I’m having difficulty that others have reported with jagged edges at the edges of the work area, followed by position error and stopping of the job. Consensus seems to be that the belts are getting too tight and I should recalibrate. That got me thinking:
What height should the Maslow be at for calibration? In Bar’s excellent videos the Maslow is all the way down with no router bit, and the calibration proceeds I followed that procedure.
But, from previous discussions I understand that Maslow has no absolute idea of its Z position. You basically have to tell it where Z=0 is. So when I put a bit in the router I then have to redefine 0 with the tip just off touching the job. How does the Maslow then know what adjustment to apply to its calibration settings which were carried out with the router all the way down? It adjusts for the spoil and job thickness (probably an inch or so) but not the tool length of approximately twice that.
I’m considering a pre-calibration step which is to zero the Z axis with the tool in, take the tool out and jog Z down to just off the bottom (if we let it hit the bottom the Z cal is lost) and then find anchor locations with the Z height showing -40mm (or whatever the length of my cutter is). Even if I don’t get the cutter back in exactly the right place I will be right to 1 mm or so, which should make no difference.
What height should the Maslow be at for calibration? In Bar¢s excellent videos the Maslow is all the way down with no router bit, and the calibration proceeds I followed that procedure.
yes, it must be down all the way. Starting calibration sets the Z-stop value
(Zm=0 standing for Z machine coordinates), telling it that the router is all the
way down
But, from previous discussions I understand that Maslow has no absolute idea
of its Z position. You basically have to tell it where Z=0 is. So when I put a
bit in the router I then have to redefine 0 with the tip just off touching the
job. How does the Maslow then know what adjustment to apply to its calibration
settings which were carried out with the router all the way down? It adjusts
for the spoil and job thickness (probably an inch or so) but not the tool
length of approximately twice that.
when you set Z-home (Z=0), the machine looks at it’s position relative to Z-stop
(all the way down) and sets the belt length based on the position relative to
Z-stop (which is also Zm-0
I¢m considering a pre-calibration step which is to zero the Z axis with the
tool in, take the tool out and jog Z down to just off the bottom (if we let it
hit the bottom the Z cal is lost) and then find anchor locations with the Z
height showing -40mm (or whatever the length of my cutter is). Even if I don¢t
get the cutter back in exactly the right place I will be right to 1 mm or so,
which should make no difference.
any time the machine crashes and you have to do the retract/extend/apply tension
cycle, or if you run the Z past the limits (all the way down or off the top),
you need to set Z-stop (in the setup → config popups) so that it knows where it
is.
as a note, the closer the belts are to flat, the less significant any error in
the Z height ends up being, but you have to make sure your anchors are solid
enough to not flex under tension.
David, thank you for that explanation. If Maslow gets itself into an unknown state I should therefore set the Z-stop with it all the way down before Zeroing the tool height?
The note on the tooltip for Maslow_calibration_grid_width says that it defines the distance between the edge of the grid and the anchors. I therefore set it to 450 mm which should imply that the grid extends over almost all 8’x4’ work area. However, in practice the calibration grid appeared much smaller than expected. Is the tool tip wrong, or have I just misunderstood? Does it mean the midpoint of the 4 anchors?
My understanding and experience suggest that setting the sizes for calibration X & Y just inside the maximum before the Maslow falls off the edge gives the best results, however I have no idea of the how the software computes internally. The values I listed are what I used and it has worked for me.
I found this explanation.
Automatically compute the recommended calibration grid size
[Maslow 4 aims to automatically compute the recommended calibration grid size based on the anchor point locations. The process involves taking six measurements to estimate anchor point locations, which are then used to generate a calibration grid. The grid width and height should be 50% of the width and height between the anchor points, and the grid size should be the least number of points that won’t trigger the “grid spacing too wide” error. Currently, users can manually enter the size and number of points for the calibration grid, but the system is working on an automatic generation feature. This feature is expected to improve the calibration process by reducing the need for manual input and providing more accurate results.]
You set the Z-Stop value from the Setup menu, then Set Z-Stop, when the Z is all the way down. You can do this without taking the bit out by raising the Maslow with a block, tall enough to prevent the bit touching the surface. This will also set the Z-Home to zero. When you set Z-Home with a bit in, take note of the difference between Z-Home and Z-Stop. You can then raise the Z to that value (from Z-STOP = 0) and Set Z-Home without having to do the full find surface again, I would suggest you do this after retraction and prior to Extend as it’s easier with the belts in.
When you do a Find Anchors it expects the Z to be right down and sets Z-Stop to zero, so if you do a Find Anchors with the Z raised it messes with the maths that tells the Maslow how much belt to extend etc.
While you are getting a feel for how the Maslow operates, I suggest you run some GCode without a bit in and the Z-Home set about 40mm up
David, thank you for that explanation. If Maslow gets itself into an unknown
state I should therefore set the Z-stop with it all the way down before
Zeroing the tool height?
The note on the tooltip for Maslow_calibration_grid_width says that it defines
the distance between the edge of the grid and the anchors. I therefore set it
to 450 mm which should imply that the grid extends over almost all 8¢x4¢ work
area. However, in practice the calibration grid appeared much smaller than
expected. Is the tool tip wrong, or have I just misunderstood? Does it mean
the midpoint of the 4 anchors?
no, the grid width was how wide the testing grid is. but in recent versions,
that should be set automatically. what firmware version are you running?
if you go to settings → test it will tell you the firmware version and the html
version. both should match
If the jagged edges looks like this it’s its probably the belt gears in the arms that are to tight. According to Bar this jagging cut can happen when two servos are counteracting each other like they are doing close to the edges. If you add up the friction of the belt gear to that the power needed to move the servos exceed the limit (4000) and are shout down to protect the electronics.
If you have this issue a recalibration won’t help. You need to get rid och the friction in the belt gear first.
As you can see I had the same issue, and have started the process of sanding down the inner side of the gear to remove the friction between the belt gear and the arm body. The first arm I tested had low friction on the lower arm half but when tested on the upper the belt gear was stuck.. I now need to reinstall it before I can say for sure that the issue is fixed but I’m hopeful.
David I’m using 1.20 on both counts, evidenced by the first screenshot. The second screenshot shows the message which describes /Maslow_calibration_grid_width_mm_X with the message “Define the distance from the anchor points to the corners of the calibration grid”. Should that be “distance from the centre point … ” (or center if you prefer)? That would be in line with your explanation I think.
@Dan_Nylen I’m definitely having this issue. Is there another thread where i can get details on a solution? I’m not exactly sure where the friction is occurring. I took it apart once and THOUGHT i knew but perhaps not because it is still happening. Any hints in the right direction would be greatly appreciated.
it is probably the spool not moving freely on each of the arm halves.
the original 4.1 arm/spool joint just didn’t have enough clearence, and it’s not
uncommon that when you bolt the two arm halves together, you bulge the plastic
out towards the spool a little bit (making it even tighter)
the fix is to use a good, high grit sandpaper to sand the inside of the spool
until it turns freely on both halves of the arm.
I just got my machine back together. Unfortunatly I still have some problem with jugging at the corner. But I think that it is better. I will do a new atempt tomorrow to see if I can get even less friction. Perhaps even lubricate it.
I was just thinking of that. Good to know. I will test to assemble the armes without the small loose gear tomorrow. Just to make sure that the screw tension aren’t the problem.
I just got my machine back together. Unfortunatly I still have some problem
with jugging at the corner.
There are a couple of things that can be a problem in the corners. depending on
the frame size, corners can be an issue, see http://lang.hm/maslow/maslow4_frame.html
but also, if there is a large Z distance from the arm to the anchor that will be
a problem if you get too close to the anchor. see if your problem is in all
corners, or if it’s worse with the upper arms than the lower arms.
