DiY 2GHz Active Probe

Pretty sure most of y’all picked up on this when Hackaday featured it. Worth further discussion, IMO. Technically useful post here.
I will admit that I have a poor understanding of composite amplifiers. So a good bit of the feedback and compensation in-the-loop is unclear to me beyond qualitative terms. The BUF802 is a nifty-as-hell JFET opamp. Impressive, even.
I do find the OpenEMS sims particularly interesting, as I’m terrible at any sort of programming/scripting. Definitely using this as a jumping-off point for an L-band design I’m working on.
Don’t have a specific line of inquiry to propose, but figured this might touch off some more discussion.

1 Like

James Wilson was kind enough to give me one of his probes at Teardown 2024.

It’s an impressive design.

I haven’t put it through its full paces yet, but I captured this time domain screenshot.

The top (purple) waveform is a Rohde and Schwarz 1 GHz active probe and the bottom is a reference trace from Wilson’s probe. tldr; they’re nearly identical. In this case, I’m measuring the 12 MHz oscillator on a Pi Pico board.

What’s even more impressive is the FFT (Spectrum) performance, shown here:

I’m using an MXO4 (provided by R&S). There is almost NO noise introduced by the probe. -90 dBm is well into this oscilloscope’s noise floor.

At Teardown, I asked Wilson what he might charge for such a probe. If he decides to to turn into a product, it’ll be an order of magnitude less than the 1 GHz active probe I have from R&S.

Good to see some independent data. Not like I have much reason to doubt his data. That’s some impressive performance for sure. I’d like to build one and characterize it independently too. I have a decent VNA and specan I can leverage. Siglent gear, nothing to write home about, but adequate for my purposes. I wouldn’t leverage the full B/W of the probe with my current scope, but the miniscule input capacitance makes it ideal for some of my finickier measurements.

Wasn’t @Stephen.K working on or built an active probe a while back?

I was. It was a hogh voltage differential probe however. I wasnt trying to hit Ghz frequencies. That just sounds like a headache.

I’m naive enough to think the number of challenges would be the same, but the types of challenges would be different for a wide bandwidth probe than a high-voltage probe–especially if you’re trying to achieve high isolation.