fCLK - where to get this value

[jiri-jirus]

Hello, i have a question - how can i get information about fCLK of particular SilentStepStick? this value (fCLK) is used in many formulas, but i dont know whether it is set by some register or hardwired by oscillator on stepstick...? thank you very much for explanation.

[teemuatlut]

Datasheet, p88

26 Clock Oscillator and Input

The clock is the timing reference for all functions: the chopper and the velocity thresholds. Many parameters are scaled with the clock frequency, thus a precise reference allows a more deterministic result. The on-chip clock oscillator provides timing in case no external clock is easily available.

USING THE INTERNAL CLOCK

Directly tie the CLK input to GND near to the IC if the internal clock oscillator is to be used. For best precision, the internal clock can be calibrated by reading out TSTEP at a defined external step frequency. It is easiest to use 1 kHz microstep frequency. With this, TSTEP shows the number of internal clock cycles per millisecond, i.e. TSTEP=1200 means that fCLK is 12 MHz in the actual IC. Scale velocity thresholds, TOFF and PWM_FREQ based on the determined frequency. Temperature dependency and ageing of the internal clock is comparatively low.

---

I can't remember if that's the corrected information or not, but using this I've determined the f_clk to be around 12.5MHz. The instructions there might be off by an order of 10.

[jiri-jirus]

thank you. so this M122 output means that i have 10.5MHz fCLK?

Send: M122 Recv: X Y Recv: Enabled true true Recv: Set current 400 600 Recv: RMS current 397 581 Recv: MAX current 560 819 Recv: Run current 12/31 18/31 Recv: Hold current 6/31 9/31 Recv: CS actual 6/31 9/31 Recv: PWM scale 34 42 Recv: vsense 1=.18 1=.18 Recv: stealthChop true true Recv: msteps 16 16 Recv: tstep 1048575 1048575 Recv: pwm Recv: threshold 79 121 Recv: [mm/s] 100.08 65.34 Recv: OT prewarn false false Recv: OT prewarn has Recv: been triggered false false Recv: off time 5 5 Recv: blank time 24 24 Recv: hysterisis Recv: -end 2 2 Recv: -start 3 3 Recv: Stallguard thrs 0 0 Recv: DRVSTATUS X Y Recv: stallguard Recv: sg_result 0 0 Recv: fsactive Recv: stst X X Recv: olb Recv: ola Recv: s2gb Recv: s2ga Recv: otpw Recv: ot Recv: Driver registers: Recv: X = 0x80:06:00:00 Recv: Y = 0x80:09:00:00 Recv: Recv: ok

[teemuatlut]

No. You can't do it in Marlin.

[jiri-jirus]

i went off: " With this, TSTEP shows the number of internal clock cycles per millisecond, i.e. TSTEP=1200 means that fCLK is 12 MHz in the actual IC." and this part of M122: "Recv: tstep 1048575 1048575" I know I cant change fCLK in Marlin, but I can get info about actual fCLK, right?

[teemuatlut]

TSTEP

Actual measured time between two 1/256 microsteps derived from the step input frequency in units of 1/fCLK. Measured value is (2^20)-1 in case of overflow or stand still. The TSTEP related threshold uses a hysteresis of 1/16 of the compare value to compensate for jitter in the clock or the step frequency: (Txxx*15/16)-1 is the lower compare value for each TSTEP based comparison. This means, that the lower switching velocity equals the calculated setting, but the upper switching velocity is higher as defined by the hysteresis setting.

---

2^20 - 1 = 1048575 == standstill

[jiri-jirus]

well, my tstep now shows 759 in the middle of move with X axis ( G1 X200 F600 I have M92 X100 = 1kHz microstep frequency) that seems to be 7.5MHz fCLK?

[jiri-jirus]

well, I am tinkering with CHOPCONF on FYSETC TMC2130 v1.1 drivers because I cannot manage to get rid of their high-pitched noise from motor coils when on spreadCycle (and I am always on 24V), the noise is a bit suppressed only at very strange hysteresis values. (-3, 0) or something like that. So it seems that i am running at very low fCLK.

[teemuatlut]

Actual measured time between two 1/256 microsteps derived from the step input frequency

TSTEP is not the time between your input STEP signals. Unless you're running with 256 ms setting.

I'd suggest you make an external Arduino sketch to test your drivers. According to the datasheet instructions you should end up with a figure something like 1.2MHz. Multiply that by 10 to get your actual f, because the information given is not quite correct. IIRC

By your calculations, you say you'd have 7.5MHz f_clk. If we divide this by 10, because of the misinformation, we get 750kHz. This is the time between two interpolated (think in terms of 256 microstepping even if that is not the setting) step pulses. The ratio between 16 (assumption) and 256 is 16. Therefore 750kHz * 256/16 = 12MHz

[jiri-jirus]

thank you. well explained.