NanoVNA V2 aka S-A-A-2

[joeqsmith]

These were OMRON G6KU-2F-RF-T.  Looks like $22/ea. 

Datasheet:
https://omronfs.omron.com/en_US/ecb/products/pdf/en-g6k_2f_rf.pdf

The first attempt used a few M/A-COM GaAs MMICs.   Both designs were not worth the effort.   Isolation was really poor.   Attached is my old Transco 28V relay, powered and controlled from a FTDI USB  TTL adaptor.    Isolation on this part is something around 60dB to 18GHz, over 80 below a GHz.  Something like that.  700C if you wanted to looked it up.     

[cdev]

The author of the article uses a shielded box partition with a wall between them for isolation, so it could easily double the size of the nanoVNA, so the idea of an external add on for bench use makes more sense I think. See the second graphic below.

This would be a useful add on for the many of us with these small VNAs.

It could also be incorporated into a nanoVNA case. If it was permanently there, it would make sense to calibrate the vna for its presence permanently and add its presence to the FW. People could plug the calibration into the routines. Not very difficult. That would eliminate another source of error.

...

I think its possible to use relays to extend the functionality of a two port VNA to make an effective four port VNA too. (Or so I read in QEX a while ago)

Joe, the same relays are used as the tiny low voltage Omron ones you use in your switch box.

I think those were $10-$20 each.  They work alright but the isolation was poor.   I have an old Transco relay that I made up a really crappy DC-DC to drive it from the USB port.  My software supports it.  I plan to mount that along with a couple of bias-t's and maybe a decent step attenuator.   But then I think, I should really add some mixers and some better couplers and maybe a wide band amp......       Reality always steps in. 

The idea is that by reversing the two ports available with the very minimal differences possible, all four sets of S-parameters of a DUT become measurable.

(Thats what I meant by "four port" - I meant that the more VNA functionality is exploitable).

"While beta-testing his VNWA2.1 vector
network analyzer, Andreas Zimmermann,
DH7AZ, came up with a brilliant idea: He
took two RF relays and built a simple switch,
which could commutate the VNWA TX and
RX ports. (see refs below) Thus, a two port device wouldn't
have to be disconnected and turned manu-
ally any longer in order to measure its two
port S-parameters".

"Surprising to me, his first veroboard prototype already showed
good isolation values. Even more surprising
to me, the relay data sheet indicated that at
1.5 GHz, isolations of 60 dB could still be
expected. 2 That's when I decided to build
such a switching unit myself, and modify
my VNWA software so it would auto-
matically control the relays. This way, the
VNWA2 is turned into a full-featured two
port network analyzer. The most charming
aspect is the simplicity of the circuit, which
only requires two relays."

This is from the article "A Simple S-Parameter Test
Set for the VNWA2 Vector Network Analyzer" in
QEX May 2009 by Dr. Thomas C. Baier, DG8SAQ

References he cites:

1 Professor Dr. Thomas C. Baier, DG8SAQ,
"A Small, Simple, USB-Powered Vector
Network Analyzer Covering 1 kHz to 1.3
GHz," QEX, Jan/Feb 2009, ARRL, pp 32 -
36. See also www.arrl.org/qex/2009/01/
Baier.pdf
2 Omron G6Y Relay, see w w w . o m r o n . c o m /
ecb/products/pdf/en-g6y.pdf
3 See, for example, Agilent AN 1287-3,
"Applying Error Correction to Network
Analyzer Measurements," Application Note,
http://cp.literature.agilent.com/litweb/
pdf/5965-7709E.pdf
4 See www.mydarc.de/DG8SAQ/NWA.html
5 The VNWA software current as of the pub-
lication date of this article is available for
download from the ARRL QEX Web site. Go
to www.arrl.org/qexfiles and look for the
file 5x09_Baier.zip

----

There is a 3.3v version of the Omron relays.

[switchabl]

I've been meaning to try something like this, but haven't gotten around to it yet. I picked up an old HP 8711 K02 switching test set for that purpose, which is a essentially just a box with 4 N-connectors on the front and a couple of high-isolation RF relays inside. It's a big 19'' unit, so maybe not the most practical NanoVNA accessory, but it was cheaper than buying the relays separately (and as a bonus has two of those unobtainium HP probe power connectors as well).

The normal NanoVNA software doesn't support this measurement mode of course, but scikit-rf will do the 2-port corrections just fine.

The accuracy may be limited (apart from the switch repetability) by the unusal NanoVNA v2 architecture. The usual error model assumes that the load match on port 2 is the same for reflection and transmission measurements. But because there is only one receiver that is switched in for R and T measurements, this will not be the case, so the error correction will not be complete. It remains to be seen how bad this is in practice.

[cdev]

Your test set may be better than you think for this. But maybe not. I guess we'll see.

If I read you right, the fact that the software may already support something like it or be close perhaps very close to it, may not apply, because of architectural differences?

Caveat!
(You can see right here how I'm speculating way above my knowledge level here. Unfortunately, not rich, but having a lot of fun and learning new stuff every day. But making a huge mess, which I need to clean up today.).

Here is what the article says about error correction.

The guy who had this idea originally was Andreas Zimmermann DH7AZ, BTW.

Advantages of Using a Test Set,
Error Correction
The most obvious advantage of the test
set is the fact that two port devices can be
fully characterized without manually turning
them during the measurement. Thus, all four
S-parameters (S 11 , S 21 , S 12 , S 22 ) can
continuously be measured. This fact yields another,
less obvious, advantage: As discussed earlier
in this article, S 11 and S 21 can be deduced
from a forward direction measurement only
if the TX port impedance, Z S o u r c e , and the
RX port impedance Z L o a d , are both identical
to the reference impedance (usually 50 Q).

If this is not the case, some of the transmit-
ted signal, b 2 , is reflected at the RX port, is
retransmitted through the DUT and adds to
the reflected input signal, b 1 , which in turn
gets reflected at the TX port again. Thus,
in this case the results of the forward mea-
surement are influenced by all four DUT
S-parameters as signals travel through the
DUT in both directions. The same holds true
for the measurement in the reverse direction.
If both directions are continuously measured,
then the effects of non ideal source and load
impedances can be mathematically corrected
in an exact manner. This technique is known
as 12-term error correction. 3
Figure 7 shows the measured S 11 and
S22 of the through calibration standard. By
definition, both reflection coefficients of the
through standard are zero. Traces S 11 and
S22 show simply corrected results, which are
identical with the detector reflection coef-
ficient caused by Z Load ^ 50 Q. Note that the
detector match is better than 25 dB over the
displayed span, which is quite good. Mem1
and Mem2 show the same data 12-term cor-
rected. The effect of imperfect load match
has completely vanished and the measured
data is dominated by noise. The rising noise
floor to the high frequency side reflects the
decreasing available signal strength of the
instrument.
One should think that the effects of
12-term correction are generally small,
but they can become large if the measured
S-parameters are used in a system simula-
tion. Figure 8 shows transmission data of a
monolithic crystal filter measured with the
VNWA at a reference impedance of 50 Q
with various error correction schemes and
then recalculated to source and load imped-
ances of 3000 Q, which are the optimum fil-
ter matching conditions for a flat pass band.
Note that the simple error correction
scheme (trace S21) shows distinct spikes in
the filter pass band, while the 12-term cor-
rected result (Mem4) is as smooth as can be.
Also shown (see the QEX article graphics in prev post)
are two simple enhanced through
correction schemes (Traces Mem1: ETC1;
Mem2: ETC2), which only require measure-
ments in one direction. They partly take into
account non perfect matching conditions, but
they do neglect multiple reflections. ETC2
can be obtained from the 12-term equations
by setting S22 = 0 for the forward terms and
S u = 0 for the backward terms. ETC1 is
obtained from ETC2 by setting the detector
impedance equal to the reference impedance
(usually 50 Q). The ETC corrections are bet-
ter than the simple one, but worse than the
12-term correction.

-----

(source: May-June 2009 QEX, article "A Simple S-Parameter Test
Set for the VNWA2 Vector Network Analyzer"
by Thomas C. Baier, DG8SAQ, (developer of the VNWA
describing Andreas Zimmermann (DH7AZ)'s innovative relay box idea.)

I've been meaning to try something like this, but haven't gotten around to it yet. I picked up an old HP 8711 K02 switching test set for that purpose, which is a essentially just a box with 4 N-connectors on the front and a couple of high-isolation RF relays inside. It's a big 19'' unit, so maybe not the most practical NanoVNA accessory, but it was cheaper than buying the relays separately (and as a bonus has two of those unobtainium HP probe power connectors as well).

The normal NanoVNA software doesn't support this measurement mode of course, but scikit-rf will do the 2-port corrections just fine.

The accuracy may be limited (apart from the switch repetability) by the unusal NanoVNA v2 architecture. The usual error model assumes that the load match on port 2 is the same for reflection and transmission measurements. But because there is only one receiver that is switched in for R and T measurements, this will not be the case, so the error correction will not be complete. It remains to be seen how bad this is in practice.

[switchabl]

Well, to my knowledge there is no software support in the popular NanoVNA QT and NanoVNA-saver for this kind of measurement. But I do most of my analysis in Python and there is a great package called scikit-rf that takes care of most of the math and has support for the kind of error correction described in the article you quote already (they call it two-port-one-path calibration). I have done some mildly promising tests where I switched the ports manually already.

The single-receiver architecture of the NanoVNA v2 is a separate issue. You would normally hope for the switching-approach to give error-correction performance similar to a true 2-port VNA, but that may or may not be true for the Nano. The reason is that the 12-term error model the algorithm is based on makes some assumptions that are no longer fully valid in our case. I am sure that the performance will still be improved over a normal transmission/reflection measurement, but to what extent remains to be seen.

Anyway, of course I'm not hoping to do metrology with the NanoVNA, I am just curious how far you can push it and where the limitations are. I see that @joeqsmith has already a lot of testing in the meantime, and I think "suprisingly useful up to 2-3GHz, horrible above" is probably a fair summary (that also matches my own experience). But I am still looking for a more quantitative picture, so the plan is to test drift, noise, linearity, corrected system performance etc. systematically - essentially everything you'd expect to find in the spec sheet of a proper VNA. The point is both to have actual estimates for your measurement uncertainties and to explore areas to improve as well as conceptual limitations. E.g. does it actually make sense to have a switching testset as an accessory (or built into a future version).

I think I have most of the things I need by now, the switching test-set, a 3.5mm precision airline, 3.5mm reference attenuator, precision step attenuator and so on. But unfortunately I have been busy with other things and it takes some time to do this properly. So it may be a while before I have anything definite to show.

[joeqsmith]

The author of the article uses a shielded box partition with a wall between them for isolation, so it could easily double the size of the nanoVNA, so the idea of an external add on for bench use makes more sense I think. See the second graphic below.

This would be a useful add on for the many of us with these small VNAs.

It could also be incorporated into a nanoVNA case. If it was permanently there, it would make sense to calibrate the vna for its presence permanently and add its presence to the FW. People could plug the calibration into the routines. Not very difficult. That would eliminate another source of error.

...

I think its possible to use relays to extend the functionality of a two port VNA to make an effective four port VNA too. (Or so I read in QEX a while ago)

Joe, the same relays are used as the tiny low voltage Omron ones you use in your switch box.

I think those were $10-$20 each.  They work alright but the isolation was poor.   I have an old Transco relay that I made up a really crappy DC-DC to drive it from the USB port.  My software supports it.  I plan to mount that along with a couple of bias-t's and maybe a decent step attenuator.   But then I think, I should really add some mixers and some better couplers and maybe a wide band amp......       Reality always steps in. 

The idea is that by reversing the two ports available with the very minimal differences possible, all four sets of S-parameters of a DUT become measurable.

(Thats what I meant by "four port" - I meant that the more VNA functionality is exploitable).

....

There is a 3.3v version of the Omron relays.

Yes, they have a few different coil voltages to choose from. 

It's still a 2-port system.     There are 4-port and more VNAs.   This technique of using a transfer relay is very old. Even my 1970's HP uses this technique.   

These are the two I made.  Again, performance was poor but the Omron relays were good enough to demo them. 
https://youtu.be/GJNMnq8eD0E?list=PLZSS2ajxhiQD2gftdurGQoyGpUM_HobNI&t=2047

[Mechatrommer]

I think its possible to use relays to extend the functionality of a two port VNA to make an effective four port VNA too. (Or so I read in QEX a while ago)

The idea is that by reversing the two ports available with the very minimal differences possible, all four sets of S-parameters of a DUT become measurable.

to measure all four sets of s-parm, you need full 2-ports VNA (to generate full s2p file), NanoVNA is not full 2-port.. is called TRL VNA iirc... ie cant measure S22 and S12 directly.. for such architecture like Nano, you'll need a  s-parm test set as you mentioned using relays to switch DUT's port automatically between port 1 and 2 to enable TRL VNA type to measure full 4 s-parm. but FW need to support it, unfortunately NanoVNA's FW doesnt. so you need to switch relays manually, or unwrench-rotate dut-wrench manually, save s1ps manually, and then photoshop the s1p -> s2p manually. 4-port or more VNA (s4p s5p ... ) is far far away from our league, and see very little to no usage in hobby arena, which usually dealt with making antenna (1 port device) or filter (2 port, only forward gain (S21) is relevant).

[switchabl]

The principle can be extended to 4 (or more) ports with more relays. Agilent did a 12-port switcher for the 871x series: https://www.keysight.com/en/pd-1000002217%3Aepsg%3Apro-pn-87050E/multiport-test-set-50-ohm?cc=US&lc=eng
Calibration does gets quite cumbersome really quick though because for full calibration you would have to calibrate each pair of ports (e.g. 4 ports: 4x SOL calibration + 12x thru calibration = 24 connections).

If you need more ports only occasionally, you can measure an N-port with a 2-port VNA by measuring all pairs and terminating the others with 50ohm loads (those don't have to be perfect, just known). Then you can calculate the full N-port parameters. Not quite as accurate and not nearly as convenient of course.

Apart from the obvious (couplers, splitters, hybrids), maybe the most intersting application of 4-port VNAs is for measuring balanced networks. For instance you can measure a differential transmission line as a 4-port, convert to common/differential-mode S-parameters and then simulate an eye-diagram from that. But I agree that most NanoVNA users would probably rather have e.g. power sweeps for amplifier measurements than a 4-port version.

[joeqsmith]

If I could afford a 4-port or more, I think I would also invest in an ecal for it.   The only time I have thought it may have been nice for home use was to look at some home made mixers and couplers.   There was a 3-port 8753 for sale a while back with the test set built in.  I've never seen one.  They were asking  around $5K.  Sold fairly quickly.   

If you wanted to use a transfer relay with the Nano, I would expect you could use a PC to drive it.  No real need for firmware changes but the software would need to support it.  Maybe the newer opensource software does.

[cdev]

One thing that I would like to understand better is when you have a need to phase shift a broad range of frequencies 90 degrees consistently. Of course if its just a single frequency, you can use a quarter wave piece of ideally higher impedance coax, for example, this works great to make a turnstile antenna for a single frequency.

But making a broadband four armed log spiral or pinwheel antenna, to get highly accurate multi-GNSS constellation positioning, requires a broadband 90 degree hybrid. Similarly with some other fairly easy to make designs. Apparently, the hybrid can be integrated into the design of the antenna and fabricated on PCB. But that's where the black magic begins and mortals like I who cant afford even the cheapest GUIfied antenna design software and struggle with understanding the math involved hit the wall, so to speak.

The nanoVNA is great because you can literally just see whats happening, and bring it with you to in my case my kitchen table, which is where I do most of my antenna experiments these days.  (usually small antennas)

It still is pretty useful in this setting for figuring out how they work.

Maybe someday soon true four port VNAs will undergo some similar value for money transition.

Who was it who said "the impossible takes a little longer"?

If I could afford a 4-port or more, I think I would also invest in an ecal for it.   The only time I have thought it may have been nice for home use was to look at some home made mixers and couplers.   There was a 3-port 8753 for sale a while back with the test set built in.  I've never seen one.  They were asking  around $5K.  Sold fairly quickly.   

If you wanted to use a transfer relay with the Nano, I would expect you could use a PC to drive it.  No real need for firmware changes but the software would need to support it.  Maybe the newer opensource software does.

A Raspberry Pi or some similar, cheap SBC with a healthy number of GPIOs would be a great platform for controlling and interacting with a VNA, particularly doing fun interesting visualizations with its output like you do with your software.

[joeqsmith]

Having your own software that interfaces with the Nano as well as a some simple I/O would be very powerful.   I'm interested in seeing what ever experiments you come up.   

There was a study done at CERN that I have been meaning to replicate for the fun of it.  I have the parts and made up a test jig for it but haven't made the time to run it yet.   
http://jeroen.web.cern.ch/jeroen/resistor/shuntC.html

[RomanWorkshop]

Is it possible to use 4" TFT touch display with ILI9488 driver instead of ST7796?

[rfclown]

To measure PCB trace impedance, better make through fixture for S21 measurements like shown here. Your S11 results show impossible - line impedance wandering between 25 and 200 Ohms. Do you know why?

It's easier to use S11 for measuring impedance than to use S21. I have some of those PCBs shown in your link, and they aren't 50 ohms. They're close enough for many uses < 2 GHz, but I wouldn't use them for characterizing components.

[edit: I take back the "aren't 50 ohms". I was remembering that they weren't good > 2 GHz, so I assumed it was impedance. It may be the cheap connectors. I brought one home from work and just measured it. It looks good, like about 48 ohms, up to 1.5 GHz, but then it goes south. A guy on eBay sells these, and they are handy... but not good for high frequency.]

I'll illustrate with a microstrip PCB that's close to 50 ohms, but not quite. If a transmission line is not terminated with its characteristic impedance, then it will act as a transformer. I'm measuring S11 of the microstrip PCB with a 50 ohm termination on the end. Since the line isn't quite 50 ohms, the line transforms the 50 ohm termination impedance to 42 ohms at 300 MHz. To do a sanity check, the board measures 126mm, which for a dielectric constant of 4 (FR4) would be a quarter wavelength at 298 MHz. The microstrip line impedance is then the geometic mean sqrt(42*50)=45.8 ohm.

Another way is to do a broadband measurement and then do an IFFT of the result, which is what the TDR math function of a VNA does. Again, this is an S11, not an S21 measurement.

[DiSlord]

The only real problem I saw when putting the plastic case that came with the V2+ together was the switches were not installed flat into the PCB.  The would interfere with the holes.   Quick touch of the iron.  Worse part is having to align the spacers when placing the covers.  They tend to move.   Extra spacers and screw were included.  I suspect people would loose them.   I left off the rubber O-ring.   

Flipper had printed a case for my original Nano.  I added a slot for the development connector.

I made case for V2 vs 4` screen and print it

Contain box for calibration kit
Allow store stylus inside
Power on switch move to top
Models avaible on
https://www.thingiverse.com/thing:4727705

[AE7OO]

Greetings,

I have the "Black and Gold" clone I got on sale from Amazon.
Attached is a screenshot from VNA-QT.  1010 data points.

I've done a SOL cal on port 0.
port 0 has the 50 ohm load attached.
port 1 is open.
Both ports have a port saver attached.

How does it look?  In other words, was it worth about $55, or did I get taken?

[thinkfat]

To measure PCB trace impedance, better make through fixture for S21 measurements like shown here. Your S11 results show impossible - line impedance wandering between 25 and 200 Ohms. Do you know why?

It's easier to use S11 for measuring impedance than to use S21. I have some of those PCBs shown in your link, and they aren't 50 ohms. They're close enough for many uses < 2 GHz, but I wouldn't use them for characterizing components.

I'm guessing it depends a lot on how close to 50 Ohms port 2 of the VNA is. Port 1 "sees" the impedance of port 2 transformed by the transmission line (which is also not exactly 50 Ohms), so if you want to measure your transmission line it's probably better to terminate it with a known-good load and only look at the reflected energy.

[xzswq21]

I have a NanoVNA v2.2 with 2.8 inch screen size, can I buy a 4 inch TFT LCD for the NanoVNA and use it? after upgrading can I use the tact buttons?

[cdev]

Recently, Ive started to see premade modules embedding RF switch chips becoming available in the usual sources. Since we're talking about switching, one good example is a 8 way (HMC253 8PST) switch I keep seeing around.

It implements a octodirectional connection. I am thinking about getting one to use for rapid comparative testing of antennas. Also maybe try a pseudo-doppler direction finding antenna.

One can use four identical whip antennas. 

Anyway, of course, I am also speculating that this might offer some advantages with computer controlled analysis with my nanovna if the response was flat enough. (as opposed to relays) I don't know. they seem to have a lot of loss. But no moving arts.

The switches on ebay use HMB series chips..
appear to be analog devices parts.

https://www.ebay.com/itm/HMC253-DC-2-5-GHz-RF-Single-pole-Eight-throw-Switch-RF-Switch-Selection-Antenna/274646042226

But there are lots more of them.

https://www.analog.com/en/parametricsearch/10723#

For HF it might be possible to just use pin diodes. (or even 1n4007's)  ?? (for the ghetto version!)

If I could afford a 4-port or more, I think I would also invest in an ecal for it.   The only time I have thought it may have been nice for home use was to look at some home made mixers and couplers.   There was a 3-port 8753 for sale a while back with the test set built in.  I've never seen one.  They were asking  around $5K.  Sold fairly quickly.   

If you wanted to use a transfer relay with the Nano, I would expect you could use a PC to drive it.  No real need for firmware changes but the software would need to support it.  Maybe the newer opensource software does.

[DiSlord]

I have a NanoVNA v2.2 with 2.8 inch screen size, can I buy a 4 inch TFT LCD for the NanoVNA and use it? after upgrading can I use the tact buttons?

See this post https://www.eevblog.com/forum/testgear/nanovna-v2-aka-s-a-a-2/msg3420062/#msg3420062
On it i put in case my test board (on it i test 4 inch display code support, V2 and 4 inch display module provided by owo)

Need just replace old LCD module on new 4 inch (but on new moduleneed remove SD card slot, since it will rest against the RF connectors)

[edigi]

I have a NanoVNA v2.2 with 2.8 inch screen size, can I buy a 4 inch TFT LCD for the NanoVNA and use it? after upgrading can I use the tact buttons?

I went also a bit further and designed and 3D printed a case for the V2 after upgrading to the 4 inch LCD (this is different from what DiSlord designed).
I'm still not quite happy with the result as both the buttons and the on/off switch get to an awkward place after the LCD upgrade.
Both remain usable but it's uncomfortable to access them as they are too deep behind the big LCD (the buttons even between the SMA that I didn't like even with the original LCD).
I haven't had time recently to think more about how it could be improved without modding the electronics but I've still not abandoned this plan.

[mcsarge]

Any tips on mounting the battery in the cavity?

I have designed a case for the NanoVNA V2. It has space for a 18650 cell, stores the stylus and has a lanyard attachment point. It is held together with 4 2.5mm countersunk screws on the bottom. Enjoy!
Edit: I just learnt that the 4 LED's on the bottom show the state of charge. Easiest would be to print the case (or at least the bottom half) out of translucent plastic. Another alternative may be to drill some holes into the side. Let's see what else we can come up with?

[ch_scr]

I just put it in with a strip of 0.2" foam to prevent it from ratteling around.

[rfclown]

I just got a NanoVNA V2 S-A-A-2. So far so good. I want to read the measured data (not control the unit) from the serial port. I tried nanovna-saver, but it just gave a black wiindow on my Win 10 machine. vna_qt runs, but I don't see how to make it save data (and I don't want to control the unit. just measure manually, and get results on my computer.

From https://nanorfe.com/nanovna-v2-user-manual.html I was able to read the FIFO data and get the S parameters. LabVIEW code to convert the string I read by sending 0x18 0x30 0x64 (read 100 FIFO elements) into S parameters is shown in attachment. What I see matches the display.

What I can't figure is how to read the frequency parameters. Addresses 0-0x23 always read 0. Seems these are write only. How can one read the frequency parameters?

my firmware says git-20200308-39438ba

... edit ... ugh! I just read that the valuesFIFO is uncorrected data.  I really didn't want to get into much software; just read corrected data.

... more edit ... updated firmware to 20200619, and I can use nanovna-saver now. tried 20201013 (latest release) but my unit was really unhappy with that. Didn't see the .bin file for 20200926. What I got going I don't like, but it worked. Cal unit, connect to nanovna-saver, enter sweep (because I guess it can't read what sweep the unit has), measure, save touchstone, but when I disconnect COM port the unit resets. I need to learn to work in nanovna-saver I guess.

... more edit ... found 20200926 binaries here: https://nanorfe.com/downloads/20200926/
my unit doesn't like 20201122 either. If all is well with 0926, it will be the version for me.

[joeqsmith]

I both the original NanoVNA and the two V2Plus VNAs.   I doubt I can be much help outside of suggesting you look at the open source software.   

On the original NanoVNA, I request the frequencies.   It could send corrected data, assuming you are referring to having the Nano use the SOLT coefficients.   I normally have it send uncorrected data and use the PC to store and apply the coefficients.
 
The Plus always sends uncorrected data.  You tell it the frequency range, you don't ask for it.  IMO, the Plus is much better to run headless.   It's decoupled to the point that that LCD, touchscreen and buttons are worthless.   Everything is handled by the PC. 

I doubt my software will support your Nano but you are certainly free to try it.

[dazz1]

Hi
The NanoVNA V2 looks like a great piece of test equipment.  I am surprised to see plastic enclosures with an LCD display over the top of the PCB. Typically I am used to seeing lots of shielding to keep out interference (eg USB) and other noise.

Firstly, is the NanoVNA performance affected by external noise? (where external is defined as outside the signal processing circuitry)  If the NanoVNA is not susceptible to external noise, then shielding would be nugatory. 

If shielding is added, it needs to be engineered as part of the circuit design.  Good shielding is often metal based (tin, copper, aluminium, silver) and these act as a mirror for RF.  If a noise source is placed inside a perfect shield, the amplitude within the shielding can high if the shielding and cavity Q is high.    This can be fixed by adding an RF attenuator inside the shield.  I use lots of cheap ferrite beads glued to the interior surface of the shield.  They soak up RF energy.  Examples can be seen in the attachments.  Shielding applied to a RF in-line amplifier.

Has anyone tried this with the NanoVNA? 
Would there be any benefit from fitting low Q shielding??