Don't buy SMA cable assemblies from CentricRF

[Smoky]

Just looking at the few samples of SMA ends in the pictures reveals differences all across the map. Thick walls, thin walls, stepped and non-stepped contacts, loose dielectrics, tight dielectrics. Now bring in all of the differences that can be found in the receptacles 

[joeqsmith]

Just looking at the few samples of SMA ends in the pictures reveals differences all across the map. Thick walls, thin walls, stepped and non-stepped contacts, loose dielectrics, tight dielectrics. Now bring in all of the differences that can be found in the receptacles 

Which is why I asked early on about if their gage could even measure their particular connectors, why I wanted to see what they looked like and why I posted about mixing the 3.5mm and SMA gages.   

Again, I have no way to know if there was a problem with the OPs measurements or not.  This is also why I had suggested an independent repeating the measurements. 

I could measure three of the eight I show with confidence that we would get something meaningful.  One that has no step.  Four with some sort of step that I doubt I could tell you anything more useful with a gage than what the photo shows.   It doesn't mean those connectors are bad but that I just can't measure them.   


A simple experiment.   I blow out the gages and zero both the pin contact and dielectric gages to the master gage.   Next measure the Digi-key cable which does not have a step.  I would expect both gages to measure from the reference plane to the dielectric.  All measurements in inches.   

Connector 1,  Dielectric 0.0045,  Pin 0.0045
Connector 2,  Dielectric 0.002,  Pin 0.002

While it meets the industry standard, it is a bit high.  The real point of this exorcise is that the two measurements basically match.

Measuring the pin of the cable I use for my X-band EPR experiments,
Connector 1, Pin 0.0025
Connector 2, Pin 0.0005
Very nice.       

Let's try the home made cable for my HP8754A.   
Connector 1, Pin 0.003 

And finally my Gore cable.   
Connector 1, Pin 0.003 
****
I checked the Gore cables dielectric and then using my 3.5mm gage, I remeasured the Pin and get 0.002875 (basically a half division off).  Consider that was a completely different master used to zero this gage, I think it's pretty safe to say we are in the ballpark. 

It seems pointless to attempt to measure the other SMAs as I have zero confidence in what the gages would show.   

[joeqsmith]

I wouldn't suggest that adapters are immune.  To conduct that T-Check experiment, I purchased some very low cost Tees.   I think I paid something like $10 for a 4 pack.  Guessing you could get these even cheaper but I picked them up off Amazon. 

Notice the pins transition to the step has a fairly dramatic radius.   Measuring the pins location:

Connector 1, Pin -0.0005
Connector 2, Pin -0.0015

I had tested this adapter using my Agilent VNA with good results to 9GHz.   There are no detectable signs where it touched off on the female pins.   Measuring one of the sacrificial adapters I used, the female pin is recessed 0.0025 providing enough clearance. 

Two wrongs don't make a right but it certainly can help.