Why a Solder-In Probe ?
When frequencies become higher and higher, probe lead inductances become more and more of a problem. The low frequency technique of alligator clips to ground and hooks to the test point simply dont work in a GHz environment. What you need is an active FET probe that imposes (almost) no load to the circuit, and you dont want any leads – welcome the active FET solder-in probe.
The requirements were:
- 26dB scale factor in 50Ohms (1:20)
- At least 150V DC isolation
- AC coupling. Frequency from some 10kHz to 1.3GHz
- Small enough to go into an actual circuit, with flexible leads
- Cheap (a few €).
The circuit used here is the same as in my other active probes for the range, its a BF998 dual gate MOSFET impedance converter, with a “gimmick” capacitive divider at the front end. The PCB design is all Christophs work, and he also made the first prototype using perfect SMD soldering (which my eyes and hands are too old for). The details can be seen here:
The Design of the Probe
The probe mechanically looks like this:
Its about 1.5x2cm in size, and it uses UFL connectors for power and output signal. For some shielding a metal can sits on top of the circuitry. Touching the cap does change the frequency response somewhat, but only above the passband, so that does not matter much.
The frequency curve is quite OK, and fairly flat until the ca. 1.3Ghz -3dB cutoff frequency:
The measurement was done on my trusty probe test fixture (PCB also designed by Christoph).
Up to 1GHz, flatness is exceptional (less than +/- 0.5dB). I like it !
Now lets try some rise and fall time measurements on a 6.3GHz scope and a 40ps Bodnar pulser. Results are here:
Quite what is expected. 1GHz would be 350ps, so we run a little below that.
We could try some less feeble connector than UFL. But on the other hand, these have the thinnest and most flexible cables we could find.