From: jl@glen--canyon.com
On Fri, 23 Jan 2026 13:59:18 +0100, Jeroen Belleman
wrote:
>On 1/23/26 12:34, Liz Tuddenham wrote:
>> john larkin wrote:
>>
>>> On Thu, 22 Jan 2026 17:38:41 +0000, liz@poppyrecords.invalid.invalid
>>> (Liz Tuddenham) wrote:
>>>
>>>> john larkin wrote:
>>>>
>>>>> On Thu, 22 Jan 2026 09:29:10 +0000, liz@poppyrecords.invalid.invalid
>>>>> (Liz Tuddenham) wrote:
>>>>>
>>>>>> What physical properties determine the velocity factor of co-ax? Most
>>>>>> of the amateur radio books give around 60% as the velocity factor for
>>>>>> 'common' types of 50-ohm co-ax.
>>>>>>
>>>>>
>>>>>
>>>>> V = c/(sqrt(Er))
>>>>>
>>>>> Solid polyethylene has Er around 2.3.
>>>>>
>>>>> Foamed stuff is lower.
>>>>>
>>>>> Polyethylene is awful. It melts when you solder it. Foamed is worse.
>>>>
>>>> Solid polyethylene isn't too bad but foamed has a nasty habit of moving
>>>> under the influence of its own 'memory'. You solder the end of a
>>>> slightly bent centre conductor and, as the heat travels down it, the
>>>> foam springs back to the straight position, leaving you with a slot in
>>>> the foam and a bare centre conductor shorted to the screen.
>>>>
>>>>>
>>>>> Your VNA measurement may be suspect.
>>>>
>>>> That was why I asked about it here, I suspected the measurement.
>>>>
>>>> However... my method of finding the first reactance swing in the
>>>> reflection from an open circuit should give me a measurement of the
>>>> electrical length of the cable that is independent of the terminating
>>>> impedances, calibration etc
>>>>
>>>> The cable was physically 6.39 metres long and the first 'resistive'
>>>> impedance point was at exactly 15.000 Mc/s. (That's another reason I
>>>> was suspicious, it really was spot-on 15.000 Mc/s, give or take
>>>> nothing.)
>>>
>>> Looking for the max resistance may not be ideal.
>>
>> Perhaps I didn't explain that very clearly. It wasn't the point of
>> maximum resistance, it was the point where the capacitive reactance
>> swung through zero to become an inductive reactance; it was quite
>> sharply defined. At that point the impedance was purely resistive but
>> it was the reactance that I was measuring, not the resistance.
>>
>>
>>> If you jack up the frequency and get multiple wavelengths in the
>>> cable, resolution will improve.
>>
>> True, but the electrical errors in measurement may increase too. An
>> accuarcy of around 1% would be good enough for the present purposes -
>> after all, where exactly is the 'end' of a piece of co-ax that is
>> splayed out for connection to something else? I also wouldn't expect a
>> length of cheap co-ax to be particularly homogenous.
>>
>>
>>> 15.000 MHz seems suspicious.
>>
>> Yes, that worried me.
>>
>>
>>> I don't have a VNA. I use TDR to measure time delays.
>>
>> My VNA will also function as a TDR and I have a telephone-testing TDR
>> for longer lines. If all else fails, a square-wave signal generator and
>> an oscilloscope will work too.
>>
>
>Having used both VNAs and TDRs for cable length measurements, I've
>always found the VNA measurements much superior to those of a
>TDR. My HP8753 would resolve 1 degree @ 1GHz with ease. That
>corresponds to a little under 3ps. TDRs are much too noisy to
>do that.
>
>I've got a picture of a TDR and a VNA-derived measurement of the
>same setup here. Note how much cleaner is the trace from the
>VNA. You can't beat a VNA for S/N.
>
>
>
>I made these measurements in the context of constructing a wide
>band beam transformer for the Proton Synchrotron at CERN.
>
>Jeroen Belleman
My 11802/SD24 is about two generations behind the times; your S-series
rig is three. The 11801 series has both smoothing and signal
averaging, and can easily resolve 1 ps of delay accurately. And has an
RS-232 interface to export waveforms.
But yes, data from a good VNA can be crunched to make a good TDR. But
TDR is DC-coupled and VNAs aren't.
John Larkin
Highland Tech Glen Canyon Design Center
Lunatic Fringe Electronics
--- SoupGate-Win32 v1.05
* Origin: you cannot sedate... all the things you hate (1:229/2)
|
