Hi all,

According to this site:

50 ohm BNC connectors are only good up to 4GHz and 75 ohm ones to 2GHz maximum. And yet - to take just one example - my best frequency counter, which accepts and can measure signals up to 22Ghz, uses 50 Ohm BNC sockets and seemingly works just fine with them. And yet it's not just the Historytools website which claims BNC is only good up to a few GHz; plenty of other sources say the same thing. What's going on here?

Thanks,
J.

开云体育

The frequency counter can't measure amplitude. Using BNC above 4 GHz causes losses due to reflection, but the frequency counter will still work even if your signal is 10dB lower. If you use BNC connectors with a spectrum analyzer, you'll see that the amplitude of your signal is now lower since you used BNC instead of SMA or N.

If you get a VNA, you can also see that the input of the frequency counter is no longer 50 ohms at 22 GHz.

I am also not aware of any frequency counters that go that high that use type BNC connectors. Perhaps it uses an N connector, and you have an adapter?

No, it's the standard 50 ohm BNC output socket. I just checked and it's actually up to 24GHz according to the front panel just above that socket. It's a Systron-Donner counter if that helps at all.

开云体育

Is it the 6054?

Are you sure there is no SMA - BNC adapter on it? Perhaps someone modified it...

开云体育

The word "limit" should not be interpreted as a fixed, discrete boundary that separates "useful" from "useless". The world is analog, so it's best to consider what is really meant in a given circumstance.

For a frequency counter to function properly, all you need is to preserve the zero crossings and some minimum amplitude, so that's a fairly forgiving application. That said, I am surprised to hear that your counter's 22GHz input uses a BNC. High-frequency counters often have a couple of connectors, one covering up to a couple of GHz and another above, but I've never seen or heard of a BNC being used in a commercial instrument for 22GHz -- could you let us know the instrument model? Or is it a homebrew counter?

That aside, in other situations you might need to worry about impedance mismatch. And that's where a typical BNC connector has unfavorable characteristics as you increase frequency. The instability of the ground connection is also a problem (particularly in mobile applications, where the modulation of impedance can cause the BNC to act as a surprisingly good vibration sensor). BNCs are also manufactured to fairly loose tolerances, so if the application demands accuracy and precision over a wide frequency range, BNCs aren't going to be a good choice.

So, it's not as if signals abruptly stop emerging from the BNC once you go above, say, 4GHz. It's more complicated than that. And whether a BNC is useful or not in a given circumstance depends very much on what performance you need in that particular case. The frequency "limit" will vary accordingly.

-- Cheers
Tom

-- 
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
420 Via Palou Mall
Stanford University
Stanford, CA 94305-4070

I thought I knew the S-D lineup pretty well, and yet I don't
recall an instance where they used a BNC on microwaves.

What model do you have?

-Chuck Harris


On Sat, 29 Jun 2024 11:22:40 -0700 "Jinxie via groups.io"
<paul666@...> wrote:

No, it's the standard 50 ohm BNC output socket. I just checked and
it's actually up to 24GHz according to the front panel just above
that socket. It's a Systron-Donner counter if that helps at all.

The 6054 has a BNC for its low band (200MHz) input, but
uses an N for its microwave input.

-Chuck Harris


On Sat, 29 Jun 2024 11:26:05 -0700 "Board via groups.io"
<board@...> wrote:

Is it the 6054?

Are you sure there is no SMA - BNC adapter on it? Perhaps someone
modified it...

On 6/29/24 11:22 AM, Jinxie via groups.io wrote:

No, it's the standard 50 ohm BNC output socket. I just checked and
it's actually up to 24GHz according to the front panel just above
that socket. It's a Systron-Donner counter if that helps at all.

First:
BNC and N WILL NOT mate properly. while the main dimensions of the center parts are the same (they are 7mm class connectors) the center pin of the N connector is lrger than the BNC center pin. If you push a N male onto a BNC frmale you can damage the BNC.

I am also very surprised that any commercial frequency counter would use a BNC connector for a 8 GHz input never mind at 20GHz. All 7mm connectors are limited to 18GHz for predictable performance. Most N types got to 12GHz but special precision versions go higher. As an example the 20GHz Marconi 2440 counter uses a precision N type for the microwave input. The otherwise identical 2442 uses a 3.5mm (like an SMA but air dielectric).

What model is the counter? Can you post a picture of the connector niclding the area around the connector?

Forgot to say, the 2442 goes to 26.5GHz.

6054B and I can't post any pictures right now I'm afraid. It doesn't appear that the BNC has been changed from SMA; looks original and why would anyone do such a thing anyway? Does seem an odd choice!

Chuck, it absolutely *cannot* have been an N type at any time in its life or there'd be a ring of some sort around the BNC as they're smaller than the hole an N type needs.

I'd expect a 22GHz counter to typically use a 3.5mm connector, but it could also use SMA or N connectors at a push.

Note that the highest (frequency) rated BNC connectors I've seen are rated to 18GHz but these are specials made by Agilent. These connectors look just like a BNC and can mate with a standard BNC connector but they are gold plated and have a precision inner connector made with better defined dimensions and a better mating technology compared to a standard BNC connector. These special BNC connectors could be used as the main channel inputs of a 13GHz scope for example.

开云体育

There was an option to move the high frequency input to the rear panel, leaving only a BNC on the front panel. Might that be what you have?

In the past, I have fed 13GHz into one of these Agilent scopes that use BNC inputs and it displayed the 13GHz signal (from an old HP 83752A generator) just fine. There was some loss, but that is to be expected in the cable and the scope front end as it was a 13GHz rated scope. It was probably less than 3dB loss but I can't remember now.

These 13GHz BW scopes (40Gsa/s sample rate) have a fixed 50 ohm input via the BNC connectors on the front panel. They are quite old scopes, there has been one at work for many years and I use it now and again for TD stuff.



I'm not sure what the upper frequency limit would be for these custom BNC connectors if they were used for a counter rather than a 13GHz scope. It's going to be a bit more than 18GHz but it might not manage 22GHz...

TNC are several orders of magnitude better than BNC IIRC. Far more phase stable (that's another aspect we haven't mentioned yet). Shame they never really caught on; bit like Betamax and VHS. :(

From what I have seen, the 6054B can have either a SMA style,
or an N style depending on the top frequency. Never BNC.

If yours has a BNC on the microwave input, then someone must
have changed it... Probably because they broke the SMA style
connector.

I say SMA style, because it is more probably an APC 3.5.

-Chuck Harris


On Sat, 29 Jun 2024 12:54:19 -0700 "Jinxie via groups.io"
<paul666@...> wrote:

6054B and I can't post any pictures right now I'm afraid. It doesn't
appear that the BNC has been changed from SMA; looks original and why
would anyone do such a thing anyway? Does seem an odd choice!

Probably only about 6 gadzillion were used on routers before
the inverse SMA became popular... But, no, they didn't catch
on in the test equipment world.

-Chuck Harris


On Sat, 29 Jun 2024 13:32:04 -0700 "Jinxie via groups.io"
<paul666@...> wrote:

TNC are several orders of magnitude better than BNC IIRC. Far more
phase stable (that's another aspect we haven't mentioned yet). Shame
they never really caught on; bit like Betamax and VHS. :(

The datasheet for that 6054B counter does state that the input is normally an SMA connector. However, it also states that other connector types are available on special order.

A standard BNC connector would be an odd choice to fit here though...

Systron-Donner 6054B – the 20 MHz -24 GHZ connector on stock models certainly looks like an SMA to me. The 20 Hz -20 MHZ input is a BNC. The S-D catalog calls out (verbatim) “SMA-3MM. Other input connectors available on special order.”

A BNC modification would nicely fit into the hole made for a SMA if a slight modification to the hole were made with no trace of any hacking.

Greg