SOLVED - Nano Biscotte V4 randomly reboot

[penpen]

Hi !

i've just bought a Nano biscotte v4 and i'm working on wiring but found some weird issue

I power up the board, have boot sound, i push "act" momentary switch, sounds are ok, light is ok, "humming" sound is idling/cycling.

If i dont move, every ok, but if i move the board, i've sound/clash/etc.. then randomly, all power off followed by the boot sound 

I've reduced my wiring to the minimum: speaker and battery only or led and battery only, nothing change, when the board move, it randomly poweroff/reboot

I also tryed others combinaisons of led/resistor, i've always the same issue: board seems ok if still, but when move and activate somes clash/moving sounds, it reboots, sometime immediatly, sometimes after several seconds :/

I don't think my soldering/wiring is too messy (to be sure, i've resoldered one time and scratch/clean a little between each pad to clean)
I've used magnification to look for short-circuits, but nothing...

The speaker is 2W 8ohm, (i've tryed several other speaker, still around 2W/8ohm)
High power leds are drived by 0.22/1 ohm 2W resistor (i've tryed with several led, low power too)
Also tryed with several battery: 18650 lion full charged, high discharge capable lipo 1s correctly charged, with and without pcb circuit

SMCs components arent specialy 'hot' touching them after reboot issue

I don't know what to do 

[penpen]

pics of board

[erv]

weak or uncharged battery. Also the SD card might not be seated well in the socket.

please share a video eventually, thanks

[penpen]

that's a little crapy movie, trying to make a better one tomorrow

cut VID 20180204 222247 - YouTube

Using only a speaker, no led, and i've put some "voltmeter" aside, showing ~steady 4.10v using 1S lipo 1000mah 25C

we can hear the sound when moving and then a blank following by the "roar" boot sound ( it sounds like a reboot to me)


I've try more pressure to "secure" SD slot but it doesn't change anything :/

[erv]

you might want to re install the default package after formatting, just to be sure it's not a sound fetching issue.
The blank indicates the board is looking for the sound to play and cannot access the SD card, then the board gives up after some time and reboots.

that can be the sign of a SD card socket problem (soldering or mechanical). Where did you purchase it and where are you located please
thanks

[penpen]

thanks, i'll try with a new SD and a fresh install this evening

( i'm working in Paris, bought it from slimsaber )

[erv]

Noted thank you. Please note that slimsabers is not an authorised dealer / retailer. I work in paris too (4ème)

[penpen]

xxx, sorry, it was out of stock elsewhere 

[penpen]

So, i've try with an other SD card and fresh install and...... it's working now ^^'   
The faulty one was a 4Gb sandisk class4 (tryed with fat32 and 16), now, with a 16go samsung class10 (haven't smaller right now) in fat32 it works

many thanks for support ! (and sorry i didn't bought it from authorized dealer/reseller, i wasn't very aware of it  )

i'm still curious in which way the sdcard can influence sound played by the chip to perform reset. I suppose mcu's ram is too small to enable in memory fetching of sound banks, so it access sd FS each time a wav must be read ? may be it would be possible to perform some kind of checksum at boot time ?

[erv]

thanks a lot for trying this and I'm glad you got it resolved. Would you mind sending the SD card back to me (I will pay for shipping) so that I can test it and diagnose it ?

the card performance is indeed important as you pointed. But we've got much slower card in the past and it was working ok. In your case it's symptomatic of a card FLASH / NAND structure issue. In the batches we buy we get some of them which are not working properly but we generally diagnose them while copying the contents on them (I built my own card replicator that diagnoses the writing speed and writing error). So I'm curious about this one.
Among the ones I trashed, most of them were incredibly slow, like if they had a degraded flash structure inside. The default package would copy on the card in about 2-3 seconds, those were taking 30 or more, no matter how many times I formatted them. Same brand, same batch.

I will replace the card by a new (and tested) one if applicable :-)

[penpen]

no problemo, PM me your address and i'll send the sdcard to you, that's the least i can do to help improving ;)

[erv]

PM incoming

[erv]

hi Pen, received your SD card yesterday, testing this morning indeed shows a reboot after attempting to play some (most) of the sounds. Sounds play ok on the card on a PC but as suspected the card formatting (as you tested on your side) is painfully slow which means the card is totally out of spec (broken NAND flash, or bad flash controller).
That's the 14th one I diagnosed (most of them during card replication) on a batch of 2000 so it's not "bad" per say but still, I don't know what is wrong with those.

I'm shipping 2 replacement one which are 1GB only but good quality (new batch) and tested on a NB to ensure it plays nice.

[penpen]

Thanks for the explanation !

Indeed the formating process was very slow.

That's funny btw that a computer still access corretly to all data. Is it the effects of a lighter/embedded optimisation of the implementation handling the filesystem ? (it is the IT guy in me who's speaking ^^)

Thanks you very much anyway ;)

[erv]

The card performed like not even a class 1 so its so slow it cannot be read in real time when accessing the sound.
It's not a problem of accessing data the card was not corrected it's a matter of performance of the card. Even the 256 MB I was using 10 years ago are working faster.