RTO1000 Fancy BNC

[jjoonathan]

My scope, a RTO1024, has some fancy high-frequency BNCs on the front (see attached). I've been using them with standard BNC-SMA adapters. Nice ones -- I don't want to damage the slotted center receptacle -- but nothing that was designed with the fancy BNCs specifically in mind. The frequency response is a little bumpy, and I suspect the adapter might have something to do with it. I haven't been able to find much information on the fancy BNCs, though. Amphenol sells something they call a "12G" BNC, but it doesn't look like a perfect match. Does anyone know if these fancy BNCs are some kind of standard that I haven't been able to find? Or are they just proprietary R&S magic?

I've seen BNC->SMA pods on the RTP scopes, but I haven't been able to scrounge up an actual listing, and I suspect they would be out of my price range anyway



FWIW, I took the attached frequency response measurement by using the sweeper / splitter / calibrated power sensor method.

[capt bullshot]

They're missing the dielectric. The only BNCs I know of with missing dielectric are 75Ohm 
Searching the internet for 12G BNC reveals 75 Ohm connectors ...

[jjoonathan]

Yes, I found that quite striking as well. I suspect either the diameters are adjusted to compensate for an air dielectric or the corresponding male fancy-BNC has extra dielectric. High performance RF connectors tend to use air/metal construction for dimensional consistency, so I'd expect something like this to happen if you asked an RF person to design a better BNC. Of course, it would need the correct mating male fancy-BNC if you actually want to achieve th extra performance

Here's S11. At 4.5GHz it looks like it does indeed tend towards higher impedance.

[jjoonathan]

The spec under 2GHz is VSWR < 1.25, which corresponds to return loss > 19dBm. It looks like fancy BNCs are needed to meet the spec.

[DaJMasta]

It's pretty common for high bandwidth scopes to use some fancy BNC, though they may be proprietary, it shouldn't be too difficult to source something since they're used in pretty much any scope with BNCs and >1GHz bandwidth.


For just measuring something, not replacing the connectors, I'd just make sure you attach either a BNC cable rated for >2GHz to the connector or a BNC to SMA (or other) adapter rated similarly.  Cables/connectors rated for the same or more bandwidth that are from a reputable manufacturer should be able to do their end of it, and as they are BNCs, they should be able to mate properly with well made BNC connectors.

Honestly, though, your measurements may be in spec.  Total uncertainty on a signal generator can easily be 1.5dB over a few GHz span on a calibrated instrument (though probably not typical), then a little loss from the cable on the higher end and the normal rolloff on the high range still puts you above -3dB at the stated bandwidth.  Do you have a spectrum analyzer that you can compare the response of the cable and generator with?  Or a power meter?

Better yet, that video you linked says you can de-embed the characteristics of your cables/adapters if they're measured on a VNA, which would give you the best picture of the actual frequency response of the frontend.


In any case, I've done some similar FFT measurements of swept signal generator sources on high bandwidth scopes with regular BNC cables and have seen similar level artifacts.  Remember that you're getting somewhat less than 8 bits of dynamic range over the full scale (as zoomed in as possible will maximize your use of the available dynamic range) with an accuracy spec of 2%.  If you zoom in a bit and compare it to an SA with similar settings you probably will be pretty close.

[jjoonathan]

Ah, yes! I think I see similar fancy BNCs on Keysight 6000X and MSO-S scopes too. That should help with sourcing, and gives me a new angle to pursue.

I took the frequency response measurement by using the sweeper / splitter / calibrated power sensor method, so the cable response and signal generator calibration were inside the leveling loop and played no role in determining accuracy. This setup calibrates a spectrum analyzer (or signal generator), not the other way around. Based on calibration history and repeatability between this meter/sensor/splitter and others, I'd expect the applied signal to be good to within 0.1dB. Stacking tolerances probably guarantees it to within 0.5dB. I'd comfortably expect similar or better from the ratioed S11 measurements -- and they're out of spec by over 2dB. In short, I still want to chase down those fancy BNCs

Unfortunately I can't just de-embed because A. that wouldn't help VSWR and B. that's a feature on RTO2000 and RTP scopes and I only have a RTO1000. I can dream, though! 

[jjoonathan]

Here we go.

https://www.keysight.com/my/faces/partDetail.jspx?partNumber=54855-67604

Now it's just a matter of waiting for one to show up on ebay at a price I can rationalize