Tuning RF circuits without a Network Analyzer?

[Mosaic]

BTW, how important is calibration when VNAs and SAs normalize conditions?

http://cp.literature.agilent.com/litweb/pdf/5989-4840EN.pdf

 

[Mosaic]

This is interesting...a new 3Ghz VNA for $1300.

 

[RadioRon]

BTW, how important is calibration when VNAs and SAs normalize conditions?

http://cp.literature.agilent.com/litweb/pdf/5989-4840EN.pdf

Hi Mosaic. I hesitated to answer because I don't think I understand exactly what you mean about "when VNAs and SAs normalize conditions?". However, I would comment that VNAs make relative measurements. In other words, they will measure the amplitude and phases of voltage and current and then compare them to something. The "something" that they compare to depends on location along the length of the test port. If you do nothing but turn the instrument on, then the default comparison reference uses voltages and currents measured at the face of the coax connector at the test port and recorded at the factory. Since we rarely want our "reference plane" to be at that same point, it is almost always necessary to perform a field calibration to the reference plane of our choosing. This has the benefit of removing the losses and phase rotation that occur in the feed cables between the test port of the instrument and the point in your circuit that you are truly interested in.

Spectrum analyzers, being devices that measure amplitude only, are similar but the situation is much simpler so field calibration is not quite as critical. The displayed value is referenced to the input connector, and external losses should be accounted for to make an accurate measurement. In many cases, however, the external losses from the coax cable are low relative to what we are interested in and so are often ignored. So, most times we don't bother with field calibration of the SA.

I hope this is helpful.

 

[Mosaic]

Thx Ron,
I have come to see the need for a proper cal kit at the end of quality ,short, cables connecting the DUT to the VNA. I have an 8753D VNA coming to me in a week or so...and I am talking to Dr. Dave Kirkby of Kirkby microwave in UK for a custom cal. kit with phase delay coefficients and known propagation calibrated adapters.

The normalize I was thinking of had to do with the way the SA can compensate for a non flat tracking gen. sweep.

 

[RadioRon]

OK, I see what you mean about normalize. This is, of course, inherently automatic in the VNA when calibration is done because it is doing the cal across a frequency range.

I wanted to mention that over the years I've seen a lot of different test cables used with VNAs. In some applications the cable becomes extremely critical and in these cases people buy specialty test cables which can be enormously expensive and overly thick, stiff and heavy. I don't like using such cables. On the other hand, just using cables that you crimp yourself, or perhaps those that you buy at reasonable prices from others suffer sometimes from premature failure of the connector (usually the crimp to the ground braid) and can also introduce phase errors as the cable is flexed. My favorite cable these days for VNA work is the "Minibend" line of cables. Minibend is a compromise of cost vs convenience as they are a bit pricier than common coax jumpers, but they have been a pleasure to use compared to others. These were made by Astrolab, a company that was bought by Huber&Suhner in 2012, so Minibend is now sold by H&S. They make a plain vanilla model called Minibend, and also models with improved phase stability (important for UHF/uWave VNA measurements) including Minibend L and Minibend CTR. I recommend these cables for VNA measurements.

**broken link removed**

Edit: wanted to add that the Minibend cables that I have been using have lasted a long time, so the initial $60 or $100 or whatever they cost (buy two, one for each port) doesn't seem so painful now.

 

[Mosaic]

I was thinking of doing a cal at the VNA connectors and then doing an assessment of self built cables S12, S21 with the VNA . In so doing I could characterize the cable itself and then have viable cables if they are good?
I have RG400, Rg142, LMR400 cables and ends to work with. Clamp and crimp.
My range on interest is to 3Ghz, perhaps 2' cables, SMA connectors.
Which cable should I go for?

https://www.digikey.com/en/product-highlight/a/amphenol-rf-division/cable-connectors

 

[RadioRon]

By all means start with what you have. However, the problem I have with the cables with the size of RG400, for example, and especially given that this one is double shielded, is that they are too stiff and heavy. The cables dictate where my test board must go on the bench in that case, and I don't like that. But thats just me. The cables I like to use are smaller diameter than 2 mm and as flexible as possible. This suits the kind of work I do, where I might have a test board on my bench with all kinds of DC and other hookups and I don't want to move it over to the VNA to make a measurement. So I keep the VNA on a rack beside the bench and use 3 foot cables to reach my test board. However, if the cables are heavy they will pull the whole board right off the bench. So I've migrated to the narrow types of coax. Another reason I like Minibend.

 

[Mosaic]

Some useful links:
https://mwrf.com/test-and-measurement/evaluating-impedance-transformers-vna
**broken link removed**
https://mwrf.com/components/tackle-wideband-rf-switching-pin-diodes