开云体育

I did made two remeasures - one with open end and one with shorted end.
They are within 300 mm from each other...

Which one is "closer to reality " ?

Since the TDR graph did identify the coax type it is not clear
if the "length" includes the VF -
hence is it
electrical or mechanical length ?


The accuracy of the TDR graph is immaterial, I am after an estimate ,
however
how do I change the horizontal ( x ) axis scale on nanoVNASaver display ?

Take a look at this nanoVNA TDR tutorial by W2AEW:


VF must be entered numerically in order to obtain correct physical length.
The X axis scale is defined by the stimulus start/stop freqs.
W2AEW presents a formula to determine stop freq based on expected cable length and VF.
Did my fist TDR measurement yesterday and got reasonable results (length, attenuation).

73, Markus HB9BRJ

---
Since the TDR graph did identify the coax type it is not clear
if the "length" includes the VF -
hence is it
electrical or mechanical length ?
---
I always had this question as well
Since you must enter a VF it must be calculating physical length.

Things that effect accuracy are stop freq. I got better accuracy when the indication filled the graph. I adjust the stop to get best fit then recalibrate. Also setting it up to do 401 pts for the sweep improved accuracy. Still its just an estimate.

...for what its worth...
Bryan, n0luf.

Before the VNA was available, I would just measure the C between center conductor + braid. Knowing how many pf per foot the cable was, it's a simple math calculation for the physical length. Knowing the VF, the electrical length is also calculated.

I always had this question as well
Since you must enter a VF it must be calculating physical length.

Incorrect
as I pointed out - the graph shows the coax cable type and its VF , it also gives the cable dielectric (how?) which in an essence effects the VF.
( again reading the source code should answer some of the concerns )

Things that effect accuracy are stop freq. I got better accuracy when the indication filled the graph. I adjust the stop to get best fit then recalibrate. Also setting it up to do 401 pts for the sweep improved accuracy.

Still its just an estimate. which was my goal in first place.

Yes, I will czech the u-tube....

BUT to answer my "length" question - if I use KNOW length of cable and then compare the nanoVNA TDR calculated "length" ... problem solved

As expected - verifying by using known length of KNOWN coax it is easily said then done ....

I found "Belden YR 16664 " "factory cable" and cannot find any data on it...

I found piece of worn-out what appears RG8 with barely readable description reading "microwave cable" .....useless

The u-tube is all about selecting menus on nanoVNA - I have given up on the menus long time ago ( my choice ).

I can get "TDR" on nanoVNASaver main display , but I cannot figure out how to "update" it....

...and coaxial cable is manual option AFTER "Time domain retroreflectormetry ... " is selected .... my error , but that does give option for VF ....

In theory - TDR can measure cable length - when the distant end is open - hence using reflection
OR it can measure "fault" AKA when there is known good coaxial cable and the distant end is shorted.

After all this is appears the measurement posted ( in nanoVNASaver) is physical length.

But still not verified, just a best guess for now.

I have added to this /g/nanovna-users/photo/286949/3628295?p=Created%2C%2C%2C20%2C2%2C0%2C0 to show the difference between "solo TDR" display and main nanoVNASaver display with "TDR" option I cannot set he nanoVNASaver display same as "solo TDR" . /g/nanovna-users/photo/286949/3628295?p=Created%2C%2C%2C20%2C2%2C0%2C0

On Thu, Jun 8, 2023 at 04:34 AM, Bryan Curl wrote:

I always had this question as well
Since you must enter a VF it must be calculating physical length.

Things that effect accuracy are stop freq. I got better accuracy when the
indication filled the graph. I adjust the stop to get best fit then
recalibrate. Also setting it up to do 401 pts for the sweep improved accuracy.
Still its just an estimate.

The NanoVNA does not calculate the physical cable length using traditional Time Domain Reflectometry (with pulses on the cable). It calculates using frequency domain S11 data and then performs an Inverse Fast Fourier Transfom (IFFT) and uses the velocity factor VF (input by the user) to calculate the physical length. It can be quite accurate if you know the VF and select an appropriate frequency range. Below are some tests I completed using NanoVNA Saver on a 9.38 meter (30.75 foot) piece of Beleden 8259 RG-58A/U cable. In the first test I selected a cable type that was similar from the drop down menu and Saver calculated the physical length as 9.628 meters (31.6 feet) using a VF of .66. I measured the actual VF using an analog TDR system and the VF for this cable was .64. I set this value in Saver using the custom option from the drop down menu and Saver calculated 9.34 meters which is very close the the actual cable length of 9.38 meters.

The frequency range used in my tests was 50 kHz. to 900 MHz. and I used the Manage button in Saver to select 401 data points for the measurement. This works well for cables up to 30 meters in length. For longer cables you need to reduce the stimulus frequency range. For much shorter cables the stimulus range should be higher if you want a better estimate.

Roger

Follow up to my previous post...

You can also use the Transform option in the NanoVNA menu to do TDR measurements without connecting to a PC. From the Display menu you select a trace and set it to S11 mode and set it to Linear display. From the Transform menu select Transform On and Low pass impulse. You set the VF (in %) which is for the cable being tested. You need to look this up in the manufacturers spec or measure it using a scope and pulse generator. Be aware that the actual VF can be different by a percent or two from the nominal published spec due to manufacturing tolerances. It also varies between manufacturers for the same cable type. Below are 3 screenshots taken on a NanoVNA-H4 using a VF of 63, 64 and 65 %. The measured VF using the pulse generator/scope method was 64%. Physical cable length was 9.38 meters and using 64% VF NanoVNA calculated 9.327 meters

Roger

A far easier and more accurate way to measure cable length is by using the
"Measure -> Cable" function in the nanovna firmware (if you have a firmware
with this function).
Just connect to an open-ended cable, and read the length and cable loss on
the left side of the screen. You still have to set the correct VF, of
course. But this method does not use TDR and transforms, with the
associated complications. Instead, it finds the quarter-wavelength
frequency of the cable by finding the first S11 phase reversal, and does
the simple calculation. It uses a good interpolation function to calculate
the exact phase-crossing point for very good accuracy. You only need to
set the frequency range such that the smith chart trace does at least a
half-circle (i.e. the range needs to contain the frequency corresponding to
a quarter-wavelength resonance). I usually just use 50kHz - 30MHz,
although you will need to go much higher for very short cable lengths. If
the smith chart trace does many circles, you may want to decrease the upper
frequency to improve accuracy.

If your firmware does not have this feature, you can use the same technique
manually, by simply displaying the S11 phase trace. When connected to an
open-end cable, it will give a "saw-blade" trace. Just move the cursor to
the first phase crossing, read the frequency, and calculate the wavelength
using the VF for correction.
Stan KC7XE

On Thu, Jun 8, 2023 at 11:22?AM Roger Need via groups.io <sailtamarack=
[email protected]> wrote:

Follow up to my previous post...

You can also use the Transform option in the NanoVNA menu to do TDR
measurements without connecting to a PC. From the Display menu you select
a trace and set it to S11 mode and set it to Linear display. From the
Transform menu select Transform On and Low pass impulse. You set the VF
(in %) which is for the cable being tested. You need to look this up in
the manufacturers spec or measure it using a scope and pulse generator.
Be aware that the actual VF can be different by a percent or two from the
nominal published spec due to manufacturing tolerances. It also varies
between manufacturers for the same cable type. Below are 3 screenshots
taken on a NanoVNA-H4 using a VF of 63, 64 and 65 %. The measured VF using
the pulse generator/scope method was 64%. Physical cable length was 9.38
meters and using 64% VF NanoVNA calculated 9.327 meters

Roger

On Thu, Jun 8, 2023 at 11:41 AM, Stan Dye wrote:

A far easier and more accurate way to measure cable length is by using the
"Measure -> Cable" function in the nanovna firmware (if you have a firmware
with this function).
Just connect to an open-ended cable, and read the length and cable loss on
the left side of the screen. You still have to set the correct VF, of
course. But this method does not use TDR and transforms, with the
associated complications. Instead, it finds the quarter-wavelength
frequency of the cable by finding the first S11 phase reversal, and does
the simple calculation.

Stan,

I have to disagree with you on this one. The method uses the 1/4 wavelength measurement method but it is prone to more error that the TDR method. Here are some actual measurements to illustrate my point...

I used the same test cable as in my previous post. This was a 9.38 meter long RG58 A/u cable with a measured 64% VF using a pulse/scope TDR technique. Using NanoVNA Saver the calculated value was 9.34 meters. Using the NanoVNA transform yielded 9.33 meters. These are both 5 cm (about 2 inches).

Using the Cable Measure feature the length depended on the calibrated frequency range. Here are the results with various frequency ranges and an open or short BNC cal terminations on the end. VF set to 64%. Screenshots attached.

50 kHz to 900 MHz. 9.240 meters (open) 9.331 meters (short)
50 kHz to 100 MHz. 9.586 meters (open) 9.608 meters (short)
50 kHz to 20 MHz. 9.588 meters (open) 9.610 meters (short)

One can clearly see that the results are much worse than the TDR method in Saver and on the NanoVNA

Looking at the screenshots you can see the following. The 1/4 wavelength frequency is about 5 MHz.. VF varies with frequency (see attached plot) and does not level off to the "nominal VF" published by the manufacturer until you are much higher in frequency. So this 1/4 wavelength method is subject to the user inputting a VF which is not correct for the 1/4 wavelength measuring frequency. The problem gets worse for longer cables because the 1/4 wavelength frequency is even lower! The second issue is that the user has to calibrate for a frequency range that does not result in too much interpolation which is what happened for the 50 kHz. to 900 MHz. measurement posted above.

In summary the Cable measure function in the nanoVNA requires fewer setup steps but is prone to greater error because the VF that is entered is different than the nominal published VF or one measured using a pulse/scope or TDR instrument. Users need to be wary of using a feature without know how it works or its limitations.

Roger

On Thu, Jun 8, 2023 at 07:34 AM, Bryan Curl wrote:

Still its just an estimate.

...which totally depends upon the correct velocity factor; any error in that will have a proportionate effect on your (physical) length measurements.
73, Don N2VGU

At 50 kHz for paired telephone cable, the the velocity of propagation is somewhat lower than at higher frequencies. For 22 gauge plastic insulated pairs, AIEE Paper 59-778, "Transmission Characteristics of Polyethylene Insulated Telephone Cables at Voice and Carrier Frequencies," lists velocities that correspond to a VF of about 57 percent at 40 kHz and about 64 percent at 1 MHz.

Coax is, of course, not identical to paired cable, but using a low end of 50 kHz makes me wonder whether the velocity might be lower than the published high frequency value sufficiently to skew the results a bit. Maybe using a higher frequency low end would be better. I don't know enough about the characteristics of coax at low frequencies to be sure, just raising a caution here.

73,

Maynard
W6PAP

Kindly allow me to restate my objective

Find out if the unknown length of coaxial cable is in multiples of 1/4 wavelength of selected frequency.
My tool of choice is nanoVNASaver and its TDR option.

I can estimate the cable VF or use preassigned cable selections which includes VF.

At this point I am making some progress using nanoVNASaver and do not want to dilute my efforts by using other tools.

Since my task is to determine the suitability of the existing unknown length of the cable to act as "in line transformer" in 14 MHz band

I like to know why other authors use "sweep " starting at 50 KHz ?

If I set nanoVNASaver to sweep from 50KHZ to 14,5MHz my cable length comes out over 22 meters,
with sweep range 13 to 14,5 MHz I get 13 meters ,

I cannot physically verify either - my cable *(dual RG58 )runs thru walls and attic...

But I can run known length "in the open" for verification purposes.

Coax is, of course, not identical to paired cable, but using a low end of 50 kHz makes me wonder whether the velocity might be lower than the published high frequency value sufficiently to skew the results a bit. Maybe using a higher frequency low end would be better. I don't know enough about the characteristics of coax at low frequencies to be sure, just raising a caution here.


BEFORE I get "analyzing" the TDR (nanoVNASaver) source code , allow me to take another stab at this .

In basic steps - how do I setup to measure length of KNOWN type of coaxial cable using TDR - in nanoVNASaver ?

Assuming that TDR uses specific frequency to perform the task (?)

Setup nanoVNASaver sweep starting frequency
setup nanoVNASaver sweep ending frequency
run sweep
open / push "Time Domain Reflectometry..." in TDR "window"
observe TDR graph and "ESTIMATED cable length:"
change / verify the cable type (!) - changes VF according to cable type
optionally fiddle with TDR graph X-axis whose parameters are in "meters" , not in nS.

Length is measured in meters (duh) but TDR calculates using "time" and reflection time is related to frequency used.

Does all this makes some sense or am I missing IMPORTANT part / option / parameter?

All this using nanoVNASaver - no other tool , for time being...

On 6/9/23 6:49 AM, Anne Ranch wrote:

Coax is, of course, not identical to paired cable, but using a low end of 50 kHz makes me wonder whether the velocity might be lower than the published high frequency value sufficiently to skew the results a bit. Maybe using a higher frequency low end would be better. I don't know enough about the characteristics of coax at low frequencies to be sure, just raising a caution here.

This would be known as "dispersion" and would cause problems with either the pulsed (time domain) or swept frequency (synthetic time domain) approach. A fast pulse (as used in TDR) has all frequencies in it: feed it into a dispersive medium (different velocity for different frequencies) and the envelope of the pulse will be stretched out - it won't be a sharp pulse any more.

The swept frequency technique (e.g. as used in a VNA) measures the amplitude and phase at a series of frequencies, does an inverse Fourier Transform to get that back into time domain, and you'll see the same smeared out pulse.

BEFORE I get "analyzing" the TDR (nanoVNASaver) source code , allow me to take another stab at this .
In basic steps - how do I setup to measure length of KNOWN type of coaxial cable using TDR - in nanoVNASaver ?
Assuming that TDR uses specific frequency to perform the task (?)

It uses a range of frequencies

Setup nanoVNASaver sweep starting frequency
setup nanoVNASaver sweep ending frequency
run sweep
open / push "Time Domain Reflectometry..." in TDR "window"
observe TDR graph and "ESTIMATED cable length:"
change / verify the cable type (!) - changes VF according to cable type
optionally fiddle with TDR graph X-axis whose parameters are in "meters" , not in nS.
Length is measured in meters (duh) but TDR calculates using "time" and reflection time is related to frequency used.

Not exactly.. it makes use of the frequency domain<>time domain duality from a Fourier transform. It measures in frequency domain, and then *displays* in time domain. So you need "all frequencies" (or at least enough frequencies).

Does all this makes some sense or am I missing IMPORTANT part / option / parameter?
All this using nanoVNASaver - no other tool , for time being...

Same idea whether done in the box or in software afterwards.

On 6/9/23 07:52, Jim Lux wrote:

On 6/9/23 6:49 AM, Anne Ranch wrote:

Coax is, of course, not identical to paired cable, but using a low end of 50 kHz makes me wonder whether the velocity might be lower than the published high frequency value sufficiently to skew the results a bit. Maybe using a higher frequency low end would be better.? I don't know enough about the characteristics of coax at low frequencies to be sure, just raising a caution here.

This would be known as "dispersion" and would cause problems with either the pulsed (time domain) or swept frequency (synthetic time domain) approach.? A fast pulse (as used in TDR) has all frequencies in it: feed it into a dispersive medium (different velocity for different frequencies) and the envelope of the pulse will be stretched out - it won't be a sharp pulse any more.
The swept frequency technique (e.g. as used in a VNA) measures the amplitude and phase at a series of frequencies, does an inverse Fourier Transform to get that back into time domain, and you'll see the same smeared out pulse.

This is true. Setting the low frequency such that the high frequency VF is closely approximated throughout the range of frequencies used would minimize the spreading and make determination of the length of the cable a bit more precise.

Radio folks are generally familiar with the high frequency approximations to the complex characteristic impedance and complex propagation constant for transmission lines, but there are also low frequency approximations used by telephone engineers that are pretty accurate at voice frequencies. 50 kHz is in the transition region of frequencies where neither the high or low frequency approximations will be very accurate and calculations may need to be made using the full set of primary constants.

During measurements, the characteristic impedance can be seen to become increasingly complex, rather than nearly real, as the frequency is lowered and the values of the attenuation and phase shift (in neper and radians per unit length) will tend to converge toward each other until their angle is almost 45 degrees at voice frequencies.

Although there is a wealth of information out there on telephone cable pairs for lower frequencies, I haven't seen anything similar for coax, maybe because coax is seldom used for voice frequency transmission. So 50 kHz brings up concerns in my mind, but it would be interesting to see something about how coax actually behaves at such a frequency.

I expressed my concerns about the low frequency of 50 kHz due to previous experience with another type of transmission line, but I can't help with nanoVNASaver because I use my nanonVNA only by itself so far.

73,

Maynard
W6PAP

BEFORE I get "analyzing" the TDR (nanoVNASaver)? source code? , allow me to take another stab at this .

In basic steps - how do I setup to measure length of? KNOWN type of coaxial cable using TDR? - in nanoVNASaver ?

Assuming that TDR uses specific frequency to perform the task (?)

It uses a range of frequencies

Setup? nanoVNASaver sweep? starting frequency
setup nanoVNASaver sweep? ending frequency
run sweep
open / push "Time Domain Reflectometry..."? in TDR? "window"
observe TDR graph and "ESTIMATED cable length:"
change / verify the cable type (!) - changes VF according to cable type
optionally fiddle with TDR graph X-axis? whose parameters are in "meters" , not in nS.

Length is measured in meters (duh) but TDR calculates using "time" and reflection time is related to frequency used.

Not exactly.. it makes use of the frequency domain<>time domain duality from a Fourier transform. It measures in frequency domain, and then *displays* in time domain.? So you need "all frequencies" (or at least enough frequencies).

Does all this makes some sense or am I missing IMPORTANT part / option / parameter?

All this using nanoVNASaver? - no other tool , for time being...

Same idea whether done in the box or in software afterwards.

On Fri, Jun 9, 2023 at 08:57 AM, Maynard Wright, P. E., W6PAP wrote:

you need "all frequencies" (or at least enough frequencies).

OK, the discussion did covered the theory , now is the time to apply it and be practical.
.

I shall repeat
the task is -
is the unknown length of cable in multiples of 1/4 wavelength @ 14 MHz band,?

hence

what should the sweep frequencies be ?

On 6/9/23 9:21 AM, Anne Ranch wrote:

On Fri, Jun 9, 2023 at 08:57 AM, Maynard Wright, P. E., W6PAP wrote:

you need "all frequencies" (or at least enough frequencies).

OK, the discussion did covered the theory , now is the time to apply it and be practical.
.
I shall repeat
the task is -
is the unknown length of cable in multiples of 1/4 wavelength @ 14 MHz band,?
hence
what should the sweep frequencies be ?

Step size (span/number of points) sets the "unambiguous range" - if the steps are too far apart, then the true length might be a multiple of 1/step size longer.

So, if you've got, say, 100 ft of coax - that's about 150 nanoseconds one way, and it's a two way measurement, so you want your step size to be smaller than 1/300 ns or around 3 MHz

The span sets the resolution (although one can interpolate) - 100 MHz span is 10 ns resolution.

Its probably best to avoid crossing the 300 MHz "change in harmonics" boundary.

With a 101 point NanoVNA, then doing 1 MHz to 201 MHz gives you 5 ns (~1.5 ft) resolution (two way time) and allows for more than 100 ft.

If you've got a 401 point NanoVNA, go for 1-300 MHz and whatever step size that gives you.

Here is a new twist ....
I replaced current , unknown type , unmarked type of coax with
known, good RG6 length plainly marked in cubits and converted them to meters - approximately 35 meters.
Now when I set sweep to 10 to 60 MHz TDR "estimated length" is around 28 meters - with open ended coax.
( Not even close to 35 meters ...)
I started looking at the software and so far it looks as there is "the original " nanoVNASaver and several "forks", some for windoze only.
It would be nice if somebody who KNOWS where " TDR" software resides help me and point me in the right direction.
YES, I can download all and then I can search but if somebody already knows where to look , let's share that info, please...
and thanks
...