Keyboard Shortcuts
ctrl + shift + ? :
Show all keyboard shortcuts
ctrl + g :
Navigate to a group
ctrl + shift + f :
Find
ctrl + / :
Quick actions
esc to dismiss
Likes
Search
Which analyzer is more accurate?
I measured my 40M OCF Dipole with my NanoVNA-H4 and it shows that the minimum SWR of 1.68 is at 6.975 MHz. I had the loan of an MFJ-259C, and it shows the minimum SWR of 1.6 is at 7.14 MHz. I also have an Arduino based SWR analyzer designed by K6BEZ and it shows the minimum SWR of 1.7 is at 6.92 MHz. The SWR figures align nicely, but the center frequency is not very close. In the 40M band, this difference represents about half of the band. I calibrated the NanoVNA, and scanned from 6.5 to 7.5 MHz so precision of the 101 steps should not be an issue. Which device(s) do I believe? How can I resolve this difference?
-- Bill AA6BD |
On 2/5/21 2:34 PM, Bill AA6BD wrote:
I measured my 40M OCF Dipole with my NanoVNA-H4 and it shows that the minimum SWR of 1.68 is at 6.975 MHz. I had the loan of an MFJ-259C, and it shows the minimum SWR of 1.6 is at 7.14 MHz. I also have an Arduino based SWR analyzer designed by K6BEZ and it shows the minimum SWR of 1.7 is at 6.92 MHz. The SWR figures align nicely, but the center frequency is not very close. In the 40M band, this difference represents about half of the band. I calibrated the NanoVNA, and scanned from 6.5 to 7.5 MHz so precision of the 101 steps should not be an issue. Which device(s) do I believe? How can I resolve this difference?Your question would be "is it a frequency measurement problem" or "is it a impedance measurement problem" - The first one could be answered by measuring a piece of coax that is shorted or open at the far end. It's pretty high Q, and would be unaffected by "fixturing". It's about a 2% frequency different you're observing, which is quite large, considering all three are essentially referenced to a crystal (timebase in the MFJ's counter, PLL ref in the Nano, whatever in the K6BEZ, but almost certainly a crystal). And a 2% error in crystal frequency is enormous. So the next question is "what is the actual R+X being measured" (as opposed to SWR). at resonance, do they give the same numbers? And then, all 3 have varying degrees of sensitivity to RFI - you're measuring an antenna that's outdoors. It's an interesting problem, but I'd rule out frequency errors first |
Bill,
I would suspect that all three of those meters are right, and that the antenna's behavior is changing according to what you connect to it! It's and OCF dipole, an antenna that is highly unbalanced. You should have a very good common mode choke on that feedline to be able to make any meaningful measurements - and to use it, too. Otherwise not just the antenna will be radiating and receiving signals, but the antenna, the feedline, the ground connection, the radio, the instrumentation, the house wiring, and yourself too! Manfred |
Mel Farrer
One thing that I did to make sure oranges were oranges and apples were apples.? I constructed a reactive termination for the analyzer.? Based on 10 MHz, I built up a series SMT structure of a 0603 50 ohm and a 0603 -j50 ohm cap to ground.? After all of the SOL calibration I put the reactive load on the analyzer and it better read 50-j50 ohm at 10.000 MHz.? What does this do?? It confirms that the calibration is linear from short to open.? Some of the analyzers have a problem here.? I sleep a lot better at night now.? I have junked a couple of analyzers that did not pass this test.
toggle quoted message
Show quoted text
Mel, K6KBE On Friday, February 5, 2021, 02:54:29 PM PST, Jim Lux <jim@...> wrote:
On 2/5/21 2:34 PM, Bill AA6BD wrote: I measured my 40M OCF Dipole with my NanoVNA-H4 and it shows that the minimum SWR of 1.68 is at 6.975 MHz.? I had the loan of an MFJ-259C, and it shows the minimum SWR of 1.6 is at 7.14 MHz.? I also have an Arduino based SWR analyzer designed by K6BEZ and it shows the minimum SWR of 1.7 is at 6.92 MHz.? The SWR figures align nicely, but the center frequency is not very close.? In the 40M band, this difference represents about half of the band.? I calibrated the NanoVNA, and scanned from 6.5 to 7.5 MHz so precision of the 101 steps should not be an issue.? Which device(s) do I believe?? How can I resolve this difference?Your question would be "is it a frequency measurement problem" or "is it a impedance measurement problem" - The first one could be answered by measuring a piece of coax that is shorted or open at the far end. It's pretty high Q, and would be unaffected by "fixturing". It's about a 2% frequency different you're observing, which is quite large, considering all three are essentially referenced to a crystal (timebase in the MFJ's counter, PLL ref in the Nano, whatever in the K6BEZ, but almost certainly a crystal). And a 2% error in crystal frequency is enormous. So the next question is "what is the actual R+X being measured" (as opposed to SWR). at resonance, do they give the same numbers? And then, all 3 have varying degrees of sensitivity to RFI - you're measuring an antenna that's outdoors. It's an interesting problem, but I'd rule out frequency errors first |
Manfred,
The OCF Dipole has a common mode choke at the feed point. I ran a coax to my shack and connected all three analyzers at the same end of the coax and at the same physical position. The only physical difference is that the MFJ analyzer has a PL259 connector on it, the NanoVNA has a PL259 to SMA adapter, and the K6BEZ analyzer has a PL259 to BNC adapter. I find it hard to believe that this much frequency difference is due to the adapters, and all three were measured at the same operating position. I have an older NanoVNA (V1?) and it gave almost the same measurement as the NanoVNA H4. Since three of four analyzers yield almost the same values, while the MFJ is off by itself, I suspect it is the inaccurate one but I will test some more. -- Bill AA6BD |
On 2/5/21 3:31 PM, Bill AA6BD wrote:
Manfred,The choke at the feedpoint doesn't mean that the coax is totally decoupled from the antenna - the fields from the antenna will couple to the shield of the coax. A symmetric dipole, with the feedline coming away perpendicular, will be decoupled, but that's a special case. If your OCF feedline has chokes every few meters, then it's not going to "participate" But I'd check the frequency accuracy of the MFJ - There's a 4MHz crystal that sets the frequency of the PIC16C73. |
On Fri, Feb 5, 2021 at 03:31 PM, Bill AA6BD wrote:
I have an older NanoVNA (V1?) and itBill, The MFJ analyzers are not that accurate. I own a MFJ-259B and results will vary with battery voltage! There is a calibration procedure for them and I suspect some leave the MFJ facility without proper calibration. I know that sounds unbelievable because MFJ is known for their rigorous quality control but it was the case on my unit. Roger |
Hi Bill,
toggle quoted message
Show quoted text
The MFJ seems to be the outlier. Consider that a man with a clock always knows what time it is. A man with two or three clocks can never be sure:) Other replies to your post point at other factors so I won't go over those. 73, Bill? KU8H t'uther Bill On 2/5/21 5:34 PM, Bill AA6BD wrote:
I measured my 40M OCF Dipole with my NanoVNA-H4 and it shows that the minimum SWR of 1.68 is at 6.975 MHz. I had the loan of an MFJ-259C, and it shows the minimum SWR of 1.6 is at 7.14 MHz. I also have an Arduino based SWR analyzer designed by K6BEZ and it shows the minimum SWR of 1.7 is at 6.92 MHz. The SWR figures align nicely, but the center frequency is not very close. In the 40M band, this difference represents about half of the band. I calibrated the NanoVNA, and scanned from 6.5 to 7.5 MHz so precision of the 101 steps should not be an issue. Which device(s) do I believe? How can I resolve this difference? |
On 2/5/2021 6:12 PM, Mel Farrer via groups.io wrote:
And then, all 3 have varying degrees of sensitivity to RFI - you'reThis is definitely something to keep in mind when you are testing an antenna at a tower site, like at a repeater site. Other transmitters at the site can affect the readings. I do antenna testing and certification with an Anritsu Sitemaster analyzer. The Sitemaster has the ability to send a coded signal that it can recognize as a valid test signal. This can eliminate some of the interference (but not always all) at a site during testing. This does affect some testing accuracy. The Sitemaster also has the ability to turn off this coded signal and go into the "CW" mode for greater accuracy. Just something to keep in mind when using the NanoVNA or other device around other live transmitters. Joe |
For what it's worth, I have a similar OFC dipole setup to what you
toggle quoted message
Show quoted text
describe, and the measurements from my nanoVNA-H4 and my MFJ-226 are almost identical. So what you are describing is not what I experience; maybe there is a problem with one of your units. On Fri, Feb 5, 2021 at 5:11 PM Bill Cromwell <wrcromwell@...> wrote:
Hi Bill, |
On Feb 6, 2021, at 2:56 PM, Joe <k1ike_mail@...> wrote:Believe he is referring to a 0603 size chip cap that has 50 ohms reactance at 10 MHz . -Ray WB6TPU |
to navigate to use esc to dismiss