3 ms·

The custom gamecube controller project I'm on (https://github.com/PhobGCC https://github.com/PhobGCC) has clobbered the supply of Teensy 4 microcontroller board

by CarVac 4y ago

The custom gamecube controller project I'm on (https://github.com/PhobGCC https://github.com/PhobGCC) has clobbered the supply of Teensy 4 microcontroller boards. We already redesigned to switch away from Teensy 3.2, which had production suspended indefinitely due to parts shortages.

rasz 4y ago
600MHz 1MB microcontroller seems a little bit of an overkill for something $0.8 can easily handle https://lcsc.com/product-detail/ST-Microelectronics_STMicroelectronics-STM32G030F6P6TR_C529330.html https://lcsc.com/product-detail/ST-Microelectronics_STMicroe...
- zamadatix 4y ago
  512,000 cycles per poll is probably plenty but I think the SRAM might be a bit small for that code. Could something be made to fit in that and still do the same things? Probably but it likely wouldn't be as easy on either the software or hardware side. Sometimes a lunch is worth more than the extra hassle of optimizing.
- CarVac 4y ago
  The Teensy 3.2 was fast enough; we're running matrix math for remapping the stick gates and a modified Kalman filter at 1000 Hz, and bitbanging the GameCube protocol meanwhile. Teensy 4 is overkill and we've been clocking them at 150 MHz to reduce power draw.
bschwindHN 4y ago
What did you switch to? I was going to recommend the RP2040 but you've probably already got things sorted out.
- CarVac 4y ago
  We already switched from Teensy 3.2 to Teensy 4 because 3.2s are unobtainable. RP2040 is what we're looking at but it has ADC issues apparently. We're going to try though.
  - bschwindHN 4y ago
    Not sure about the ADC issues, but if you're bit-banging the gamecube protocol, you'll gain back a lot of CPU by using the PIO peripheral.
  - TaylorAlexander 4y ago
    I looked in to the ADC issues and decided it was fine. It has slightly lower effective resolution because certain ADC values are more likely to occur. It looks like in the end not a huge deal.
999900000999 4y ago
Are you planing on selling these ? I'd love to buy one ( under 100$ )
- CarVac 4y ago
  I convert existing GCCs for $120 for my local scene (NYC).

by CarVac 4y ago

rasz 4y ago

600MHz 1MB microcontroller seems a little bit of an overkill for something $0.8 can easily handle https://lcsc.com/product-detail/ST-Microelectronics_STMicroelectronics-STM32G030F6P6TR_C529330.html https://lcsc.com/product-detail/ST-Microelectronics_STMicroe...

zamadatix 4y ago
512,000 cycles per poll is probably plenty but I think the SRAM might be a bit small for that code. Could something be made to fit in that and still do the same things? Probably but it likely wouldn't be as easy on either the software or hardware side. Sometimes a lunch is worth more than the extra hassle of optimizing.
CarVac 4y ago
The Teensy 3.2 was fast enough; we're running matrix math for remapping the stick gates and a modified Kalman filter at 1000 Hz, and bitbanging the GameCube protocol meanwhile. Teensy 4 is overkill and we've been clocking them at 150 MHz to reduce power draw.

bschwindHN 4y ago

What did you switch to? I was going to recommend the RP2040 but you've probably already got things sorted out.

CarVac 4y ago
We already switched from Teensy 3.2 to Teensy 4 because 3.2s are unobtainable. RP2040 is what we're looking at but it has ADC issues apparently. We're going to try though.
- bschwindHN 4y ago
  Not sure about the ADC issues, but if you're bit-banging the gamecube protocol, you'll gain back a lot of CPU by using the PIO peripheral.
- TaylorAlexander 4y ago
  I looked in to the ADC issues and decided it was fine. It has slightly lower effective resolution because certain ADC values are more likely to occur. It looks like in the end not a huge deal.

999900000999 4y ago

Are you planing on selling these ? I'd love to buy one ( under 100$ )

CarVac 4y ago
I convert existing GCCs for $120 for my local scene (NYC).