Vector Network Analyzer Probe Test Information

Hi, everyone. I have some concerns about using the probe to test the microwave system by network analyzer.

Since my design is at a few Giga hertz, so I cannot use the SMA connector for the network analyzer connection. I am gong to use the probe kits to test the circuit directly.

My question is what kind of probe kits should I use, and type of pads I should use in the microwave circuit design to attached the probe.

Any information about the probe kit, test procedure and design notes are appreciated. Thanks.

SMA connectors commonly can be used to 18 GHz. Related K connectors are good to 40 GHz.

The concerns you have may be in the transition from your device to a coax rather than the connector.

You do not provide details so any additional speculation on my part would just be a guess and might not be appropriate or helpful.

Hi, Azulykit. Thanks for your reply. Actually I am going to design an on-wafer dipole antenna array working on 60GHz.

The problem is that the dipole itself is balance structure (differential), so there is no clear definition for the ground.

I am wondering by using the probe station calibration kits with such as G-S-G pads, is it properly representing the balance(differential) feed, or I should design a balun to convert the unbalanced signal from the G-S-G pads to the dipole feeding.

If you truly have a differential feed, something like differential CPW or some variant of edge coupled traces (think of twin line) you have a balanced feed structure.

A G S G probe is like single ended CPW not balanced. I would think that you need a balun structure to convert to what I assume is a coax to connect to your VNA.

HFSS can help you get a handle on all this and examine your antenna, feed, probe and the various interactions. Getting into all this at 60 GHz makes it even more difficult. I have worked microwave and mm wave bands to 200 GHz but when I was at 60 GHz the transmission lines were (small) rectangular waveguides.

Good luck with your PCB.

Added after 5 minutes:



Just another thought: You might want to consider terminating one side of the line and using the G S G probe on the other side. You would lose 3 dB but may be able to get enough information about the antenna without the complexity of a more appropriate probe.

Sometimes a hack is good enough.

Regards,

Eb

Hi, Azulykit. Thanks very much for your information. It seems you are an expert in mm-wave range.

I am wondering whether you can give me more details about your last thought. Taking a edge-coupled traces as example, so how can I terminating one side of the line, shorting it to the ground or add 50Ohm resistor to the ground?

Is it the structure something like these：

－５０Ｏｈｍ－
｜　　　　　｜
｜　　　　　｜　
｜　　　　　｜
｜　　　　　｜
｜　　　　　｜
Ｇ　　　　　Ｓ　　　　　Ｇ

？　　　　

i can say just one thing termination means proper matched termination so you definately cant just short it to ground it would reflect whole signal with 180 phase change so use proper 50ohm temination

The hack I was thinking about was for a differential. grounded CPW. Remember, this is a hack and not good practice.

The line would be G S S- G all over a groundplane and via stitching. One might consider termination one side of the S S- and then connecting a single ended probe to the other side.

A real balun would be the responsible technique. I really don't think that this is good practice, but it might be a way to get a view into your circuit, even if it is a little corrupted.

Good Luck

Azulykit