On Thu, Apr 2, 2020 at 06:14 AM, DiSlord wrote:
Use 10kHz IF freq (5kHz before)
============================================
DiSlord,
I haven't installed the new 96kHz update yet but I assume that with the new 10kHz IF you had to move the minimum start frequency back to 50kHz, right?
-Herb
|
Gabriel,
Can you tell me what the difference is between my firmware version git-20200117- 4c067eb and firmware version git-20200229-aea0dc4 reported by another member?
Thanks,
-Herb
|
On Thu, Apr 2, 2020 at 05:59 AM, Christoph Kopetzky wrote:
So what sense does it make to use a SOLT kit, when the firmware does not support a short?
==============================
Chris,
You have to read back starting from the subject of internal calibration. An internal Open/Short is used to provide temperature drift compensation after each sweep. Its was mistaking referred to as a calibration but is really part of the correction factor process.
A normal OSLT calibration is required at CH0 and CH1 to use the S-A-A-2 with any accuracy. I believe the S-A-A-2 arrived uncalibrated with the frequency range set to a range not usable for most amatuers, so a OSLT calibration was immediately necessary.
If you are planning on purchasing the S-A-A-2 and don't already have the accessories, get the bundle. For only an additional $10 it is a bargain. Even if you already fave the accessories I would pay the extra $10 just to have a spare set.
- Herb
|
Thanks
Here new update:
Change AIC3204 dac from 48kHz to 96kHz (it allow more fast measure)
Use 10kHz IF freq (5kHz before)
Add 2kHz bandwidth option
Move bandwidtth setting to config (now allow save setting config->save)
As result:
2kHz bandwidth sweep speed increased on 2x (yes little noise exist)
1kHz bandwidth sweep speed increased on 1.5x (noise level not change)
Better see sweep speed non use Smith or Polar grid.
Add both version to H (tested, but can`t measure real bandwidth and noise level, only on my work) and H4
toggle quoted message
Show quoted text
On Thu, Apr 2, 2020 at 03:09 AM, Jos Stevens wrote:
Hi DiSlord,
If you like,I 'm always willing to test your beta versions of firmware
on my -H4. You can send i\\them to my e-mail address.
Regards
Jos
Op 2-4-2020 om 01:56 schreef DiSlord:
Font size not change, change only some glyph like 'T', 'Y', 'V', 'M', 'K', 'R', 'C', '0', '3', '.', ':', ',' (and some small lile 'b', 'x', 'g', 'v')
About points count, i test on my NanoVNA-H variable points set (51 and 101)
all worked. For H4 a try complete this then i recive it. Difficult write code
and not test it on device. I test some part on my H version, but...
On Wed, Apr 1, 2020 at 03:45 PM, hwalker wrote:
On Wed, Apr 1, 2020 at 10:17 AM, DiSlord wrote:
I hope updated font look better
===============================================
DiSlord,
The larger AA fonts in the previous beta firmware version might be
preferable to some users. Perhaps that version could be the Nano-VNA-H4
antenna analyzer version, and the current beta version the default VNA
release.
All other changes appear to be working. What are your thoughts about
increasing the measurement points from 101 to a higher number in the
future?
More points would not be backwards compatible to the NanoVNA-H, but would
take
advantage of the larger flash memory and faster processor on the
NanoVNA-H4.
- Herb
|
Hi Herb and Eric,
You are right. There is a short at the e-cal, but it is used to generate the reference signal, so it cannot be used as an independent short calibration. With the addition of an extra switch you could have an independant path from the TX signal to the RX multiplexer. The signal should be attenuated as much as possible. The switch is only needed to reduce crosstalk from the reference signal to the other receiver signals. The major problem of a single receiver design is the crosstalk in the multiplexer. Much better to have multiple mixers and multiplex after the mixer. Signals are low frequency there so crosstalk is no issue. The current design is totally optimized for low cost. Gabriel & team did an amazing job there. But I'd prefer a more performance optimized VNA, even if it would cost 4x as much. But the market is probably 10x smaller...
Reinier
Op 2-4-2020 om 14:47 schreef hwalker:
toggle quoted message
Show quoted text
On Thu, Apr 2, 2020 at 02:03 AM, Reinier Gerritsen wrote:
I just checked the schematics, the e-cal switch does have open, load and short.
----------------------------------------------------------
Reinier
As Gabriel, the designer replied, "short" is not currently supported in firmware. That is why I said internal OSL is currently unavailable to use as you suggested.
- Herb
|
Hello Herb,
I am wondering why the seller of the V2 offers with the second badge two versions of the V2. A single V2 and a V2
with two SS405 cables and a SOLT calibration parts kit (male). So what sense does it make to use a SOLT kit, when the
firmware does not support a short?
May be the newer firmware (link to their GitHub repo is in the tindie shop site) supports now a full SOLT calibration...
best regards
Chris
---
'Everything should be made as simple as possible, but not simpler.' -- Albert Einstein
Am 02.04.2020 um 14:47 schrieb hwalker:
toggle quoted message
Show quoted text
On Thu, Apr 2, 2020 at 02:03 AM, Reinier Gerritsen wrote:
I just checked the schematics, the e-cal switch does have open, load and short.
----------------------------------------------------------
Reinier
As Gabriel, the designer replied, "short" is not currently supported in firmware. That is why I said internal OSL is currently unavailable to use as you suggested.
- Herb
|
On Thu, Apr 2, 2020 at 02:03 AM, Reinier Gerritsen wrote:
I just checked the schematics, the e-cal switch does have open, load and short.
----------------------------------------------------------
Reinier
As Gabriel, the designer replied, "short" is not currently supported in firmware. That is why I said internal OSL is currently unavailable to use as you suggested.
- Herb
|
The S-A-A-2 is a single receiver 2 port VNA. So this single receiver needs to get 3 input signals. reflection, thru and reference. If you look at the schematic you will easily see the switch to select between reflection and thru but there is no obvious reference switch path The reference is "switched" to the receiver by selecting reflection and using the ecal switch to bring the bridge in a defined state of unbalance that effectively requals selecting a reflection signal. Of course once you have a switch between port 1 and the bridge you can start to imagine various things you can do with the switch but for now the ecal switch is functioning as half of the reflection switch. -- NanoVNA Wiki: /g/nanovna-users/wiki/homeNanoVNA Files: /g/nanovna-users/filesErik, PD0EK
|
Hi DiSlord,
If you like,I 'm always willing to test your beta versions of firmware on my -H4. You can send i\\them to my e-mail address.
Regards
Jos
Op 2-4-2020 om 01:56 schreef DiSlord:
toggle quoted message
Show quoted text
Font size not change, change only some glyph like 'T', 'Y', 'V', 'M', 'K', 'R', 'C', '0', '3', '.', ':', ',' (and some small lile 'b', 'x', 'g', 'v')
About points count, i test on my NanoVNA-H variable points set (51 and 101) all worked. For H4 a try complete this then i recive it. Difficult write code and not test it on device. I test some part on my H version, but...
On Wed, Apr 1, 2020 at 03:45 PM, hwalker wrote:
On Wed, Apr 1, 2020 at 10:17 AM, DiSlord wrote:
I hope updated font look better
===============================================
DiSlord,
The larger AA fonts in the previous beta firmware version might be
preferable to some users. Perhaps that version could be the Nano-VNA-H4
antenna analyzer version, and the current beta version the default VNA
release.
All other changes appear to be working. What are your thoughts about
increasing the measurement points from 101 to a higher number in the future?
More points would not be backwards compatible to the NanoVNA-H, but would take
advantage of the larger flash memory and faster processor on the NanoVNA-H4.
- Herb
|
I just checked the schematics, the e-cal switch does have open, load and short.
Reinier
Op 2-4-2020 om 10:51 schreef hwalker:
toggle quoted message
Show quoted text
On Thu, Apr 2, 2020 at 12:22 AM, Reinier Gerritsen wrote:
In theory (I think), it could be used to calibrate the VNA if you know the exact reflection coefficients of the not so perfect internal open, load and short.
================================================
Gabriel responded earlier that:
"Currently the corrections only use OPEN and LOAD ports of the ecal switch; In the code there is an option to do a full SOL ecal, but this is turned off in all firmware versions and is only there for experimentation."
So there is currently no internal "short" standard provided to do what you suggested. My feeling is that a lot of user's are going to try to make measurements, without a performing a proper OSL calibration, after changing frequency parameters, and then blame the S-A-A-2 because its impedance values don't match their antenna analyzer. No interpolation is performed as done on the NanoVNA.
VSWR measurements are about the only ones I would trust without re-calibrating.
- Herb
|
F.y.i.:
The HCXQS group ontindie.com has refilled their nanovna V2's. So you can order again!
There is also now a version with RF cables and SOLT calibration parts for plus 10$!
If you do not find it: here <> is the link for the offer...
Have fun!
best regards
Chris
---
'Everything should be made as simple as possible, but not simpler.' -- Albert Einstein
Am 02.04.2020 um 02:27 schrieb johncharlesgord via groups.io:
toggle quoted message
Show quoted text
I don't remember this being mentioned before:
When the S-A-A-2 (NanoVNA-V2) frequency is changed, it does a "self cal" using an open, short, and load on its Port 0 input switch. The result is not great, but it is not terrible for some purposes. It seems to do a pretty good job near the center of the Smith chart, but not so well at the edges. Other than the expected phase differences due to the different calibration plane, it works quite well from 20 MHz to 700 MHz. I assume that the "raw" data it reports over the USB port actually uses this self-cal.
My unit has firmware version git-20200229-aea0dc4.
--John Gord
|
On Thu, Apr 2, 2020 at 12:22 AM, Reinier Gerritsen wrote:
In theory (I think), it could be used to calibrate the VNA if you know the exact reflection coefficients of the not so perfect internal open, load and short.
================================================
Gabriel responded earlier that:
"Currently the corrections only use OPEN and LOAD ports of the ecal switch; In the code there is an option to do a full SOL ecal, but this is turned off in all firmware versions and is only there for experimentation."
So there is currently no internal "short" standard provided to do what you suggested. My feeling is that a lot of user's are going to try to make measurements, without a performing a proper OSL calibration, after changing frequency parameters, and then blame the S-A-A-2 because its impedance values don't match their antenna analyzer. No interpolation is performed as done on the NanoVNA.
VSWR measurements are about the only ones I would trust without re-calibrating.
- Herb
|
On Wed, Apr 1, 2020 at 09:13 PM, Gabriel Tenma White wrote:
if you are only measuring antennas at VHF, you can get away with not calibrating, but you must then treat it as a SNA and not trust any "impedance" values on the screen
====================================================================
Gabriel,
That agrees with my testing. Although uncalibrated, I couldn't even treat it as a moderately accurate SNA (scalar network analyzer) for S21 measurements because no channel to channel correction is performed via the ecal switch. My S21 measurement error without re-calibrating after a frequency change was 3 dB or more.
Do you know what the difference is between my firmware version git-20200117- 4c067eb and firmware version git-20200229-aea0dc4 reported by another member?
- Herb
|
I have the Nanovna-H V3.2, which finally has a new display and have added diode D2 plus updated the firmware.
Are there any of the previously described component changes/mods that might also be appropriate to this version (I looked for a schematic of the V3.2 but couldn't find one) that should/need to be done?
--
Terry VK5TM
( )
|
In theory (I think), it could be used to calibrate the VNA if you know the exact reflection coefficients of the not so perfect internal open, load and short. Now the question is, can you derive these reflection coefficients from measurements with a known cal kit? My feeling says yes, but I did not spend any time on it trying to do the math.
My SAA2 tracking status is stil "Shanghai, delivered to air transport" for almost 2 weeks now...
Reinier
Op 2-4-2020 om 06:13 schreef Gabriel Tenma White:
toggle quoted message
Show quoted text
Currently the corrections only use OPEN and LOAD ports of the ecal switch; In the code there is an option to do a full SOL ecal, but this is turned off in all firmware versions and is only there for experimentation. However this can not replace the user calibration, as the biggest unknown that has to be corrected for is the SMA connector transition, and this has a big unit variance (I think it's hand soldered). In my experience though if you are only measuring antennas at VHF, you can get away with not calibrating, but you must then treat it as a SNA and not trust any "impedance" values on the screen.
|
Gabriel,
That agrees with what I am seeing. I suppose I should say "self-correction" instead of "self-cal". The self-correction alone does look pretty good for return loss or SWR in the VHF to 500+ MHz range.
--John Gord
toggle quoted message
Show quoted text
On Wed, Apr 1, 2020 at 09:13 PM, Gabriel Tenma White wrote:
Currently the corrections only use OPEN and LOAD ports of the ecal switch; In
the code there is an option to do a full SOL ecal, but this is turned off in
all firmware versions and is only there for experimentation. However this can
not replace the user calibration, as the biggest unknown that has to be
corrected for is the SMA connector transition, and this has a big unit
variance (I think it's hand soldered). In my experience though if you are only
measuring antennas at VHF, you can get away with not calibrating, but you must
then treat it as a SNA and not trust any "impedance" values on the screen.
|
There is no "self cal". There are corrections to track temperature drift, but this is not a full SOL calibration. The corrections are always happening and every sweep some small percentage of the time is allocated to it, except for the first sweep after changing frequency range which performs a full correction sweep, which is why the first sweep is much slower. If you compare the V2 to V1 you should see a big difference in temperature drift.
|
johncharlesgord wrote: I don't remember this being mentioned before: ... When the S-A-A-2 (NanoVNA-V2) frequency is changed, it does a "self cal" using an open, short, and load on its Port 0 input switch... ... I assume that the "raw" data it reports over the USB port actually uses this self-cal. My unit has firmware version git-20200229-aea0dc4.... ================================================================== John, Gabriel never mentioned any "self cal" feature in the S-A-A-2 user manual. It just says the The S-A-A-2 is a single switched receiver VNA. Two channels are selectable by the receive mixer through SW_RECV, and a third channel, the reference channel, is provided by setting SW_ECAL to the ¡°open circuit¡± position. By controlling these two switches the receiver is able to observe reference, reflected, and thru signals. In message, /g/nanovna-users/message/12051 , Gabriel was asked a bout the USB data transfer and she replied. "The USB interface always outputs raw, uncalibrated values, as is in the user manual. The reason is the assumption that the PC software always has superior calibration routines (e.g. supporting arbitrary cal kit parameters) and it is far less error prone if you can't be confused whether you are reading calibrated or uncalibrated values." In my own testing if I change frequency parameters, calibration is turned off and no interpolation or correction is applied. Most of my testing is S21 and after changing frequency, without re-calibrating, the results are terrible. If "self cal" is applied for S11 measurements I don't see the benefit since as you said the measurement plane would be different. I could see such a feature being used as a self-test during manufacture. My S-A-A-2 just arrived last week and has firmware version is git-20200117- 4c067eb. Your version git-20200229-aea0dc4 is newer but I'm not sure what the difference is. The latest release at says that it is the "First production release" so I assume there is no advantage to me updating. - Herb
|
On Wed, Apr 1, 2020 at 04:56 PM, DiSlord wrote:
... Font size not change ...
About points count, i test on my NanoVNA-H variable points set (51 and 101) all worked. For H4 a try complete this then i recive it. Difficult write code and not test it on device. I test some part on my H version, but...
.================================================================
DiSlord,
I 'm happy you have been experimenting with variable points set. I hope after you receive your H4 you will be able to port the experimental variable point code to it.
The font size for the display menus had more density and stood out better on the March 31 beta. That is what I meant by larger font. I could make out the characters at a greater distance than I could on the April 1 beta. Some users might prefer the fonts used for the March 31 display menus.
I think you have been doing a great job with the beta firmware and wish I could buy you a beer.
- Herb
|
How do I remove the screen cover sheet. Do I have to disassemble top cover?
|
dekagon
OwO