RF Metrology

Sep 29, 2024 Last reply: 1 year ago 18 Replies

Greetings, gentlemen,



I bought one of those TinySA Ultras recently and have spent some time testing it today. I have to say it's amazing what they've done for the price. I mean, truly amazing. I do have two other 'proper' lab-grade RF spectrum analyzers, but I think I'm going to be mostly using the TinySA in future as it's just *so* convenient and doesn't weigh a ton. Anyway, to get back to the point of this post, having checked out the TSA and establishing I didn't buy one of the fake versions that are out there, I turned my attention to my mid-level RF SA, an HP who's model number escapes me (not the 8566B I've posted here about before which is now fully working, but a newer model that's about 1/3 as heavy. I say I can't recall the model number but it's not relevant to this question anyway. "So what is the f****ng question, CD??" I hear you not unreasonably cry. Well, it's this: When I'm feeding an RF signal into the SA, I'm seeing differing amplitudes at different frequencies. So I've programmed in a sweep from 10Mhz to 5.4Ghz at -49dBm using my Aeroflex RF signal generator and I'm seeing the displayed amplitude vary as it sweeps through the range. But this only happens when I'm using a cheap, Chinese N-type to SMA adaptor at the signal generator output. Would I be right to suspect some imperfection in the manufacturing of the adaptor could cause such an effect? I do have a VNA I could characterize the adaptor with but it's a bit of an effort to do. It would seem like the SA is showing the adaptor's shortcomings in the frequency domain. But is that a feasible hypothesis?



Your pal,



CD.


Hard to tell. How big are the whoopdedoos, and roughly where?

Posting a screenshot would help.

Cheers

Phil Hobbs

What specific tiny SA did you buy? How's the image rejection?

Connectors can cause passive harmonic and IM distortion. That's a big deal in cell phone towers.

Aren't we getting a teensy bit lazy?

Jeroen Belleman

I have seen transmission (S21) losses of around 1dB at about 1.2GHz just from failing to tighten an N connector sufficiently. We do need to know the magnitude and frequency of the problem in order to give helpful comments! John

Not really. If you just think of it as a slight impedance mismatch in the cable between the SA and the signal generator. However, this model has the facility to be able to save files of the screen display (although I very much doubt it'll save a real-time video of the sweep). And John - it's the 'Ultra' version. I already had the base model and this is a worthwhile upgrade. Jeroen - it's more a problem of available time!

I collected some TDR traces of various connectors, adapters and terminators a few years ago:

formatting link
. There were a few surprises. I was taken aback by the poor results for H+S BNC and LEMO-00. For BNCs, Radiall is better than H+S, but for N-connectors, it's the reverse.

After having seen what LEMO-00 hardware did to my signals, I was pleasantly surprised by the performance of SMC.

SMA was the best I had around, hands down, no surprises there.

Jeroen Belleman

Ultra what? Got a link?

SMBs are surprisingly good, and easier to use.

Shining Star makes nice, crazy cheap, edge-launch SMAs and SMBs.

formatting link

formatting link
(sorry, no time to use tinyurl at the moment!)

Yes! That is what I believe is happening. Nothing much occurs between

1Mhz and 100Mhz, but it becomes more and more apparent as I go from 1Ghz to 2Ghz, 2 to 3, 3 to 4 and so on - as does the attenuation. You can't expect miracles from cheap Chinese crap, but even crap works well up to high VHF/low UHF.

It has its own home page:

formatting link

Measurement of good vs bad cables:

formatting link

Very precient of you, Klaus. I was just about to ask about others' experiences of cheap Chinese RF connectors. I'm reluctant to hook them up to my expensive Western analyzers in case they get damaged by poor fit or some such issue. I have to say, though, that the only issues I've heard about in this regard are damage to the Chinese connectors' center pins (being pushed in) during insertion. I could live with that, so long as my expensive test equipment remains unharmed.

I solved a billion-dollar showstopper last year. Some country started demanding that some expensive semi fab gear be EMI tested, and our units failed. After great hassle we traced it to bad ground crimps on shielded CAT5 cables that our customer bought cheap.

I bought a set of phase-matched cables from Taiwanese company JYEBAO because my usual supplier Huber & Suhner failed to answer a quote request. I was pleased to find that JYEBAO's cables were *better* than H+S, and cheaper too.

Jeroen Belleman

So ar, the generic cheap Amazon SMA and SMB cables have been excellent. As have been the edge-launch connectors.

The Shining Star connectors are great too.

Probably made next door to each other, or in the same sweat shop on different days.

RL

I don't think so. I compared H+S matched sets with JYEBAO's and there were clear differences. The JYEBAO cables were clearly better matched, and of overall better quality too.

Another positive experience I had with them concerned the delivery of cables with a 75 Ohm C connector at one end. Nobody makes 75 Ohm C connectors anymore, but they delivered.

Jeroen Belleman

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required