Electronics > Projects, Designs, and Technical Stuff

PDN impedance test board - what to include?

(1/1)

jmw:
I want to make a PCB for familiarizing myself with the two-port shunt-through method for measuring power distribution network (PDN) impedance. I've made a common-mode choke transformer by winding coax around a ferrite that's a hopefully a good clone of the Picotest J2102B. This evaluation board would have several isolated modules, each one with two neighboring coax ports, and then some connected network to simulate the PDN.

What kind of configurations should I put on this board? So far, I was thinking of the following:

* precision shunt resistor, like 1 or 10 mΩ, for testing the effectiveness of the common-mode choke
* bunch of 0603 + 1206 footprints, connected with traces only, and at various lengths away from the pair of ports to visualize importance of placing capacitors close to bypassed parts, use of decade capacitors, and effect of tantalum/higher ESR caps
* same as above, but with power and ground planes
What else?

For coax connectors, I was thinking to use compression-mount SMA connectors. They're expensive, but if this board has many sections that need two ports each, it may end up being cheaper than buying several SMT or through-hole connectors, and I can just unscrew and move the connectors around to make measurements.

Picotest app note about the method: https://www.picotest.com/measurements/images/download/the%202-port-shunt-through-measurement-app-note_REV1_073118.pdf

jmw:
This is sketch of the layout I'm thinking of: put one capacitor ACAP between the connectors, then have several open footprints at various distances. Haven't yet added via stitchings or a version with power and ground planes.

EEEnthusiast:
You might want to make that center conductor a bit more thicker. The goal here is not to reach 50 ohms impedance, but to keep the series impedance of the probe as small as possible. This helps when you are measuring PDN impedance of few milli ohms.
I assume that you have an SMA calibration kit and the calibration would be performed till the end of the cable. That still leaves out the SMA connector from your calibration. It will add some uncertainty, unless you de-embed them as well.
It would be worthwile to have a coupon with the 2x thru, which enables you to de-embed the SMA to SMA structure.
Compression mount SMA always adds some milli ohms of contact resistance at low frequencies. Use soldered connections to avoid this contact resistance.

Navigation

[0] Message Index