https://github.com/mihtjel/nanovna-saver/releases/tag/v0.2.0 Zoom, zoom, zoom! New feature: Zooming This release adds the ability to "zoom" on the frequency-based charts: Use your mouse's scroll wheel to zoom in and out - or hold the control button and drag a box on the chart to zoom into. New chart type: Group Delay Group delay can now be displayed as a chart, and can also be enabled for the marker data display for both S11 and S21. New chart type: Permeability A new chart, R/ω & X/ω, is available for determining permeability characteristics of toroids. Thanks to David, F4HTQ, for requesting and helping out with this. Logarithmic Y-scale: The VSWR and Permeability charts now have the option of having the Y-scale shown as Linear or Logarithmic. Changes to markers: Now able to display Group Delay for S11 and S21 Markers can be filled or "hollow" Markers can be shown centered, or with the point at the tip Markers can be locked from mouse control by selecting the radio button next to "Show/Hide data" Time-delay reflectometry: Min/max impedance values can be manually set There's a movable marker to measure distance in the TDR chart New scaling function for values: Holger Müller, DG5DBH, contributed new code to show values more correctly, using SI-prefixes. David Hunt provided further changes to avoid losing precision when displaying frequencies for sweep settings. Also included: Several bug fixes! As ever, I look forward to hearing about any feedback, suggestions and bugs!
While the ebuild is opened, could you please improve the DESCRIPTION? I had no clue what NanoVNA is. But the screenshots look very promising. "GUI to record data with a NanoVNA vector network analyzer"
NanoVNA is an gadget/hardware tool to analyze electronic networks at radio frequencies. It's a hardware device. I suspect the best analog to the electronics-audio world is a distortion analyzer perhaps, but that's not a great analogy either. nanovna-saver was originally a software tool to pull data and save data from the device. However it now also can control the NanoVNA device through USB to do additional functions the hardware tool can't do on its own (yet), due to computational limits of the microcontroller on the device. I'm still waiting on receiving my NanoVNA, sorry for adding a comment, but was having problems merging nanovna-saver-0.1.5 as it refuses to install with any setting of python_targets_python3_7 for some reason...
Found an additional issue that should be addressed in the ebuild: 0.1.5 does not seem to work as is - it merges but does not run. Ebuild needs dependency on sci-libs/scipy - and dev-python/numpy would then be redundant as scipy needs it too. sci-libs/scipy has its own merge issues for the current stable package's dependencies, I think this may need its own bug report so I'll not add details here. After merging sci-libs/scipy , NanoVNAsaver will pull up its GUI. I still do not have my hands on my NanoVNA yet so i can't fully test it, it's still in the mail somewhere. I sure hope the distributor wasn't lying that it was originally less than 500 miles away (instead of 8000 miles...) I also can't wait until python3_7 is stable on Gentoo (though it may be worth it to see why python3_6 will choke on it.)
Created attachment 600308 [details] Tested ebuild for nanovna-saver-0.2.0 slightly modified ebuild from 0.1.5 that will work for 0.2.0 (and 0.1.5). SciPy requires NumPy to be built with [lapack], but portage will give the correct diagnostic. 0.2.1 which has also been released needs more work with the ebuild...
Note: I still have not received the hardware yet to fully test functionality so it should still be ~. As for the 0.2.1 release not working with the ebuild, the diagnostic it gives implies test mode was being installed and portage blocked it, but this is beyond what I know how to handle at this point. On a soapbox here: I would think such tools like nanovna-saver should be in sci-electronics/nanovna-saver instead of net-wireless - though it is strongly related to wireless communication, its main use is for antennas and general wireless communications electronics, applicable equally to voice radios (two way and broadcast) and television. It's not necessarily used for computer networking - its low bandwidth (most versions top out at 900MHz) is insufficient for 802.11b/g/... analysis. Also note that a new version of the NanoVNA hardware that supports in excess of 3GHz is imminent. A similar tool will be required, and hope that the same software tool can be used with either hardware version...
Tested 0.2.0 with the actual hardware device seems to work okay. The software is a bit clunky but that's what's to be expected with a 0.2.0 version. The only additional comment I'd have is that it would be nice to add a udev rule for this specific usb device to let consolekit/elogind/systemd console users have access to this device. It is a standard cdc-acm interface, but unfortunately with a vid/pid of 0483:5740 there's no way to tell that one of these devices are actually plugged in versus some other device having the same ID - this is a generic ID for any of these STM devices using the virtual serial port in ChibiOS. I will need to flash upgrade my unit, perhaps newer ones have a unique identifier for the device.
Created attachment 602520 [details] ebuild for nanovna-saver-0.2.1 I got nanovna-saver 0.2.1 to build/install by telling the ebuild to zap the test stuff that's in the tarball. Seems to work fine with my NanoVNA (using NanoVNA-H firmware build, 0.4.0-3-ge04d3af (November 25, 2019 - 16:22:14)).
v0.2.2 of nanovna-saver just got released. I tested it on x86, works with the 0.2.1 ebuild (it has the same test software in the tarball.) So the main things that need to be changed: - SciPy as a dependency instead of (or in addition to) NumPy - Remove the test stuff from the tarball. I have to say that nanovna-saver does not work well with a low resolution display (1024x600) as one needs to scroll around a lot. I think the author intended some sort of 1280x960-ish display. Alas, it's what it is.
Currently, in portage, version 0.2.2 is available. This is currently the latest version, marking this bug report obsolete. Closing.
Looks like the new ebuild still does not depend on SciPy, I'll submit another bug when I run across it again (i.e. try to install on another machine that does not have SciPy already installed...)
