Almost there: Laser reacts strange around 1% power (and traps for beginners)

Discussions and file drops for Auggie
Post Reply
Roman
Old Timer
Posts: 31
Joined: Wed Nov 06, 2019 4:07 am

Almost there: Laser reacts strange around 1% power (and traps for beginners)

Post by Roman »

Hi there
I am new here and I came from Mach 3 because the M03 and M05 commands are to slow for laser and M10/11 are not supported by the Pokeys57CNC.
After quite some twists and long hours I have now the laser running and can control it from Auggie.
One strange thing though:
Using the Pokeys GUI the Laser works perfect, 0% to 100%.
However, in Auggie after M03 the laser turns on at about 1% power, a light dot is visible.
When I fire anything below 1% it switches fully off, no dot any longer. Firing anything from 1% and above it works normal as it should.
I could not find it, Motor Tuning for PWM is set to no tickle, zero minimal, 100 maximal.
Why is this effect between 0% and 1%? I would like to get rid of it.

Traps for newbies:
- If you come from Mach 3: Do not forget to activate 8 Axes in Pokeys57CNC, not the 6 it normally has activated. Else you will only get it to work in Pokeys GUI but Auggie does nothing as it uses the 8th Axis to control the PWM (the price was about 8 hours of searching  :(  )
- If you happen to have an OptLaser (cool Polish brand) with an Interface box for x-Carve and others using an Arduino based contoller: Remove the diode in the PWM signal cable in the box. Or the laser reacts eractic.
 
User avatar
ArtF
Global Moderator
Global Moderator
Posts: 4586
Joined: Sun Sep 05, 2010 6:14 am
Contact:

Re: Almost there: Laser reacts strange around 1% power (and traps for beginners)

Post by ArtF »

Roman:


>>Why is this effect between 0% and 1%? I would like to get rid of it.

  The way the pokeys works is it gives me a control word that is integer from 0 - 100, 0 is fully off, 100
is fully on. However, it is integer, so 0 is off, 1 is on at 1%, you cannot do .5% for example.

  I get around this here with an arduino programmed as a PWM convertor, so my PWM from auggie goes
into this arduino and it has a screen that allows me to select 0 - 100% of the incoming. So if I need .5%
I can set the arduino to 50% of incoming and use 1% in auggie. If I need 100 levels from 0 - 5, I can set auggie
to 100% and set the arduino post processor to 5%, I now get 0-5 in 100 levels..

  So your 1% dot is simply the lowest Auggie can do and is your laser at 1%.
How powerful is your laser? Usually 1% is low enough for a start power. Its unfortunate
that the power word is only 0-100 instead of 0-255 , but thats in the pokeys hardware
for the realtime pwm control.


 
Art

Roman
Old Timer
Posts: 31
Joined: Wed Nov 06, 2019 4:07 am

Re: Almost there: Laser reacts strange around 1% power (and traps for beginners)

Post by Roman »

Hello Art

Thank you for your answer.
My Laser has 6 Watt so it does not burn anything at 1%. It is just a 'cosmetic' problem, I wanted to understand if I do something wrongly(and learn).
If I understand you correctly then it seems that M03 switches the Laser to the minimal 1% (I see the dot). If I Testfire with a setting of 0.9% or below it sets the laser for this time to 0% (temporarily 'overrides' M03 so to say).
Correct?

Thanks a lot for all your time to develop this and to help the users to get it to work. This is highly appreciated.
Best greetings from Switzerland.

PS: regarding the 8th axis:
The point where the penny dropped came when I was googleing to find more info on my possible error. I found your video 'AuggieLaser1' in your YouTube channel which emphasized the 8th axis usage. This was so good that it finally made 'klick' in my head.
It would be of advantage for other newbies to put this link on the https://www.gear2motion.com/auggie  page as well.   
User avatar
ArtF
Global Moderator
Global Moderator
Posts: 4586
Joined: Sun Sep 05, 2010 6:14 am
Contact:

Re: Almost there: Laser reacts strange around 1% power (and traps for beginners)

Post by ArtF »

Hi Roman:

>>If I understand you correctly then it seems that M03 switches the Laser to the minimal 1% (I see the dot). If I Testfire with a setting of 0.9% or below it sets the laser for this time to 0% (temporarily 'overrides' M03 so to say).
Correct?

  The M03 does a couple things. Many lasers need a  pre-ignition pulse to keep them
ready to fire. This is called a tickle, and Auggie does a 1ms tickle while m3 is on,
you can change this to 0 I think in the settings.

Couple things worth knowing.....

  If M3 is on, it also tells Auggie to fire the laser whenever a G1 move is in progress. If an
augment image is loaded ( The Aug light is flashing ), then that firing during G1 will not
happen if the motion is outside the image field, and when firing the power will be a scaled
grey scale value. A 255 in the image will fire at the Power Sliders value, and less than 255
will fire at pixel/255 * SliderPower.

  The button "Dist. Cor" under the laser power sets Distance Correction mode, in this mode
the power of the laser will be corrected by speed. Its important to know how this works..

  Lets say you command F1000 as your feedspeed. Also lets assume your power slider
is set to 50% for this example. Lets say the next line executed in your code is
G1X100 , the speed at start is obviously 0, so the laser power is zero as well. As the axis
speeds up to do the move, the power will be  ("instant speed" / "requested speed" ) * PowerSlider.
  So if the axis hits the 1000mm/min called for the power output will be 50% as requested
by the slider. If you simply turn on a laser and do the move without distance correction, the
acceleration zone of the move will be very burnt as the slow parts of the move get more
radiation. This is why many CNC programs burn at the corners of squares. Distance Correction
looks instead to how far one would move at full speed of F1000, it then takes the distance
we will actually move in the next ms, and scales the power to make sure the same power
is spent on each ms of the move. As lasers aren't completely linear in output this correction
isn't perfect but definitely worth using in most situations.

  An important note here, if you command F5000 for example, and the fastest your axis
can do on this move is 500, the power will never exceed 10%. It is important to make sure
your Feedrate commands can be achieved if you wish to get proper power. I often run without M3
for a few minutes and watch the actual motions feedrate to see what I should be commanding
to make them match as best I can. This will make the entire burn more even and gives me very
good results. 

  It is also important, if one wishes best performance to understand Auggies trajectory planning,
unlike Mach3, Auggie is 6th order. This makes a difference in the way one must tune their axis
to get best performance. Because 6th order planning  doesnt use acceleration in the same way
as Mach3, your accel setting will normally not be as Mach3's is. In Mach3 you set a maximum
accel to make sure you dont stall during motion, in Auggie you set it as high as you can for
jogging only. If you use the keyboard jog you can see the effects of accel. The higher you can make
it the better so raise it while testing jogging between changes and make it so its fast and peppy
at startup but not quite to stalling from trying to start too quickly.  The reason you do this
is because Jog doesnt use 6th order, so your setting the max accel your motor can handle
without correction. Auggies Jerk controller will then use whatever accel it needs , even if much higher
than mach3 can accept and limit the speed and accel by jerk formula's.

  After doing this you run Gcode while adjusting MaxJerk setting between test runs. If set too low
on a small stepping Gcode, something that moves in jerky directional changes, the axis will
jerk around slowly as it tries to keep jerk low. As you increase the max jerk, the axis will all
start to smooth out and go faster. This jerk setting can be large or small depending on the system,
my galvo for example is set to 25,000,000 and moves very very fast. Typically 25000mm/min
is what I run in Gcode. My Cartesian table though, needs a setting of 50,000 to make it move
quick and smooth , but not jerk. Too low and some jobs go very slow and jerky. So if you find
motion slow and jerky, raise the maxjerk setting, dont be shy about how much, start with 1000
and make the changes smaller or larger by result.  The biggest mistake users of Auggie make is
not understanding how the jerk limit works, and it makes a huge difference in performance. Its
not unusual for a user to report the smoothest motion they have ever seen on their system, others
find no difference from Mach3, its a matter of tuning and of the system mass and such.

  I do occasionally cut a job and find it too jerky and raise the Jerk setting or too rounded on corners
so I lower the setting. Unlike Mach3 Auggie will not blend motions together to get speed, instead
it will put a small circular radius between lines sized as per jerk settings to keep the energy
of the move low as it goes around each radial bend.

  Another thing to consider is the lookahead, I set mine to 500 at times, on small line code this
speeds up things considerable. Heres why. Auggie is a replanning planner. Imagine you send
a G1X100 move. Auggie on reception plans the move entirely from start to end making sure
velocity is zero at the end and all jerk constraints are satisfied. It then starts to move. The max
speed after all the calcs is say..100mm/min.

  Imagine now that the next line of code is G1X200. Auggie is now stepping into the start
of the X move as it recieves this move. It will now go back and replan the all the moves including that X move. It
now knows the end position of that move will be 200. The only way to ensure a smooth motion
is to now replan everything forward taking jerk into account to make a new motion waveform.
  After the first move the top speed was going to be 100, but now, it knows the increased distance
so the move ends up doing the full F300 commanded. So if you have complex code say a wiggly
contour of something, the more lines it see's , the higher the end speed of all of them as
Auggie can create a very long, complex motion waveform that takes into account all the necessary
blending and added curvatures.
  (there is a statistics button that switches the toolpath to an oscilloscope view of the jerk waveforms.
It wont make much sense to you normally, but if you switch it on and command a G1X5 or something
youll see the jerk waveform of the moves if you wish to play with what your system is doing
jerkwise and how moves blend. )

  My 10 watt laser ( cartesion table) did the following cut .. as an example of what a 10 watt can do
with all corrections and doing photo engraving.

Art

Attachments
Laser10Watt.jpg
Roman
Old Timer
Posts: 31
Joined: Wed Nov 06, 2019 4:07 am

Re: Almost there: Laser reacts strange around 1% power (and traps for beginners)

Post by Roman »

Hello Art
Thanks a lot for your insight and help. It works and I am tuning.. tuning ... tuning and get better results every time.
Therfore I paid for a licence today as announced to appreciate the dev's efforts and the help.
Remaining guestions:
- I can not set the Lookahead above 150, the system resets it immediately to 150?
- Is the button 'Image Dither' performing a real time dithering of the picture code generated with Augs? Or how has this to be used?

Thanks a lot
Best
Roman
User avatar
ArtF
Global Moderator
Global Moderator
Posts: 4586
Joined: Sun Sep 05, 2010 6:14 am
Contact:

Re: Almost there: Laser reacts strange around 1% power (and traps for beginners)

Post by ArtF »

Hi Roman:

>>- I can not set the Lookahead above 150, the system resets it immediately to 150?

  My apologies for that. I will look to see if I can find a reason why. I set mine commonly to 500
or more if the segment size is small. (Basically you set it higher as you encounter those Mach3 Gcode
type of files that caused Mach3 to stutter move as it does a contour, or move very slowly. A contour
in mach3 , if it is sufficiently small segment and changing direction constantly causes Mach3 ( and Auggie)
to do that countour slowly as it has to change direction at every segment. Mach3 adds segments together
but if it can only see 100 segments, that may encompass only a couple inches. ( or less), it slows things
down or makes it stutter. Auggie replans on very additional G1, so after it processes a couple hundred
(if the Jerk limit is suffiently high) the top speed gets much higher for that line. Auggie makes the line
internally to be like a complex wave for motion, so the top speed gets much higher with high lookahead.
  Its possible I have a debg in there that allows me to raise it higher than you, if so I will post a
fix for that. I find 500 works very well for me in small segment code.
  This is the end is a complex dance between JerkLimit and lookahead, so its important you have
that upped lookahead for those times you need it, and a high enough jerk setting for that
run. Ill fix this as soon as possible.


> Is the button 'Image Dither' performing a real time dithering of the picture code
generated with Augs? Or how has this to be used?

  Image dither is a test mode I was playing with, its function may change but its use
was intended to cut down on the burning of wood in slower code with lots of corners and such.
One of my largest complaints with lasers in CNC is the tendency to burn in corners
due to the slow speeds it much go down to inorder to make tight corners. Dither
was turned off recently as I added PWMMin as a method to help with nonlinearity
in laser power. The augment mode now automatically adjusts for the slow speed
by averaging the distance in the current millisecond vs the distance you would have
been moving if at full speed and setting laser power to that % of power. So if
your moving at 100%, you get 100% of the sliders power setting, 50% , you get 50%.
  Now this can be a problem at very slow speeds as you get no burn in the corner
as it goes too slow, say .1% and your laser at .1% has no burning power.

  So in the setting there is a PWMMin(power) setting for that that tells if as long as its moving
if the % power is to be set below that PWMIN(power), set it to PWMMIN(Power), so a burn occurs.
This setting then shoudl be set at the PWM under which you are able to just see an effect on your
material.

  This all has the effect of a linear regression to the first order for the power curve
as that PWMMin is in the scaling formula for all other power scaling during a variable
power burn.

Dither was an attempt to bring it to a second order regression, but has been turned
off till I find a better way. So ignore it, Ill post if I add its proper capability. But I can
say that all my tests have been great so Im not sure its needed. That , however, may
not be true for everyone. I have 2 lasers at different powers running under Auggie
and just because they work well, doesn't mean everyone's will. Linearity can vary,
but I haven't seen one vary so highly that a simple first order regression isn't enough.

  If your into laser engraving, have lower power or wish to try 3d, I suggest you do
some experiments with 3d engraving. It can help even in 2d really, so setting 3d
multipass with a shallow depth can help with photos as well. Get used to the extra
menu under the special DRO's tab which has DRO's for pass, number of passes,
and Z motion per pass. This mode doesnt simple repeat each pass, it calculates
the total laser energy per pixel and ensures it totals to the correct grey scaled
average per pixel of density. SO if you request 5 passes, dont be surprised if
it looks very bad and strange in pass 1, 2, or 3, the average has to build. You set
your power so the first pass doesnt burn too much( or at all), as the "darkest black" in the
image will get full slider power each of the 5 passes. The 127 ( or middle grey )
will get full power for 2 passes, and then 1/2 power on the 3rd pass and no power
on passes 4 or 5. ( for a total of 1/2 slide power over 5 passes. ). My laser has enough
power I can set it to a point where it doesnt burn at full speed, but removes a
thin layer of wood cleanly, so in multi-pass it just gets deeper each pass where
it must.

Sorry for the verbosity, figured Id do an archival explanation of this function,
I've found it the most fun to play with as it makes my 10 watt laser put out
engravings I've only seen 100 watt lasers to be capable of. There seems to exist
in most lasers a speed at which a full power output will not burn, but only remove
wood or other material cleanly. Finding that speed I have found to be a great
move forward in doing 3d depth engravings. In normal black/white photos you want
to find a speed/power where full speed at full power leaves a black line. To find it
I simple use a test piece of wood, set my slider, do a single line MDI of G1X100
and see if it burned, set the slider for power to just before it does.

  When I find that speed in the material I want to 3d burn, I measure the depth
of the removed material from that test, that is the depth per pass you will request
for auggie to put out the 3d Gcode. Multiply that depth by the total depth you wish
to get and it works amazingly well in many more linear materials.  So some lasers
may have a .1mm per pass, where I use up to 1-2mm per pass on mine. Get it right
and youll get the same engraving I get at 50 watts, just more passes if
lower power or fewer if higher. 

    I just tested the 150 limit, and sure enough the release has one implemented. I
will remove it and notify when the update is ready.

Art

User avatar
ArtF
Global Moderator
Global Moderator
Posts: 4586
Joined: Sun Sep 05, 2010 6:14 am
Contact:

Re: Almost there: Laser reacts strange around 1% power (and traps for beginners)

Post by ArtF »

Hi:

  150 limit has been updated to 1000. Keep in mind though that as a replanner, Auggie can take a lot of CPU to keep
up if you use too much. I find I can use 500 with no trouble. The updated version is now online.

Art
Roman
Old Timer
Posts: 31
Joined: Wed Nov 06, 2019 4:07 am

Re: Almost there: Laser reacts strange around 1% power (and traps for beginners)

Post by Roman »

Hi Art
I installed the new version. If one tries to load GCode then Auggie program does nothing for about 10 seconds and then stops, restarts and shows the profile selection window.
There must be a new bug.  :o
Roman
Old Timer
Posts: 31
Joined: Wed Nov 06, 2019 4:07 am

Re: Almost there: Laser reacts strange around 1% power (and traps for beginners)

Post by Roman »

FYI ... now installed last version after saving away Profiles folder. Moved Profiles folder back. It works, GCode can be imported. Preview limit back to 150 max-
User avatar
ArtF
Global Moderator
Global Moderator
Posts: 4586
Joined: Sun Sep 05, 2010 6:14 am
Contact:

Re: Almost there: Laser reacts strange around 1% power (and traps for beginners)

Post by ArtF »

Roman:

  Sorry, I for some reason didnt get notified of this message, Ill fix that up immediately..

Thx
Art
User avatar
ArtF
Global Moderator
Global Moderator
Posts: 4586
Joined: Sun Sep 05, 2010 6:14 am
Contact:

Re: Almost there: Laser reacts strange around 1% power (and traps for beginners)

Post by ArtF »

Roman:

Can you try that again? I just installed the latest installer, then ran the release version and loaded a
272K GCode file with lookahead set to 300. Seemed to load quickly and fine.

  Check your lookahead, perhaps the switch set it to a non number in error or something..

Art
Roman
Old Timer
Posts: 31
Joined: Wed Nov 06, 2019 4:07 am

Re: Almost there: Laser reacts strange around 1% power (and traps for beginners)

Post by Roman »

Hello Art
Will be back in a week or so only. Can not test now as I am 'on the road' (Not because of Thanksgiving).
Happy Thanksgiving!
User avatar
ArtF
Global Moderator
Global Moderator
Posts: 4586
Joined: Sun Sep 05, 2010 6:14 am
Contact:

Re: Almost there: Laser reacts strange around 1% power (and traps for beginners)

Post by ArtF »

Same to You.
:)

Art
Roman
Old Timer
Posts: 31
Joined: Wed Nov 06, 2019 4:07 am

Re: Almost there: Laser reacts strange around 1% power (and traps for beginners)

Post by Roman »

Hello Art
I tried it but no effect: 150 lines maximum.
The only idea I have now is to erase it totally and start a clean install from scratch.
Will be back.
Best regards
Roman
User avatar
ArtF
Global Moderator
Global Moderator
Posts: 4586
Joined: Sun Sep 05, 2010 6:14 am
Contact:

Re: Almost there: Laser reacts strange around 1% power (and traps for beginners)

Post by ArtF »

Roman:

Let me know how it works out , and what version the installer claims its installing...

Art
Post Reply

Who is online

Users browsing this forum: No registered users and 3 guests