10 Oaks updates: 8 success and 2 not communicating

[driffster]

Update 2: Now 8 are working The one that failed to register on particle worked once I manually put its ID on the website, the other just worked after retrying a few other time the update (always at max speed).
The 2 failed ones simply refuse to communicate on a protected wifi (even once updated) and wont register on particle on a guess network, (see my other thread for all what I tried: http://digistump.com/board/index.php/topic,2050.0.html)

Update: Updated for my second round of test: 6 success; 1 success but failed to register to particle, one  bugged update (fast flashing) and the same 2 not responsive.

This time I used My Serial to USB cable (to feed 5 volt using the pins (didn't have my sole micro USB around). Is the quality of the cable/power that might be in play here? 
 
Old message:
Decided to try to update all my Oaks using v6 updater (aside from 2 boards with headers, all the other are fresh with no header soldered yet). 7 of them were untouched Oaks.

I just had an updated modem/wireless router from my telephone company so I needed to redo all of them.

I had 3 that updated on the first pass. 5 others shown activity but failed on the first and other subsequent passes (in general I did 2 each). I didn't try all of them as the delay of the new slower mode was more of annoyance when it is clear that the communication had broken down. These 5 I should be able to connect eventually, as they all show up on my router, I just don't have the time today to try further.

I got 2 boards that shown no connection activity (never registered on my router despite multiple tries). This includes the board that I had problem in my previous thread. (http://digistump.com/board/index.php/topic,2050.0.html)

Not a great start.. I will try changing router settings as suggested, but I wanted to see if a default setting would work.

One big suggestion for the updater is to put an option to see your password (since  I would doubt that might be the reason I failed to connect on these 2 boards.. but after doing 8 others pretty sure it is not the password).

[jwhendy]

I couldn't ever get my router to work and ended up using my phone as a hotspot as suggested elsewhere. Could you try that?

I found it a little hokey, as my hotspot triggers wifi to turn off. Thus I had to do the following:
- pre-setup the phone hotspot to have no password
- turn off hotspot, wireless off, phone internet on
- go to the softap site
- turn on wifi, (making sure phone data isn't active) and connect to the Oak
- enter the name of my phone AP (AndroidAP is my default which I've never changed)
- hit "Next" which pushes to the Oak, then quickly go into settings and kill wifi/turn on the hotspot

That was the only way that worked for both of my oaks.

[driffster]

Well despite my multiple oaks I don't have a cellphone. Problem is even if it does work, I need to be able to connect to them via my protected wifi router after the update if I want to use them for anything they are designed for.

Hopefully only 2 seems not to be able to log on at all to my network, all the other seems to works, they just didn't update well. Probably retrying a few times will make them work, but this was just a fast test, since I don't have the time to repeat multiple update attempts (or solder all these headers). I tested the first one that I had difficulty quite a lot, but after a week-end spent only on testing that one unit and not being ever able to login I had quite my fill on testing

Today I retried all boards (using my USB to serial adapter for power), 4 updated after one or two attempts, one which failed the update but now flash rapidly and one that was successful but failed to register to particle. The 2 that refused to connect still refuse.

Is the quality of power (both max current and voltage) can play a part in failed updates? I been using the same cable for all my boards (except when using the serial adapter) and the same USB port of my desktop..

[PeterF]

I would think that as long as your power supply actually does put out 5v, and at least 500ma (preferably 1A capable), and your USB lead isn't a rediciously long one - say a 30cm cable (up to 1M), you should be fine. Some people find the flat micro usb cables better than the normal round ones for power (as have I)... which is strange as the conductors should be smaller... but I digress

Have you seen the alternative update process where you can run up your own update server? If not, give that a whirl, and see how you go. That fixed up my other oaks that were having the usual update socket errors - they both updated first time.

Pete

[jwhendy]

@driffster: I used the phone just for updating firmware. After that, I went back to the SoftAP site which let me select an alternative wireless network. I used my home router and the oak stayed happily connected for a long time after that. Unfortunately, I have another post in which I conclude this "success" story with the fact that mine no longer stay connected. I power cycle, see them on Particle for a few minutes, and then they're gone and don't come back. Not sure!

I've seen other suggestions on router settings: B-mode, no security (at least for the flashing), and not channel 1. Just a few ideas to toss out there. Sorry I can't help more!

[driffster]

Thanks for the hints!

I am now at 8 registered on particle!!    One of which I had to manually ad it on the particle web page, the other one just needed more tries.

For the 2 more problematic ones that failed to communicate to my router, I retried again using the open guess network.
Both were able to connect to that and updated to v6, but they both failed to show on Particle (despite being connected on the guess network). I can't log on them again until I disabled the guess network. Once I did that and changed the wifi settings on them to connect on the protected network they both fail to connect to it (never see them on the wifi router).
This is like I already reported, but note that I actually changed router in between those tests and I have 2 Oaks that show the same behavior.

@pfeerick: Ok.. seems unlikely its the cable fault.. just tried  with a raspberry pi power supply cable rated for 2 A so that should rule out any current problem.

I didn't try running an update server and I am not desperate enough to try, wth 8 working Oaks, I rather start playing around with them

[jwhendy]

Yay! Glad you had success

Regarding switching networks, you can unplug the oak, jump pin 1 to GND, plug it in, and then remove the jumper. This returns the oak to wifi-scanning mode. If you go back to the rawgit SoftAP site, it will take you to the router configuration process. I believe it detects that you're on the most up to date firmware, as mine never tries to re-flash it. You just get the network setup bit.

Out of curiosity, could you let me know which of the "tips" you think helped? I'm curious myself, and it might benefit others who stumble here! I'm guessing you mean one of these?

I've seen other suggestions on router settings: B-mode, no security (at least for the flashing), and not channel 1.

[driffster]

Honestly hard to say if anything helped at all, USB cable might be an issue, but got no real proof on that.

Unprotected network usually is failsafe to update (can still need a few tries) but for me at least this didn't save the Oaks that still refused to connect to particle after the update. I personally think these 2 Oaks might have a defect of some kind, since they really didn't behave like the others.

In general though, if you try and it starts flashing rapidly (showing that it is downloading) even for a few seconds, your Oak should be able to update eventually, it is a matter of retrying.

Also I personally never had one that failed on the first time work on the slowest download mode. I tried in the beginning but it was just too slow (with 10 to do I was not in the mood to retry 3 time at  slow speed..). What work was just power cycle the Oak, and choose change setting (and just keep it the way it was) to avoid going into that mode, after 2 to 5 tries it usually worked.

Also problem might be due to internet connection itself, so retrying later at a different time of day is a good idea instead of just trying 20 times in a row and getting frustrated in the process.







 

I can tell
unprotec

[PeterF]

That's why the little update server thingy that jldeon made which I linked to earlier is so handy - it's pretty much an unzip and run affair, and then you just use a custom version of SoftAP (unless Erik updated the offical one) to choose to update from your own server, instead of the digistump official one. Knocked my two last two virgin oaks into shape first attempt, instead of serveral retries on my normal wifi, or via a hotspot.

As long has you have a few the play with, that's all the matters! I have a couple that have been powered 24x7, and running particle publish or subscribe code, and working great. Looking forwards to seeing what other projects people use them for.