensors as well? My understanding is that those too are essentially magic capacitors that sometimes move charges on their own
Yes, you can. In fact bootstrapping is a useful technique for a variety of circuits involving some small input signal, and even large ones, like power supply design.
This article is a little weird in framing the concept solely around the author’s journey in photodiode circuits, but that’s blogging for you I guess.