[time-nuts] Don't use cheap cables -- a cautionary tale (Bob Camp)

David Kirkby david.kirkby at onetel.net
Wed Mar 6 08:02:17 EST 2013


On 6 March 2013 11:33,  <lists at lazygranch.com> wrote:
> At what frequency do these "problems" appear?  And how bad it bad?

My interest for VNA calibration and verification kits is probably well
outside most time-nut uses. I normally test N connectors from 50 MHz
to 6 GHz. 50 MHz is the lower limit of my HP 8720D (20 GHz) VNA and 6
GHz is the upper limited of my N calibration kit (HP 85032B).

To give a couple of figures.

1) Huber and Suhner female-female barrel, which is the most expensive
I tried (15.26 GBP), has a worst case return loss of just 17.2 dB.
2) TE Connectivity the cheapest I tried at just  £2.84 GBP, had a
return loss of 35.1 dB.
2) Telegartner, mid-priced at 6.68 GBP had a return loss of 38.9 dB.

So a mid-priced one was better than both the expensive and cheapest
choice. All are based on a saaple of two, and worst case results
quoted above. I did not find much variation between two of the same
manufacturer.

Although I don't have exact data in front of me, as I'd need to go
through the individual data files which I can't be bothered to do, all
were very good at low frequencies (say 500 MHz or less).

The worst case is not always at the highest frequency, although the
performance at 6 GHz is always worst than at 50 MHz.

All data was colllected by use of

* HP 8720D VNA within its calibration period.
* Agilent 3.5 mm test port cables on the VNA (over $5000 a pair)
* Metrology grade 3.5 mm to N adapters
* HP 85032B N calibration kit.

Since the N calibration kit was after the metrology grade 3.5 mm -> N
adapters, any issues with those adapters would be removed by the VNA
calibration routine.

I've seen some Americans make comments like "if it does not have
Ampenol printed on it, then avoid it". I'm not German, but the
mid-priced German made Telegartner F-F adapter was the best. See
pictures at the bottom of

http://www.vnacalibration.co.uk/support/NCV032B/HP/

For qualtiy of the threads, the Chinese made Huber and Suhner device
was the best, but it was by far the most expensive.

The 85032B comes with 4 APC-7 to N metrology grade adapters. They can
be assembled to make any N adapter you want (male-female, male-male,
female-female). These always provide better peformance than any
commerical grade connectors I mention above, but such devices are far
too expensive for the kit I use, which I sell at less than the price
of the adapters.

I would never use an adapter for a critical high frequency application
without testing it first.

BTW, I used to work in a Ministry of Defence calibration lab in the
UK. They always removed the rubber washers in N connectors. I notice
the metrology grade adapters don't have these, and Joel Dunsmore,
Agilent top VNA guru, reccomeneds in his book on VNAs to do the same.

Dave


> -----Original Message-----
> From: David Kirkby <david.kirkby at onetel.net>
> Sender: time-nuts-bounces at febo.com
> Date: Tue, 5 Mar 2013 18:16:54
> To: Discussion of precise time and frequency measurement<time-nuts at febo.com>
> Reply-To: Discussion of precise time and frequency measurement
>         <time-nuts at febo.com>
> Subject: Re: [time-nuts] Don't use cheap cables -- a cautionary tale (Bob
>         Camp)
>
> On 5 March 2013 17:08, Gregory Muir <engineering at mt.net> wrote:
>> When I compare the typical easily-obtained "Jameco"
>> or "Marlin P Jones" type of connector to a better made product such as a
>> true MIL item on a network analyzer, one can see a noticeable difference
>> between the impedance and VSWR characteristics. It has been long known that
>> the simple nickel-plate connector can cause problems over time in an
>> installation from changing contact resistance and such.  The manufacturer
>> "RF Industries" still uses silver & gold plating on their types of RF
>> connectors as compared to the cheaper nickel-plated types and I have found
>> these to be very reliable with a cost slightly higher than the cheaper
>> nickel plated types.
>
> I've recently been playing with some N adapters from various
> manufacturers. I was particularly keen to find a decent male to male
> adapter to include with a low-cost N calibration and verification kit
> I am sellling.
>
> http://www.vnacalibration.co.uk/sales/
>
> If you read the specs on that, I state if the male to male adapter is
> used, it limits performance. I've tried various makes, from the cheap
> to fairly expensive (Huber and Suhner), although I am not considering
> metrology grade devices, given the price I am selling the kit for.
>
> I don't have all the data in front of me, but certain things stick in my mind.
>
> 1) The Huber and Suhner female-female N barrel was the most expensive
> of any I tried. The qualtiy of the threads was excellent. But when I
> stuck it on the VNA, the performance was really poor. I was intending
> to phone H+S to ask what was wrong, but before doing so I looked at
> the specification sheet. The spec was indeed poor, and the one I had
> met it.
>
> 2) Someone recently sent me a couple of RF Industries male to male N
> barrel adapters. RF performance was higher than I had seen on any
> connector of its type with the exception of metrology grade devices.
> However, it felt as though there was sand between the threads when
> tightening these up.
>
> 3) I found some pretty decent female-female Ns that were really cheap
> (< $5). RF performance was good, mechanically well made. They did not
> have the best RF performance of those I found, but they were cheap.
>
> For a male-male adapter in that cal kit, I have settled on the Radiall
> component for now, but I'd like to find a better one. The RF
> Industries unit would do, as long as I supply a lubricant to make
> screwing the things on better.
>
> So to me, there does not appear to much correlation between cost and
> performance. I've seen poor expensive items and fairly decent very
> cheap items.
>
> Dave
> _______________________________________________
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
> _______________________________________________
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.


More information about the time-nuts mailing list