Spelling the name of the file wrong doesn't suggest a problem with numpy,
no ;-) I don't know exactly what Linux version you are using, but the
instructions given normally work, and will install any required
dependencies via pip.
I hope you manage to get it to work in any case :-)
--
Rune / 5Q5R
toggle quoted message
Show quoted text
On Sat, 26 Oct 2019 at 16:24, <subs@...> wrote:
"I have different problems with every version tried.
For using nanoVNA standalone,
I install Hugen's most recent AA dfu version from here:
When I fist saw that _AA_ version I thought that a 274kb file would not
fit on a 128kb device! I am familiar with Arduino process where we upload
.hex files with an AVR programmer so was drawn to that file first. Looking
at the manual section 8, I see that I need dfu-utils, which is available in
Ubuntu so I can download that and use the .dfu file when a new update is
released. For now, since my device has been supplied with the most recent
version, I see no reason to change the firmware yet. However, what is the
difference between the _AA_version and the non-_AA_ version?
"the file name is "nanovna-saver.py". So the error you got seems to have
been because you tried to run a file that did not exist?"
Yes, absolutely, which suggests that the first time with the python-numpy
package installed, the installer did not complete correctly. The second
time with python3-numpy installed evidently it did. But that didn't get rid
of the "bdist_wheel" problem.
Sorry, I had listed all of the commands in my post but a copy and paste
error meant that I had to re-type everything in haste. I basically followed
the readme exactly. Here are the steps:
- unpack the .tar.gz file into a directory
- enter the directory
- check all dependencies are present
$ sudo apt install python3.7 python3-pip
$ python3.7 -m pip install .
Once completed:
$ python3.7 nanovna-saver.py
These were present:
* `python3-serial`
* `python3-pyqt5`
This one does not exist:
* `numpy`
However, python-numpy exists, and I installed python3-numpy for good
measure.
"I would also suggest you are too close to the antenna you are trying to
test. With a 20cm cable your body is very close. If you are also testing
indoors it can also scew your results."
The guy in the video connected the antenna directly to the SMA connector
and had his hands onver the antenna as well as the nanoVNA yet (apparently)
had no problems. Maybe some creative editing? I am testing indoors and both
antenna and nanoVNA are close to me. I found that if I take note of where
the peak is when my hand is away from the unit and then move the marker to
approximately that point on the screen and then move my hand away to see
here it landed, then eventually with two or three more corrections it I can
get the marker pretty close to spot on.
I had a closer loor at the unit today. Its seems that the front and back
are not made of metal as I had thought, but of thin PCB material. The
corners have plated through holes, but there doesn't seem to be any metal
layer across the board, so reaaly, there is no outer shielding. The 4 screw
mounts are not connected to the ground plane either so any shielding would
not have been electically connected anyway. Is this intended by design? I
notice that the more expensive version in the black box appears to come in
a plastic case?
