OTA update problem

[torwag]

Hi,

I tried to update my wipy board to release 1.1
First downloaded the binary.zip file form the release page
https://github.com/wipy/wipy/releases/tag/v1.1.0
Unpacked it and moved mcuimg.bin to /flash/sys/ via an ftp client (filezilla to keep things in sync with the doc).
First of all, the file is visible as long as you do not change the folder. However, that could be a caching thing of the ftp-client. Switching to another folder and back... its gone as described in the doc.
Resetting the wipy board and checking for the current version via a telenet connection
>> import os
>> os.uname()
I still end up with release 1.0.0.
That was to be expected since micropython welcome message refer to v1.4.6 where it should be 1.5 according to the release notes.

I tried to compile my own version from the source, however I noticed the source code zip-package from the release page does not contain the micropython source code.
Using my local git repro I could build the wipy firmware but it did not work either.

Any ideas what is going wrong?

Two side notes:
It seems my boot.py is not read during boot-up. I added to code to access the REPL via UART but it did not work. If I enter the code manually it works.


Within the docs for the UART-REPLY Echo there is a small type in the code part... os..... damn cancel that it was fixed as I was writing this...

[danicampora]

Seems like your WiPy is always safe booting. Do you have anything connected to GP28? If not, check the soldering, maybe there's a short...

[torwag]

Thanks for the quick reply.
Yep, that seems to be the problem. I just recognized the different blinking periods during boot.
The problem seems to be Pin 28. Reading it on REPL it is high. Albeit nothing is connected to the expansion board or wipy itself.
I tried to set it to low but it did not work. I could not see any short by naked eye. I would need to check for shorts with a meter tomorrow.
However, I even bridged Pin 28 and GND to pull it low. But even then it seems not to work (still safe booting after reset).
I frighten the wipy board is kind of defect.

Thanks for your help!.

BTW. I would like to congrats you and your team for the achievement to get the wipy-boards out. Its great to see a successful project like this.
.

[danicampora]

Hi torwag!

Thanks , and sorry to hear you are having those issues. It's really strange, because after flashing every WiPy we run this test:

https://github.com/micropython/micropyt ... s/smoke.py

Which also checks for short circuits.

Please try to see once more what could be the problem, and if it persists, just let me know and we'll ship another WiPy to you.

Cheers,
Daniel

[torwag]

Hey Daniel,
seems I face the same problem as some others here. I am not aware that I displaced the wipy board. But the symtoms are very much the same as those who added the Vin to the Safeboot pin.
Since I gave the board to some of my students to tinker with it for a day, I cannot say for sure that the failure lays not on our side and might be a PEBKAC.
Any way to rescue the board?

Anyhow I assume that this could become a bigger problem. The wipy and the extension board are not protected against accidentally wrong assembly. Esp. bad luck that it hits the safeboot pin. Maybe for a second generation it might be good to make sure that Vin can't destroy anything in case of wrong assembly.

All the best

Torsten

[andrew]

I wondered about some kind of physical keying that stops you plugging them together wrongly. Or maybe just an outline of the WiPy on the expansion board to show how it goes.

Andrew

[danicampora]

Hi Torsten,

If you are lucky your board will recover itself as some others have experienced. Well, it is indeed back luck that the safe boot pin ends up in the 5V position and hence that one it is destroyed when you plug it upside down. I think the only way to really protect against it is by making the board asymmetric so that it can only be plugged in in one way, but then again people can still make mistakes when wiring their own components or when having the WiPy in a breadboard. The expansion board has labels that must match the pin numbers on the WiPy, and that should be enough to indicate the correct orientation, however, it seems that this is not enough.

Still, point taken to make the next revision more robust against incorrect wiring (which is not easy).

Cheers,
Daniel

[Roberthh]

Simply printing the outline of the WiPY board on the expansion board could help to attach WiPy in the proper orientation.

[Sebastian]

Is it perhaps possible to change the SafeBoot Pin to another GPIO?

[pythoncoder]

+1 for @roberthh's solution. I reckon that would fix the problem.