measuring quarter and half wave coax line lengths

[w5un]

I'm a brand new user. I want to be real certain about how to cut exact 1/4 and 1/2 wavelength coax lines at 144.1 mHz. I have an old Autek RF-5 VHF analyzer that I used for 2 meters. I noticed that AIM4170 tells me a 1/4 wl length for my LMR400 coax is around 17.5 inches. The Autek has a Z null at 15.5 inches. So I'm a little confused as to what to believe.

I made a calibration file using the BNC-UHF adapter to cancel out the adapter physical length. The way I measured with the AIM 4170 was to put a PL259 on the coax under test, connect it to the adapter, run SCAN and prune until the first resonant frequency read 144.1 mHz. I then measured length from the front of the PL259 shell to the end of the coax. Is this the correct way to do it? I need a little reassurance here. Please guide me.

[dave]

Hello:
My name is Dave (VE3KL):
You might want to get some data from Times Microwave about LMR400.

I have calculated the the length for a one quarter wavelength piece of LMR400 as 14.2 inches at a frequency of 144.1 MHz. This is based on a velocity factor of 0.695.

Your measurement of 17.5 inches seems to be too high by about 20 %.

73 Dave

[Bob]

The location of the measurement plane is very important for coax stubs. If additional connectors/adapters are used, they should be included in the calibration path. The open, short and load resistor should be connected at the same point where the cable will be connected. This cancels out the path length and stray capacitance/inductance due to the interconnections.

Recently it was noticed that the early version of the "short" cal device included with the AIM4170 package had 6 or 7 nH of extra inductance that affected the readings of stubs. This stray inductance can be greatly reduced by a slight modification of this load device, as shown in this picture:
w5big.com/Short_cal.jpg
The wire is removed and the ground tab is bent over to make contact with the center conductor directly.

Right now I'm evaluating the calibration procedure and I'll have an update in a few days.

73/ Bob

[w5un]

Thanks for the response. I have measured actual LMR400 coax velocity of propagation as low as .65, and as high as .85, so using vp as a reliable method of cutting exact stubs is a risky non-starter.

Bob, how much repeatability can I expect at 144 mHz? I can get variances of as much as one or more mHz using the same calibration file, but measuring the same stub at different times. Is this normal, or is something not quite right? I assume that distance to fault method, 1/4WL method, and regular scan method will all give 1/4wl frequency for an open stub. Is this correct? Sometimes all three methods will give nearly exact results. Most of the times they do. At other times I will see something as far away as 1 mHZ. Hard to know what to believe when this happens. The same unaltered stub, 17 inches of LMR400, is being tested every time. and gives 1/4 wl frequency of 144.6XX mHz, most of the time. I have seen 143.7 mHz for this same stub using the same cal fine, but at a different test time (say an hour later).

I've got to believe the analyzer is much more accurate than my old Autek RF-5 VHF analyzer. Have you made any comparisons between them?

By the way, I used the BNC-UHF adapter to insert open, short and standard resistor. I am being very careful to keep lead lengths as near to zero as possible.I assume this will then cancel out the added length of the adapter, leaving the cable plus connector under test as the only length being tested. Let me know if there is more to it than that.

[Bob]

w5un said:

Thanks for the response. I have measured actual LMR400 coax velocity of propagation as low as .65, and as high as .85, so using vp as a reliable method of cutting exact stubs is a risky non-starter.

Bob, how much repeatability can I expect at 144 mHz? I can get variances of as much as one or more mHz using the same calibration file, but measuring the same stub at different times. Is this normal, or is something not quite right? I assume that distance to fault method, 1/4WL method, and regular scan method will all give 1/4wl frequency for an open stub. Is this correct? Sometimes all three methods will give nearly exact results. Most of the times they do. At other times I will see something as far away as 1 mHZ. Hard to know what to believe when this happens. The same unaltered stub, 17 inches of LMR400, is being tested every time. and gives 1/4 wl frequency of 144.6XX mHz, most of the time. I have seen 143.7 mHz for this same stub using the same cal fine, but at a different test time (say an hour later).

I've got to believe the analyzer is much more accurate than my old Autek RF-5 VHF analyzer. Have you made any comparisons between them?

By the way, I used the BNC-UHF adapter to insert open, short and standard resistor. I am being very careful to keep lead lengths as near to zero as possible.I assume this will then cancel out the added length of the adapter, leaving the cable plus connector under test as the only length being tested. Let me know if there is more to it than that.

The repeatability of the AIM should be around 0.01 to 0.02MHz at 150MHz. You can double check the freq by listening to it with your receiver.
Although the calibration procedure affects the accuracy, it does not affect the repeatability. It looks like your cal procedure is good.

To see if there is a mechanical influence when handling the cable, set up the scan limits so you can see the quarter wave freq, say 140 to 150MHz. Set the step size for about 50 points. Average=4. Then use the Recycle mode and look at the Resonant freq in the upper left corner of the display. Watch this value as you move the cable around.

I just tried this with a stub that is resonant at 50MHz and the variation is around 0.5% due to flexing. There may also be a relaxation effect where the cable properties change slowly over time after it has been flexed.

Since your measurement seems to vary over a long period of time, look at it every once in a while during the day to see if it might be correlated with the room temperature.

I don't have an Autek analyzer. Maybe someone can comment on this.

73/ Bob

[w5un]

Thanks Bob, that did the trick. The position of the 1/4 wl LMR400 stub, and its proximity to other objects is causing the variation. I isolated the stub from surrounding objects, and then when scanned. The results of each scan agreed: 144.1 mHz for 17.3 inches length

Whenever I moved the stub close to other objects, the frequency would change as you said, moving up in frequency a bit.

I computed the physical length using LMR400's vp of 0.85. The computed length was exactly the same: 17.3 inches at 144.1 mHz. I am satisfied the analyzer is giving VERY GOOD results.

Lesson learned: keep the measured coax as far away from other objects as possible, and don't let it flex during measurement.