Nano VNA Reads Low SWR

[edtyler]

Hi,

I have read https://www.eevblog.com/forum/rf-microwave/difference-in-swr-readings-between-nanovna-and-swr-meter/50/

and this question is slightly different than what was described there.

My test setup:
NanoVNA (black open case, small screen) - purchased in September 2019. Sweeping from 144-148 MHz, calibrated with lab quality standards.
Homebrew tape measure Yagi fed with 25 feet of Andrew C400 and assorted adapters terminating in a BNC female.

When I measure this antenna using the NanoVNA at 146.565 Mhz, the SWR is reported as 1:1.85

This seemed pretty good, so I attached a transmitter and a Bird 43 with a 5W element just to verify transmitter output as I usually do in a new setup. The output was lower than expected and the reverse power was much higher, Now I understand about the 5% full scale accuracy of the Bird, but still the reverse power was 1.3W with a forward power of 3.7W.

So, I hooked up a MFJ-259B and it reported SWR of 1:3.5 at 146.565. To verify, I hooked up a HP 8753C, sweeping from 130-150MHz, did a one port calibration and observed a SWR of 1:3.27 at 146.565 MHz. To eliminate the cabling, I replaced the antenna with a 50 ohm terminator at the end of the feed line and the result was an SWR of 1:1.12 on the 8753C. The MFJ also reported a similar SWR. The NanoVNA reported 1:09.

I then redid the NanoVNA calibration and got nearly identical results to the first trial using the antenna and the dummy load. I  removed the feed line and connected a 150 ohm carbon resistor, which read 1:1.68 on the Nano VNA. The 8753C read 1:3.002. An open on the NanoVNA reads 1:2.52, with a stable Smith Chart. A short reads 1:infinity and a 50 ohm standard reads 1:1.0, both with stable smith charts.

It appears the Nano VNA's reverse detector is inaccurate on my unit.  My question: Is this a common problem and can it be corrected, or should I throw away this Nano VNA and buy another? I don't recall if I had seen this before and with this type of inexpensive tool, one should expect high precision, so prior, minor variations of +/- 1:1.2 did not worry me. 1:1.18 vs 1:3.27 is not acceptable.

Thanks for any help or ideas you might have.

[joeqsmith]

If you forget all that antenna and feed line stuff and start with some basics.  Make up say a 25 and 100 ohm load, cal the Nano and measure them.  Repeat the measurement with the HP.  At 150ish MHz and lower, I would expect very similar results between the two and I would expect the SWR to come out very close.   

For the resistors, I would just tack down a couple of 0805s or 1206s on the backside of what ever connector type you are using.  I wouldn't be concerned with with reference plane.   

Here you can see me run a similar test on the Nano. 
https://youtu.be/mKi6s3WvBAM?t=2165

[radiolistener]

It appears the Nano VNA's reverse detector is inaccurate on my unit. 

Add ferrite RF-choke on the feeder between NanoVNA/Transmitter and antenna. Without good RF choke you're trying to measure antenna with different counterpoise. It means different antenna, different VSWR. This is why you get different results.

My question: Is this a common problem and can it be corrected

yes, common mode currents on the feeder is a common problem. Especially when you feeding non-symmetrical antennas.

In order to fix, you're needs proper coupling between feeder and antenna and good RF-choke.

[Bud]

An open on the NanoVNA reads 1:2.52, with a stable Smith Chart.

Here is your problem. Your Open is more like 126 Ohm which is far far far from Open. Open should have Infinite impedance and Infinite SWR (not ideal Infinite but a large number nevertheless)

[edtyler]

@joeqsmith Thanks for the reply. I tried a 150 ohm carbon resistor as I said in my email. The SWR should have been 1:3 but was 1:1/68.

I can try other values, but the result at 150 ohms appears to indicate a problem with the NanoVNA, especially since the 8753 showed SWR of 3.001 on the same resistor.

Is this a known problem with the Nano VNAs?

[edtyler]

@radiolistener - I did a test with just a 150 ohm carbon resistor as described in my original description. The Nano VNA measured a SWR of 1:1.68, the 8753 measured 1:3.001.

The antenna setup, including the feedline and all adapters was exactly the same for the NanoVNA and the 8753. Why should a RF choke be needed with one measurement device and not another? Also recall that I used a MFJ-859B and it agreed within a reasonable tolerance with the 8753, as did the Bird 43, considering its 5% full scale spec.

Can you explain further why 3 instruments mostly agree and the NanoVNA should need a choke?

Thanks,

[edtyler]

@Bud - I agree, this appears to indicate a problem with the nanoVNA. Especially considering the 150 ohm resistor test I also described.

Is this a common problem and is there a fix or should I recycle the NanoVNA and get another one or something else (NanoVNA2?)

[DH1AKF_Wolfgang]

Before you give it back, you should recalibrate!
I'm a member of the NanoVNA newsgroups e.g. https://groups.io/g/nanovna-users/topics and there was no error like yours.
Most errors grounded in misunderstanding of the manual.

[joeqsmith]

@joeqsmith Thanks for the reply. I tried a 150 ohm carbon resistor as I said in my email. The SWR should have been 1:3 but was 1:1/68.

I can try other values, but the result at 150 ohms appears to indicate a problem with the NanoVNA, especially since the 8753 showed SWR of 3.001 on the same resistor.

Is this a known problem with the Nano VNAs?

I  removed the feed line and connected a 150 ohm carbon resistor, which read 1:1.68 on the Nano VNA.

In  the above statement there was no mention of calibration prior to making the measurement.   Not being there to watch what you are doing, every detail would need to be stated.   

I've made many measurements with the Nano and as long as I stay below 300MHz, for the most part it works really well for the $50.  If you were asking how many times it requires a power cycle after it locks up from trying to use the poor UI, I would say a lot.   

I then redid the NanoVNA calibration and got nearly identical results to the first trial using the antenna and the dummy load.

It's very possible you had made a mistake, then repeated the same mistake.  Again, it's difficult to tell from where I am sitting.

I am only suggesting that you gain a basic understanding of how to use the tools prior to attempting to make more complex measurements.  You may be juggling more than you know.   

... calibrated with lab quality standards

I borrowed a couple of sets of some lab grade standards not too long ago.  The SMA kit wasn't too bad.  Maybe $4000.   The N, closer to 10.  Not something I would use with the Nano.   For 1GHz,  I have always used the PCBs I show to perform the calibration.   For higher frequencies, I plan to still use home made standards (which was the reason for borrowing the lab grade sets, in order to characterize them).   

No matter.  Cal the Nano with your known lab standards and then attach your 150ohm unknown load and repeat this process with the HP using the same standards.  I am not suggesting that what ever your unknown 150 ohm  carbon resistor is or how you have it attached to the VNA will behave like a pure 150 ohms but I would expect the same results (within reason) from both VNAs.     

Once you get results that make sense, then try adding your 25' of coax into the mix.

[radiolistener]

@radiolistener - I did a test with just a 150 ohm carbon resistor as described in my original description. The Nano VNA measured a SWR of 1:1.68, the 8753 measured 1:3.001.

You said that calibration is done. It means that device shows correct values for Open, Load and Short. Since your device shows incorrect values for known load, this is indication that calibration is wrong. So, you're needs to perform proper calibration before measurement.

[edtyler]

@DH1AKF_Wolfgang and joeqsmith

"My test setup:
NanoVNA (black open case, small screen) - purchased in September 2019. Sweeping from 144-148 MHz, calibrated with lab quality standards."

I am sorry that I did not list the step by step sequence, but it seems like when one gets into a problem deeply sometimes the basic assumptions are missed. To be clear,I calibrated before *each* test and also usually power cycled the NanoVNA, even though the calibration should have been good, at least short term. I did the open and 150 ohm tests 3 times each with nearly identical results.

The lab standards I have are R&S for SMA and HP for N. They were used, but verified for correct operation on very high end VNAs. New cost on each set was > $10K, but I bought them used. On the Nano VNA, my unit does not provide feedback when going through the open, short, iso, load steps. It just moves on to the next step. However, when it is done and the cal set stored in memory 0, the unit does *not* report infinite or near infinite SWR.

I used the same SMA standards for both the NanoVNA and the 8753 (single port cal only). The NanoVNA showed an incorrect SWR (1:1.68) while the 8753 showed 1:3.001, which is fairly close to ideal.

It may be true that I am repeating the same error over and over with the NanoVNA, but the techniques are pretty simple and nearly identical between the NanoVNA and the 8753.  You enter cal mode, then follow the steps listed on the display. Not very complicated and not likely to be done wrong. That said, I wish I knew what error I could be making. Can either of you confirm your NanoVNAs show near infinite SWR with an open?

I have been using higher end VNAs and other RF tools for a few years. The NanoVNA is about 2 years old and when I initally got it I made a series of measurements to determine how precise and useful it was. As you say, below 300MHz, it is not bad. Actually the part used in my unit is spec'd for < 200MHz, so I limited my testing to that range. SWR, and complex impedance was within about 20% of a lab quality VNA up to about 1:4.0. Through measurements suffered greatly from lack of dynamic range, which measured about 45dB at 150MHz, if I recall. This makes it barely useful for filter design and characterization at lower frequencies. But it was $50 and fits in my pocket, so no complaints, until now.

It does sound you both are saying that the expectation of near infinite SWR on an open and roughly 1:3.0 on a 150 ohm resistor are reasonable on a NanoVNA. Sadly, my NanoVNA seems to have packed up and died. It is probably not worth trying to fix, but removing the covers and looking for leakage paths and other issues like cold solder joints and flux  deposits might be worth trying.

I really appreciate the time all of you have taken to respond to my questions.

[joeqsmith]

@DH1AKF_Wolfgang and joeqsmith

"My test setup:
NanoVNA (black open case, small screen) - purchased in September 2019. Sweeping from 144-148 MHz, calibrated with lab quality standards."

I am sorry that I did not list the step by step sequence, but it seems like when one gets into a problem deeply sometimes the basic assumptions are missed. To be clear,I calibrated before *each* test and also usually power cycled the NanoVNA, even though the calibration should have been good, at least short term. I did the open and 150 ohm tests 3 times each with nearly identical results.

The lab standards I have are R&S for SMA and HP for N. They were used, but verified for correct operation on very high end VNAs. New cost on each set was > $10K, but I bought them used. On the Nano VNA, my unit does not provide feedback when going through the open, short, iso, load steps. It just moves on to the next step. However, when it is done and the cal set stored in memory 0, the unit does *not* report infinite or near infinite SWR.

I used the same SMA standards for both the NanoVNA and the 8753 (single port cal only). The NanoVNA showed an incorrect SWR (1:1.68) while the 8753 showed 1:3.001, which is fairly close to ideal.

It may be true that I am repeating the same error over and over with the NanoVNA, but the techniques are pretty simple and nearly identical between the NanoVNA and the 8753.  You enter cal mode, then follow the steps listed on the display. Not very complicated and not likely to be done wrong. That said, I wish I knew what error I could be making. Can either of you confirm your NanoVNAs show near infinite SWR with an open?

I have been using higher end VNAs and other RF tools for a few years. The NanoVNA is about 2 years old and when I initally got it I made a series of measurements to determine how precise and useful it was. As you say, below 300MHz, it is not bad. Actually the part used in my unit is spec'd for < 200MHz, so I limited my testing to that range. SWR, and complex impedance was within about 20% of a lab quality VNA up to about 1:4.0. Through measurements suffered greatly from lack of dynamic range, which measured about 45dB at 150MHz, if I recall. This makes it barely useful for filter design and characterization at lower frequencies. But it was $50 and fits in my pocket, so no complaints, until now.

It does sound you both are saying that the expectation of near infinite SWR on an open and roughly 1:3.0 on a 150 ohm resistor are reasonable on a NanoVNA. Sadly, my NanoVNA seems to have packed up and died. It is probably not worth trying to fix, but removing the covers and looking for leakage paths and other issues like cold solder joints and flux  deposits might be worth trying.

I really appreciate the time all of you have taken to respond to my questions.

It's possible that the pretty simple techniques are where you are missing some step.  For example you never mention resetting the unit prior to the cal.  Even the best standards in the world will not overcome a mistake during the calibration.  I agree, it is simple from my perspective but again, not being able to see what you are doing, hard to say. 

The question of an infinite SWR is new.  I showed that 200ohm resistor, which is a poor mismatch, reading correctly. 

Again, I would make sure you understand how to use it first.  Making up some basic standards and testing them is one way to build confidence in the Nano (or any measurement) and cement how to use it.  Of course, there is no substitute for reading.   

"I have been using higher end VNAs and other RF tools for a few years. The NanoVNA is about 2 years old and when I initally got it I made a series of measurements to determine how precise and useful it was. As you say, below 300MHz, it is not bad. " "SWR, and complex impedance was within about 20% of a lab quality VNA up to about 1:4.0."

If the above statements are true, I am surprised there would be any question.  You know the Nano is working.  You know how to properly calibrate it.  You have the equipment and expertise to verify your results.  I'm missing something.     

It is very possible that once you have a known working setup that you may have problems with your feedline.  I have a section of coax that is loaded up with ferrite along with several clip on parts.   

I was playing around trying to make a crystal filter with some canned software package.  These have all been <20MHz.   
https://www.eevblog.com/forum/rf-microwave/fun-with-crystal-filters/msg3058240/#msg3058240

A few of us were playing with some home made attenuators.   Here you can see a 50dB unit swept to 1.5GHz with the Nano.   Next post I show the same part running on my old HP.   
https://www.eevblog.com/forum/projects/20db-rf-attenuator-seeking-feedback-to-improve/msg2904926/#msg2904926

[edtyler]

@joeqsmith

"If the above statements are true, I am surprised there would be any question.  You know the Nano is working.  You know how to properly calibrate it.  You have the equipment and expertise to verify your results.  I'm missing something"

And that's my real question. Either I am doing something stupid or the NanoVNA I have is broken. Your help and continued feedback are appreciated.

Here is what I did this morning:
1) Make sure NanoVNA battery is fully charged.
2) Turn on unit. Range is 144-148MHz as previously set
3) Select CAL, then RESET
4) Select CALIBRATE, then step through OPEN, SHORT, LOAD, ISOLN (2x 50ohm loads), THRU (using supplied 6" cable), then DONE. I used my R&S SMA lab standards and saved the calibration result in memory zero.
5) After Pressing DONE, with nothing attached, the display shows the SWR trace above the first horizontal division and the value for SWR is 1:2..38. The impedance is shown at 118 ohms with 3.02nH. Smith chart is stable.
6) reinstalling a 50 ohm standard, gives 49.9 ohms and 15.2 nF and a SWR of 1:1.00 SWR trace is at the bottom of the display. Smith chart is stable.
7) reinstalling a short standard gives 944miili Ohms and 4.15 nH and SWR of 1:131 (varies from 128-133) SWR trace is at the top of the display. Smith chart is stable.
a 150 ohm, non inductive 5% leaded resistor gives 81.8 ohms 4.98 nH and a SWR of 1:1.64 Smit chart is stable.

For reference, I recheched the same resistor on my 8753 and it measured an SWR of 1:3.002.

Nothing else was done to the NanoVNA and I listed every action taken, This was done on my bench with an ESD pad. The RF environment is typical, with no high power transmitters operating nearby, as shown by a spectrum analyser with a whip antenna that I left running in peak hold mode spanning 1-3200MHz while these tests were in progress. No signal greater than -60dBm, except in the 2.4GHz and 900MHz bands, which had the usual Wifi, BLE and Celluar traffic.

Any ideas?

[radiolistener]

Can either of you confirm your NanoVNAs show near infinite SWR with an open?

yes, NanoVNA with proper calibration should show almost infinity VSWR for open input.

If your NanoVNA shows low VSWR with open input, this is an indication that your calibration is failed. It may be due to calibration mistake or malfunction of some component.

Check that calibration is ok before measurement.

You can test calibration it in the following way:
1) Connect OPEN and make sure NanoVNA shows infinite VSWR, and Smith Chart dot on the right side of circle
2) Connect SHORT and make sure NanoVNA shows inifinte VSWR, and Smith Chart dot on the left side of circle
3) Connect LOAD and make sure NanoVNA shows VSWR=1.00, and Smith Chart dot in the center of circle

also you can perform additional test:
4) Connect 150 Ohm load and make sure NanoVNA shows VSWR about 3.

If calibration test is failed, it means that calibration is wrong and you will get incorrect measurements.

[joeqsmith]

I think that last response pretty much says that your two year old Nano is now dead, unless I am missing some other clue.

Personally, I suggest you treat yourself to a new V2 or just buy another V1.     

If you want to know why it was damaged, hard to say.  I'm not there to watch over you.  If you are connecting it to an outside antenna and not grounding the system prior to attaching the unit, maybe ESD.   Maybe you don't always handle it with a ground strap.  Maybe you tried to measure an amplifier with it at one point.   I've seen SA's and VNAs that were a mess after being overloaded.

Still, no big deal to replace it. There are places now that stock them.   $50 is hardly worth the time.  Just replace it an move on.  Or, use the 8753C.   

[radiolistener]

There is possible bad contact on SMA connectors.

[joeqsmith]

I wouldn't expect it to provide repeatable results.   Sounds more like it's leaking.

[ogden]

Most vulnerable part of NanoVNA - mixers. You may consider following test: reset calibration of your Nano and look at S11 logmag (uncalibrated). Up-to 300MHz it shall indicate around -20dB .. -23dB [or better] for open port. If you see way lower value like -30dB or -40dB at let's say 150MHz - port1 (ch0) mixer is dead. Repair is comparably simple BTW.

[iMo]

..
4) Select CALIBRATE, then step through OPEN, SHORT, LOAD, ISOLN (2x 50ohm loads), THRU (using supplied 6" cable), then DONE. I used my R&S SMA lab standards and saved the calibration result in memory zero.
..

Is that ISOLN cal a correct way how to do it? I do it with ch0 open and ch1 loaded with stock 50ohm (I always use the H4's stock cables attached to the ch0 and ch1 terminals).
For the antenna swr measurements I do not do the ISOLN and THRU cal steps, btw.
Try to calibrate as radiolistener has written above..

[ogden]

During ISOLN calibration ch0 must not be reflecting any signal back into VNA, thus shall be loaded using 50Ohms. Ch1 must not act as an antenna, thus it is good idea to "plug" it using 2nd 50Ohm load or short.

[cdev]

SMA savers or whatever they are called might be a good investment if they dont impact the response. Keep your SMA connectors clean and never rotate them hard if there is resistance, unscrew it and redo it. I also clean mine with IPA.

But I agree with the others, the nanovna 2 is so cheap and much better, I would just buy one. Then handle the SMAs with care. I always leave the two fairly decent short cables it came with on, and keep ferrite split beads on those extension cables to decouple them from possible common mode currents.

[joeqsmith]

..
4) Select CALIBRATE, then step through OPEN, SHORT, LOAD, ISOLN (2x 50ohm loads), THRU (using supplied 6" cable), then DONE. I used my R&S SMA lab standards and saved the calibration result in memory zero.
..

Is that ISOLN cal a correct way how to do it? I do it with ch0 open and ch1 loaded with stock 50ohm (I always use the H4's stock cables attached to the ch0 and ch1 terminals).
For the antenna swr measurements I do not do the ISOLN and THRU cal steps, btw.
Try to calibrate as radiolistener has written above..

You can test calibration it in the following way:
1) Connect OPEN and make sure NanoVNA shows infinite VSWR, and Smith Chart dot on the right side of circle
2) Connect SHORT and make sure NanoVNA shows inifinte VSWR, and Smith Chart dot on the left side of circle
3) Connect LOAD and make sure NanoVNA shows VSWR=1.00, and Smith Chart dot in the center of circle

Using the same parts that you cal with to check the cal is not a real good idea.   For example, swap the open and short during the cal, then recheck them and the unit will show what appears to be the correct data.  You could swap a 200 ohm for the open and cal to that and it will show it as an open.   You could leverage this. 

also you can perform additional test:
4) Connect 150 Ohm load and make sure NanoVNA shows VSWR about 3.

I would run some sort of secondary check like this. 

****
...

5) After Pressing DONE, with nothing attached, the display shows the SWR trace above the first horizontal division and the value for SWR is 1:2..38. The impedance is shown at 118 ohms with 3.02nH. Smith chart is stable.
6) reinstalling a 50 ohm standard, gives 49.9 ohms and 15.2 nF and a SWR of 1:1.00 SWR trace is at the bottom of the display. Smith chart is stable.
7) reinstalling a short standard gives 944miili Ohms and 4.15 nH and SWR of 1:131 (varies from 128-133) SWR trace is at the top of the display. Smith chart is stable.
a 150 ohm, non inductive 5% leaded resistor gives 81.8 ohms 4.98 nH and a SWR of 1:1.64 Smit chart is stable.

5 is pretty odd though as I would expect the math to show the port as open, even if there were a problem.  If say the port had a 1K resistor on it normally, and you went through the cal.  Once you open the port, even though it has a 1K, it better read open.  That or there is a problem with the math. 

There was no mention of playing with the firmware prior to the problem and after owning it for two years and getting good results up until now is all really odd.   

Then again, I am clueless of what a stable Smith chart would be.   

[Bud]

Does this dinky gizmo ( i do not have one ) has settings to specify calibration standards coefficients or electrical delay? Maybe that numbers are incorrectly entered or corrupted which may lead to the problem OP is experiencing.

[joeqsmith]

Does this dinky gizmo ( i do not have one ) has settings to specify calibration standards coefficients or electrical delay? Maybe that numbers are incorrectly entered or corrupted which may lead to the problem OP is experiencing.

There was no way to change he delay or enter coefficients in the older FW.  Maybe this has changed.  It's been a while since I looked at new firmware. 

[radiolistener]

Is that ISOLN cal a correct way how to do it? I do it with ch0 open and ch1 loaded with stock 50ohm

Your way is acceptable. Technically both ports should be loaded, but the difference is not significant. So, you can leave CH0 open and load CH1 with stock 50 Ohm.