Connector VSWR

[bjcuizon]

Hi, I found this SMA Adapter/thru on RS (This one) and as you can see, its specifications on its page below says its VSWR is derived from the formula:
1.07 + 0.07 x f(GHz)
So my computation for it went out like this: 1.07 + 0.07 x 18(as it's the max frequency specified)
It turns out the VSWR at 18GHz is 2.33? 
Is this the real value or am I doing something wrong?

Thanks

[TheSteve]

Your math is correct. There is rarely anything cheap with connectors that work really well at 18 GHz. At the price they aren't too bad.

[bjcuizon]

Your math is correct. There is rarely anything cheap with connectors that work really well at 18 GHz. At the price they aren't too bad.

So, that means at 6GHz, the VSWR is ~1.5?

How do they characterize these things..I mean like how do they test and see its performance. Do they just measure it with a VNA on one end and no load on the other end? Or do they put some sort of load with extremely low VSWR at extremely high frequencies on the other end?

Thanks

[Tony_G]

Usually the process is to get those parameters from the simulation software they used during design, build a batch of them and test to confirm that everything is statistically within spec. Depending on the price/target market of the connector they might individually test each one but at that price I can't imagine they do that for pass/reject - They might, I can't see the economics working but I'll be happy to stand corrected.

TonyG

[CopperCone]

is there a general rule of thumb that connectors made of steel are generally better then the plated brass ones in terms of VWSR and insertion loss? for the second hand market

[bjcuizon]

Usually the process is to get those parameters from the simulation software they used during design, build a batch of them and test to confirm that everything is statistically within spec. Depending on the price/target market of the connector they might individually test each one but at that price I can't imagine they do that for pass/reject - They might, I can't see the economics working but I'll be happy to stand corrected.

TonyG

So that means these given parameters are only for the connector itself?, without the terminating resistor to prevent standing waves, and all other fancy stuff.

[Tony_G]

Yes - The assumption these numbers make is that if you were to measure the connector then what you will have done before starting is to effectively calibrate away all the errors introduced by the test gear (cables, analyzers, loads, etc) so that you're only measuring the mismatch introduced into the system by the connector.

There is a good set of App Notes on this on the Keysight website:

http://www.keysight.com/main/editorial.jspx?ckey=272209&id=272209&nid=-35176.0.00&lc=eng&cc=US

Part 3, sections 3 & 4 are probably most useful to you.

TonyG

[bjcuizon]

Yes - The assumption these numbers make is that if you were to measure the connector then what you will have done before starting is to effectively calibrate away all the errors introduced by the test gear (cables, analyzers, loads, etc) so that you're only measuring the mismatch introduced into the system by the connector.

There is a good set of App Notes on this on the Keysight website:

http://www.keysight.com/main/editorial.jspx?ckey=272209&id=272209&nid=-35176.0.00&lc=eng&cc=US

Part 3, sections 3 & 4 are probably most useful to you.

TonyG

Thanks very much Tony_G. I will have a good read of it..most likely tomorrow.(I have to go to sleep now. hehe)

[G0HZU]

1.07 + 0.07 x f(GHz)
So, that means at 6GHz, the VSWR is ~1.5?

I'd expect an SMA F-F adaptor from a reputable brand to have a VSWR below 1.1:1 at 6GHz. Probably better than 1.05:1 typical.

It would have made more sense to me if the equation was something like 1.07 + 0.007*f(GHz)

Either it's a typo in the spec or the quality of that connector is dire.

[bjcuizon]

So, that means at 6GHz, the VSWR is ~1.5?

I'd expect an SMA F-F adaptor from a reputable brand to have a VSWR below 1.1:1 at 6GHz. Probably better than 1.05:1.

Umm..its actually an M-M adapter. And yeah, I believe it should have a really low vswr (probably with the correct load impedance.) as it has a decent brand name.

[G0HZU]

Sorry, I always get the sex backwards when describing adaptors. I'm used to thinking in terms of VNA cal kits. But I think your spec probably has a typo in it.

It would have made more sense to me if the equation was something like 1.07 + 0.007*f(GHz)

[G0HZU]

Your math is correct. There is rarely anything cheap with connectors that work really well at 18 GHz. At the price they aren't too bad.

With a VSWR of 2.33:1 at 18GHz I'd argue that it wouldn't be fit for purpose. A VSWR of 2.33:1 would give a mismatch loss of 0.75dB alone.... crazy.

Their equation for the insertion loss seems to be much more realistic. It predicts 0.17dB insertion loss at 18GHz. This doesn't stack up with a VSWR of 2.33:1 so I think there must be something wrong with the VSWR equation?.

[PA0PBZ]

http://docs-europe.electrocomponents.com/webdocs/11de/0900766b811de94e.pdf

VSWR:  1.23 max.   

[bjcuizon]

Your math is correct. There is rarely anything cheap with connectors that work really well at 18 GHz. At the price they aren't too bad.

With a VSWR of 2.33:1 at 18GHz I'd argue that it wouldn't be fit for purpose. A VSWR of 2.33:1 would give a mismatch loss of 0.75dB alone.... crazy.

Their equation for the insertion loss seems to be much more realistic. It predicts 0.17dB insertion loss at 18GHz. This doesn't stack up with a VSWR of 2.33:1 so I think there must be something wrong with the VSWR equation?.

Yeah, that's why I was wondering. And so I started the thread.

[bjcuizon]

http://docs-europe.electrocomponents.com/webdocs/11de/0900766b811de94e.pdf

VSWR:  1.23 max.   

Thanks PA0PBZ, I did not see that on their website. That's really helpful.

[G0HZU]

Maybe a zero is missing in the equation because if you amend the equation to the one I suggested earlier then:

VSWR = 1.07 + 0.007*f(GHz) would give 1.2:1 at 18GHz. This seems close to the data from PA0PBZ.

Obviously, this is just a crude equation but it fits better with what I would expect of a cheapish SMA adaptor like this when measured up at high frequencies.

[bjcuizon]

Maybe a zero is missing in the equation because if you amend the equation to the one I suggested earlier then:

VSWR = 1.07 + 0.007*f(GHz) would give 1.2:1 at 18GHz. This seems close to the data from PA0PBZ.

Obviously, this is just a crude equation but it fits better with what I would expect of a cheapish SMA adaptor like this when measured up at high frequencies.

Yep, I agree with you. It might be really be a typo. After all it is what you'd expect from a $12 connector from a good source. It's not like its $1 from an unknown source in aliexpress/china with unspecified details.

Thanks very much for all replies.