These are chat archives for Makuna/NeoPixelBus

16th
Aug 2015
sticilface
@sticilface
Aug 16 2015 00:07
cheers
sticilface
@sticilface
Aug 16 2015 11:58
@makuna ... doing a lot more testing.. have 6 eps all running the same animation and the same binary...
i get significantly more resets... like significantly more when I have the mDNS included... I'm not sure it is related to the interrupts now...
do you use mDNS in any of your sketches...?
The example sketches for mDNS do not include an update() call in the loop, but in the text for the mDNS.cpp it says to include that... if i put MDNS.update() in my loop my ESPs go bananas with lmac 622.... rebooting almost instantly...
sticilface
@sticilface
Aug 16 2015 12:16
although.. i recompliled with debugging for the mdns, and its fine! i'm confused
sticilface
@sticilface
Aug 16 2015 12:24
i might have found something... i think multiple arduinos running the OTA interact...
Michael Miller
@Makuna
Aug 16 2015 15:08
That will be annoying to confirm, you would need to remove Ota from sketch and try them all.
sticilface
@sticilface
Aug 16 2015 15:10
i tracked it down more... i enabled the mDNS debug and it was going mental.. spewing out received requests, and replies... too fast for the serial to keep up with.. it all stopped when i shut the IDE down.
So for some reason the IDE was requesting mdns too fast for the ESP to keep up... since i've restarted it the eps have been stable for hours.. vs... max 10min... next time i get rebooting i will look and see if it is the IDE causing it... an interesting find none the less
sticilface
@sticilface
Aug 16 2015 22:24
OK.... I'm getting somewhere but its not too good news. in a nut shell having nointurupts in neopixelbus... even just xt_rsil(1); is enough to cause the lmac error. it is made significantly worse by having the mdns running at the same time. if i remove the no inturrupts, the LEDs work but very poorly lots of flashing, but no lmac after 2 hours. with xt_rsil(1); + mDNS it lasts on average 5-10min. Without the mDNS it lasts 1-2 hours... guess this is going to require some lower level workarounds....
I should add that this is whilst running fairly heavy animations. What was driving me nuts was that it always involved my heavy animations, with lots of computing...now however, they are stable as... just with the interrupts disabled...
sticilface
@sticilface
Aug 16 2015 22:40
(What i meant by that, was that it was due to some coding error or bug on my part, that was causing the resets) I'm now fairly confident it is due to the no interrupts...