I need a low cost VNA

[xzswq21]

Hello
I want to buy a low cost VNA bcoz I'm working with filters. I have found some VNAs such as VNWA, TinyVNA,...
which one is more reliable and better? I have to publish the results in IEEE.
(my budget is $500~$600)

Thanks

[tautech]

Hello
I want to buy a low cost VNA bcoz I'm working with filters. I have found some VNAs such as VNWA, TinyVNA,...
which one is more reliable and better? I have to publish the results in IEEE.
(my budget is $500~$600)

Thanks

Frequency requirement ?

[bob91343]

The great bargain is the nanoVNA.  It is very low in price and does an amazing job.  It has its limitations of course but for what they cost it's unbeatable.

For operation at low frequencies there is a converter from China that shifts the 0 - 2 MHz input signal up to 10 - 12 MHz.  I am tempted to buy one.  The nanoVNA goes down to 50 kHz, low enough for many applications.

[hendorog]

Some nanoVNA's are very poor at low frequencies. I have done a test here between the two different nanovna's I have:

The second to bottom chart shows the differences at low freq.
https://github.com/hendorog/nanovna_test/blob/master/NanoVNA%20comparison.ipynb

The white gecko branded one is much worse that the hugen supplied device. Hugen is this guy:
https://github.com/hugen79/NanoVNA-H

In general the hugen nanovna was good under 300MHz which is in the fundamental range of the PLL. Above that they both get progressively worse, and whether that matters depends on your use case.
Getting a good one isn't that easy. There are details about where to get them in hugen's github repo.

Also, the cal kits in these nanovnas do not come with characterisation data. You will need that or else your error correction is just in the 'near enough' bracket.
The bottom chart on this page shows this error when measuring a reference attenuator - the green trace is correct, and as measured by my cal kit supplier. The blue trace shows what a nanovna will display. The red trace is what I got from my 8753 when I removed the characterisation data from it - just as a double check.

https://github.com/hendorog/nanovna_test/blob/master/NanoVNA%20test.ipynb

[xzswq21]

I have to publish the results in IEEE.

Irrelevant. If you publish a journal, maybe they need repeatability, and maybe traced calibration. For a conference proceeding, no one cares.

Actually working with filters is a part of my project. for example I make 50MHz filter so I want to know the response (s11 and s21) and print it. then I use the filters in my real signal generator and sampling device. I have heard VNWA 3 is very accurate but I'm not sure.

[TheSteve]

How much dynamic range do you require?

[TheUnnamedNewbie]

I have to publish the results in IEEE.

Irrelevant. If you publish a journal, maybe they need repeatability, and maybe traced calibration. For a conference proceeding, no one cares.

I doubt any of the more high-ranking conferences will let you in if you show a measurement setup using such a VNA. (I don't want to get into the debate on if that is acceptable or not - just pointing out my experiences with reviewing conference papers myself).


To the OP: If you want to publish with the IEEE, what is your background/field/reasoning? Do you work at a research group that is not well funded and thus doesn't have access to a 'real' (sorry for the term) VNA?

[radiolistener]

Some nanoVNA's are very poor at low frequencies. I have done a test here between the two different nanovna's I have:

you're needs to update with more fresh firmware. Old version have instability issues. My NanoVNA works great at low frequency. The best dynamic range from 10 kHz to 300 MHz, about 70 dB for S11 and about 80-90 dB for S21.

[xzswq21]

I have to publish the results in IEEE.

Irrelevant. If you publish a journal, maybe they need repeatability, and maybe traced calibration. For a conference proceeding, no one cares.

To the OP: If you want to publish with the IEEE, what is your background/field/reasoning? Do you work at a research group that is not well funded and thus doesn't have access to a 'real' (sorry for the term) VNA?

I want to measure the quality of my current source such as output impedance, SNR, passband and etc.
I've seen some researchers built signal generator using DDS and FPGA (even with low cost DDS ICs and Rigol Oscop!) and used them in their research with the IEEE or IOP Science (master or PHD thesis). so I built a Signal generator and Sampling device with Avent's board and Analog devices components. I can show the results for them (HD2 and HD3 and SNR and etc)
now I want to test the filters. when I build or montage a filter I should test it many times so I need a low cost VNA. the filters are used as Anti-aliasing filter. cut of frequency is 10MHz or 40MHz.

[radiolistener]

now I want to test the filters. when I build or montage a filter I should test it many times so I need a low cost VNA.

If you're don't need more than 1 GHz, NanoVNA works good for such purposes. It cost 30-45 usd on aliexpress.

Here is NanoVNA low frequency measurement for 24 kHz ultrasonic transducer:

[tautech]

now I want to test the filters. when I build or montage a filter I should test it many times so I need a low cost VNA. the filters are used as Anti-aliasing filter. cut of frequency is 10MHz or 40MHz.

Do you have a DSO and a sig gen that can sweep ?
Learn how to do a Bode plot which can give you basic cutoff info.

https://youtu.be/uMH2hGvqhlE

[radiolistener]

Here is my home-made BPF, tuned with NanoVNA for 145.3 MHz 

[radiolistener]

It looks that my BPF works a little bad due to randomly untuned coil. A little tuning and now it's better



Amazingly, the price for device that allows to do that, is just 30-40 USD 

[MarkF]

Beware.
You may want to see this about the nanoVNA (starting @ 2:05):

And discussion:  https://www.eevblog.com/forum/chat/nanovna-shielded-ebay-sellers/

   

Better choice (more expensive though):

[radiolistener]

Beware.
You may want to see this about the nanoVNA (starting @ 2:05):

I already seen that. I think this is pretty normal behavior. In my opinion the root of cause for that is the common mode current on the outer braid of coax cable, because he uses unbalanced antenna with no RF choke.

Different VNA have different ground, for example one has bigger metal case, another has smaller metal case. When you connect antenna to VNA with no RF choke, your VNA metal case will be a part of antenna. So, the antenna parameters will depends on metal case size and geometry.

In order to avoid that, he needs to use RF choke. I think with RF choke on coax cable he will get the same results on both VNA

[xrunner]

you're needs to update with more fresh firmware. Old version have instability issues. My NanoVNA works great at low frequency. The best dynamic range from 10 kHz to 300 MHz, about 70 dB for S11 and about 80-90 dB for S21.

(Attachment Link)

If you want that non-working red battery indicator to function you can install the missing diode on the board of the NanoVNA (D2 - SD103AWS-7 SCHOTTKY DIODE ).

[radiolistener]

If you want that non-working red battery indicator to function you can install the missing diode on the board

Thanks, I know that. I just don't have proper size SMD diode, needs to buy...

Did you tested if this diode affects power consumption in power off mode?
Currently my nanovna eat 50 uA in power off mode.

[xrunner]

Thanks, I know that. I just don't have proper size SMD diode, needs to buy...

Did you tested if this diode affects power consumption in power off mode?
Currently my nanovna eat 50 uA in power off mode.

I will test it and report back here ... 

[hendorog]

Some nanoVNA's are very poor at low frequencies. I have done a test here between the two different nanovna's I have:

you're needs to update with more fresh firmware. Old version have instability issues. My NanoVNA works great at low frequency. The best dynamic range from 10 kHz to 300 MHz, about 70 dB for S11 and about 80-90 dB for S21.

(Attachment Link)

Will do, that test was on the shipped firmware IIRC. But you should state which nanovna you have, as there are hardware differences too.

[hendorog]

Some nanoVNA's are very poor at low frequencies. I have done a test here between the two different nanovna's I have:

you're needs to update with more fresh firmware. Old version have instability issues. My NanoVNA works great at low frequency. The best dynamic range from 10 kHz to 300 MHz, about 70 dB for S11 and about 80-90 dB for S21.

(Attachment Link)

I updated the firmware to the latest edy555 on the google drive. It is still poor compared to my other nano below a few MHz (+/- 0.5 dB immediately after calibrating with port open) and is unusable below about 1MHz. Effective directivity is only about 25dB at 1 MHz - tested by leaving the load on after calibrating.

I also tested with the latest hugen AA firmware, and it is the same.

Possibly the decoupling caps are too small on this white gecko version I have. I got this very early, before many people were aware of them.
There is a clear roll off at low frequencies.

Which version do you have?

[radiolistener]

I updated the firmware to the latest edy555 on the google drive. It is still poor compared to my other nano below a few MHz (+/- 0.5 dB immediately after calibrating with port open) and is unusable below about 1MHz. Effective directivity is only about 25dB at 1 MHz - tested by leaving the load on after calibrating.

I also tested with the latest hugen AA firmware, and it is the same.

that's strange, I have about -70 dB for S11 from 10 kHz.
Below 10 kHz dynamic range drops down. At 2 kHz it's about 30-40 dB. But above 10 kHz it's ok.

Which version do you have?

I'm using hugen79 hardware with custom firmware based on edy555 code (some fixes to make measurements more stable and get sweep with unlimited points from PC)

[xrunner]

Did you tested if this diode affects power consumption in power off mode?
Currently my nanovna eat 50 uA in power off mode.

It's about the same - it fluctuates from 47 to 60 uA off.

On, it runs at around 150 mA.

[hendorog]

I updated the firmware to the latest edy555 on the google drive. It is still poor compared to my other nano below a few MHz (+/- 0.5 dB immediately after calibrating with port open) and is unusable below about 1MHz. Effective directivity is only about 25dB at 1 MHz - tested by leaving the load on after calibrating.

I also tested with the latest hugen AA firmware, and it is the same.

that's strange, I have about -70 dB for S11 from 10 kHz.
Below 10 kHz dynamic range drops down. At 2 kHz it's about 30-40 dB. But above 10 kHz it's ok.

(Attachment Link)

Which version do you have?

I'm using hugen79 hardware with custom firmware based on edy555 code (some fixes to make measurements more stable and get sweep with unlimited points from PC)

Aha. Then it is not strange at all. The point of my tests was to compare my hugen hardware with my white gecko hardware.

If you revisit the link in my original post (added again below) you will see a flat line in the middle of the chart - that is the hugen hardware. The wiggly blue line is the white gecko hardware. This test is with the port open and is from 10kHz to 6MHz.

Second last chart on this link
https://github.com/hendorog/nanovna_test/blob/master/NanoVNA%20comparison.ipynb

Not all nanovna's are the same. On groups.io there are some people who are affected by confirmation bias, and insist that their white geckos perform the same as the hugen unit. Not true.

Is your stability fix is to improve the issue with wide spans and large step sizes that I identified in the joeqsmith thread?

[xzswq21]

can I buy a second hand VNA on eBay?
wish I had a setup that could generate a signal and  capture the response!
https://www.ebay.com/sch/i.html?_from=R40&_nkw=network%20analyser&_sacat=0&_udhi=700&_sop=16&rt=nc

Why NASA used a VNWA in their experiment?! watch @26m50s
 

 

[virtualparticles]

The ultra low-cost, "cheap" VNAs tend to be very drifty and of course there is no metrology to back up any claims of accuracy. Without accuracy it is difficult to make any claims at all about measurements that are made. I believe the best value on the market is the TR1300 from Copper Mountain Technologies. These are backed by an ISO17025 certified lab with traceability to NIST.

https://coppermountaintech.com/vna/tr1300-1-2-port-1-3-ghz-analyzer/

I don't usually do blatant marketing here but I believe I am answering the question.

Thanks!

Brian