Nico Tonnhofer

Ok, it's done. My simulations looks quit good. https://github.com/Traumflug/Teacup_Firmware/commit/9daa942f6e Also the speed, acceleration etc. are the same on different axis-steps-per-m. https://github.com/Traumflug/Teacup_Firmware/commit/b75bb5d3 One axis looks like a bit noisy in gtkwave,...

Any time same gcode, same acceleration. First one is branch acc_ramp_len with 120,000 steps/m for X and 80,000 step/m for Y. Second one the experimental branch with 80,000 steps/m for...

const PROGMEM for the divisor is faster and needs same space. On the STM32 https://github.com/Traumflug/Teacup_Firmware/commit/9daa942f6e = 10 cycles https://github.com/Traumflug/Teacup_Firmware/commit/05b0c7be34bb1c0 = 7 cycles btw. I test the previous inline code. That...

I'm simulating some HEAD~x and find an issue. But this should not be a problem for ACCELERATION_RAMPING in the last stand. But regarding to [this comment](https://github.com/Traumflug/Teacup_Firmware/blob/master/dda.c#L845): `// A "-=" would...

I'm getting closer...

Ok, what is the difference between inline in a .h-file and not inline-ing in a .c-file? I take the code from dda_math.c: ``` uint32_t acc_ramp_len(uint32_t feedrate, uint8_t fast_axis) { static...

@triffid Thanks for explanation. I thought this was the issue. But in the end it seems to be a missing pgm_read_dword for the const array. Also the move_duration needs some...

Any idea what this magic '2' does? [First commit](https://github.com/Traumflug/Teacup_Firmware/commit/1e6c74096eb) [Line 57 of first commit](https://github.com/Traumflug/Teacup_Firmware/blob/1e6c74096ebce70884b5620e7142afeac7d7e859/mendel/dda.c#L57) [Current code](https://github.com/Traumflug/Teacup_Firmware/blob/master/dda.c#L289) ``` if (distance < 2) distance = delta_um[E]; ```

Ah, yes. It's 2um ~ 0.

@phord yes I thought also something in that direction. Think about distance is 3 and delta_um[E] is huge.