loopa mess

0a082a70ce9c84e2

hi There,

a long story short: with a few simple build projects and repairs behind me, i was brave enough to start building loopa for myself. i enjoyed the process indeed, but at the end it did not passed the OLED, DIN and DOUT tests.

i contacted Peter, than trusted builder Michael, and after a few months and the addititonal expense of €320 i received the assambled working version!

then i started using it and after a few rehersals the screen went blank. Michael suggested that i should open a topic here, so i may get help. i disassambled the unit, took pictures of the boards. you can also find a video of the problem.

i am looking for a solution, please advice!

thanks,
Imre

e8abac78cdf65c56

hei imre. pictures are missing.

0a082a70ce9c84e2

Phatline this is the video of the malfunction

please let me know if you need more infos

f8d38342ee587dc8

The OLED header pin soldering looks a little suspicious - some may possibly be dry - two in particular (which may explain why it was working for a short while). It may be worth reflowing them (with some flux) - just as a suggestion.

0fc5608d995b44ff

forestcaver yup, looks like there’s possibly a dry solder joint on the display headers.

e8abac78cdf65c56

forestcaver you mean the J15C socket on the main-pcb?
or on the oled-pcb itself? or on the UI-PCB where the oledscreen is soldered in?

f8d38342ee587dc8

Phatline The soldering on the loopa pcb where the OLED headers are soldered (not the headers on the OLED itself)

e8abac78cdf65c56

Reflow : take some solder and reflow the solderpoint - so it hase a good connection again…
You need more time and heat for Ground-Pins - since they are mostly connected to heat-absorbing ground-planes…

Measure : on the Bottum side off the Core-Main-PCB at J15 to the pins on the display itself (with assambled PCB Sandwitch off course)

see also this: display-only-show-up-after-softreset

pinout from J15 to the OLED-Pins

PS: DONT OVERHEAT THE OLEDS PINS!!! I had a defect OLED-Screen in the past - because off that.

e8abac78cdf65c56

i got this device today for repair - will not charge money for this debugging… but i fear it will cost a lot off material again…

while all is working in testmode (sd-card removed) it seems that some system or disk operations in normal loopa mode lead to a reset… no problem with the screen at all…

i openened the loopa, removed the UI-PCBs, so i had the main pcb in front off me.
i noticed that the backside was hot…

By touching the STM32F407 itself it gets “non touchable” hot in 2secs after plug in the Power (dont matter in bootload mode or in normal loopa mode).

So i remove all SOICs from Main PCB > changed nothing.
Checked all Capacitors, Diodes, Resistors and remove the ESD-Diode - changed nothing.

then desoldered all “jumpers” on the Waveshareboard and Set the 5V Switch to USB and connected directly a USB to Waveshareboard > changed nothing > extremly hot.

So i removed Waveshareboard from the Main PCB which destroyed again the Main PCB - these Holes for the Pin headers are way to small in order to suck all the Solder from it! Tryd it with tons off flux, with PB-Solder with hot air with Pump with Copper-Nets, with big and small Tips and so on > in anycase ( i had 3 cases in the past) i want to remove a potentially Dead Microcontroller - i have to buy a new Main-PCB - - and that there is not enough space -even with small formfactor Pinheaders (from the SEQV4+ set) - to solder in a Female Pin header in the main PCB - because off Height > sucks! for me i have to think: should I ever solder these loopers for other people again - when the maintainance sucks in this way….

-beside this frustrating fact:
i removed the waveshareboard from the now Dead Main-PCB T>>> Traces, Vias and SolderStopMask are off now.
Then Connected the waveshareboard directly to USB > STM32 is extremly hot again - so changed nothing…

So what can lead the Waveshareboard to get extremly hot? - if i ever buy a new Main-PCB and plug the old UI- (which i didnt solder - since it was a repair) - do i have to fear that something will demage the new STM32 again?
Can some accident on the UI PCB demage the STM32F4?
What can kill the stm32F4? - aka kill not really because it works, until its so hot it resets itself or hangs… so its working but gets very unhealty hot, even without any peripheral (main pcb) is connected - wtf?

… dont know what to do… if i now buy a new core pcb and all its parts, and then again all is for the trash… bad situation. if i just could swap the stm32-board, with a nice pinheader it would be less problem in anycase, but that its not the case, because the loopas case was designed to small - i dont get it…

0fc5608d995b44ff

Did this LoopA at any time run longer than 10 seconds before shutting itself down?

The MCU can heat up and could get permanently damaged, if there’s an unwanted short on an MCU pin pair, or of an output MCU pin shorted to ground - this happens often i.e. due to solder residue on an SMT IC or a “scratched” PCB trace that shorts out a signal. This short could be on any board, the UI board, the core PCB and even on the Waveshare MCU board itself.

Please check that there is no solder residue around the BOOT1 “solder jumper” on the backside of the Waveshare MCU board - i had seen a similar problem a few years ago, which also lead to overheating of the microcontroller - it should look like this: GND should be “jumpered” via a solder blob to BOOT1 - if there’s the tiniest bit of solder residue between 3.3V and BOOT1, the MCU would overheat:

Regarding headers and desoldering - this can be done successfully (i’ve done so multiple times) without lifting traces on the Core PCB with a vacuum desoldering iron, e.g. a ZD-915 clone.

Best regards,
Peter

e8abac78cdf65c56

https://www.midiphy.com/en/shop-details/120/42/mcu-core407v-1mb-flash-bootloader-

https://www.waveshare.com/core407v.htm

is this the right replacement? on Midiphy Picture - the STM32 Chip says its a VET6. which on waveshare orginal site says its a 512KB Flash Ram Model. you are writing “mcu-core407v-1mb-flash-bootloader” aka 1mB Flash?

PETER: do you have a link for a ZD-915 clone - thx.
so measure each GPIO Pin on the UI-Boards and check if it is grounded or hot to 3V? (J8/9, J15…) if nothing at all - then the UI-Board should not be the MCU-Killer.

I am pretty sure that the board was running longer then 10sec in this Hot-temperatur state - so yes i guess permanent damege.

0fc5608d995b44ff

@Phatline you can run LoopA without a problem on the 512KB VET6 cores. We’re using the 1MB VGT6 cores for future compatibility with larger SEQ v4+ firmwares, but these are not required for LoopA. LoopA uses a lot of RAM, but not too much flash.

As it has been such a long debugging process with this LoopA: if you order a new LoopA Core PCB from the shop, i can offer to add a free 512KB Waveshare VET6 Core to the parcel - there’s no bootloader installed on it yet, as it is a factory-sealed unit, but if you have an ST-Link v2 programmer to flash the bootloader, it will work nicely.

This desoldering station will do the job nicely:
https://eleshop.de/desolderingstation-zd-915.html

Needless to say, that yes, you should try to find the problem of the short - it might have been on the core, then you’re good when the whole board is replaced - it might have also been on the other boards as well. If you have a second LoopA with some boards available for testing, it might help to find the problem.

Viele Grüße,
Peter

e8abac78cdf65c56

@Hawkeye
Thx for the free Waveshare VET6 - i have ordered the PCB now.

Yes i have a st-link programmer- factory sealt is good.

also thx for the ZD915 link.

danke danke!