SKR 3 - Spindle Sync

[CBR900cc]

Hello, I'm retrofitting an old lathe, it as encoder onboard and I've wired everything to SKR3, but encoder not reading, this is my platformio env:

`[env:btt_skr_30_h723_tmc5160]
board = generic_stm32h723vg
board_build.ldscript = STM32H723VGTX_FLASH.ld

build_flags =
${common.build_flags}
${usb_h723.build_flags}
;${sdcard.build_flags} ; disattivalo per test flash
-D BOARD_BTT_SKR_30
-D HSE_VALUE=25000000
-D DRIVER_SPINDLE_ENABLE=7 ; 1|2|4 = SPINDLE_ENA|SPINDLE_DIR|SPINDLE_PWM
-D VARIABLE_SPINDLE
-D SPINDLE_ENCODER_ENABLE=1
-D SPINDLE_SYNC_ENABLE=1
-D SPINDLE_ENCODER_USE_QEI=1
-D SPINDLE_SYNC_USE_QEI=1
-D SPINDLE_ENCODER_CLK=1
-D PROBE_ENABLE=0
-D SPINDLE0_ENABLE=11
-D N_SPINDLE=1
-D MY_MACHINE
-I Inc
-include Inc/my_machine.h

lib_deps =
${common_h723.lib_deps}
${usb_h723.lib_deps}
;${sdcard.lib_deps}
motors

lib_extra_dirs =
${common.lib_extra_dirs}
${usb_h723.lib_extra_dirs}
;${sdcard.lib_extra_dirs}

upload_protocol = dfu
monitor_speed = 115200
monitor_echo = yes
`

old encoder has 5 wires: 5v, gnd, out1, index, out2

[terjeio]

You have modified the board map and added pin mappings for the encoder?

BTW the H7 driver is by @dresco, I do not have any H7 boards available at the time.

[CBR900cc]

yes, pin mapped, there is some thing to check if spindle sync is really active by grbl command, like $sync?

[terjeio]

$I output contains SS in the NEWOPT string when enabled and the $PINS output should contain the encoder pins.
The related settings will also become available. $38 has to be set > 0 to enable encoder input which is required for spindle sync.

[CBR900cc]

ok, these settings are available and I've setted $38 to 1

[terjeio]

Ok, you should then be able to use the $SD command to output encoder data, pulse count, index count, error count and angular position. Use $SR to reset the data.

[CBR900cc]

ok, no one input

[dresco]

I've bench tested on my own prototype board, so the basic functionality should be working on the H7 driver. The commit where support was updated is here.

You'll note that there are a couple of clock input options, and I did hardcode an internal pullup on the spindle sync input (not optimal - but don't think there was a setting for this & I never got around to asking).

Can you share your pin mappings etc. I also have an SKR3 here, so can have a look..

[CBR900cc]

I'm using PB4 for SPINDLE_PULSE_PIN and PD7 for SPINDLE_INDEX_PIN
if can help, my encoder is 1250ppr and goes to gnd when pulsing

[CBR900cc]

ok, using PC6 and PD14, rotation is recognized but is very oscillating, and when I try PID Tuner, after pressing "Get Data" (IOSender), console print:

error:20 - Unsupported or invalid g-code command found in block.

Any suggestion?

[CBR900cc]

@terjeio @dresco please help me, I need to solve as soon as possible, i can pay for a faster help if you need!

[dresco]

Am using my own prototype board - map here, pins are PG14 & PB4.

If you turn the spindle manually, does your $SD output change as expected?

[CBR900cc]

$sd change not good, now I'm replacing encoder + wiring, update you soon
But is correct $sd show numbers update simply with cable connected (motor off!) but no update with cable disconnected (motor off)?

[CBR900cc]

uhm...... with cable connected and VFD ON (motor off) $SD change, but with VFD OFF(motor off) $SD stay unchanged!

[terjeio]

The F7 and H7 uses (or may use) incorrect clock settings (prescaler) for the RPM timer that may result in incorrect RPM returned.
I have fixed the F7 code (copied over from the F4 driver) and will commit it later. When testing with my simulator the RPM reported is stable regardless so this "bug" is not the source of instability. The F4 code also supports 16 bit RPM timers via counting overflows - I'll copy this code over as well.

One way to check that the $SD output is approximately correct is to run the spindle for e.g. 30 seconds and check that returned values are within range. At 60 RPM the index count should be around 30 depending on the time taken to accelerate/decelerate the spindle. The pulse count should be close to the index count multiplied with the actual PPR.

[CBR900cc]

ok, thank you, I think there are high noise from vfd, working hard to solve