search

bigtreetech / BIGTREETECH-SKR-V1.3

32bit board with LPC1768, support marlin2.0 and smoothieware, support lcd2004/12864, On-board TMC2130 SPI interface and TMC2208 UART interface no additional wiring is required
1.01k stars 1.15k forks source link

SKR 1.4 turbo E0 TMC 2209 driver = magic smoke. Is it the code or the board? Interesting! #508

Open Balduc opened 3 years ago

Balduc commented 3 years ago

Hi everyone. I've order an SKR 1.4 turbo and follow instructions of Chris Basement and Teaching Tech YouTube channel to compile and upload MARLIN 2.0.x. The board came with the TMC2209 driver, so I've enable UART mode, use sensorless homing for X and Y and cut the DIAG pin of the Z, E0 and E1 on each other TMC2209. Didn't use the E1 on that printer, but install the driver anyway.

Everything was perfect was working flawlessly. After a month the board kind of... died. After a quick troubleshoot, I'm unable to say what happened, but i was unable to communicate via serial(USB), and the reprap 12864 LCD full graphic controller was receiving the sufficient voltage, was lit, but nothing was happening. The board didn't want to boot. But before that, it was totally fine.

So, I've re-order a new one, re-compile Marlin (the same code as the other one), put on the drivers, turn it on to upload Marlin on it, and... snap crackle pop, MAGIC SMOKE from the E0 driver. As soon as it received voltage. The little chip underneath the driver on the E0 burned. I was like "ok maybe the driver was not good, and the other board died because of that driver".

So I've check the jumpers, and my code, put another TMC2209, and power it up. Hurray, everything looks fine. the Marlin logo shows up. It asked for a restaure default, did it, was like "Ok cook, it was the driver" and after 30 seconds, maybe less, the E0 driver (not the same one, but the same emplacement) snapped, crackled and popped with magic smoke again. I've re-check my compilation, it was fine, for me... Now I'm afraid that I may have change the code by mistake.

What i can do to avoid that kind of issue in the future with any kind of boards? Or to go further: is there a setting that can lead to that kind of issue? Nothing was connected (bed, hot ends, steppers or sensors). Only 24 (23.97) volt from a bench power supply, and a LCD 12864 smart graphic controller. The current draw was 0.64 amps and ramp up just before the burn of the drivers, but i don't remember the number, I was concentrated to unplugged everything.

Why on first boot the driver burned immediately, and on the second boot, it took a minute? A capacitor issue? Note that the drivers was working fine on the other boards. But i got 2 different issues, with boards butwith the same code. So i want to know if it's me, or the boards I've ordered.

I just hope it was not my code, but just in case, I need your opinion. Thank you for your time. So here's my MARLIN 2.0.x source that I've compiled with VS / plateformIO:

https://drive.google.com/file/d/11_SPINRwsQfv3pVEPf4aSFKXqCz7qdgO/view?usp=sharing

IMG_20210304_150827 IMG_20210304_150722 IMG_20210304_150714 IMG_20210304_150707

radek8 commented 3 years ago

I can't open your configuration file. I can't imagine a Marlin configuration that would cause such behavior. The first TMC controller could be defective and by burning it contaminated the SKR board with combustion products that are conductive and subsequently the second TMC controller also burned down.

Balduc commented 3 years ago

After a lot of posting in different forum, it might be the board itself. I've noticed that purchase those are like the lottery. I hope the next one will not have any quality issue. After testing with limited current, it is booting, but without access of the the h bridge of any of the drivers. Dead dead.

https://drive.google.com/file/d/11_SPINRwsQfv3pVEPf4aSFKXqCz7qdgO/view?usp=drivesdk

radek8 commented 3 years ago

I own one SKR1.4 board and one SKR1.4 Turbo. I have TMC2208 and TMC2209 drivers, everything works flawlessly. Only on the SKR1.4 board did I not clean the solder flux when soldering the connector on the print head. soldering flux propijiola temperature sensor with 24V pin, and the pin on the processor measuring the temperature was destroyed. But that was my fault. Fortunately, there was another pin to which it could be connected.

In my experience, if a component burns in the same way as for you, part of the metal conductors will evaporate, which will then condense in the immediate vicinity. This will create a conductive surface on everything close to the point of ignition. Even if you manage to remove the combustion products, it is quite likely that some pins on the processor leading to the driver will be destroyed. The processor pins are sensitive to voltages above 3.3V The driver is powered by 24V If a short circuit occurs ....