NanoVNA Custom Software

[joeqsmith]

I was curious how the LiteVNA would handle the PDN measurements.  Shown with the homemade wideband common mode transformer measuring a 100mOhm, 25mOhm and 1mOhm resistor.   I will admit, I was concerned about the firmware causing a very poor low end response.  Keep in mind they only spec it to 50kHz.     

[joeqsmith]

We had ran the unit much lower.  Here I have it set to 2kHz.   I then attach the 1mOhm and swept from 2kHz to 1MHz.   

I had looked at the noise floor and system dynamic range all the low cost VNAs I have.   I plan to repeat that with this unit and will post the results.  OWO was also asking about temperature stability so I'll drag out the meat packing box.

[joeqsmith]

Looking at the noise floor, once we get beyond 100MHz the LiteVNA is worse than the V2Plus4.  As we get closer to 1.2GHz, the two are similar.   

The system dynamic range for the V2Plus4 is better than the LiteVNA.   Jan's VNA offers further improvements. 

[joeqsmith]

Sweeping from 1.5GHz to 3GHz, the noise is comparable.    The V2Plus4 seems to have better dynamic range.  Again, Jan's VNA offers further gains.   

[joeqsmith]

Sweeping from 3GHz to 4.4GHz, the V2Plus4's noise is very comparable with the LiteVNA.   

The LiteVNAs system dynamic range is better as we get beyond 3.3GHz or so.   Note, data for the V2Plus4 was collected two ways.  One with the VNA powered from the PC, the other with it running from the internal battery.   I had ran a few tests and found any time I use any external power source with the V2Plus4, the noise goes up.   Even with a highly filtered linear supply.  I doesn't seem to matter and am guessing it is a problem with the design.   The LiteVNA does not appear to have this same problem. 

[Alextsu]

Hi, Joe,
Thanks for sharing this results, very interesting!
Just a short remark concerning Jan's VNA port isolation tests, they are not up-to-date.
There had been found a workaround solution for the port isolation improvement at higher frequencies.
Also, there's a new discussion group, and actual info is posted there.
https://groups.io/g/LibreVNA-support

[joeqsmith]

Jan had announced here when they had moved the group and about the problems they had with the account.   Seemed like a really bad business choice (using an unknown account if that's really what happen) but at least they now have it sorted.  I never joined the group after reading of all the shortcomings (no direct way to control the VNA without their software layer,  new hardware announced, higher cost, low frequency performance...).   

When we bought the two LiteVNA's, they were $130/ea on sale.  I think that's what I paid for the V2Plus4 when it first came out.   The price for the V2Plus4 was $200 last I checked.   I would assume the price for the LiteVNA will also increase over the next year.

After running the LiteVNA for about 4 hours last night,  I did not experience a single problem with it.  I was really concerned with Dislord's firmware.   I noticed they advertise for donations in the version menu for those wanting to help them out.   The version of firmware in this VNA is 1.0.70.   

I was unable to change the attenuator and assume it still doesn't have support for it which does seem a bit odd after Dislord's comment about adding it.   I'm guessing the problem is with my software and lack of documentation on how to implement it.

[Alextsu]

Jan had announced here when they had moved the group and about the problems they had with the account
 I never joined the group after reading of all the shortcomings (no direct way to control the VNA without their software layer,  new hardware announced, higher cost, low frequency performance...).   

Ok, understood.
How do you think,
In case I ask Jan to add support for standard V.2 command protocol to his SW, would You consider it possible to add it in Your SW?

[joeqsmith]

Jan had announced here when they had moved the group and about the problems they had with the account
 I never joined the group after reading of all the shortcomings (no direct way to control the VNA without their software layer,  new hardware announced, higher cost, low frequency performance...).   

Ok, understood.
How do you think,
In case I ask Jan to add support for standard V.2 command protocol to his SW, would You consider it possible to add it in Your SW?

As long as the interface is well documented, then yes.  Obviously, it would require the hardware.

[DiSlord]

>I was unable to change the attenuator and assume it still doesn't have support for it which does seem a bit odd after Dislord's comment about adding it

You mean output power? This my temp V2 protocol format description:

V2 Binary protocol:
V2 contain 256 registers for read/write

On write to registers data put to reg[address] and V2 grab data and process (depend from address)

All values after in hex format!!!!
Commands for read/write
0x0D - read version (return "2")
0x10 - read byte
0x11 - read word
0x12 - read dword
0x13 - read qword

0x18 - read measured data (use 30 address for count)

0x20 - write byte
0x21 - write word
0x22 - write dword
0x23 - write qword
28 - write buffer (next byte data size, and data)

address:
0x00 - (qword) start frequency
0x10 - (qword) frequency step
0x20 - (word) points count
0x22 - (word) values pre frequency (need set to 1,

0x26 - (byte) mode (0 - USB mode, 1 - send raw data, 2 - normal mode)
0x30 - (byte) points for measure, or up to (measured values)

0x40 - (byte) average
0x41 - (byte) si power
0x42 - (byte) adf power

0x50 - (dword) color value
0x54 - (byte) color index (use H/H4 indexes)

0xee - (byte) any value, on write V2 send screenshot

0xD0 - (dword) SN[0] (serial number 0)
0xD4 - (dword) SN[1] (serial number 1)
0xD8 - (dword) SN[2] (serial number 2)

0xF0 - (byte) device variant (2 vor V2)
0xF1 - (byte) protocol version (1 vor V2)
0xF2 - (byte) board revision (2 vor V2, 3 - V2Plus, 4 - V2Plus4)
0xF3 - (byte) firmware major
0xF4 - (byte) firmware minor

Example:
Set average = 5 command:
0x20 - write byte
0x40 - address (average)
0x05 - value
Need send : 204005

Example:
Set SI5351 (<140MHz) power = 2 command:
0x20 - write byte
0x41 - address (Si power)
0x02 - value (0 - 3 value range avaible)
Need send : 204102

Example:
Set ADF/MAX(>=140MHz) power = 1 command:
0x20 - write byte
0x42 - address (ADF/MAX power)
0x01 - value (0 - 3 value range avaible)
Need send : 204201 (in hex bytes)

[joeqsmith]

>I was unable to change the attenuator and assume it still doesn't have support for it which does seem a bit odd after Dislord's comment about adding it

You mean output power? This my temp V2 protocol format description:

V2 Binary protocol:
V2 contain 256 registers for read/write

....

Example:
Set SI5351 (<140MHz) power = 2 command:
0x20 - write byte
0x41 - address (Si power)
0x02 - value (0 - 3 value range avaible)
Need send : 204102

Example:
Set ADF/MAX(>=140MHz) power = 1 command:
0x20 - write byte
0x42 - address (ADF/MAX power)
0x01 - value (0 - 3 value range avaible)
Need send : 204201 (in hex bytes)

Yes, these both appear correct.   The version of your firmware supplied with the LiteVNA is 1.0.70. 

Are there other steps beyond having the frequency set that are required to change it?   Have you verified it works?

[DiSlord]

I not change V2 protocol in LiteVNA (only extend it) for allow existing software work.

I think about use NanoVNA v1 protocol, but not ready do this

[joeqsmith]

That's fine.  Does the two examples shown actually work with the LiteVNA?   Have you tested them? 

[DiSlord]

I test LiteVNA control in NanoVNA-App software (add this feature to this app) all work. Avg setting / power for Si5351 (LiteVNA use MS5351 analog) and MAX2871

PS LiteVNA in < 400k work slow, it use less IF and need more time wait (i try think how possible improve this).

[joeqsmith]

Thanks for confirming you have tested it.  I will have another look at my code and see if I can sort out where the problem is.     

[joeqsmith]

I test LiteVNA control in NanoVNA-App software (add this feature to this app) all work. Avg setting / power for Si5351 (LiteVNA use MS5351 analog) and MAX2871

PS LiteVNA in < 400k work slow, it use less IF and need more time wait (i try think how possible improve this).

Set the start frequency to 200MHz followed by setting the step to 0 (200MHz CW).   Now try and change the  ADF's power level.    Repeat this at 100MHz with the SI's power level.   Now try repeating the above with the step of 1MHz rather than 0 to force the sweep. 

It seems that the power may only be set if the VNA is sweeping.  Let me know if you can replicate this. 

[joeqsmith]

Looking at the home made 70mm stepped airline with a sweep range of 1M-6.4GHz. 

Not sure of a good way to benchmark the temperature stability of the LiteVNA compared with the V2Plus4.   We want the contribution of the VNAs, not the loads and cables.   Seems like we would want to include both ports.   

[DiSlord]

>It seems that the power may only be set if the VNA is sweeping.  Let me know if you can replicate this.
Yes, if nano not change frequency, new power setting not apply (i fix this in future fw update). So this command not work in CW mode.

[joeqsmith]

It appears that the 23 +/-5C may be a common range they are tested over.   

For testing the two VNAs, I am thinking to place them together in the meat packing box, run two instances of the software and collect the data with a short section of Teflon coax between the two ports.  Maybe average the data down enough so we can see something.     This was a request from OWO so maybe they will chime in.   

***
They are suggesting S11 with a high return load (open or short) as they suspect there would be little error in S21.   We loose the other circuit.   I may have a look both ways. 


https://www.testforce.com/pub/media/sparsh/product_attachment/anritsu-MS2036C-datasheet_compressed.pdf
 

[tungsten2k]

Sweeping from 1.5GHz to 3GHz, the noise is comparable.    The V2Plus4 seems to have better dynamic range.

One obvious difference with the V2plus4 is the addition of a metal case which should reduce external noise floor somewhat, but would also increase internal reflection of noise from components within the VNA. The V2plus4 combats this by adding a layer of Radio Absorbing Material (RAM) to the entire back of the unit. In researching possible improvements in the LiteVNA to achieve system dynamic range parity with the V2plus4, I found this paper detailing the use of very common carbon fiber-filled 3D printer FFF filament for absorbing energy but at much higher frequencies that present in our VNAs, namely 63GHz-215GHz and observed in the neighborhood of -20db difference:

https://arxiv.org/pdf/1808.00820.pdf

Does anyone have a 3D printable case design that they would consider modifying with treatment similar to the paper (they used an open Hilbert curve infill, but I suspect even no infill but just using a 3D printable case with the appropriate CF-filled FFF filament should be enough to vet the idea).

Thoughts ?  If this is not on topic, I can start a new thread, but I'm really most interested in finding out if the LiteVNA can be made as good as the V2plus4 in the ranges the V2plus4 supports.

Thanks !

-=dave

[joeqsmith]

Sweeping from 1.5GHz to 3GHz, the noise is comparable.    The V2Plus4 seems to have better dynamic range.

One obvious difference with the V2plus4 is the addition of a metal case which should reduce external noise floor somewhat, but would also increase internal reflection of noise from components within the VNA. The V2plus4 combats this by adding a layer of Radio Absorbing Material (RAM) to the entire back of the unit. In researching possible improvements in the LiteVNA to achieve system dynamic range parity with the V2plus4, I found this paper detailing the use of very common carbon fiber-filled 3D printer FFF filament for absorbing energy but at much higher frequencies that present in our VNAs, namely 63GHz-215GHz and observed in the neighborhood of -20db difference:

https://arxiv.org/pdf/1808.00820.pdf

Does anyone have a 3D printable case design that they would consider modifying with treatment similar to the paper (they used an open Hilbert curve infill, but I suspect even no infill but just using a 3D printable case with the appropriate CF-filled FFF filament should be enough to vet the idea).

Thoughts ?  If this is not on topic, I can start a new thread, but I'm really most interested in finding out if the LiteVNA can be made as good as the V2plus4 in the ranges the V2plus4 supports.

Thanks !

-=dave

OWO had thought the performance of the LiteVNA would be worse in this region.  While I agree that the LiteVNA should have performed at least as well as the V2Plus4,  I thought that the V2Plus4 should have at least performed as well as the original NanoVNA.   They had asked for a few photos as well and eventually I plan to pull both apart for them.  As of today, I have not had them apart outside of inserting a battery in the V2Plus4.   

[joeqsmith]

The LiteVNA is boxed up and heading to 18C.   SOL cal with Open inserted.  I'll let it set for an hour at each temperature (18 & 28). 

***
The meat packing box has added insulation and an RTD that I use for feedback.  Temperatures are fairly stable.   

[joeqsmith]

The LiteVNA, swept from 18 to 28C.  Looking at S11 after SOL with open attached.   

[joeqsmith]

Data for the V2Plus4 zoomed in to get some idea how the drift looks.    The LiteVNA appears much more stable as we get above 3GHz. 

[joeqsmith]

I had also captured a Touchstone file for each VNA at the both temperatures just prior to changing temperatures.   No cables were used during this test.  The open was attached directly to the VNA's port1.