[Amps] All the recent UHF connector discussion

Carl km1h at jeremy.mv.com
Thu Nov 14 09:44:18 EST 2013

This reminds me of all the discussions and outright arguments we had when I 
worked in the 12-75 GHz region at a few companies and well above 200 GHz at 

At one place the 2 PhD's in R&D called in HP to settle some things and they 
only made it worse by admitting that the cal kits were not as precise as 
they wished but as good as could be repeatably manufactured.

So smoke and mirrors and magic ruled again in the RF world.

This discussion with UHF connectors is mostly becoming academic nit picking 
as I dont know of anyone that is serious at 400-500 MHz would even think of 
using them.

Use the proper devices and cables such as SMA connectors at that frequency 
when tuning for best NF, etc.


----- Original Message ----- 
From: "David Kirkby" <david.kirkby at onetel.net>
Cc: "amps" <amps at contesting.com>
Sent: Wednesday, November 13, 2013 10:03 PM
Subject: Re: [Amps] All the recent UHF connector discussion

> On 14 November 2013 00:48, Larry Benko <xxw0qe at comcast.net> wrote:
>> Dave,
>> I am also not aware of any UHF calibration standards.
> I don't think there ever will be, as the connector has an undefined
> impedance, so its hard to see how one could make any sort of standard.
>> I used the HP cal kit
>> which is an offset (both open and short) cal kit with the open also 
>> having a
>> fringing capacitance.
> The short does have some inductance too, though you can't enter that
> on an 8753. The file here
> http://na.tm.agilent.com/pna/caldefs/PNA/85033DE.htm
> for the PNA does give the inductance.
>> The VNA removes the cal standards anomalies during
>> calibration giving a reference plane at the edge of the 3.5mm connector.
> To be more precise, it is at the outer conductor mating plane.
>>  I
>> think the offset is as you mentioned 32-33ps.
> Looking at above, they are 29.243 ps on the open and 31.80799 ps on
> the short, so my guess was not too far off.
>> I have hundreds of the RG-142
>> jumpers which came from some UHF aircraft diplexers and are manufactured
>> from the same batch of RG-142B.  I did the cal with one of these cables 
>> and
>> then cut the connector off one and soldered the PL-259 on.  I was only
>> interested in SWR so losing the reference plane was not important and 
>> could
>> still measure SWR.
> Yes, but subject to the fact you are not just measuring the PL-259
> mated pair, but the BNC->SMA adapters too.
>> I came into a buy of a large number of very high quality new adapters a 
>> year
>> or so ago from a company who designed PCS type cell equipment and was
>> closing a design center.  The BNC to SMA adapters were doubled up and 
>> showed
>> SWR @ 500MHz below 1.05:1.  I believe what I did to be accurate at 
>> 500MHz.
> In that case, I think you should add that fact to your web page, as
> looking at now, it is clear to me the results could be skewed by the
> SMA to BNC adapters.
> I've tested a fairly wide range of SMA & N adapters, mainly up to 6
> GHz, but some to 20 GHz, and it is clear to me that there are
> variations among them, and that it is significant, with some real
> surprises - like return loss of female-female N from H+S being worst
> than some really cheap adapters at 10% of the price. But the cheap
> ones normally suffer from poorly cut threads, so will wear out what
> they mate to if you use them a lot. I would expect variations of
> PL-259 / SO-239 combinations might have far more variability, but
> unlike you I have never tested PL-259s. I've measured adapters with N,
> SMA, 3.5 mm, APC7 and WR90 waveguide, but never PL259 or BNC.
>> Larry, W0QE
> Anyway Larry, it is an interesting page. I think if you add the fact
> you consider the effect of the adapters insignificant, it gives it a
> bit more credibility.
> There is another way you could get better data, but it would need a
> higher spec VNA with the TDR option
> 1) Measure up to a much higher frequency
> 2) Do an inverse Fourier Transform of the frequency domain data to get
> time domain data
> 3) Put a window around the discontinuity in the time domain
> 4) Transform that back the window to the frequency domain with a
> Fourier Transform
> A VNA with option 010 (time domain) could do that, but to get
> sufficient spatial resolution, one would need a lot higher frequency
> VNA. Then one could eliminate the effect of the BNC/SMA adapter. As
> the Window in the time domain would remove it.
> Anyway, an interesting exercise, but I'd like to know if it repeatable
> among difference combinations of connectors. That said, since the
> electrical length is so small at 500 MHz, it probably is.
> Dave, G8WRB
> _______________________________________________
> Amps mailing list
> Amps at contesting.com
> http://lists.contesting.com/mailman/listinfo/amps
> -----
> No virus found in this message.
> Checked by AVG - www.avg.com
> Version: 10.0.1432 / Virus Database: 3222/6335 - Release Date: 11/14/13

More information about the Amps mailing list