Cables, Connectors (N/SMA/BNC) & Adapters with Measured Results for Comparisons

[Rubo]

Also compare the prices to the Sucoflex 104 or higher, and they're definitely worth it.

No arguments there 

[KungFuJosh]

I'm tempted to add support for the SNAs to TestController for sweep comparisons.

Do it 

 

I probably will. I've been procrastinating a driver for the Hioki IM35XX LCRs that I'll probably do first.

Do you have an SNA?

[tautech]

I'm tempted to add support for the SNAs to TestController for sweep comparisons.

Do it 

 

I probably will. I've been procrastinating a driver for the Hioki IM35XX LCRs that I'll probably do first.

Do you have an SNA?

Show AI the SNA programming guide and some TC code and set it loose. 

[KungFuJosh]

Show AI the SNA programming guide and some TC code and set it loose. 

No need, I know how to do it. I've tried using AI in the past with it, and sometimes it helps a little. Mostly it screws up and guesses stupid shit and tries to force it to be something else. Some things AI is good at, and everything else makes shit take longer just correcting the AI slop.

[Rubo]

Do you have an SNA?

Not as of right now (but you never know  ).
I have access to HPAK stuff and have the Agilent E7495A & the SV4401A here.

[KungFuJosh]

Do you have an SNA?

Not as of right now (but you never know  ).
I have access to HPAK stuff and have the Agilent E7495A & the SV4401A here.

Oh man, I'm looking forward to the points from that when you do get an SNA and blame me. 

The driver I (might) make for the SNA series would only work for that series. Your HPAK stuff would need their own drivers (if somebody didn't already make one). I do have an SV4401A, but I don't think I'd make a driver for it, unless it's close enough to the SNA programming (assuming both use standard SCPI stuff). I haven't checked into either at all yet.

[Rubo]

Oh man, I'm looking forward to the points from that when you do get an SNA and blame me. 

No blame will be assigned  . I've been on the fence for quite some time, but I keep buying used gear  (just recently got a non-working 53220A to see if I can fix it).

SV4401A is just a toy, not even sure if it has ANY support for SCPI???

[KungFuJosh]

Oh man, I'm looking forward to the points from that when you do get an SNA and blame me. 

No blame will be assigned  . I've been on the fence for quite some time, but I keep buying used gear  (just recently got a non-working 53220A to see if I can fix it).

I'm still gonna collect the points in the TEAS war.

SV4401A is just a toy, not even sure if it has ANY support for SCPI???

Yeah, definitely a toy, but kinda fun. No idea if it has SCPI or what protocol it uses over USB. Not really exciting to explore with that one.

The UI in the SNA series needs some more development IMO, but it's generally awesome and fairly intuitive.

[Rubo]

I might have missed it, which SNA do you have?

[KungFuJosh]

I got my feeBay Sucoflex 104 cables in today. One of them is definitely trash, the other not sure.

Here's the first one... There appears to be a defect at one end of the cable, and I'm guessing I had the wrong end in Port1 for the tests.

This one is getting returned.

[KungFuJosh]

Here's the 2nd cable...

It was listed as N to N, but it's obviously N to SMA with an N adapter. Is the adapter why the DTF spike is wider?

With the adapter, it's not better than the M&P cables.

Should I keep this one or return it?

[KungFuJosh]

I might have missed it, which SNA do you have?

SNA5002A with the 8.5GHz upgrade and the other options.

[Rubo]

I would return both. Your M&P are performing better.

That's why I don't buy used cables from eBay.

SNA5002A with the 8.5GHz upgrade and the other options.

Nice 

[KungFuJosh]

I would return both. Your M&P are performing better.

That's why I don't buy used cables from eBay.

I mean, this one has that adapter on it, and that might be the reason it's performing poorly...and I only paid $22 total including shipping. I'm gonna test with another adapter. 🤔

ETA: The adapter is a Pomona 4297, which costs $37. Even if the cable is terrible (it isn't thaaat bad), it's probably worth it for the adapter if it isn't bad.

[KungFuJosh]

This is that better Sucoflex cable with the Pomona SMA-N adapter on it. It's very sensitive to bends, but this is the best I got with it so far.

The other end also feels like SMA-N but is probably original to the cable (under heat shrink). Is that normal with these, or is this a weirder cable?

[Rubo]

The other end also feels like SMA-N but is probably original to the cable (under heat shrink). Is that normal with these, or is this a weirder cable?

The ones I've see have a direct N connector crimped to it, no adapters under the heatshrink.

What I don't like is the S11 plot. S21 will depend on the lenght of the cable. How long is it?

[KungFuJosh]

The other end also feels like SMA-N but is probably original to the cable (under heat shrink). Is that normal with these, or is this a weirder cable?

The ones I've see have a direct N connector crimped to it, no adapters under the heatshrink.

What I don't like is the S11 plot. S21 will depend on the lenght of the cable. How long is it?

Yeah, the S11 looks weird for sure.

The DTF was accurate ~1.8m:

[Rubo]

Just looked up the datasheet for Sucoflex 104 (I think that's what you have). It's supposed to be around 0.4 db/m @ 3GHz.
Maximum insertion loss @ 26.5GHz (for 1 meter cable) should be no more than 1.77db. That's with 3.5mm connectors, N will be a little worse.
So 1.3 db for 1.8m seems a little much. I think the adapters on the ends are screwing up the results.

[KungFuJosh]

Yeah, I think so too. How can I remove the adapter from the equation? I have the auto fixture removal option, but I've never tried it and I don't know if it will be sufficient, or if it's even appropriate lol. SOLR is supposed to be better, isn't it?

I looked back on my M&P results, and they were significantly better. I forgot they were under 1dB. If the adapter is okay, I might keep it anyway, but I dunno. $22 shipped is cheap enough for the cable + adapter that it's not a big deal either way. I mean, that adapter has to be better than the free crap that came with my nanoVNA, right??

[Rubo]

I have the auto fixture removal option, but I've never tried it and I don't know if it will be sufficient, or if it's even appropriate lol. SOLR is supposed to be better, isn't it?

Not having hands on experience with the SNA line, I can't help there. In theory SOLR should do better.

If the adapter is okay, I might keep it anyway, but I dunno. $22 shipped is cheap enough for the cable + adapter that it's not a big deal either way. I mean, that adapter has to be better than the free crap that came with my nanoVNA, right??

The datasheet I was able to find for Pomona 4297 is a joke, the only "spec" they list is 0 - 12.5GHz frequency range.
For $22 I would personaly keep the cable, but I would either buy better adapters (which you probably want to have either way) or buy good connectors for that cable and re-crimp it.

Just my $.02 

P.S. You will be paying WAAAY more than $22 for decent set of adapters 

[KungFuJosh]

Oh, I know, this shit gets expensive fast.

BTW- Aren't I doing this wrong anyway? SOLR, where R is Reciprocal for the "Unknown Thru" method. Searching says: "The Thru must be reciprocal (S21=S12)." But that's not the case here using an adapter.

Correct me if I'm wrong since I'm new AF to this stuff, but wouldn't I need to auto remove the adapter fixture, and then use two separate cal kits (N + SMA) to measure the cable correctly?

I'm guessing there's an easier way to do it. I dunno. 🤷

Either way, I think I'll keep the cable whether it's good or not. Eventually, hopefully, I'll be good enough to figure that out. 

Attached is that same cable S11 with OSL correction, and nothing attached to the SMA end. I know, I know, I need to waste money on an SMA cal and load shit too.

Thanks,
Josh

[gf]

BTW- Aren't I doing this wrong anyway? SOLR, where R is Reciprocal for the "Unknown Thru" method. Searching says: "The Thru must be reciprocal (S21=S12)." But that's not the case here using an adapter.

Almost every passive network is reciprocal. Your cable+adapter certainly is.
If you have a high-quality, characterized and insertable Thru standard, you can do a regular SOLT calibration, of course.

Correct me if I'm wrong since I'm new AF to this stuff, but wouldn't I need to auto remove the adapter fixture, and then use two separate cal kits (N + SMA) to measure the cable correctly?

That’s correct. Attach the adapter to the VNA port (N-type) and calibrate at the SMA reference plane.
This effectively removes the adapter from the measurement.
[ Of course, this assumes the adapter is electrically stable; if you suspect it's faulty, the calibration may not be reliable. ]

De-embedding would be an alternative, but only if you already have the adapter's S-parameters.

I doubt AFR (Automatic Fixture Removal) will work here. It typically requires a fixture length of at least four wavelengths, and at your f_max of only 3 GHz, the adapter is simply too short.

If you want to isolate the cable, Time Gating out the connectors is straightforward, but you won't be able to isolate the connectors alone without including a few cm of the cable as well. Keep the resolution in mind: at only 3 GHz, even the shortest time-domain window is relatively wide.

[KungFuJosh]

De-embedding would be an alternative, but only if you already have the adapter's S-parameters.

What parameters should I measure?

Is there a good tutorial somewhere to profile adapters and cal kits?

I have a shitty SMA cal kit that came with a nanoVNA, but if it's profiled well with the good N cal kit I have, then it should be usable, right?

It's clear I have a lot to learn.

Thanks,
Josh

[gf]

De-embedding would be an alternative, but only if you already have the adapter's S-parameters.

What parameters should I measure?

Simply measure the S-parameters of the adapter (S11, S21, S12, S22).
[ The catch, however, is that you need calibrated reference planes where the adapter can be mechanically inserted.
E.g. For an N to SMA adapter, the VNA needs one calibrated SMA port and one calibrated N port ]

Is there a good tutorial...

The 'VNA Bible' is the Handbook of Microwave Component Measurements: with Advanced VNA Techniques by Joel Dunsmore.
You'll find a lot of useful information on the web, too.

I have a shitty SMA cal kit that came with a nanoVNA, but if it's profiled well with the good N cal kit I have, then it should be usable, right?

How can you characterize the SMA kit if the calibrated VNA ports are N-ports where you cannot attach them?
You need a "golden standard", either a characterized SMA kit, or a characterized adapter.

Well, up to, say, 300 MHz, you can probably live with 'ideal' standards—meaning you treat them as ideal and just measure the R_DC of the Load. But at 3 GHz, the parasitic effects and the electrical length of the standards are no longer negligible, making proper kit definitions essential.

[KungFuJosh]

How can you characterize the SMA kit if the calibrated VNA ports are N-ports where you cannot attach them?
You need a "golden standard", either a characterized SMA kit, or a characterized adapter.

Well, up to, say, 300 MHz, you can probably live with 'ideal' standards—meaning you treat them as ideal and just measure the R_DC of the Load. But at 3 GHz, the parasitic effects and the electrical length of the standards are no longer negligible, making proper kit definitions essential.

Thanks for all the info. I plan on getting a proper kit + adapters, probably from Roger. I do want to learn how to characterize stuff eventually.

Thanks,
Josh