Sub $1000 VNA for antenna matching

[dermeister]

I am looking to buy a VNA for determining the filter network components for matching antennas to low-power embedded systems RF SoCs. The max frequency I work with is 2.4 GHz for BLE and 802.15.4. The VNA will only ever see bench use, so it does not need to be portable or have an integrated screen.

There are a few low cost (<$1000) VNAs I’ve been researching like NanoVNA, LibreVNA and LiteVNA, but I’m a beginner at using VNAs so I’m not sure what performance specifications are required for my application outside of the max frequency. Ideally I would like the antenna matching process to be as simple as possible, so if any VNAs automate it more than others, that would be a consideration I would take into account.

Can anyone recommend a VNA for what I’m trying to do?

[nctnico]

An alternative could be to buy a used 4G base station analyser that has VNA abilities but many of them only seem to support a Smith chart and no capacitance / inductance cursors. For your purpose and budget I would take a long hard look at LibreVNA. I have owned a NanoVNA but found it too cumbersome to use with the small screen.

[wasedadoc]

The original nanoVNA only goes to 1.5GHz. Other similar devices do go higher.  I have the "black and gold" V2 which goes to 3GHz (using harmonics).  That has a 3.2 inch screen.  Small screens on these device can be mitigated by linking to a PC.  However I've never used mine to do antenna matching so cannot say with certainty that one of the low cost VNAs will adequately meet your need.

[nctnico]

One of the problems with the NanoVNA is that there are so many versions out there and -AFAIK- different people working on firmware. And no overview on what is what.

[Mechatrommer]

For your purpose and budget I would take a long hard look at LibreVNA. I have owned a NanoVNA but found it too cumbersome to use with the small screen.

why dont you get larger screen 6GHz LiteVNA? its only $100+ do you really need full 2 port feature? LibreVNA has serious internal mismatch issue. i have both Nano and Lite.. Nano seems to become useless (and probably my Deepace KC901V too) once i have LiteVNA.

[joeqsmith]

I assume with the low budget it's a home project.  That and with you planning to build it, I assume you are avoiding having to have it certified.

A while back I attempted to characterize a purchased BLE radio's output with a VNA.  The manufacture's supplied models for the transmitter and antenna.   I was planning to remove the transceiver IC and probe the board directly to measure the return loss of the matching network and antenna.   The radio with antenna was about the size of a quarter.   That board had a mezzanine board mounted to it with the actual radio and antenna.   The antenna and matching network made up just the end of the board.  Sorting out how to interface to it and calibrate.... far exceeded my abilities.  The IC manufacture had a demo board with a test point located at the matching network.  Consider this where the transmit and receiver nodes come together, not at the IC.   The production radios I looked at had no test points.   The attached photo showing the matching network.   A small sewing needle placed at the top.  Any coax or connectors I have would be MUCH larger than that. 

I did get something together but doubt the data was meaningful.   The entire circuit was about the size of an SMA connector.   I thought I had saved it and was going to post a picture of my poor workmanship but may have been pitched.   The cost of probing this board may far exceed your budget.  I assume you are designing the boards and have considered how to probe them to make these measurements.   

To measure S11 return loss, I would expect 20dB would be good enough.  Maybe 30 for those with ADHD and hyperfocus. 

Consider posting some details on your radio and layout as it may help people guide you.   

I played around with some of the free demo simulators but everything I looked at was too limited to be very useful.  ADS is expensive but you may want to look at Pathwave:
https://www.keysight.com/us/en/products/software/pathwave-design-software/pathwave-rf-synthesis-software.html

Sound like a fun project.  I'm interested in following it if you are able to post about it's progress.

[joeqsmith]

Just for a reference, here is the smallest coax I have on hand.  The upper left corner of the photo is the edge of the coax.  This is roughly 2.4mm dia.   

I'm thinking something like a TRL standard on the board with some sort of probing station would be required.   

Maybe have a look here to get some ideas..
https://www.microwaves101.com/encyclopedias/rf-probing

[JohnG]

LibreVNA has serious internal mismatch issue.

Can you please elaborate on this? What specifically is the mismatch issue? I had not heard about this, and I am looking at purchasing a LibreVNA for my job.

Thanks,
John

[Mechatrommer]

LibreVNA has serious internal mismatch issue.

Can you please elaborate on this? What specifically is the mismatch issue? I had not heard about this, and I am looking at purchasing a LibreVNA for my job.

Thanks,
John

please follow this post... https://www.eevblog.com/forum/rf-microwave/clone-85033e-vna-cal-kit-measured-against-pico-cal-kit-using-librevna/msg4475422/#msg4475422 i have yet to get a good answer on why the ringy plot of LibreVNA output file. the last time we were discussing among VNA users that the ringy plot is due to internal mismatch (tracks are not 50 ohm impedance)... the way i see it, its internal problem and difficult to cancel from external tricks, unless the designer takes care of this internally or in FW. ymmv.

[joeqsmith]

LibreVNA has serious internal mismatch issue.

Can you please elaborate on this? What specifically is the mismatch issue? I had not heard about this, and I am looking at purchasing a LibreVNA for my job.

Thanks,
John

please follow this post... https://www.eevblog.com/forum/rf-microwave/clone-85033e-vna-cal-kit-measured-against-pico-cal-kit-using-librevna/msg4475422/#msg4475422 i have yet to get a good answer on why the ringy plot of LibreVNA output file. the last time we were discussing among VNA users that the ringy plot is due to internal mismatch (tracks are not 50 ohm impedance)... the way i see it, its internal problem and difficult to cancel from external tricks, unless the designer takes care of this internally or in FW. ymmv.

Agree, it does seem a bit odd.  I ran a similar test with the LiteVNA, using the ideal model and standards supplied with the V2Plus4 (sorted load from minicircuits) and home made cables.   First LiteVNA was calibrated then I then collected S11 with nothing attached to the cable.  I then attached my home made FR4 open standard.  Finally, a 200mm section of semi-ridged coax.   

Their post stated:

I have attached the raw measurement data of the clone 85033E kit

and I would assume that is what they meant.   May not be the VNA but something changing in their setup.  Bad expensive cables moved?    Someone who owns the LibreVNA would need to chime in. 

[joeqsmith]

Calibrating the LiteVNA with my home made FR4 standards (really bad) then measuring the open supplied with the V2Plus4 and with nothing attached to the cable. 

[ogden]

You may find this useful to measure small RF stuff:

https://www.pasternack.com/sma-female-unterminated-lead-sexless-pe-020sr-cable-assembly-pe3ca1107-3-p.aspx

When cut straight - it becomes sexless

Other option:
https://www.fairviewmicrowave.com/test-probe-pigtail-sma-female-straight-cut-lead-cable-.020-coax-fmca1187-3-p.aspx

[joeqsmith]

If I place a mismatch (20dB RL) inline with the cable and then calibrate the LiteVNA using my FR4 standards.  Then measure with nothing attached and the V2Plus4 open standard,  we get the following.    Because we are looking at systematic errors, the SOL will take care of much of the error. 

[samofab]

The OP question was basicly "what are the requirements for VNA used for tuning PCB antennas up to 2.4GHz".
Members, please correct me if anything below is incorrect or misguided.

The process for tuning the PCB antenna goes like this (with many adaptations possible):

• attach SMA connector before before the matching network. That can be done in a few ways: a.) if you planned for this you'll have a PCB connector availible, but this can be tricky becuase could be interrupting a transmission line (TL) b.) more likely you'll solder a micro-coax (rigid or semi-rigid pigtail) to the point where transciever would be or you'll cut the TL and attach the micro-coax to the first component in matching network or close by
• calibrate you VNA + cable
• with no matching network in place (or with short on first element) measure the electrical length of your micro-coax
• enter this number to VNA
• populate the matching network with reasonable initial components (e.g. from datasheet)
• measure S11 (in final mechanical environment) and display as return loss around your frequency of interest or wide enough to see a drop in response
• verify that your antenna is radiating (block it with hand to see a difference on RL)
• if the drop is lower than -10dB at your target frequency, (my) congratulations! Also, remember.. a lot of things closer than lambda to your antenna will move this a lot.
• more likely you'll need to modify the matching network. Display S11 as Smith chart at frequency of interest.
• iteratively modify components until the dot is in the center of the Smith chart. Absolutely keep lab notes on the the procedure; iterate and triple check everything. Don't move your cables. Wait for your VNA to heat up (=stabilize).

Based on this process, here are my thoughts on the VNA:

For measuring return loss you have two numbers that are important: frequency and drop in response (level / SWR). If your antenna is very mismatched, the drop in response can be far away, so it's useful to have VNA that covers larger frequency range than your interest (say x1.5 or x2). In my opinion: The accuracy of frequency will never be a problem. The accuracy of the level is also not very important unless you're trying to have an exact measurement how good your antenna is. In other words -10dB or -11dB will not make any difference.

For matching the antena (which you'll be doing on Smith chart) the most important thing is not absolute accuracy, but precision within one measurement/matching session. Also you can do matching on the very top of the frequency range of your VNA. If you fire up any Smith chart matching calculator https://www.will-kelsey.com/smith_chart/, you will see that errors accumulate very quickly. Even a 2% error in original impendance measurement followed by 2% in component tolerance (let's say for two steps) can move the impendance far far away from 50. Be especially careful when the measured antenna impendance is close to border between different matching strategies.

Once you get unexpected results you will doubt (in no particular order): your VNA, your cables, your calibration kit, your matching componentes, your PCB design, you pigtail, your knowledge and your freespace environment. If you pay more money for the things you can buy from the prevoius list, you'll reduce the amount of things that will be the source of your doubt. Not sure how all of the above helps in purchasing decision :-)

[joeqsmith]

I've never looked into it but it does seem like the VNA/software could calculate the values based on what model you select for the network.   Maybe there is standalone software that does this today with Touchstone files. 

with no matching network in place (or with short on first element) measure the electrical length of your micro-coax

Or just leave the cable open, you just need a high reflect.   You could just build it up and use gating to remove it?   

Good point about the micro-coax.  UT047 is about 1.2mm OD.   You could get something fairly cheap from Digikey assembled:
https://www.digikey.com/en/products/detail/microchip-technology/RN-UFL-SMA6/2094296
Or
https://www.digikey.com/en/products/detail/taoglas-limited/CAB-058/5287277
https://www.digikey.com/en/products/detail/amphenol-rf/095-902-462-003/7386033

Paper describing some of the basics.   
https://usb-vna.com/2020/08/26/wi-fi-bluetooth-pcb-tuning-and-antenna-testing/

Software
https://coppermountaintech.com/video-antune-antenna-impedance-matching-and-antenna-measurement-demonstration/
Or maybe look through some of these:
https://www.rfpage.com/free-tools-to-analyze-and-plot-s-parameters-file/

[dermeister]

Thanks a lot for all of the replies! I am tending towards getting a LiteVNA64 now. It seems like it will be sufficient for my current needs and also good tool to learn VNA basics.

[joeqsmith]

Just for another reference,  here is some magnet wire with an OD of 1.12mm.  This is slightly smaller than UT047 at 1.19mm (OD of outer conductor). 

[Bud]

Get a VNA that has Port Extensions feature. Not just Electrical Delay correction. Some VNAs call the simple electrical delay adjustment "Port Extension" which is confusing. Those are different features. Electrical delay is a correcting value which is applied to the measured phase. It does not correct for the connection loss and  the calibration plane remains where it was. Port Extension shifts the calibration plane to the DUT terminals by taking in account the delay and loss, recalculating the error terms for the test port.

[joeqsmith]

Thanks a lot for all of the replies! I am tending towards getting a LiteVNA64 now. It seems like it will be sufficient for my current needs and also good tool to learn VNA basics.

IMO all of these low cost VNAs are great tool for learning and you can hardly go wrong with any of them.   Between the LiteVNA and the original NanoVNA, you pretty much have the basics covered.  Below 300MHz, the original NanoVNA offers some advantages I have yet to see matched.  The Lite on the other hand is fast, higher frequency range, somewhat adjustable output level.   Dislord has done a very good job with the Lite's firmware.  While I have no use for 90% of the features, it's proven to be very stable.   

Downside with all of these are the output is not level and it's a squarewave.   You don't get a lot of the features you would with a higher end system (built-in bias T's, input and output attenuators, more complex calibration schemes, two port one path...).  You also are not going to get the dynamic range, they have higher noise, poor connectors....   For passive circuits, much of it won't matter.  Active is another story.  I made a video showing an simple circuit to demonstrate the possible problems with the squarewave drive (attached if interested).    Obviously, if you are trying to measure an amplifier and the VNA's output is not known and level, you can run into problems quick.   

FYI on the Lite.
The first one I bought worked well and met the suppliers specs.  A friends had very high noise.  They bought a second and shipped me theirs.  Their second one also had high noise.  After some testing, I found the Lite was very sensitive to the input voltage level from the USB port.  Basically, to get around it, I made a USB cable with a diode in series with the power to lower it.  This solved the problem.   Thanks to Dislord, you can use it to measure narrow band devices (crystal filters) but it is dog slow.   It's low frequency performance is poor (much higher noise) compared with the original NanoVNA.   One of the Lite's my friend purchased came with a bad USB cable.   It supports a CF card.  I don't have a use for it but did try it out.  First thing that happened, the card fell inside the case and I had to take it apart to fetch it.  $120....   You may find several posts on the standards supplied with them.  I've never looked at what was supplied with the Lite.   Any test data I have shown with these low cost VNAs has used the ideal model (standards are considered perfect), and using the short and open supplied with the V2Plus4.  For the load, I bought a lot of 8 mini-circuit terminators that I characterized using a set of Agilent standards and my Agilent PNA.  One had really good RL which I set aside.  The second best is what I use. 
 

[points2]

... Below 300MHz, the original NanoVNA offers some advantages I have yet to see matched.

hello joeqsmith,
you mention that over low frequencies, the nanoVNA is OK.
I have a nanoVNA-H4, I confirm that point, when I just want to see the self resonant freq of a MLCC X7R (sorry for this basic example :-) )

Recently I get bored to wait for the nanoVNA to sweeeeeeeep (too long), so I bought a LibreVNA => much faster ! Cool ! The LibreVNA freeware : top cool too (requires some time to understand the features but... top ! )
Problem : when I tested a basic X7R, having a SRF @ 300kHz => the LibreVNA was blind vs its SRF due a too high noise (wavvy curve around magnitude=-40dB) over the weird frequency range : mini 100kHz to 1.29Mhz, then, the measurement was in line with the product characteristics (strange).

Is it a common drawback of LibreVNA & others (except the nano) to struggle to do clean measurements over the low-end frequencies ? or if my vendor replaces my LibreVNA by a new one, I'll have no issue ? (a bad item, it can happen) ; I'm waiting for feedback from him, he's waiting for feedback from the manufacturer...)

After reading this thread & the other in link above, I'm a bit confused to discover that the ability of VNAs to do clean measurements over low frequencies is "tricky".
Fortunatly, my nanoVNA does the job. But it is so slow...

When I look at the other VNAs on the market, I end up to the conclusion that, regarding low freq measurements, we have (vs price) : the nanoVNA (100eur) and then... the cheapest seems to be the Siglent SNA5002A (8000eur). What a gap ! :-)

[joeqsmith]

The LiteVNA is so much slower than my original NanoVNA.  Of course, with the Lite you can sweep very fast if you don't care about the data.  So when I suggest slower, I'm setting the IF to something reasonable to get comparable noise with the NanoVNA.  I have tried a few experiments to improve the Lite's noise but so far, my original Nano is lower. 

I don't own a LibreVNA and don't know too much about it.  I suggest joining their group and asking the designers any questions you have. 

Is it a common drawback of LibreVNA & others (except the nano) to struggle to do clean measurements over the low-end frequencies ?

CMT offers a VNA that would fit my use very well but was more than I wanted to spend for my hobby.   When I was looking at replacing my 70's VNA with something newer,  I had asked in the HP group about lower frequency use (see below) and suspect it's not a common use case.

Does one really need such sophisticated instrumentation to work
below 1MHz?  50 ohm transmission lines scarcely even exist at such low frequencies.

The VNA was built to make it possible to characterize components so that you could more easily build circuitry by stringing together such characterized components, lumped along transmission lines.

They are necessary in the microwave region, where there is really nothing that even resembles a pure resistance, pure capacitance, or pure inductance.

At low frequencies, you can very much rely on your parts being what you think they are.  Capacitors are capacitors, and not a mixture of capacitance, inductance, and resistance.

Unless physics has changed, I don't see a VNA being necessary, or even useful at such low frequencies.

-Chuck Harris


Joe Smith via Groups.Io wrote:

     Does one really need such sophisticated instrumentation to work below 1MHz?

I appreciate your concern.  I would say in my particular case, placing the question mark after "instrumentation" is appropriate and the answer to which is no.   I certainly have no need for any of my equipment.  It's a hobby and I enjoy it.


Chuck Harris
1/16/20   #101895  

True, but wanting your Microwave VNA to go down to DC is like wanting your Ferrari to have tank tracks.

It just isn't right.  Better would be to have a Ferrari and a tank.  They serve different functions.

-Chuck Harris

[joeqsmith]

Don't laugh too hard.  Attached photo showing my attempt to tap into the matching network on the BLE board.  This is a semi rigid silver plated copper coax with an OD of 2.2mm (0.086).   Much too fat for this application.   

Link to antenna data sheet.
https://www.yageo.com/upload/media/product/productsearch/datasheet/wireless/An_Ceramiz_2450_3216_L00_0.pdf

[mawyatt]

Looks like you did an OK job, what's readily available with OD less than 2.2mm?

Best,

[Mechatrommer]

Looks like you did an OK job, what's readily available with OD less than 2.2mm?

coax 1.13mm.. china come.. the problem is... did the coax deembedded? or proper way/calibration made right on the coax's tip? and then there's parasitics from input tracks that might differ from actual operation (when input driver/ic populated)

[joeqsmith]

Looks like you did an OK job, what's readily available with OD less than 2.2mm?

coax 1.13mm.. china come.. the problem is... did the coax deembedded? or proper way/calibration made right on the coax's tip? and then there's parasitics from input tracks that might differ from actual operation (when input driver/ic populated)

The UT-047 MICRO-COAX I had linked above is 0.047".   For calibration, I had made two probes which I had phase matched.  The one was shown in the previous picture.  The other has a sewing needled attached to the shield.  This probe is used to cal the system attaching it to a section of copper PCB and in my case, just leaving it open.  Again, the ideal model was used.  A better solution would be to toss some major cash into it but it was just for fun.  The radio must be placed onto the circuit board.  I had just hand cut a ground plane and soldered to that.    One other problem was because the coax is so thick I was not able to install the shield.  All these details add up.

It's one thing to have datasheets for a single component.  In this case, I have no way to verify my results. 

I set it all up with the LiteVNA. 

[joeqsmith]

While looking for software that would make it easier for OP to match their antenna, I came across this VNA/software.   I didn't check but the description talks about being able to download the software and run it without the VNA using their test data.  Maybe it supports Touchstone.   

https://www.megiq.com/blog/item/47-vna-match-function-has-adjustable-components

[joeqsmith]

Attached showing the calibration probe.  Keeping in mind that the OD of the coax shield is 0.086".   A small sewing needle was placed along the coax and then wrapped with 30AWG wirewrap wire.   Then the whole assembly was soldered.   These two probes have a 90 degree bend in them, which is why the shrink tube is slanted.   

I may chop up a PCB to get the correct ground plane and then cut the shield so it can be installed and just see what difference we get. 

When looking at the radiation pattern, you can imagine this big chunk of coax hanging out the side and large ground plane behind the board along with the missing shield is going to have an effect.    A better measurement would be to look at the end product running and mounted to a turntable.  Then measure the pattern with a receiver.   With the OP limited to a $1000 budget, that's about an hour of lab time.  Guessing that's not part of their plan.   

[joeqsmith]

I've mounted the radio to a sectioned board with the proper layout.  Attached photo shows my hack job on the shield.  I found a bit of Arlon (not sure the part number) and epoxied the sectioned board with radio to it.  This will allow easy mounting to the rotor's mast.   Once this dries, I will repeat the calibration and measurements with the LiteVNA.   

[joeqsmith]

I've attached a photo showing the setup.  Note that the shield was soldered to the coax.   

Looking at the return loss, the Red trace was just after calibrating the VNA.  The Yellow trace is using yesterdays calibration.  There's a bit of difference which I'm not too surprised.  This time I was using some copper foil tape rather than the PCB.    The bigger change was due to all the other changes in the setup. 

The 2D plot would be how I would normally use the radio.   

[joeqsmith]

please follow this post... https://www.eevblog.com/forum/rf-microwave/clone-85033e-vna-cal-kit-measured-against-pico-cal-kit-using-librevna/msg4475422/#msg4475422 i have yet to get a good answer on why the ringy plot of LibreVNA output file. the last time we were discussing among VNA users that the ringy plot is due to internal mismatch (tracks are not 50 ohm impedance)... the way i see it, its internal problem and difficult to cancel from external tricks, unless the designer takes care of this internally or in FW. ymmv.

I was reading the LibreVNA group yesterday and it seems they have new hardware.  Someone posted an AB compare that may be of interest.  

https://groups.io/g/LibreVNA-support/topic/what_is_the_spec_defrance/95474170?p=,,,20,0,0,0::recentpostdate/sticky,,,20,2,0,95474170,previd%3D9223372036854775807,nextid%3D1666293968019920350&previd=9223372036854775807&nextid=1666293968019920350

[Mechatrommer]

please follow this post... https://www.eevblog.com/forum/rf-microwave/clone-85033e-vna-cal-kit-measured-against-pico-cal-kit-using-librevna/msg4475422/#msg4475422 i have yet to get a good answer on why the ringy plot of LibreVNA output file. the last time we were discussing among VNA users that the ringy plot is due to internal mismatch (tracks are not 50 ohm impedance)... the way i see it, its internal problem and difficult to cancel from external tricks, unless the designer takes care of this internally or in FW. ymmv.

I was reading the LibreVNA group yesterday and it seems they have new hardware.  Someone posted an AB compare that may be of interest. 

https://groups.io/g/LibreVNA-support/topic/what_is_the_spec_defrance/95474170?p=,,,20,0,0,0::recentpostdate/sticky,,,20,2,0,95474170,previd%3D9223372036854775807,nextid%3D1666293968019920350&previd=9223372036854775807&nextid=1666293968019920350

still look ringing to me on both plot, is that a Through test? not really clear..

[joeqsmith]

Agree, not at all what I would expect.  With the other posts showing difference in cal standards causing it, compared with what I show using various loads, I wonder if it's a user problem.  Maybe a problem with their math.  

They really should have sent one to someone that knew what a VNA was.

[Smokey]

...
FYI on the Lite.
The first one I bought worked well and met the suppliers specs.  A friends had very high noise.  They bought a second and shipped me theirs.  Their second one also had high noise.  After some testing, I found the Lite was very sensitive to the input voltage level from the USB port.  Basically, to get around it, I made a USB cable with a diode in series with the power to lower it.  This solved the problem.   
...

I'm looking at getting one of these LiteVNAs.  What should I be looking at, performance wise, to determine if a unit has an issue with the USB voltage? 

[joeqsmith]

...
FYI on the Lite.
The first one I bought worked well and met the suppliers specs.  A friends had very high noise.  They bought a second and shipped me theirs.  Their second one also had high noise.  After some testing, I found the Lite was very sensitive to the input voltage level from the USB port.  Basically, to get around it, I made a USB cable with a diode in series with the power to lower it.  This solved the problem.   
...

I'm looking at getting one of these LiteVNAs.  What should I be looking at, performance wise, to determine if a unit has an issue with the USB voltage?

I've posted data showing the problem, how I collected it and also how to mitigate it.  Did you have a specific question about that?  I understand they released an update to the hardware to improve its performance.  Someone had asked how it compared with what I had shown.  There was no response from the designers and all went silent.  You can find that on the LiteVNA groups.io.   

[Smokey]

Cool.  Thanks for the info.  I think this is the forum post regarding the power supply update:
https://groups.io/g/liteVNA/topic/the_new_litevna64_which/93138117

The last messages are from 11/11/22 and claim the distributors like randl.com have the 64.0.3.1 version.

[joeqsmith]

Yes, that is the thread.   Looks like it grew a bit. 

If you buy one, I would be interested in having you run a simple test just to get some idea how much they improved it compared with my stock one. 

[Smokey]

Yes, that is the thread.   Looks like it grew a bit. 

If you buy one, I would be interested in having you run a simple test just to get some idea how much they improved it compared with my stock one.

It's on it's way.  I'll update when it gets here.

[joeqsmith]

Sounds good.

Did you order the large display or small?   I'm not sure if there would be a difference in performance between the two.   Also the firmware is not compatible between them. 

You wouldn't think that would be a big deal but I found with the P2Plus/4, I had no problems with the firmware for the large display version.  The small display version took a lot of trial and error to find something stable.   Based on what I have seen from Dislord, I doubt we would see the same level of problems but I have no experience with that smaller Lite.

I do not know how any of the open software handles the calibration.   If you want to start with what ever software you like, we can start there.  If there is a major improvement, I may  install the same software you are using just to remove that variable.

[Smokey]

LiteVNA64 from here:
https://store2.rlham.com/shop/catalog/product_info.php?products_id=75669

[joeqsmith]

Perfect.

[Smokey]

Arrived.  Got the new "64-0.3.1" label style on the case (image from store site):



Firmware version it shipped with = v1.3.07
Build time = Aug 30 2022 - 10:14:02

[joeqsmith]

Good deal.

Once you have what ever software you plan to use installed and running, set the Lite to full sweep and calibrate it.  Terminate the two ports and measure S21.  Then upload the touchstone file for it.   Basically, I would like the Touchstone file for what is shown in that photo you attached except you may as well run it down to 50kHz.  1001 points is fine. 

I had tried this with my friends Lite using the SD card and running it stand alone before I sorted out the problems with the USB voltage.   You could try running yours standalone as well.  This may provide us with even lower noise.   

[Smokey]

I feel like I'm missing something about firmware versions.

This looks like DiSlords's github, but there is no liteVNA repo: https://github.com/DiSlord

The liteVNA groups.io doesn't appear to have the wiki setup and I'm not seeing a firmware section: https://groups.io/g/liteVNA

There was a link to a google drive in one of the posts, but it's not clear what versions are on it, and all the files appear quite old: https://drive.google.com/drive/folders/10lU5xnwmHDnPX98z21ky09N8nb_cEOk4

There is a nanoVNA beta groups.io page, but it's invite only: https://groups.io/g/nanovna-beta-test

What am I missing?

[joeqsmith]

ZeenKo links the GoogleDrive you provided, which should have the latest release from the manufacture.

https://www.zeenko.tech/litevna?spm=a2g0o.detail.1000023.1.b5a97ff8tHYEOr

Dislord is a member here and has uploaded test versions for us from time to time.  This was their last upload:

https://www.eevblog.com/forum/rf-microwave/nanovna-custom-software/msg4579603/#msg4579603

If you like testing code, it may be worth your time to look into the beta group you mention.