开云体育

Bad Coax...feedback requested

Robin Midgett All Messages By This Member #9185 I've calibrated my NanoVNA & connected it to an approximate 350' length of coax with a known good dummy load on the opposite end. The S11 plot is attached...not pleasing at all. Can anyone on the list say what the cause may be? I have ideas, but I don't want to sway anyone's opinion.. 350_ coax with dummy load.png
Clyde Spencer All Messages By This Member #9186 Did you put the load on the output of the VNA? This looks somewhat normal but the return loss is a bit marginal. I am curious about the return loss of the load by it self. Do you have the VNA saver program to look at the line in the TDR format? toggle quoted message Show quoted text On Sun, Jan 5, 2020, 7:28 PM Robin Midgett <K4IDC@...> wrote: I've calibrated my NanoVNA & connected it to an approximate 350' length of coax with a known good dummy load on the opposite end. The S11 plot is attached...not pleasing at all. Can anyone on the list say what the cause may be? I have ideas, but I don't want to sway anyone's opinion..
Clyde Spencer All Messages By This Member #9187 Most rf lines like low density foam has a spec return loss of about 20 db. toggle quoted message Show quoted text On Sun, Jan 5, 2020, 7:28 PM Robin Midgett <K4IDC@...> wrote: I've calibrated my NanoVNA & connected it to an approximate 350' length of coax with a known good dummy load on the opposite end. The S11 plot is attached...not pleasing at all. Can anyone on the list say what the cause may be? I have ideas, but I don't want to sway anyone's opinion..
Robin Midgett All Messages By This Member #9189 Additional info...it's a piece of 1.25" hardline, and the response is much worse going up in frequency; ~-5 return loss at 146 MHz.
W0LEV All Messages By This Member #9191 Port 00 or Port 01, that used for one-port measurement of s11, may not be a good 50+j0 source over your frequency range. You exhibit everywhere better than 15 dB return loss which is pretty good, but the periodicity of the measurement is a bit disturbing. Try recalibrating and remeasuring with a 6 dB attenuator on the s11 port. That will assure a return loss in excess of 12 dB and stabilize the impedance of that port. Dave - W?LEV toggle quoted message Show quoted text On Mon, Jan 6, 2020 at 12:28 AM Robin Midgett <K4IDC@...> wrote: I've calibrated my NanoVNA & connected it to an approximate 350' length of coax with a known good dummy load on the opposite end. The S11 plot is attached...not pleasing at all. Can anyone on the list say what the cause may be? I have ideas, but I don't want to sway anyone's opinion.. -- Dave - W?LEV Just Let Darwin Work Just Think
Robin Midgett All Messages By This Member #9193 Hi Clyde, Thanks for the quick response. Here's a plot of the dummy load across the same span....looks good to me, but I'm here to learn from others. I haven't tried VNA Saver yet...would have to spend some time on that; hopefully this week. dummy load same sweep.jpg
Robin Midgett All Messages By This Member #9194 Thanks Dave, I'll try that soon..hopefully tomorrow evening. toggle quoted message Show quoted text On Sun, Jan 5, 2020 at 06:47 PM, David Eckhardt wrote: Port 00 or Port 01, that used for one-port measurement of s11, may not be a good 50+j0 source over your frequency range. You exhibit everywhere better than 15 dB return loss which is pretty good, but the periodicity of the measurement is a bit disturbing. Try recalibrating and remeasuring with a 6 dB attenuator on the s11 port. That will assure a return loss in excess of 12 dB and stabilize the impedance of that port. Dave - W?LEV On Mon, Jan 6, 2020 at 12:28 AM Robin Midgett <K4IDC@...> wrote: I've calibrated my NanoVNA & connected it to an approximate 350' length of coax with a known good dummy load on the opposite end. The S11 plot is attached...not pleasing at all. Can anyone on the list say what the cause may be? I have ideas, but I don't want to sway anyone's opinion.. -- Dave - W?LEV Just Let Darwin Work Just Think
David J Taylor All Messages By This Member #9200 From: Robin Midgett I've calibrated my NanoVNA & connected it to an approximate 350' length of coax with a known good dummy load on the opposite end. The S11 plot is attached...not pleasing at all. Can anyone on the list say what the cause may be? I have ideas, but I don't want to sway anyone's opinion.. Attachments: 350_ coax with dummy load.png: /g/nanovna-users/attachment/9185/0 ============================== Robin, You might try a TDR plot and see whether there is an obvious discontinuity? Cheers, David -- SatSignal Software - Quality software for you Web: Email: david-taylor@... Twitter: @gm8arv
Dr. David Kirkby, Kirkby Microwave Ltd Member Profile All Messages By This Member #9203 On Mon, 6 Jan 2020 at 00:28, Robin Midgett <K4IDC@...> wrote: I've calibrated my NanoVNA & connected it to an approximate 350' length of coax with a known good dummy load on the opposite end. The S11 plot is attached...not pleasing at all. Can anyone on the list say what the cause may be? VNA measurements require phase stables cables. The ones I use for microwave use (upto 26.5 GHz) are 24” long and cost $5000 a pair. The changes in phase you will observe with very long cables is going to be much higher than short cables of the same quality. Flexing, temperature changes are really going to screw you up with long cables. When I do critical measurements I try to eliminate cables completely. Measurements of my VNA calibration standards sold in kits are made There are some ways to test cables for their suitability for VNA use. I will add some details later on my website under the FAQ section. There are other issues too, but they might not be relevant on the NanoVNA. As the frequency of the source is swept, the detector looks for that frequency. If there’s a very long cable, by the time the reflection is seen back at the VNA, the source has changed frequency and its detector is then looking for higher frequency than what it receives. The solution on some VNAs is to slow the sweep speed down. Since the NanoVNA does sweep that fast, it is unlikely to be the problem. Reducing the span to the absolute minimum you can get away with might help, as the frequency shift between each point will be much lower. What’s the temperature stability of your load like? If you take a load from room temperature and change its temperature much, it is likely to change in value. I have ideas, but I don't want to sway anyone's opinion.. I wonder if any of your ideas are the same as mine! Dave, G8WRB. -- Dr. David Kirkby, Kirkby Microwave Ltd, drkirkby@... Telephone 01621-680100./ +44 1621 680100 Registered in England & Wales, company number 08914892. Registered office: Stokes Hall Lodge, Burnham Rd, Althorne, Chelmsford, Essex, CM3 6DT, United Kingdom
Dana Whitlow All Messages By This Member #9207 Robert, It would be desirable if you could also measure the transmission through the cable, including both magnitude and phase versus frequency. Plus any info you have about the supposed characteristic impedance and velocity factor of the cable. Extend the lower frequency limit down to 50 kHz as well. A second set of equivalent plots but with the nano's scan set for a span of about 10 MHz and a center frequency well up in the VHF regime (~150 MHz would be fine) might also be pretty informative. These should also include the scan of the termination that you're using by itself. The overall shape of the plot is somewhat consistent with two point reflections of roughly equal magnitude, separated in round-trip time by a bit under 500 nsec. However, the deteriorating overall return loss at increasing frequency is not really consistent with that simple model. But if your connections to the hard line were a bit crude, creating added impedance bumps at the two ends, the curve I've seen so far would be better-explained. Could you also send a photo showing exactly how you are physically interfacing to the hard line from the small cable environment? What I've seen so far suggests a cable whose impedance is not really 50 ohms, combined with some other issue such as (perhaps) impedance bumps at the two ends. The period of ~2 MHz seen in the S11 plot is not consistent with a physical length of 350 ft- in fact it's off by almost a factor of 2 if one assumes a VF of 0.67. Dana K8YUM
Tim K4SHF All Messages By This Member #9209 I missed what was actually being tested.... Also, if one is testing cables themselves, is there a finite distant limit? BR Tim K4SHF toggle quoted message Show quoted text On Mon, Jan 6, 2020 at 6:28 AM Dana Whitlow <k8yumdoober@...> wrote: Robert, It would be desirable if you could also measure the transmission through the cable, including both magnitude and phase versus frequency. Plus any info you have about the supposed characteristic impedance and velocity factor of the cable. Extend the lower frequency limit down to 50 kHz as well. A second set of equivalent plots but with the nano's scan set for a span of about 10 MHz and a center frequency well up in the VHF regime (~150 MHz would be fine) might also be pretty informative. These should also include the scan of the termination that you're using by itself. The overall shape of the plot is somewhat consistent with two point reflections of roughly equal magnitude, separated in round-trip time by a bit under 500 nsec. However, the deteriorating overall return loss at increasing frequency is not really consistent with that simple model. But if your connections to the hard line were a bit crude, creating added impedance bumps at the two ends, the curve I've seen so far would be better-explained. Could you also send a photo showing exactly how you are physically interfacing to the hard line from the small cable environment? What I've seen so far suggests a cable whose impedance is not really 50 ohms, combined with some other issue such as (perhaps) impedance bumps at the two ends. The period of ~2 MHz seen in the S11 plot is not consistent with a physical length of 350 ft- in fact it's off by almost a factor of 2 if one assumes a VF of 0.67. Dana K8YUM
[email protected] All Messages By This Member #9213 You might try rapping the connectors on the ends of the line and check for any changes in S11 or S21 response. S21 loss should be less then -0.5 dB @ 30 MHz for 350' of 1.25" hardline. If connectors are good and cable has not taken a lightning strike in its past life you should be more than 20 dB return loss on S11.
Clyde Spencer All Messages By This Member #9244 What does the VSWR trace look like? Clyde K. Spencer toggle quoted message Show quoted text On Sun, Jan 5, 2020 at 7:28 PM Robin Midgett <K4IDC@...> wrote: I've calibrated my NanoVNA & connected it to an approximate 350' length of coax with a known good dummy load on the opposite end. The S11 plot is attached...not pleasing at all. Can anyone on the list say what the cause may be? I have ideas, but I don't want to sway anyone's opinion..

Previous Topic Next Topic