RF probe for network analyzer measurements

[Leiser]

Hi,

I´m working with small RF amplifiers designed for production. That means that when I have to take lab measurements of intermediate stages, where there are no cables, I have to solder on the PCB coax cables all the time. I was wondering if anybody could advise me on RF probes or something similar to a probe which I can connect to the network analyzer so I can put the signal at any PCB point without soldering a cable. (like an osciloscope probe but for a network analyzer)

Thanks a lot.

 

[logbook]

This is a common and difficult problem, especially difficult in the GHz range. It is nice to have packaged amplifiers with SMA or K connectors so you get well defined input and outputs.

If you just solder on a coax you have added another 50 ohms in shunt at that point in the circuit. If you disconnect the load so you only have one 50 ohm load you then have an open circuit stub driving the open pads and you have changed the load conditions on the amplifier under test. What you want is a coplanar waveguide output so you can make signal and ground connections easily to the amplifier. If nothing else you should be able to get a repeatable measurement, although the absolute accuracy is questionable.

You can get wafer probe stations for microwave circuits. Take a look at those for clues.

 

[biff44]

If you need to test each stage for some reason in production, then you should design in some sort of test point for doing that. A 20 dB directional coupler in microstrip with the coupled port attached to some sort of gpo (snap fit) surface mount connector would do as a probe point.

You turn on the amplifier chain, and you can probe the output power of each stage, minus the 20 dB coupling.

 

[biff44]

It might be possible to design some sort of loosly coupled air-line directional coupler in a test fixture to lower near to the amplifier PCB. You would have some sort of coupling line approximately quarterwave long, and lower it close to the microstrip line at the output of each amplifier stage. If you lowered it only close enough to get maybe 25 db down coupling, it would not seriously impact the interstage tuning. Probably an air gap of maybe 3 substrate thicknesses or so?

Of course, You would have to carefully calibrate it somehow to make sense of the data. Also, you would need some straight lines interstage to mate up with the coupled section.

 

[Higgler]

What's your frequency range?
Common is a 3 point probe, it requires you to bring up ground points on both sides of your microstrip line. The 3 pint probe grounds touch these points and the center of the probe is on your transmission line.
At higher frequencies, you need to break your line, test your items, then solder the lines back together.

You could just add the grounds and make measurements on a good unit and record the data as a standard, then compare to the standard.

Tricky measurement if you're winging it. Hard to get the pass fail requirements with the probe some times.

kch