How do I interpret these results? #nanovna-h
|
Greetings All,
I've attached two sweeps of my comet GP-1 antenna. It's mounted on a 10ft pole above my second story chimney (~35ft above ground). It's grounded and bonded, etc. The 2m performance is as I would expect from the antenna, however the 70cm graphs are showing an unexpected sinusoid curve.. What should I make of this result? Is this an indication of a failed nanoVNA? Or is the antenna at fault? Cable is LMR400. Thanks! I'm running DiSlord 1.1 on the NanoVNA-H. And I calibrate before each sweep range change just to make sure. |
|
Ben,
Your 2M plot looks good but there is something wrong with the 70 cm plot. You should not see that much "ripple" with LMR-400. The actual impedance of this cable is very close to 50+j0 at 440 MHz. . You are using NanoVNA Saver for your plots. You need to do the calibration correctly or you will get poor results. Some things to try. 1. I suggest you calibrate on the NanoVNA-H and do a stand-alone measurement first . Do you see the "ripple" on 70 cm. ?? 2. Next connect to NanoVNA Saver and do a sweep without calibrating in NanoVNA Saver. This works because you have already calibrated on the nanoVNA. You should see the same results. 3. You can calibrate a specific range in NanoVNA Saver BUT the NanoVNA must be calibrated over a wider range than the range you want to calibrate over in NanoVNA Saver. I always use calibrated slot 0 over the range 50 kHz to 900 MHz. as my default settings so it is ready to go when used with Saver. If you are still seeing large ripples in Step 1 above try putting a non-inductive 50 ohm load (very short leads) in place of the antenna. A SO239 to SMA adapter and the cal load will work well if you are using PL259 connectors. . You should see a SWR near 1 with only a little ripple (due to transmission line not being exactly 50 +j0 ohms). Try a 75 or 100 ohm resistor if you have one. These tests will help you verify that the NanoVNA and LMR400 are OK Roger |
|
On 1/3/22 12:09 PM, Ben Cranston via groups.io wrote:
Dave, That's a pretty typical thing. What you're seeing is a small mismatch, probably at where the feedline mates with the antenna, but potentially elsewhere (do a TDR and it might be revelatory). If you look at the S11 plot, you can figure out how much it is. I've attached a plot of a 75 or 100 ft cable with the far end open. ideally, you'd expect to see a smooth decrease in S11 magnitude with frequency, but it isn't because at some frequencies the reflection is in phase and others its not, at some the 2 or 3 times around reflection is in phase at others it's not. |
|
Hi Ben,
переключити цитоване повідомлення
Показати цитований текст
as Jim describes it is due to a small mismatch, probably from a bad connector or a bad bend in the koax. I simulated your antenna with a cable and a mismatch (0.7 pF shunt C) at the start of the cable with an old program rfsim99. At 2m the influence of the mismatch is not so big as at 70cm. This is due to the fact, that I simulated the mismatch by a shunt capacitor. See results in my picture. 73, Guenter, DK5DN Am 03.01.2022 um 15:32 schrieb Ben Cranston via groups.io: Greetings All, |
|
The ripples you are seeing are primarily the result of the finite directivity (even after correction) of the RF bridge in the source plus any degradation caused by adapters, etc. added on after calibration. The directivity is the ability of the bridge to reject forward power while sensing reflected power. One thing you can do to make a better measurement of cable loss is to average a measurement with an open cable end with one with a short circuit attached.
|
|
Ben,
The reason I suggested that you check your calibration and measure the cable first was to determine if your test setup was working correctly before you tried to see if your antenna measurements were correct. As others have pointed out seeing ripples in the SWR plot is common and can be due to several reasons like impedance of the transmission cable, kinks in the cable, directivity of the bridge, termination connectors, adapters used etc. At higher frequencies this becomes more evident. The Comet GP-1 uses an SO-239 on the base for mating with a PL-259. These connectors work well at HF and are passable for VHF but have serious shortcomings at UHF. PL259's are basically a banana plug with a metal shell and do not have an impedance of 50 ohms. Poor quality ones have a dielectric that degrades performance at higher frequencies. What they have going for them is they are inexpensive and fairly rugged. If you are trying to calibrate your NanoVNA to measure at UHF you have to use good calibration loads and adapters and even then you will have ripples in the measurement. Poor adapters will result in worse measurements. To illustrate this point I used 50 feet of RG-213/U cable and two different adapter scenarios. I calibrated on the NanoVNA-H4 connectors with the provided SMA cal loads. In the first experiment I attached a short 1 foot adapter cable with a SMA on one end and a SO239 connector on the other to the coaxial cable and terminated the far end with a rigid SO239 to SMA adapter fitted with a 50 ohm cal load. For the second experiment I attached a small SMA to SO239 adapter to the NanoVNA, connected it to the PL259 on the cable and terminated with a barrel connector and a PL259 cal load made from a non-inductive resistor. You can see on the screenshots below that the ripple is considerably worse for the second setup. In summary you should verify your test setup first before measuring the antenna especially when dealing with higher frequencies. Roger. |
|
You'll find that the longer the transmission line is, the more ripples will be present. The attached compares UHF band VSWR test; one test was done on a line 82 feet long, the other around 380 feet long.
It occurs because of the time delay in the transmission line from one end to the other and how it relates to the frequency range of the test. This occurs on both the Nanavna, which is where the photos came from, but it also occurs on commercial VNA's (Anritsu) costing 18K or so... As was pointed out by previous gentlemen, he was able to verify your cable length by measuring the distance of these ripples - when a TDR test is done with a VNA, they are basically "zooming in" on these ripples to determine round trip delay from one end of the line to the other... . As others have mentioned, it is very important to re-calibrate your VNA (short, load, open) when you change frequencies, or test jumper cables/connectors, especially on the higher frequencies. Brian - KI8KY |
|
Good morning everyone. Rather than replying to each individual message, I'm going to summarize my responses in one message to the list. Hopefully that does not break etiquette.
Brian - KI8KY, Ok, so ripples are to be expected in UHF analysis. VHF not so much. Reading your two scans I see worse case VSWR of 2.3:1 and 1.7:1. Am I reading that correctly? As it pertains to the calibration I always re-calibrate with any change of connectors, cables, or stimulus ranges. that’s been my practice from the beginning. Thanks for the help! Roger Need, Thank you for the guidance. I’m thinking it thru and have a question. When I calibrate to test the antenna, should I be putting the termination at the end of the feed line near the antenna? I’d think the calibration point should be at “end” of my adapters and any strain relief cable extensions so I can capture the resultant system of feed line and antenna as it would be presented to the back of a radio. I have tried several combinations of adapter/short extension cables. Interestingly enough, and to your point, I found one combination which introduces other issues. That cable has been tossed out. All the other combinations have lead to the same analysis output I originally shared. Donald Brant Jr, Good idea on the open/short testing. So at the end of the feed line on the antenna side use the open and short calibration standards and use the resultant to form averages for analysis? Jim Lux, Looking at the raw data I don’t believe I’m getting advertised VSWR in the 70cm band. At the low end of the band I’m seeing as high as 2.8:1. Above 440MHz i see it start to go below 2:1. I was anticipating the curve to be below 2:1 across the 70cm band. Perhaps that’s my error in thinking? I still want to re-visit the connection to the actual antenna. The GP-1 (https://www.dxengineering.com/parts/cma-gp-1) was replacing a Tram 1411 (https://www.amazon.com/Tram-1411-Discone-Scanner-Antenna/dp/B00QVNI1V0) which was starting to have issues. Maybe something is askew on the connectors at that end… W0LEV, Attached is a sweep from 430-450MHz. Interesting enough I have an FM broadcast station really close by (High School station) which is a 5/8 harmonic of 440.5MHz. the very place in my graphs where there is noise present. I think I did the math right… Maybe I put the FM block filter I have into the chain to see what happens… ALL Thank you so much for the help and guidance on my questions! I’ve learned quite a bit from everyone. I love the spirit of Amateur Radio and how folks dive in to help each other out. I think I'm all set from the nanoVNA side of the world. Once we have a warmer day here in Michigan I'm going back up the ladder to the roof to inspect the connections just for good measure. 73, Ben KE8CGS |
|
On Wed, Jan 5, 2022 at 07:51 AM, Ben Cranston wrote:
Hi Ben - yes, you are correct - the worse case VSWR of 2.3:1 and 1.7:1 on the two images I provided. For comparison, I have attached a VHF sweep from a transmission line that is also 380 feet long - the ripples are still there, but they are stretched further apart, so to speak, because of the longer wavelength vs. transmission line length. |
|
Feedline interaction. Nothing to worry about. Here is a broad sweep of my Comet SS460 over 4 radials.
-- 73 es God Bless de Ron KK1L <>< KK1L.COM ( http://kk1l.com ) |
Roger Need
schweppe
Donald S Brant Jr
Ron Rossi
