NanoVNA V2 aka S-A-A-2

[TimNJ]

 @OwO

Just a general comment that the SMAs on my board are quite crooked...I checked the NanoRFE webpage, and lo and behold, seems like those connectors are crooked too.

Not sure the angle, but pretty obvious. Consider tightening up the holes a little bit?

[OwO]

Thanks for bringing that up, I'm just about to do another prototyping round for the 4 inch PCB and also realized many other holes are too loose too. I should add this to my DFM checklist...

[rf-loop]

NanoVNA V2 have also "zero span" mode as CW mode. Even more there is selection for stop sweep.

It can asssume in this mode it generate normal CW signal as what ever RF generator or as example common SA's tracking generator in zero span mode.

It was quite big surprice to me that Nano works in this mode more like noise transmitter than anything else. Useless.  Now I wonder why. There is quite ok TCXO reference and circuits what can generate much more clean signal than now - I believe. 
Where is this problem. Is it possible to repair just adjusting things in FW for generate more clean CW than this do now. Why there is CW mode if it is totally useless for anything but playing fun.

Here is one SA image. 
Signal is 146.5MHz out from Nano.
Mode CW and then also selected sweep stop.

Jitter, FM, residual FM, hopping, phansenoise or AM mod or what ever mess there is....

I know this is VNA. In this use it is not so problem but same HW can use also for some kind of RF generator. There is even selection for CW.

Is in any "to do" list so that may have some hope about small somehow working tiny RF generator because there is hardware already...
Kind wish to development team, please look if it can make usable with just programming without any HW mod.


Lets give picture talk...how it is now.  All understand it is now nearly useless...
Yes it can also see that TCXO is far off but somehow acceptable stable in this class. "SA" used here is based to better reference and it is least inside 0.1ppm.

[edigi]

Jitter, FM, residual FM, hopping, phansenoise or AM mod or what ever mess there is....

I you check it with DSO (that can handle reasonable high frequency and has big enough memory) it'll become pretty clear that it's not a continuous signal (which I'd also prefer) but more like when it's sweeping so a bit of signal then silence for PLL to lock to the next frequency (which is just the same here) and so on (I've checked it a long time ago but if my memory serves me correct, stepping is something like 5ms).
While I did not check the code, I suspect that the same loop is used to sweep.
It's OK for VNA functionality but for CW it would be indeed better to emit real CW signal like the original Nano (even though it may eat a bit more battery etc.)

[rf-loop]

Jitter, FM, residual FM, hopping, phansenoise or AM mod or what ever mess there is....

I you check it with DSO (that can handle reasonable high frequency and has big enough memory) it'll become pretty clear that it's not a continuous signal (which I'd also prefer) but more like when it's sweeping so a bit of signal then silence for PLL to lock to the next frequency (which is just the same here) and so on (I've checked it a long time ago but if my memory serves me correct, stepping is something like 5ms).
While I did not check the code, I suspect that the same loop is used to sweep.
It's OK for VNA functionality but for CW it would be indeed better to emit real CW signal like the original Nano (even though it may eat a bit more battery etc.)

Yep.
As I write no prob in VNA mode but in CW mode this is totally bad. CW is Continuous Wave.
And when it pulse after then it is modulated. Now with this modulation there is also quite high phase noise and perhaps we can see some residual FM after this pulsing is stopped but then we can talk is it ok... yes perhaps due ti its nature as it is, mainly tiny cheap VNA what is better than its price.

[switchabl]

Yes, this has come up a few times, I think. Here is what the ouput power looks like in the time domain (@1GHz). The stimulus is turned off every ~5ms. I believe the device regularly performs some kind of self-calibration to deal with drift. The intervals are not quite regular either (likely depends on what other tasks it has to perform), so the spectrum will look rather wild.

I don't think "CW mode" is intended for use as a signal generator. It is simply a zero span mode for the VNA (this is called "CW time" on Agilent VNAs as well). I agree that the name is somewhat misleading.

Adding a proper CW output in the firmware should not be particularly hard. It would still be of somewhat limited use, since it is unlevelled and there is no filtering. This is probably why no one has bothered so far. It is open-source after all.

[rf-loop]

I don't think "CW mode" is intended for use as a signal generator. It is simply a zero span mode for the VNA (this is called "CW time" on Agilent VNAs as well). I agree that the name is somewhat misleading.

Adding a proper CW output in the firmware should not be particularly hard. It would still be of somewhat limited use, since it is unlevelled and there is no filtering. This is probably why no one has bothered so far. It is open-source after all.

Yes. But there is more than just one zero span mode. For what all these need use this pulse mode for measurement.  Least one can be just plain Continuous Wave. I have in my homelan workshop "trucload" of RF generators. But I think peoples who have this Nano, what is cheap, pocket size handy tiny tool. Even if its RF CW output is not good quality and without level adjustment and not even known level and or it is far away from  flat, not even close, from low to high freq.
Still I think it is not at all bad feature for many peoples if it have just this RF quality but without this "pulse modulation" what just make it totally useless if try use it as some kind RF generator. And because all is ready in HW for it. 
In menu there can select start, stop, center, span,  cw freq, and pause sweep. Is it enough that one setup can result CW without pulsing.
No matter if Nano do not "listen and measure" anything in this CW output mode (CW and Pause selected example). Now in CW+Pause it still is pulsing and this is what I wonder. Without pause, it also measure in CW mode so perhaps it need pulse as it also do when sweep (stepping sweep).
 
Of course if known level and need adjust level, nano do not this, but for many hobbyists there is available many cheaper and more cheap low grade attenuators and some day he or shee may check different frequencies true level and after then he can use it knowing something more about level. For things what need medium or high quality RF generator, this is not.
Also its RF out is quite "dirty" but still better than no gen for some hobby purposes and in machine what can take in pocket to everywhere.

I am not ranting about this and that like I am if this is true commercial business model product, but it is not, this is just open source as you told.

It was wish that if developing team see it is easy to do change it is perhaps good to do. Even with all cons what it have as rf generator. Just like, better than nothing for perhaps many peoples. Least it do not shoot not down this NanoVNA value if there is this "CW only out" mode.
It also mean that in this mode it do not measure anything like it do now. In CW mode now,  it do measurements also of course.
But then there is Pause Sweep selection. Now it do not update any displayed measurements in CW mode but still it continue Pulsing. For what.
In this state it can, imho,  send just Continuous wave out. Least I somehow think that when it do not need measure anything in CW mode then pulse modulation can also stop if reason for pulsing is measuring.
 
Perhaps example OwO can answer what is real reason for pulsing in state/mode when it do not any measurements.
Or is it just..."no one care" reason.

[cdev]

I don't think "CW mode" is intended for use as a signal generator. It is simply a zero span mode for the VNA (this is called "CW time" on Agilent VNAs as well). I agree that the name is somewhat misleading.

Adding a proper CW output in the firmware should not be particularly hard. It would still be of somewhat limited use, since it is unlevelled and there is no filtering. This is probably why no one has bothered so far. It is open-source after all.

Yes. But there is more than just one zero span mode. For what all these need use this pulse mode for measurement.  Least one can be just plain Continuous Wave. I have in my homelan workshop "trucload" of RF generators. But I think peoples who have this Nano, what is cheap, pocket size handy tiny tool. Even if its RF CW output is not good quality and without level adjustment and not even known level and or it is far away from  flat, not even close, from low to high freq. Still I think it is not at all bad feature for many peoples if it have just this RF quality but without this "pulse modulation" what just make it totally useless if try use it as some kind RF generator. And because all is ready in HW for it. 
In menu there can select start, stop, center, span,  cw freq, and pause sweep. Is it enough that one setup can result CW without pulsing. No matter if Nano do not "listen and measure" anything in this CW output mode.  Of course if need adjust level, nano do not this, but for many hobbyists there is available many cheaper and more cheap low grade attenuators and some day he or shee may check different frequencies true level and after then he can use it knowing something more about level. For things what need medium or high quality RF generator, this is not.

I am not ranting about this and that like I am if this is true commercial business model product, but it is not, this is just open source as you told.

It was wish that if developing team see it is easy to do change it is perhaps good to do. Even with all cons what it have as rf generator. Just like, better than nothing for perhaps many peoples. Least it do not shoot not down this NanoVNA value if there is this "CW only out" mode.
It also mean that in this mode it do not measure anything like it do now. In CW mode now it do measurements also. But then there is Pause Sweep selection. Now it do not update any displayed measurements in CW mode but still it continue Pulsing. For what. In this state it can send just Continuous wave out.  Least I somehow think that when it do not need measure anything in CW mode then pulse modulation can also stop.
 
Perhaps example OwO can answer what is real reason for pulsing in state/mode when it do not any measurements.
Or is it just..."no one care" reason.

It would be worth it to code an interface to allow scripting of the signal generator. ,I have some very small 3.3 volt omron RF relays I would like to use them to be anle to do a RF switchbox so I could script and speed up testing and workflow.

[OwO]

The pulsing isn't because of internal calibration, it's because the reference channel is measured by shorting port 1 to ground. The two synthesizers do not have known phase relationship (ADF4350 phase resync is not usable for the full frequency range) so the reference measurement is required. It is possible in CW mode to only measure reference once, but then you will have a lot of drift and anything that causes a ADC buffer overrun (like writing to flash) will cause it to lose sync and phase will become random.

[rf-loop]

The pulsing isn't because of internal calibration, it's because the reference channel is measured by shorting port 1 to ground. The two synthesizers do not have known phase relationship (ADF4350 phase resync is not usable for the full frequency range) so the reference measurement is required. It is possible in CW mode to only measure reference once, but then you will have a lot of drift and anything that causes a ADC buffer overrun (like writing to flash) will cause it to lose sync and phase will become random.

Afaik, in normal "CW" aka pulse modulated CW it still do normal S11/21 measurements. It pulses and measure continuously, example if I connect load or what ever it give measured results.

I mean now, outside of this functionality, just different "CW blind output" mode what do not measure anything but push raw RF out, if something drift or not just "so what".
 
It is now same pulsed output if I select start and stop same, if I select center and zero span or if I select just CW. All same. But even more, if I select Pause Sweep independent of  selected Span or just CW it also generate pulsed output.

Example it can do so that normal CW out as now can be just as now including it do measurements. But if CW is selected and then user select sweep pause. It do not anymore measure anything or least it do not display anymore new results. In this mode when CW selected and PauseSweep selected  Now it can just blind push selected RF frequency out without anything else. If there is some drift or not, just out with blind eyes and ears. Then this generator output drift like it drift, more or less and it is as it is, take it for use or not.
I believe still this feature add some small value to Nano, it hardly reduces the value.

[switchabl]

I have patched the firmware so the measurement is turned off when the sweep is paused. I am not too interested in the S-A-A2 as a signal generator, but I want to some more detailed performance characterizations and for that purpose it can be useful to have a CW output.

https://github.com/switchabl/NanoVNA-V2-firmware/tree/cw

From a quick test it seems to work fine, but it is probably not ready to be included in the mainline firmware. There are other events (like changing display settings) that unpause the sweep and in that case the ECAL switch may be left in an unexpected position and throw off the next measurement. Also this is based on the current master, so may contain other unexpected bugs (and features). Still, if you feel adventurous, feel free to try the binary (https://github.com/switchabl/NanoVNA-V2-firmware/releases/tag/cw-test1, usual disclaimers apply).

I am not entirely convinced that this is the right way to go about it. I think the least surprising behaviour would be to turn the stimulus off entirely when the sweep is paused. If desired, there could be a separate mode (or at least menu) for use as a generator. But on the other hand that would be a lot more work.

[ch_scr]

Now here's a case.  It's a wonder what you can do with Covid-19 spare time.


Angled so you can see it.
Holds SOLT, attenuators and adapters in a drawer.
Holds a standard 18650 Li cell.
Lots of air holes for cooling.

Your drawer has bugged me for a while, but I wanted it optional and tiny. Today was the day!


The black PLA seems to get only warm, not full on hot on the bottom so I opted out of cooling holes.
This will not fit the case I posted earlier, so I have not included an stl for now.

[switchabl]

I have patched the firmware so the measurement is turned off when the sweep is paused. I am not too interested in the S-A-A2 as a signal generator, but I want to some more detailed performance characterizations and for that purpose it can be useful to have a CW output.

https://github.com/switchabl/NanoVNA-V2-firmware/tree/cw

Somewhat more polished version, where you can change the frequency without having to pause again afterwards:
https://github.com/switchabl/NanoVNA-V2-firmware/releases/tag/cw-test3

I think it shouldn't break anything else (sweep pause is now separate from UI refresh). So if there is interest in having it in the official firmware, I can make a pull request.

[DiSlord]

I would really like that device with larger and higher-resolution screen, because normally I look at them from some distance (to have enough space for DUT directly in front of me). I wonder if it's possible to re-design the display part to utilize something like 7" RGB24 screen (perhaps via something like EVE controller - FT813 or BT815). Or use beefier main MCU which supports parallel RGB LCD panels directly.

I test 7`display for H4, and it show good perfomance on only SPI bus (on display test i use 401 point measure) (yes if use 8bit or 16bit parallel connection can more faster), but 8bit color mode + hardware bitblt and fill allow get good speed.



After all optimization (on video my first run after add touch, after i increase update speed ~2-3x), at this moment update speed ~as on old firmware H4 (full screen update need only ~90ms)
Code can ported to V2 (it use some codebase for display as H or H4)

Additional possible add 5x4 keyboard (20 keys)
Capacitive touch allow use finger for control
Backlight control

Minus is only cost for display module ~40$

[Mechatrommer]

to avoid hijacking the other thread, i just post this here...

That's the one I bought; look back at my posts to see how I changed it to 4" display.
I have to say that the original 3.2"? LCD was close to unreadable for my old eyes, the 4" display I fitted is really good and I have no problems using it now.  I haven't got around to printing my 3D case for it yet though

thanks for the tip. requoted here for future reference... i'm thinking if 3d printing cant do what i want, maybe cutted wood/acrylic/plastic/metal can may do the job as well for the enclosure / test rig... i'll try to do whats best for my "way", not necessarily need 3d printed enclosure.

So I got round to upgrading the firmware to the version that handles the 4 inch display.  For those who haven't done it before:
1. Get the new .bin file ready.
2. Plug in your NanoVNA and check which port it comes up on.
3. Turn the NanoVNA on and put it into DFU mode via touch menus.
4. Plug into USB port.
5. Run up NanaVNA_QT
6. Select the port under device
7. You should now see a pop up window that says found device in DFU mode, do you want to upgrade firmware (or something close to that)
8. Click [Yes], browse to .bin file, and click [OK]
9. When it's finished unplug and power off NanoVNA
10. Fit new display and power on, it should work.

I recalibrated my touch screen after the upgrade but it was close anyway so the menus worked.  I removed (carefully) an SD card reader that was fitted to my display as it interferes with the fit.  I must say that the larger display and font has made this very usable for me, much easier to read now.

I found this one on Amazon https://www.amazon.com/480x320-Display-Module-ST7796S-Driver/dp/B08C5FFX9P/ref=sr_1_2?dchild=1&keywords=480x320+ST7796S&qid=1596014221&s=electronics&sr=1-2
[EDIT] I also just found this one on AliExpress, please check but I think it's the right one. 

The git version has 4 inch display support but there is no binary released yet; I'll go do a release right now. https://github.com/nanovna/NanoVNA-V2-firmware
The battery is any li-ion battery (3.6V and 3.7V cells are the same thing, they are all 3.0 - 4.2V). The dimensions should be 6*40*60 mm to fit in the device, usually labeled as "604060", or use that enclosure that supports a 18650 cell (I would go this route because 18650s are generally safer).

edit: i just noticed someone upgraded it with 7" LCD. i got one here for rasp pi orange one, not sure if its compatible, will check deeper when my unit arrived..

[OwO]

can existing version be made faster sweep too? is it simply updating FW,  or there using new IC?

There are both circuit and firmware changes, I will post more info some time after it's released, or DM me.

[OwO]

Pictures of the V2 plus4 in metal enclosure, with 3200mAh battery inside

After calibration the dynamic range still reaches 70dB to 3GHz, so plus4 will come with a metal enclosure.

[joeqsmith]

I still haven't pulled the trigger on the V2 but you keep enticing me.

Will this one support the faster sweep times you mention?  How fast on the screen, and how fast across the USB port?   Was there a hardware change to support the faster sweeps, or will this firmware run in the original V2, allowing for the same faster sweep rates?

Whats the number of samples per sweep?

Will it support more complex SOLT?  At 3GHz, I'm surprised it wouldn't have something better than the Nano.  OF course, this leads to the question, what's the progress on the 6GHz unit?   

[OwO]

Yes this one is 400 points/s (compared to 100 points/s for V2_2), same speed on-screen and over USB. I will post a comparison video later. There are substantial hardware changes to support the higher sweep rate, and a new directional coupler with high native directivity (>18dB for the coupler itself, ~17dB in circuit with SMA connectors, 50kHz to 4GHz).

Number of sweep points is adjustable up to 201 points on-screen and unlimited over USB (compared to 1024 points in V2_2 and V2 plus). More advanced calibration may be added to the firmware later. 6GHz VNA is developed by a different person and I'm not sure what the timeline is.

[Miti]

Pictures of the V2 plus4 in metal enclosure, with 3200mAh battery inside

After calibration the dynamic range still reaches 70dB to 3GHz, so plus4 will come with a metal enclosure.

Where and when will this be available? Price? I'm sorry if this is discussed somewhere in the thread, TLDR.

[OwO]

Probably 2 more weeks, I think the PCBs are already finished and they are waiting on PCBA. I don't control the pricing but it'll be set based on cost, so plus4 won't be as cheap as the V2, but I do know V2 plus costs the same as the V2 so it will be priced the same.

[thinkfat]

I'm starting to loose track of the different models now.

V2 plus4 is the upgrade with better dynamic range and faster sweep with more points and 4" screen?
V2 plus is - what?

EDIT: going back through the thread, "V2 plus" seems to also have lower noise and faster sweep, but not more points and small screen. Can you perhaps provide some differentiation between the models? Since both seem to improve in similar areas, but there's no numbers on the improvement the V2 plus brings over the original V2 and how it compares against the V2 plus4.

[cdev]

Pictures of the V2 plus4 in metal enclosure, with 3200mAh battery inside

After calibration the dynamic range still reaches 70dB to 3GHz, so plus4 will come with a metal enclosure.

Looks gorgeous!

What is the new hole to the left of the three buttons, also am I corrent in that you now are using a full size USB connector?

[OwO]

I'm starting to loose track of the different models now.

V2 plus4 is the upgrade with better dynamic range and faster sweep with more points and 4" screen?
V2 plus is - what?

V2 plus - 2x sweep speed and slightly lower noise, this version replaces the V2_2 which will no longer be manufactured (but will continue to receive firmware updates).

V2 plus4 - 4 inch display, 4x sweep speed, 70dB dynamic range to 3GHz, frequency range extended to 4GHz, and metal enclosure. This is the only hardware version that can be put into a metal enclosure without messing up dynamic range because signal isolation is much improved.

[Kosmic]

woo I'm clearly not following, I thought the V2_2 was the latest and greatest. Ordered one yesterday