calibration step-by-step

[Mark Van den Borre]

Hi list,

Calibrating our lasersaur was a really frustrating experience. While
ours is not 100% perfectly calibrated yet, we got it into a relatively
usable state some time ago after several days of attempts.

This is more or less the useful part of what we did:
https://github.com/nortd/lasersaur/wiki/Calibrating-mirrors-and-lens:-step-by-step

I started documenting this in excruciating detail. Making it an almost
brainless, ticking a list experience should help. It takes a lot of
time, but it almost guarantees you a somewhat working lasersaur
afterwards.

Really tempted to get a high qualtiy red dot laser and a beam combiner
though when this laser tube needs to be recalibrated...

Kind regards,

Mark
--
Mark Van den Borre
Hogestraat 16
3000 Leuven, België
+32 486 961726

[Ray Debs]

I have been thinking that a nice feature to have would be a beam power compensation over the area of the bed. This could be as simple as this:

output_power = file_specified_power * (((xpos + ypos) / (xmax + ymax)) * scale_factor) + 1.0)

The scale_factor can be adjusted from 0-1 to provide equal power at a point (0,0) as at a point (xmax,ymax). This should produce an output power of 100% at the farthest point, and something less everywhere else, decreasing as the beam path gets shorter. This assumes a somewhat linear reduction in power with distance, which may not be true, but it may be a good enough approximation. I would probably start calibrating with a small scale_factor of around 0.05, then adjust as necessary. The measurement of power can be dome by shooting a short burst into an acrylic block and measuring the relative depth of the crater in the min and max positions.

Of course, this would mean varying the power while moving the cart, not just at the start of a segment.

This is relevant to the discussion about calibration because a final step would be added to measure power at both locations so the scalefactor could be determined.

The math is off the top of my head, so I may have messed it up, but I hope I can get the general idea across.

What do you think?

Ray

Sent from my iPad

> --
> You received this message because you are subscribed to the Google Groups "lasersaur" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to lasersaur+...@googlegroups.com.
> To post to this group, send email to lase...@googlegroups.com.
> Visit this group at http://groups.google.com/group/lasersaur.
> To view this discussion on the web visit https://groups.google.com/d/msgid/lasersaur/CAEoHNyB8Z5q0iiE2zA-YMutZODHWRNgGtmXfku661_TrGxwQwQ%40mail.gmail.com.
> For more options, visit https://groups.google.com/d/optout.

[Mark Van den Borre]

Hi Ray,

2015-02-24 14:13 GMT+01:00 'Ray Debs' via lasersaur
<lase...@googlegroups.com>:

> I have been thinking that a nice feature to have would be a beam power compensation over the area of the bed. This could be as simple as this:
>
> output_power = file_specified_power * (((xpos + ypos) / (xmax + ymax)) * scale_factor) + 1.0)

That looks like a great idea! Right now, our beam power really decays
_a lot_ at (1200,600). We started at (0,0) with 7 or 8mm and got no
visible marks at (1200,600) under about 28-30mm 15% power and
2000mm/min speed...

I have added it to https://github.com/nortd/lasersaur/issues/4 for
future reference.

Kind regards,

Mark

[Steve Baker]

Getting the thing aligned the first time is a major pain - but once it's
roughly aligned, it's much easier because you can just adjust at the four
corners. Most of the time, I only need to realign after I replace a
mirror, or if one of the Y cart belts or the Y-axis drive shaft slips or
has to be removed for some reason. In those cases, it's really only the
rotation of the mirrors that has to be adjusted - not the position of the
mount itself.

The problem with that is that the design of the lasersaur's mirror mounts
pretty much guarantees that any time you loosen the mirror mount bolt to
reposition it, you'll screw up the rotation.

So in practice, you want to get the position of the mirror mount itself
sorted out first - and only then start messing with the rotation...but you
have to get the rotation right in order to figure out how much you still
need to move the mirror mount by. This is really frustrating - but once
your Y-axis motion is working without the belt skipping, you'll probably
only have to do it if you ever move your lasersaur - or after a few years
when your laser tube needs replacing.

I use laser-cut plywood targets that are the same diameter as the mirrors
and which are etched with three concentric circles and a cross-hair with
an arrow showing which way is "up". Sadly, you need a working laser
cutter in order to make those targets! Doing this saves a TON of time.

I've attached the SVG file I used to make a bucketload of them from a 30cm
x 30cm sheet of plywood. Remember to make a new batch as the first thing
you do after you've realigned your laser!

If you want an utterly mindless alignment algorithm, then it's this (sorry
for the pseudo C code!):

int mirror ;
double Xnear, Ynear, Xfar, Yfar ;

void measure ()
{
/* Measure the position of the laser spot both close to, and far from
the mirror */

move the target as close to the mirror that you're adjusting as
possible ;

zap the laser and adjust the power until you can just see a clear
circular spot ;
measure where the spot hits the target in the X and Y directions ;
those distances are Xnear and Ynear ;

move the target as far from the mirror that you're adjusting as
possible ;

while ( true )
{
zap the laser (you may need a little more power than before) ;

if ( the laser leaves a mark )
{
measure where the spot hits the target in the X and Y directions ;
those distances are Xfar and Yfar ;
break ;
}
else /* You're a long way off-alignment! */
move the target closer to the mirror ;
}
}

void main ()
{
for ( mirror = 1 ; mirror <= 3 ; mirror++ )
{
do
{
do
{
measure () ;

if ( Xnear >= Xfar ) slide the mirror horizontally by -Xnear ;
if ( Ynear >= Yfar ) slide the mirror horizontally by -Ynear ;

} while ( Xnear >= Xfar || Ynear >= Yfar ) ;

do
{
measure () ;

if ( Xnear < Xfar ) then rotate the mirror horizontally until
Xnear==Xfar ;
if ( Ynear < Yfar ) then rotate the mirror vertically Ynear==Yfar ;

} while ( Xnear != Xfar || Ynear != Yfar ) ;

} while ( Xfar != 0 || Yfar != 0 || Xnear != 0 || Ynear != 0 ) ;
}

/* Hooray! */
}

...well, that's not quite right because X and Y are signed numbers...but
you get the idea.

The trick there is to understand that if you slide the mirror by (say)
5mm, then the spot will move by 5mm on the target no matter how far it is
from the mirror - but when you rotate the mirror, then the distance that
the spot will move by will be larger at greater distances.

-- Steve

> 3000 Leuven, België

> +32 486 961726
>
> --
> You received this message because you are subscribed to the Google Groups
> "lasersaur" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to lasersaur+...@googlegroups.com.
> To post to this group, send email to lase...@googlegroups.com.
> Visit this group at http://groups.google.com/group/lasersaur.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/lasersaur/CAEoHNyB8Z5q0iiE2zA-YMutZODHWRNgGtmXfku661_TrGxwQwQ%40mail.gmail.com.
> For more options, visit https://groups.google.com/d/optout.
>

-- Steve

[Mark Van den Borre]

Steve,

> The trick there is to understand that if you slide the mirror by (say)
> 5mm, then the spot will move by 5mm on the target no matter how far it is
> from the mirror - but when you rotate the mirror, then the distance that
> the spot will move by will be larger at greater distances.

This is exactly the explanation I have failed to read anywhere until
now. Thank you! It's obvious if you know it, but not so much if you
don't. It really helps to have it explained as eloquently as you just
did. I will add it to the English language written out calibration
algorithm I dumped in the wiki.

Kind regards,

Mark


--
Mark Van den Borre
Hogestraat 16

3000 Leuven, België
+32 486 961726

[hepfive]

When I aligned for the first time the lasersaur I used your target, Steve. I printed out and laminated the targets and cute all with scissor lol.
The first thing I cute with laser was this targets :)

[Steve Baker]

I've been thinking about that too.

The trouble is that the adjustment numbers vary depending on how much
smoke there is along the beam path...which depends on what you're cutting,
how fast, whether you're cutting all the way through with air assist blowing
the smoke through the kerf or just etching.

It even depends on the direction the head is moving (going in the direction
of the airflow leaves smoke in the wake of the head, going in the opposite
direction means that there is generally clear air.

I've quite often found that the west-to-east direction cuts have failed when
the east-to-west cuts have come out marginally OK.

I'm also not convinced that the Y axis should play much part in the
compensation math because (at least on my machines) I pull the smoke
away from the end of the machine where the beam is traveling in the Y
direction - so the air along that path is pretty clear.

When we have things out of whack (poor focus, dirty optics, dying laser tube,
warped material), it's generally the right-hand end of the machine
that cuts poorly, not the front edge.

That said, I think a small amount of operator-adjustable compensation
might be worth having.

HOWEVER: What's mostly a problem is failure to cut all the way through
the material at 100% power that bothers me. Subtle etch variations across
the bed at 30 to 50% power aren't even noticeable.

So what we need is not beam POWER adjustment, it's SPEED. Ideally,
slowing the head down a little on X==big side of the machine.

(That said, my second lasersaur is a mirror image of the standard design, so
I need more cutting performance on the left side of the bed!).

But it's tougher when people vent smoke in different directions...and the
venting system isn't a part of the standard design, so this isn't a simple
matter.

-- Steve

>> 3000 Leuven, België

>> +32 486 961726
>>
>> --
>> You received this message because you are subscribed to the Google
>> Groups "lasersaur" group.
>> To unsubscribe from this group and stop receiving emails from it, send
>> an email to lasersaur+...@googlegroups.com.
>> To post to this group, send email to lase...@googlegroups.com.
>> Visit this group at http://groups.google.com/group/lasersaur.
>> To view this discussion on the web visit
>> https://groups.google.com/d/msgid/lasersaur/CAEoHNyB8Z5q0iiE2zA-YMutZODHWRNgGtmXfku661_TrGxwQwQ%40mail.gmail.com.
>> For more options, visit https://groups.google.com/d/optout.
>
> --
> You received this message because you are subscribed to the Google Groups
> "lasersaur" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to lasersaur+...@googlegroups.com.
> To post to this group, send email to lase...@googlegroups.com.
> Visit this group at http://groups.google.com/group/lasersaur.
> To view this discussion on the web visit

> https://groups.google.com/d/msgid/lasersaur/9004EAED-6BEA-49DF-BD25-4FFA720001E5%40aol.com.

> For more options, visit https://groups.google.com/d/optout.
>

-- Steve

[Mark Van den Borre]

Steve,

There's something I don't understand 100% about the mirrors and
targets. If you place your mirrors in the nortd labs holders as we
did, they are approximately 3.5mm deep in the holder. According to my
old friend Pythagoras, that means if I put the targets flush with the
thop of the holder, I need to aim for the M1 and M2 targets 3.5mm
right of center, and for M3 3.5mm above center to reach the center of
the actual mirror. I can hardly imagine you take out the mirrors and
replace them by targets, because that would cause serious deviation,
right?

How do you deal with this?

Kind regards,

Mark

> To view this discussion on the web visit https://groups.google.com/d/msgid/lasersaur/0292ba6de6124183ef4415eaaff1faa2.squirrel%40webmail.sjbaker.org.

> For more options, visit https://groups.google.com/d/optout.



--

Mark Van den Borre
Hogestraat 16

3000 Leuven, België
+32 486 961726

[Steve Baker]

If your power is decaying that badly across the bed then either:

1) Your smoke extraction system is performing badly.

...or...

2) Your beam alignment is off and the beam is just clipping the edge of
the nozzle at those far distances.

If either of those things are your problem, then this kind of compensation
would merely hide a much more serious issue!

There is always going to be some small variation in performance - but it
shouldn't be that noticeable when the system is working normally.

-- Steve


Mark Van den Borre wrote:

> --
> You received this message because you are subscribed to the Google Groups
> "lasersaur" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to lasersaur+...@googlegroups.com.
> To post to this group, send email to lase...@googlegroups.com.
> Visit this group at http://groups.google.com/group/lasersaur.
> To view this discussion on the web visit

> https://groups.google.com/d/msgid/lasersaur/CAEoHNyB28ZoL4u31B%2BHd7O67Y91nJV5UbY%3DauMC9vTRfLdoWYQ%40mail.gmail.com.

> For more options, visit https://groups.google.com/d/optout.
>

-- Steve

[hepfive]

The final adjustment I used the mirrors, the targets I used just to have one ideia, some people use tape in the holders...

[Mark Van den Borre]

Hi Steve,

2015-02-24 16:53 GMT+01:00 Steve Baker <st...@sjbaker.org>:
> If your power is decaying that badly across the bed then either:
>
> 1) Your smoke extraction system is performing badly.

This is while testing _before_ smoke extraction. Just a 15%,
2000mm/min pulse for 7mm makes a visible dot on the paper masking tape
covering M3 at (0,0) , and we need ~30mm at 15% 2000mm/min at
(1200,600). I only activated smoke extraction after basic calibration.
When working with ply, we are using a "slightly" overkill dual Soler &
Palau industrial venting system with plenty of power and only 1m to
bridge to the outside through adequately wide tubes, in an arrangement
as suggested by you and other: pull in air in the middle of the right
side panel, suck it out from the left side panel (two holes in our
case).

> ...or...
>
> 2) Your beam alignment is off and the beam is just clipping the edge of
> the nozzle at those far distances.

This is even before going through the nozzle! The beam is centered
within 1mm of the center of the lens on top of the nozzle.

Mark

> To view this discussion on the web visit https://groups.google.com/d/msgid/lasersaur/a3dedb8bf458d8e46beb14c892ff6b6d.squirrel%40webmail.sjbaker.org.

> For more options, visit https://groups.google.com/d/optout.



--

[Steve Baker]

No, I take the mirror out of the holder and use the wooden target instead.

Remember that you're going to adjust the #1 mirror completely (with the
wooden target in the #2 holder), then - only when you're hitting the center
of the target 100% perfectly do you put a mirror into the #2 lens holder
and put a wooden target into the #3 lens holder...and then when that's done,
I unscrew the nozzle, remove the lens and tape a wooden target at the
bottom of the lens
tube to adjust the #3 mirror.

I certainly wouldn't place paper or tape over the mirror because of the risk
of scratching the very delicate surface, or getting smoke and/or ash from the
burning paper contaminating the mirror's surface...so even with the paper
tape
approach, you have to take the mirror out of the holder!

Furthermore (as you say) you have to mentally compensate for the plane of the
paper target not being exactly where the mirror's surface is.

My wooden targets (which I cut from 1/8" or 3.5mm plywood) are almost the
exact same size and thickness as an actual mirror. They fit into the holder
perfectly and can be held in place with the regular set-screw.

In theory, the plane of the wooden target might not be PERFECTLY where the
mirror will ultimately be - but that's OK because I don't put the mirror
in there
until the beam is hitting the wooden target dead center over the entire
distance
of travel. Once I've re-installed that mirror, I'm not taking it out
again until the
next time I have to realign my machine!

I used to use the painters' tape approach, but when I had the idea to cut
wooden
targets, life got a million times easier. Having the concentric circles
and the
cross-hairs etched into them means that there is no marking of the edges of
the lens holder, no measuring of the distance to the center - and no
problem of
forgetting which way up the tape was when stuck onto the holder!

Trust me, I've been aligning my two lasersaurs once a month or so for two
years,
and the wooden targets are VASTLY easier than the alternatives.

>>> 3000 Leuven, België

> https://groups.google.com/d/msgid/lasersaur/CAEoHNyDb2rzKPNJtKSEBODkdQi4Yt6hcd9qYHf7ygRpSTFSAmQ%40mail.gmail.com.

[Mark Van den Borre]

Hi Steve!

Ok, this makes a lot of sense! Thank you for your suggestions. Our
cutter is cutting sufficiently well to do those. That should improve
our calibration another little bit. I'll update the wiki with those
instructions, so that others can benefit from it too.

Kind regards,

Mark

> To view this discussion on the web visit https://groups.google.com/d/msgid/lasersaur/1187c8ab1dcd0d82ec06df33e55780d6.squirrel%40webmail.sjbaker.org.

> For more options, visit https://groups.google.com/d/optout.



--
Mark Van den Borre
Hogestraat 16

3000 Leuven, België
+32 486 961726

[Steve Baker]

The fact that the beam is going through the center of the lens tells you
almost nothing about where it's going to hit the nozzle. Remember
that until you get the laser to emerge from that tiny hole it might
not be heading vertically downwards.

One saving grace here is that the lens does tend to focus light
inwards, so a small positional error where the beam hits the
lens is tolerable PROVIDING it's travelling vertically downwards
at the time!

But you're right, for a quick zap, smoke in the path is unlikely
to be an issue, so the vent system shouldn't matter much.

In practical use, you should generally be cutting materials like wood
and acrylic at 100% power and cranking the speed up as high as
it will go with you still cutting through the material.

For etching, and for thin materials like cloth and paper, you need
to work at full speed and adjust the power level to get the effect
you need.

But small variations in power over the bed really don't affect the
etch/thin-materials performance where you have plenty of power.

The problem is with cutting at the extreme right-hand end where
you really need to be cutting more slowly than over on the left side
of the machine.

I don't do much acrylic work - but with wood, the problem is that you
can't crank the speed up to right at the point where it only just
cuts the material - because wood varies in density with the grain
and things like glue variation and interior knot holes with plywood.

So running within even 10% of the maximum possible cutting speed
is a bad idea - you'll get all sorts of patchy cutting because of the
inconsistancies of the material.

Compared to that margin of safety, the variation of laser power across
the bed of a properly aligned and well-ventilated machine is really
kinda negligible.

Our two laser cutters are run for 10 hours a day, 5 or 6 days a week
and we've come to know them VERY well.

At the settings we use, the variation in performance across the bed is
negligible compared to the problems of cutting through variable
material density and the interaction of the cut direction and the airflow.

We do, on occasions, see that we're getting poor cutting at the right hand
end of the machine. When that happens, we check our air filters, clean
lens and mirrors and check alignment and focus. When everything is
clean and aligned, the problem goes away.

This issue is simply telling you that your machine needs work.

So "fixing" this with a software adjustment is at best a third-order issue.

It *might* allow me to drive my lasersaur's cutting 5% faster than we do
now. Sadly, because the head has to accelerate through curves and corners,
that wouldn't translate into 5% more productivity...probably more like 1%
to 2%.

It's certainly not a magic bullet for fixing the issues you're seeing.

> 3000 Leuven, België
> +32 486 961726

>
> --
> You received this message because you are subscribed to the Google Groups
> "lasersaur" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to lasersaur+...@googlegroups.com.
> To post to this group, send email to lase...@googlegroups.com.
> Visit this group at http://groups.google.com/group/lasersaur.
> To view this discussion on the web visit

> https://groups.google.com/d/msgid/lasersaur/CAEoHNyDB7QYGKEtPU2GfZ%3D0RZmnpGTDpi5_%3DYPoWiX_PVrWBzw%40mail.gmail.com.

[Steve Baker]

Did you see the alignment advice document that I wrote there?

If you have improvements or clarifications, it would probably be better to
keep them all together in one place rather than confusing people with two
separate procedures...but beware, if you're doing this for the first time,
it's probably not wise to be leaving advice for others to follow!

>>>>> 3000 Leuven, België

> https://groups.google.com/d/msgid/lasersaur/CAEoHNyAaA2YOCxt7e_41w5-WUss%3DK-s_vjmxkfATv5fWVcXSZQ%40mail.gmail.com.

[Mark Van den Borre]

Hello Steve,

2015-02-24 17:27 GMT+01:00 Steve Baker <st...@sjbaker.org>:
> Did you see the alignment advice document that I wrote there?

Yes. I actually pasted your advice there and explained so on the
mailing list some time ago.

> If you have improvements or clarifications, it would probably be better to
> keep them all together in one place rather than confusing people with two
> separate procedures...but beware, if you're doing this for the first time,
> it's probably not wise to be leaving advice for others to follow!

I'll rework the info, label it "work in progress" and prominently
point people at the other document. I'm trying to listen very
carefully to all knowledge you (and other list members!) built up, and
make it into a more of a step-by-step calibration guide. So the focus
of that work-in-progress document is slightly different. I'd rather
put it there in the open than in a private document that benefits
noone...

Mark

> To view this discussion on the web visit https://groups.google.com/d/msgid/lasersaur/38174c98db8c3de76fe65a709815da4c.squirrel%40webmail.sjbaker.org.

> For more options, visit https://groups.google.com/d/optout.



--
Mark Van den Borre
Hogestraat 16

3000 Leuven, België
+32 486 961726

[Steve Baker]

Wow! That's very enthusiastic of you!

It occurs to me that I should probably be selling a bags of 100 targets
to new lasersaur owners from my website!

:-)

Maybe we can persuade Stefan to cut some and provide them with the
lasersaur parts he sells at the Nordt labs site? That would make more
sense.

-- Steve

> --
> You received this message because you are subscribed to the Google Groups
> "lasersaur" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to lasersaur+...@googlegroups.com.
> To post to this group, send email to lase...@googlegroups.com.
> Visit this group at http://groups.google.com/group/lasersaur.
> To view this discussion on the web visit

> https://groups.google.com/d/msgid/lasersaur/00e8f5ee-09a9-4a2e-9c91-3fa111e4531d%40googlegroups.com.

[Mark Van den Borre]

Hey Steve,

2015-02-24 17:22 GMT+01:00 Steve Baker <st...@sjbaker.org>:
> The fact that the beam is going through the center of the lens tells you
> almost nothing about where it's going to hit the nozzle. Remember
> that until you get the laser to emerge from that tiny hole it might
> not be heading vertically downwards.

I caught that as _the_ major issue we still need to improve. Thank
you. I'm sure it will help a lot!

[...]
I cut the other useful usage hints. Thanks for those though! I might
develop them into components for a lasersaur user manual.

Mark