Journal - Reprogramming my Launchpad Mini

You can find updates on my progress regarding repairing/reprogramming my launchpad mini here.

Introduction

One day I got the idea to read out the serial data sent over USB by my launchpad with an arduino. I just thought it could be a cool way to use the launchpad as a controller, as it only sends and receives MIDI data. After a bunch of fiddling with open wires I crossed the wrong wires and bricked the device. A second-hand one would cost me about 50 euros. But why buy a new one when I can try to fix it myself?

Nov 2 2024 - Resoldering

So today I bricked my launchpad. Seemed like the microcontroller was the problem. (spoiler, it wasn't)

Nov 9 2024 - Resoldering

Some big progress! Today, with some help, I managed to replace the STM32 on the board with my new one from the blue pill board. Even after all this, though, no luck powering it. After some tests I discovered that the problem might've just been a voltage regulator I fried... That sucks. Next to that, I also lost the original controller, so I'm pretty much forced to reprogram it myself. (No, I'm not buying a new one) Next week I'll try and screw around with that.

Nov 23 2024 - Connectors

Not much progress today. I did a lot of research about JTAG connectors and STLink connections, and I'm still not sure whether I should get a JTAG programmer. They are quite expensive, and according to some people I should be able to program my board via STLink through the JTAG terminal if I wire it up correctly. I'll probably attempt to do that next week.

No luck with powering the board either.

Nov 30 2024 - Powering up

Some progress today! After some tests earlier this week, I realised that the broken voltage regulator shorted, which was why I couldn't hook up my STLink or power the board. Removing this voltage regulator allowed me to hook up my STLink and power the board properly. When I did so, some lights turn on, seemingly at random. After a while it seemed to "stabilize" into powering all LEDs. Somehow the components are receiving some signals that trigger the LEDs. I still cannot reach the STM though, whatever I try.

Something weird I noticed is that PA13 and PA14 are not actually connected to anything on the board, even though the datasheet does claim so. I painstakingly soldered some small wires onto those pins and hooked up my STLink pins directly to them. Still no results. Kind of afraid I fried it because of something I did earlier, whether it was trying to reach it with faulty wiring or overheating it due to the soldering.

Dec 7 2024 - Entering & Breaking

Some more progress today. Previous week I stated that the PA13 and PA14 pins weren't connected to anything, but I was wrong. I fished up an old picture where the MCU was desoldered and there were clear traces underneath it, which coincidentally connected to PA13, PA14 and some other pins. This means I can disconnect the very tiny wires I painstakingly soldered onto the MCU and just use the boards terminals to program it.

Another discovery I've made is that I was using the wrong side of the JTAG terminal. I was using the one on the MCU's side, but that's all bridged to ground. I was confused as to why and cut all those traces, but on the bottom side of the board, all terminals are connected to the right pins on the MCU so I tried hooking up my wires to that. And it actually worked! I was able to reach my MCU a few times and read its flash, but there were issues. The programmer gave various errors like it was unable to read the flash, or that it had some security features enabled and I couldn't reach it. After a while I couldn't reach it at all anymore and I'm pretty sure it's bricked now. While screwing around trying to erase the flash, I think I've edited some memory that just doesn't allow me to communicate at all anymore. I'll be ordering some MCU's of the same type (STM32F103C8T6) off Aliexpress and I'll try to replace it again. Now that the wiring is connected correctly, I think the next attempt at connecting will be much easier. I've already found a batch of 5 MCU's so I have 4 more chances to fuck up.