Correct Laser/Spindle issues

[shitcreek]

When using non PWM pin with just on/off, laser/spindle setPower uses SPEED_POWER_STARTUP instead of SPEED_POWER_MIN.
Changed inline_enabled call to inline_ocr_power to inline_power.

Fixed an issue that I had created in the other PR: laser wasn't turning off between moves. Solution: relocated translate_power back over to spindle_laser.cpp from spindle_laser.h.

[mentaldemise]

Looks like I have this commit already in my local but the laser is occasionally not turning off. It doesn't seem to matter what I do with the S0/G0/M5 commands the laser will still stay on for a few of the first movements and then work like normal. The same will happen at various times through the etch but I'm having issue defining exactly where. Fire pin is tied to 5V with a 10K resistor.

The minimum I can manage to get to reproduce it is:

; LightBurn 0.9.11
; Marlin device profile, absolute coords
G21
G90
; Scan @ 80 mm/sec, 15% power
M106 S0 
G0 X181.39 Y40.4 F4800
G91
G1 X-0.03 F4800 I S38.3
M5 I
G0 X-7.13 Y-0.1 F4800
G1 X0.16 F4800 I S38.3
G0 X6.93 F4800
G1 X0.11 F4800 I S38.3
G0 X6.94 F4800
G1 X0.16 F4800 I S38.3
G0 X0.13 Y-0.1 
G1 X-0.31 I S38.3
G0 X-6.88 F4800
G1 X-0.19 F4800 I S38.3
G0 X-6.87 F4800
G1 X-0.32 F4800 I S38.3
M5 I
G1 X-0.13 Y-0.1 I S0
G90
M5 I
G90
; return to user-defined finish pos
G0 X0 Y0 F12000

[shitcreek]

No, I don't mind it.

I was experiencing a similar thing before where the laser doesn't turn off for the first couple cuts, but it's all working now.

First off G0 and G1 don't take the I parameter. If you are using LASER_POWER_INLINE, LASER_MOVE_POWER and LASER_MOVE_G0_OFF then M3/M5 aren't even needed.

[mentaldemise]

OMG I love you! This is broken right now in the main bugfix branch though, so heads up.

Is there something I should tell them as to when to include the I or not?

[shitcreek]

I'll add a note to the documentation

[thinkyhead]

git fetch origin
git checkout max-power-non-pwm
git reset --hard origin/max-power-non-pwm

[shitcreek]

Tested. With CUTTER_POWER_DISPLAY set to PWM255, power sent via gcode is being interpreted as percent but shown as PWM on the LCD, ie M3 S10 shows laser power at 25 on the LCD which is 10% of 255. Power set via the LCD is off by one to three degrees except for certain power such as at 204 and 255 which were spot on.

[thinkyhead]

Corrected some unit conversions, and maybe even all of them.

[shitcreek]

Tested, same exact issues. I didn't mention before but power beyond min and max range are not corrected to min/max. And effective power range (where the laser actual power output is affected) appears to be between 0 and 25 via the LCD.

There were a couple obvious mistakes. The main cause here I believe is that there is a confusion between Unit power, Display power and what is set for min, max and startup power in config_adv. I think I am close with my local patch.

[shitcreek]

'cpower' = power set in config_adv ie SPEED_POWER_MAX - for spindle it is in RPM, for laser it is in percent.
'upower' = unitPower = set by CUTTER_POWER_UNIT in PWM255, PERCENT or RPM
'power_to_range()' will take any value and puts it within the min and max power range.

In CUTTER_POWER_RELATIVE mode, 'M3 S0' won't turn off the spindle/laser except if unit power is in RPM.
mpower = menuPower = laser/spindle control via LCD menu.
Note: The dial can set the laser/spindle while in the off state. When the laser/spindle is on, the dial can adjust the power, and turn off the power when dialed to zero.

Todos/Notes:

• M18 G28 at the end of file turns the pin on again.
• Enabling SPEAKER interferes with laser/spindle - timer issue?
• At and above 20Khz PWM frequency on my K40, the spindle/laser power is no longer modular, it is always on max. This is probably the limit of my laser power supply. Between 2.5Khz and 18Khz works well, but higher frequency demands faster moves.

[mentaldemise]

I'll have to pull latest and merge later. It's interesting that I'm running mine on an LPC1768 at 50kHz, was on the RAMPS as well and had no issue with power scale. I did on the LPC prior to adding a level shift.

Here's some of my videos of your code running(thought last time I had to update something to have it build lol) - This is a Mega2560(Ramps)
https://www.youtube.com/watch?v=vAA65vcwuIo

Raster(LPC required):
https://www.youtube.com/watch?v=EFzJmyRZ6YY

[shitcreek]

Added LASER_MOVE_G28_OFF to disable the laser when homing.

[gminadak]

I don't know if this is the right place to report this.
After the latest merge i have the following errors
Compiles fine if i download it from @shitcreek or overwrite the files mentioned after every error (this is the first).

In file included from Marlin\src\MarlinCore.cpp:124:

Marlin\src\feature/spindle_laser.h: In static member function 'static void SpindleLaser::inline_disable()':
Marlin\src\feature/spindle_laser.h:228:37: error: no match for 'operator=' (operand types are 'power_status_t' and 'int')
  228 |       planner.laser_inline.status = 0;
      |                                     ^
In file included from Marlin\src\MarlinCore.cpp:44:
Marlin\src\module/planner.h:124:5: note: candidate: 'constexpr power_status_t& power_status_t::operator=(const power_status_t&)'
  124 |   } power_status_t;
      |     ^~~~~~~~~~~~~~
Marlin\src\module/planner.h:124:5: note:   no known conversion for argument 1 from 'int' to 'const power_status_t&'
Marlin\src\module/planner.h:124:5: note: candidate: 'constexpr power_status_t& power_status_t::operator=(power_status_t&&)'
Marlin\src\module/planner.h:124:5: note:   no known conversion for argument 1 from 'int' to 'power_status_t&&'
In file included from Marlin\src\MarlinCore.cpp:124:
Marlin\src\feature/spindle_laser.h: In static member function 'static void SpindleLaser::inline_ocr_power(uint8_t)':
Marlin\src\feature/spindle_laser.h:262:55: error: no match for 'operator=' (operand types are 'power_status_t' and 'int')
  262 |         planner.laser_inline.status = ocrpwr ? 0x03 : 0x01;
      |                                                       ^~~~
In file included from Marlin\src\MarlinCore.cpp:44:
Marlin\src\module/planner.h:124:5: note: candidate: 'constexpr power_status_t& power_status_t::operator=(const power_status_t&)'
  124 |   } power_status_t;
      |     ^~~~~~~~~~~~~~
Marlin\src\module/planner.h:124:5: note:   no known conversion for argument 1 from 'int' to 'const power_status_t&'
Marlin\src\module/planner.h:124:5: note: candidate: 'constexpr power_status_t& power_status_t::operator=(power_status_t&&)'
Marlin\src\module/planner.h:124:5: note:   no known conversion for argument 1 from 'int' to 'power_status_t&&'
Compiling .pio\build\LPC1768\src\src\feature\cancel_object.cpp.o
*** [.pio\build\LPC1768\src\src\MarlinCore.cpp.o] Error 1

[shitcreek]

@gminadak I can confirm this new issue and have resolved it with this PR: #18237

Thanks for reporting the issue.

[gminadak]

I have the following issue but i can't be sure if it is Marlin related or from my machine/configuration.
The laser starts out of the designated path as you can see from the pictures.
This also happens with squares but its not visible always(on wood at least if i run the same gcode on other thinner material is visible).
On circles you can clearly see it every time.
I have change many parametres on my configuration microsteps, speeds, acceleration, laser delay, arc segment, junction deviation, tighten belts but the problem persists.
I find it hard to be mechanical because happens every time at the same shape spot where the laser starts firing but i could be wrong.
below is my configs and the Gcode.

Thank you.

pictures
config.zip

[shitcreek]

It should be M5, not M05 and G0 and G1 do not take the I parameter because they are inlined already.
I corrected those, then ran your gcode and cannot reproduce your issue and your config seems fine after a quick glance.

[shitcreek]