For those on this list first venturing into the v.h.f. and u.h.f. regions
and who would look for repeatable results down around minus
50dBm and quibble over a tenth of a dB, BNC is not the way to go. FWIW, calibrating 50kHz to 100MHz with these 34cm SMA-to-BNC pigtails: .. and a generic BNC terminator and female coupler yields lower CH0 LOGMAG when subsequently remeasured (-68 to -81dB) than do included SMA cable and terminator (~ - 63dB) For 500-900MHz CH0 LOGMAG with BNC pigtail remeasured after cal yields -47 to -51 dB while remeasuring SMA pigtail after calibration yields -68 to -73 dB For frequencies of interest >>to me<<, BNC appears better..
|
Hi Bo,
now I've gotten to your long list of ideas! Thanks for providing them!
Screenshot/image saving is definitely interesting! Currently, the
graphs/plots are all painted directly on a Qt widget, so I would have to
find something that could save those. I think it should be possible, though.
Exporting a general frequency/amplitude (or even amplitude/frequency) file
shouldn't be much of an issue. But, maybe it relates to your next
suggestion: The auto markers for filters. I'd *love* to do that! My biggest
issue is finding a way to 1) identify the filter, 2) find out what the
passband is, and what the passband level/ripple is, and then identify the
3dB/10dB etc. points. The latter part, finding the points, is "easy". What
worries me is the first part :-)
The marker architecture is fully set up to allow for more than 3 markers -
but to fit in 1366x768, it's currently limited to 3. I think I'll add it as
a setting so you can have more. :-)
Thicker trace lines should be an easy setting to make.
The cosmetic points are all sensible, and I'll endeavour to have them
included soon! The only exception is the COM port bit: The current
functionality already finds the correct device if possible, so a dropdown
menu would only really be required if you have more than 1 NanoVNA
connected... I don't think I'll present the users with devices that the
software can't identify as compatible, for now at least, but maybe the
connection functionality needs an overhaul at some point.
Segments: This was previously called sweep count, and is really the base
functionality of the software, the sole reason I wrote it: It refers to how
many segments the sweep span is broken into, each of 101 sample points, in
order to increase resolution over the default the NanoVNA provides. So
putting in 10 provides 1010 points, 50 gives 5050 points, etc. The label
next to it presents the distance between each data point, as that's proven
useful for me, at least. Hz/step is my term for the number of Hz between
two data points .. maybe that's just my own terminology bleeding through, I
don't know. I'm open to suggestions :-)
Thanks again for taking the time to write out all these suggestions!
--
Rune / 5Q5R
toggle quoted message
Show quoted text
On Fri, 27 Sep 2019 at 18:47, Bo, OZ2M <groups.io@...> wrote: Hi
I am relatively new to the NanoVNA thus the NanoVNA-Saver but I have spent
a number of hours in front of an 8753C :-) I really like the NanoVNA-Saver
and therefore have some ideas for further improvements. If I didn't like it
I wouldn't even bother to make suggestions :-D
Ideas rel. to 0.0.12
- make it possible to save the full NanoVNA-Saver window and/or each plot
without having to do a screen dump via another tool
- It would be ideal if there is a save dialog with an input field that is
used as preamble in the file name(s) and/or also for caption in the
pictures. In the file name please use underscore instead of spaces
- the dialog box could have check boxes for which pictures to save,
perhaps even a check box for adding a time stamp YYYYMMDD_HHMM to the file
name(s)
- in addition to the S-parameter files also to have a set of files for
plain amplitude/attenuation vs. frequency
- x dB auto markers, e.g. 1 dB, 3 dB, 6 dB, 30 dB, 60 dB is difficult in
general on the NanoVNA, and buttons for finding these points on LPF, HPF,
BPF and Notch fillters
- more than three markers in general
Cosmetics
- the lines on my monitor are rather thin. It might be related to my
screen resolution being 2560 x 1440 pixels on a 32" monitor. But it might
be an idea to allow for thicker traces lines
- reduce the number of decimals, e.g. fF and pH resolutions are not
important
- there is a missing space between the impedance value and the Ohm-symbol,
1234.5ohm-symbol -> 1234.5 ohm-symbol (ISO 80000)
- there is a missing space between the band value and the meter, 2200m ->
2200 m (ISO 80000)
- show the marker color in the caption of the marker value panes
- auto 1000 separators on frequency input boxes
- COM port dropdown box listing COM-ports and the logical names if
available. It will make it easier to find the device. Perhaps even auto
connect if logical name is found
Segments
- I am not sure I understand "segments", however, I do understand points
which might be an old habit. The Hz/step unit of the segments is confusing
to me
Bo
|
Re: nanovna Battery Specifications
Thanks Warren.? Mine has those shields.? I have heard that the shields increase the dynamic range above 300 MHz and especially above 600 MHz.? Is that not true?
Mike WY6K
toggle quoted message
Show quoted text
On Friday, September 27, 2019, 01:29:24 PM CDT, Warren Allgyer <allgyer@...> wrote:
This is the bottom side of the shielded version.? The shields are on the right and cover the RF portions. Many..... most maybe.... do not have these shields. If yours does not, do not fret. They make no significant difference in practical use.
The space on the lower left is where the battery goes, secured by a piece of double sided tape or a dab of contact cement. Battery solder tabs are in the upper left corner.
WA8TOD
|
Bo,?
HOw thick is that FR4 board?? Thinner material would improve the isolation, right?
Mike WY6K
toggle quoted message
Show quoted text
On Friday, September 27, 2019, 02:48:03 PM CDT, Bo, OZ2M <groups.io@...> wrote:
Hi
From a market volume point of view it might be an idea to divide the RF boards into:
1) Si5351A, as now. This is more than fine for the vast majority of radio amateurs
2) Si5351A + ADF4351, lower volume. Covers my frequencies of interest
3) Si5351A + e.g. ADF5355, very low volume and very high cost
They may all share the same digital processor and display boards. Could the digital processor board be a RPi? If so, then this may have a huge impact on the price, and there is a much higher volume to drive the development of this platform. It might be a bit more clunky though.
One thing is loss, but also isolation is an issue. I have a circuit on FR4 on my desk right now, where the isolation above 1,7 GHz cannot go higher than around 55 dB. Heavy shielding may help, but shielding and tooling are cumbersome and expensive.
Bo
|
Re: Si5351A max fundamental frequency
On Fri, Sep 27, 2019 at 10:04 PM, Bo, OZ2M wrote: Should it be an option, in the S/W, to set the max fundamental frequency? I am
not thinking of the 1,5 GHz possibilities as such. There is no such option. You can modify firmware but the logic for frequency boundaries is not easy, needs to deal with this. By the way, edy555 firmware uses different frequency segments boundaries. You can try edy555 firmware, may it will help to solve your issue with no need to change firmware code.
|
Hi John,
if you experience crashes, please try:
- Running the program for a command prompt, which preserved the crash
message,
- Running using "-D logfile.txt" as a parameter, which saves debug data to
a file, which you can then pass to me for further analysis.
It *may* happen if you have not calibrated the NanoVNA itself, and it sends
values that are so unrealistic, the program doesn't know how to parse it.
Generally, measurements indicating severely positive gain (more than 30dB
maybe?) are rejected as corrupted. If the same measurement shows bad values
more than 20 times in a row, the program stops (and reports an error on the
console). It's not supposed to *crash* when it does this, but ... I think
sometimes it does. :-(
--
Rune / 5Q5R
toggle quoted message
Show quoted text
On Fri, 27 Sep 2019 at 21:47, John AE5X <ae5x@...> wrote: Tried it on two different Windows 10 machines - it seems some combinations
of sweep settings cause the program to close suddenly. This occurred on
both PC's.
Calibration was done (400kHz - 5MHz), amateur bands selected to show and
sweep was set to 3.5 - 15 MHz.
|
Re: Experimental 256 point FFT Firmware
qrp.ddc,
I purchase cables of unknown quality at hamfests and the nanoVNA has paid for itself in helping me avoid selecting cables that physically looked fine, but from both cable ends looked like virtual shorts using the nanoVNA TDR function. With your access to better equipment, the nanoVNA's TDR function probably does seem like a toy, but if the user realizes its limitations and doesn't expect laboratory quality from a $50 device, the function is still a worthwhile tool and learning feature. If the TDR function tells me that a reel of cable marked 25 meters is 24.5 meters that's close enough for me to feel confident purchasing it. Carrying a laptop with me to do the same TDR measurements is in-convenient, although at home I do use NanoVNA Saver to perform TDR measurements. With the multiple firmware versions that have appeared on the scene, maybe one of the developers will branch off a version that removes the TDR function and implements the features you would rather have. That's the great thing about open source projects, the source code is open to modification however it suits each of our individual needs.
Herb.
|
Re: Si5351A max fundamental frequency
did you tried original edy555 firmware?
Firmware from hugen79 and from edy555 have different frequency modes. So they have different behavior at different frequency boundaries.
|
Re: Experimental 256 point FFT Firmware
In my previous post, I was referring to coax from my hamshack to the antenna - not a piece of transmission line a foot or 2 long.
Please remember that from a practical viewpoint, LOW resolution in a handheld device is much better that NO resolution (ie: no TDR function).
...Larry
|
Hi
From a market volume point of view it might be an idea to divide the RF boards into:
1) Si5351A, as now. This is more than fine for the vast majority of radio amateurs
2) Si5351A + ADF4351, lower volume. Covers my frequencies of interest
3) Si5351A + e.g. ADF5355, very low volume and very high cost
They may all share the same digital processor and display boards. Could the digital processor board be a RPi? If so, then this may have a huge impact on the price, and there is a much higher volume to drive the development of this platform. It might be a bit more clunky though.
One thing is loss, but also isolation is an issue. I have a circuit on FR4 on my desk right now, where the isolation above 1,7 GHz cannot go higher than around 55 dB. Heavy shielding may help, but shielding and tooling are cumbersome and expensive.
Bo
|
Re: nanovna Battery Specifications
Don't sell yourself short, Frank: we hams can sometimes contribute
on this list. It's a fuzzy faced young engineer or academic who remains confident
that he has nothing to learn from a genuine ham of long and diverse experience. Not all of us are simply warmed over chicken banders. Some of us have ended up embarrassing engineers. Please continue to contribute.
John
at radio station VE7AOV.
On 2019-09-27 5:00 a.m., Frank Dinger , EI7KS wrote: Yes ,indeed a 5V USB powerpack ,readily available ,even from supermarkets , will do fine.
For those needing a Li-Ion cell (nominally 3.7V) , I get these from discarded Laptop battery packs . These packs are discarded because usually only 1 of the 6 cells is no longer OK.
I take these packs to bits and charge each cell individually with an appropriate charger (usually good for charging 4 cells at the time) I then fit cells in 18650 cell holders available from Banggood and other on-line retailers.
There are 2 types of cell holders , for 1 cell and 4 cells.
WIth the cell holders I make multiple cell battery packs to feed my Yaesu FT-817 , QRP transceiver
For this transceiver I use 4 multiple cell units in series ,ensuring that the cells will be re-charged when the radio input voltage has dropped to 4*2.8V=11.2V To ensure a long life a Li-Ion cell should not be discharged below 2.8 V
The above just for info.
After all I am just a radio AMATEUR.
Frank , EI7KS
... --
|
Tried it on two different Windows 10 machines - it seems some combinations of sweep settings cause the program to close suddenly. This occurred on both PC's.
Calibration was done (400kHz - 5MHz), amateur bands selected to show and sweep was set to 3.5 - 15 MHz.
|
Hi Mike, I've heard a number of users mention antivirus problems. I have tried submitting it to VirusTotal, and it seems the "only" things it reports is a set of a few antivirus programs worried about Python programs being trojans. I think maybe it's a case of ophidiophobia. ;-) Thanks for your report! -- Rune / 5Q5R On Fri, 27 Sep 2019 at 18:13, mike watts via Groups.Io <wy6k= [email protected]> wrote: Rune,
The only problem I had was that Norton AV thinks it is malware and
quarentines it.
Mike WY6K
"... somewhere in the distance, there's a tower and a light, broadcastin'
the resistance, through the rain and through the night..."
On Friday, September 27, 2019, 6:05:03 AM CDT, Rune Broberg <
mihtjel@...> wrote:
I just released 0.0.12:
This release of NanoVNA-Saver offers a number of new features, all of which
have been widely requested.
First of all, there's now the option of scaling the plots: right click them
for a menu, where it's possible to set the maximum and minimum values for
the frequency and data axes. Sadly not yet ready for the polar plots, and
mouse control of zoom is also pushed to a future release.
Second, it's not possible to have "bands" displayed in the frequency based
plots: Select "Display setup" to find the option for this. The default data
is for amateur radio bands - or you can put in your own.
Thirdly, the calibration procedure has received a new "calibration
assistant": A series of popup messages prompting you to switch between
calibration standards, and code to automatically sweep them for you and
store the results.
Finally, there's as always a number of bugfixes and stability improvements.
With more than 800 downloads of the 0.0.11 version, I can't wait to hear
what you all think of it this time - and I can't thank this community and
mailing list enough for the support, encouragement, suggestions and testing
you have provided!
I am interested in knowing what platforms you are getting this software
running on, *particularly* if you had had to jump through hoops or do
anything unexpected to get it to run: Do email me at mihtjel@... if
you can tell me what you had to do to make it work, so that I may update
the documentation, or make things easier to install in the future.
Thank you!
--
Rune / 5Q5R
|
Re: Experimental 256 point FFT Firmware
On Fri, Sep 27, 2019 at 10:23 PM, Larry Rothman wrote:
If the Nano TDR shows shows me a reflection a foot or two away from a
connector, don't you think I will look at the connector before anything else?
If I'm looking for discontinuities in the circuit, which total length is 1 foot total and TDR can show it one or two foot away from real point, I think it will be useless. To be more clear what I'm talking about, here is pictures of TDR measurement for the same S1P file captured with NanoVNA, with different FFT size. This TDR implemented on PC side, so it doesn't limited with memory and don't requires TDR support in the firmware and allows to use any size FFT for tests. Actually I capture this S1P with old firmware which doesn't have TDR in the firmware at all.
|
Re: Si5351A max fundamental frequency
This is why hugen came out with his 800MHz version of the firmware.
|
Hi Kurt,
thanks for all your help with the calibration! I'm still working on it, and
as you say, I need some better functions for saving the resulting
calibration, that it may be used later.
You are right that I haven't implemented scaling for the Y-axis of the
phase display: I'm not entirely sure if it makes sense, so I disabled it
for now. If it's requested and wanted, I'll add it in. :-)
The R+jX scaling is getting another look, as it's clearly not entirely
functional at the moment. There's also some rounding taking place in some
of the charts where the software attempts to show "nice" values for the
tick marks. This might be interfering with the user settings. The scaling
is, clearly, a first attempt. :-)
Thanks again for your help, and for your feedback on the software! I hope
it proves useful for you!
--
Rune / 5Q5R
toggle quoted message
Show quoted text
On Fri, 27 Sep 2019 at 17:33, Kurt Poulsen <kurt@...> wrote: Hi Rune
Version 0.0.12 great news.
I have tested your calibration using the Delay and L/C coefficients
together with my HP83033C female calibration standards and tested the
calculated data for equivalent L and C and at 100MHz they are OK.
The scaling of the Data Axis is super but some problem still exist. The
R+jx is if setting to + - 11 is showing 10. I suppose it will come but the
Dara Axis for any of the phase displays are greyed out. That was what I
stumbled over.
I did a T-Check manipulation where I copied the S11 to S22 and S21 to S12
after I subtracted the Thru adaptor delay (42ps) from S21/S11. Result
attached but is not nice simply because there is no 12term error correction
performed. Adding a 10 or 20dB attenuator in front of Port1 would help as
linearizing the port1 input impedance to be closer to 50ohm
Keep up the good work very much appreciated.
Kind regards
Kurt
-----Oprindelig meddelelse-----
Fra: [email protected] <[email protected]> P? vegne af Rune
Broberg
Sendt: 27. september 2019 13:05
Til: [email protected]
Emne: [nanovna-users] NanoVNA-Saver 0.0.12
I just released 0.0.12:
This release of NanoVNA-Saver offers a number of new features, all of
which have been widely requested.
First of all, there's now the option of scaling the plots: right click
them for a menu, where it's possible to set the maximum and minimum values
for the frequency and data axes. Sadly not yet ready for the polar plots,
and mouse control of zoom is also pushed to a future release.
Second, it's not possible to have "bands" displayed in the frequency based
plots: Select "Display setup" to find the option for this. The default
data is for amateur radio bands - or you can put in your own.
Thirdly, the calibration procedure has received a new "calibration
assistant": A series of popup messages prompting you to switch between
calibration standards, and code to automatically sweep them for you and
store the results.
Finally, there's as always a number of bugfixes and stability improvements.
With more than 800 downloads of the 0.0.11 version, I can't wait to hear
what you all think of it this time - and I can't thank this community and
mailing list enough for the support, encouragement, suggestions and testing
you have provided!
I am interested in knowing what platforms you are getting this software
running on, *particularly* if you had had to jump through hoops or do
anything unexpected to get it to run: Do email me at mihtjel@... if
you can tell me what you had to do to make it work, so that I may update
the documentation, or make things easier to install in the future.
Thank you!
--
Rune / 5Q5R
|
Re: Experimental 256 point FFT Firmware
OK, no problem then.
However, please remember - considering the cost of the NanoVNA - it puts more capability in your shirt pocket than most of its predecessors used by Amateurs.
As I mentioned, it takes no effort to flash different versions of firmware onto the device. We aren't doing microwave design were pico-seconds matter.
And you can take it up an antenna tower - without a PC.
If I need to check out a length of cable from my Tx to the antenna, it is more than adequate to show the approximate location of any issues - and that includes connectors.
If the Nano TDR shows shows me a reflection a foot or two away from a connector, don't you think I will look at the connector before anything else?
It is good enough to play with and get a feel for how it works and that in my opinion is not a waste of time.
Others are using the Nano to learn from - from both a programming and an RF point of view.
There has been some great dialogue in the forum regarding TDR and RF theory in general.
You are sharing your TDR knowledge as well - that's great!
I've worked in the RF field (no pun intended) for 40 years and I'm still learning.
All I am saying here is that the various diagnostic functions that are being graciously developed for the Nano by very talented individuals, may not be for everyone but, they are NOT a waste of time for everyone.
Cheers,
Larry
toggle quoted message
Show quoted text
On Fri, Sep 27, 2019 at 02:21 PM, <qrp.ddc@...> wrote: There is no question if TDR needed or not. It is definitely must have. The
question if it worth to make measurement more worse, cut-off useful
measurements just to add low-res TDR on firmware side? Or if it's better to
make NanoVNA more precise, more stable and more useful with more measurements,
but with TDR implementation on PC software side (with much better resolution
and more usable)?
|
Hej Bo,
I'll address the port numbering first: Port 0 and Port 1 should probably be
"Ch0" and "Ch1", and are taken from the port numbering on the physical
device:
While it's quite confusing to have S11 be reflection on port 0, and S22 be
reflection on port 1, that's a physical decision design ... ;-)
I'll address your other comments (thank you for them!) as I go through the
thread.
--
Rune / 5Q5R
toggle quoted message
Show quoted text
On Fri, 27 Sep 2019 at 20:34, Bo, OZ2M <groups.io@...> wrote: One more thing.
It might be an idea to change the name of Port 0 to Port 1 and Port 1 to
Port 2 in the S/W GUI so it matches S1x and S2x etc. instead of e.g. S10 ...
I acknowledge that in programming the starting point if often zero (0) and
not one (1). But from an RF point of view it will probably be good karma.
Bo
|
Dr. David Kirkby from Kirkby Microwave Ltd
LOn Fri, 27 Sep 2019 at 19:34, Bo, OZ2M <groups.io@...> wrote: One more thing.
It might be an idea to change the name of Port 0 to Port 1 and Port 1 to
Port 2 in the S/W GUI so it matches S1x and S2x etc. instead of e.g. S10 ...
I acknowledge that in programming the starting point if often zero (0) and
not one (1). But from an RF point of view it will probably be good karma.
Bo
I agree 100% with you. Dave -- Dr. David Kirkby, Kirkby Microwave Ltd, drkirkby@... Telephone 01621-680100./ +44 1621 680100 Registered in England & Wales, company number 08914892. Registered office: Stokes Hall Lodge, Burnham Rd, Althorne, Chelmsford, Essex, CM3 6DT, United Kingdom
|
Si5351A max fundamental frequency
Hi group
I have measured a low pass filter, but experience some, for me expected, oddities around the maximum usable fundamental frequency of the Si5351A. Please see the attached picture that clearly shows an abrupt change in the attenuation between ~300 MHz and ~303 MHz. The NanoVNA should be able to run in 3 x 300 MHz segments, i.e. fundamental plus second and third harmonic. So far so good. However, my experience in working with the Si5351A, and the RFzero project in particular, is that the fundamental frequency from the Si5351A can be used up 280 MHz with 100% certainty, and up to somewhere below 300 MHz depending on the actual device. But once the frequency increases, then the PLL will suddenly not lock, the frequency is of course unstable and the S/N is VERY poor.
The PLL in the Si5351A I have on my RFzero right now can lock at 299,6 MHz, but not at 299,7 MHz. But I have seen Si5351As that cannot lock above ~287 MHz. I don't think it is possible to bet on the max freq. even vs batch.
So why can the NanoVNA say it has 900 MHz, 3 x 300 MHz, usable bandwidth and not e.g. 840 MHz, 3 x 280 MHz, or use the fourth/fifth harmonics? I am a bit troubled with measurements in the 280 MHz - 300 MHz, 560 MHz - 600 MHz and 840 MHz - 900 MHz segments. Should it be an option, in the S/W, to set the max fundamental frequency? I am not thinking of the 1,5 GHz possibilities as such.
Bo
|
Oristo
WY6K Mike Watts
Bo, OZ2M
