7i92, laser power velocity controlled

Hi,
i have a few questions regarding the hal configuration for my current project: laser vector engraving.
I already managed to hook up a 2,5W laser to my 7i92, changed a few lines in a Fusion360 Postprocessor and can engrave with on/off commands.
The Postprocessor uses the 2D-Contour CAM module for plasma/waterjet/laser and simply turns on/off the laser with M62/M65. Postprocessor attached. Heres an example of my first laser engraving:


As you can see, the rectangles look ok, but the numbers are way too burnt. I think the problem is that there are too many lines with too many accelerations/decelerations in one spot. I think i can just model with fewer lines, but the acceleration and deceleration will still be a problem. To compensate for slower movements, i need to lower the laser power.
My next step would be to use a PWM to control my laser power. I would net my velocity to that PWM, If the velocity is high the laser power is high - if the velocity is small, the laser power is low. Everytime M62/M65 would come up the PWM gets enabled/disabled.
Since this is my first fairly complex hal experience, i would like to discuss it here. Does this hal-snippet looks somehow of what i described?
For testing i will set pwm_scale to 0.0017 since my max linear velocity is 34mm/s, which should result in the maximum PWM frequency and maximum laser power.

it looks close though the PWM scale should be something like maximum velocity
(in machine units/second) so you get 100% duty cycle at full speed.
( a scale of 0.0017 would mean full PWM at any velocity >= 0.0017 mm/s)

It might be helpful to use the scale component because it has an offset
and I would expect the PWM control of the laser vs marking to be quite
non-linear. If more complex linearization is needed, the lincurve component
could be used.

This could be easilly managed via custom component or as PCW says lincurve by monitoring motion.current−vel
The trick though you will need to add a remap on the F command so you can save the commanded velocity using M67/M68 as currently Linuxcnc has no way of telling HAL what that was.

So then its a simple matter to apply a PWM adjustment based on the percentage of commanded velocity if current_vel < commanded velocity.

What is the advantage of making the laser exposure (inversely) proportional to the feed rate?
Once you have the PWM vs velocity correct for a given burn depth it seems like you would not want mess with it...

PCW wrote:

What is the advantage of making the laser exposure (inversely) proportional to the feed rate?
Once you have the PWM vs velocity correct for a given burn depth it seems like you would not want mess with it...

Well I might be a bit out of my depth here but its very similar to a problem we encountered with plasma machines pre-plasmac which was compounded by errors in the documents (since corrected on our feedback). The issue was around THC velocity lock which we initially tied to current-vel instead of commanded-vel so we got dives when encountering slow segments becasue the THC velocity lock never came on or was re-enabled at slow velocities unexpecredly

If only someone could push state tags branch there would be much cleaner solutions as we would know commanded-vel.

Wouldn't you have different PWM settings and feedrates for different materials? If so, wouldn't this be more flexible?

I suspect you might need different velocity/PWM curves for different materials and speeds,
depending on how fussy you are.

I don't think you have the same issues as corner lock since there is no height control to manage.

PCW wrote:

I suspect you might need different velocity/PWM curves for different materials and speeds,
depending on how fussy you are.

I don't think you have the same issues as corner lock since there is no height control to manage.

Understood there is no need for Corner lock but the same variables are in use and you need to remap the Gcode F command to capture the commanded velocity.

We might be at cross purposes here Peter. I was thinking you would calibrate based on straight line cuts at commanded velocity for each material and then adjust the PWM downwards to some algorithm based on current velocity....

My point was that I did not see the need for the commanded velocity, only the current velocity
since the absolute laser exposure could be preset with just that information

Thanks! I finally got to my shop and tested it. Some lines were incorrect or unnecessary, here is my current hal snippet:

Here is the result of that:

I need to do way more testing and fiddling, but that second engraving looks quite promising.

What i noticed:

• If my machine is stopped while it engraves, the laser turns "off". Thats because my velocity is 0 -> PWM is 0 -> laser power 0
• That may lead to dangerous situations where i emergency stop the machine, the laser looks like it's off, but if i jog afterwards it turns on. See the straight lines between the engravings. In short term i will use a external power switch for the laser for that
• The laser power looks like it's way weaker than the original engraving.
• I had a scale of 34, a max speed of 34mm/s but only engraved with 17mm/s, so there was only a 50% PWM output
• I will need to do more testing with my speeds, scale, maybe the offset parameters or lincurve.