Is it always BR that is over 12mm?
One thing you can try is lowering the retract value to 200 then extend, retract (which it won’t do as the value is too low) then extend again. repeat until belt is fully extended. Reset your Retract value to it’s normal value and retract again while keeping pressure on the belt. What this will do is repack your belt. Inspect the belt for any sign of damage to the teeth.
What is the retract value? Can you put a copy of the maslow.yaml file here?
I had a similar problem, and it turned out to be a damage spool cog which would mean pulling the BR arm apart to inspect it. We will try everything else first.
The Maslow seems to be more accurate if the retraction value is as low as possible when Find Anchors is run. This may help, but I don’t think its your problem.
What you can try here is set retract force to 700, extend and retract. If all belts retract, excellent! Otherwise increase it by 100 and try again. What we are aiming for is the lowest possible value where Maslow retracts all belts consistently.
I haven’t cut anything thicker than 12 mm (accept the spoil board when the Z went nuts, but that’s another story) and I haven’t bothered with setting work piece height or spoil board thickness, but my frame is extended 4.5m x 4m and the Anchor Points are attached to the walls and set at the same height as the arms so it hasn’t been needed yet.
I note from your maslow.yaml file that the Z values are non-standard.
If you have raised these to get closer to the height of the arms, it is important that they have no movement. Can you post a photo of the BR Anchor Point please.
tlZ: 53.500000
trZ: 41.500000
blZ: 19.500000
brZ: 63.500000
It’s getting late here (I’m in Australia) so I will be closing down soon.
I release tension to place materials under the Maslow - put them on the table and put the Maslow on top. This is when I would reset anything that needs resetting prior to tensioning.
I will try with lower force when back.
Yes my Anchor points are raised at different height, An I don’t think they move, I have film them as suggested by Dlang and it is stable
Frame is made of 2x4 fully covered with 5/8 ply
The rising piece is about 12” x 12” (minus the arc cut in it) adding 12.5mm in height
The steel triangle anchor (2 Steel corners welded together) is screwed in with 3 x 3 inch screws through the 2x4 and ply. It does not move
The spoil board is 18.4mm thick on top
Her is BR
BL
You are very very close to passing so something is just very slightly off.
What we’re looking for isn’t something majorly wrong, just very slightly.
Is this accounted for when setting the spoil board thickness? Setting the spoil board thickness to 18mm in the settings is telling the math that the bottom of the sled is 18mm above the anchor points, but looking at your setup it looks like the anchor points might be 12.5mm above the bottom of the sled, is that right? In that case you would want a spoil board thickness of -12.5mm
Syl wrote:
Yes my Anchor points are raised at different height, An I don’t think they move, I have film them as suggested by Dlang and it is stable
I think these Z height issues are a significant part of the problem.
if all your anchors are at the same height, you can use the spoilboard thickness
to adjust all of the Z offsets at once and leave the Z offsets in the
maslow.yaml at their default values
if your anchors are at different heights, then you should adjust the Z offsets
in the maslow.yaml and NOT set a spoilboard thickness value (the Z offsets you
set account for this)
having both modified the Z offset values and used the spoilboard thickness, you
are probably double correcting and causing issues.
the total of the Z offset for an arm + spoilboard thickness + workpiece
thickness = the height difference between the belt end when fully retracted on
the arm and the belt end attached to the anchor
the default Z offsets are:
tlZ: 100.000000
trZ: 56.000000
blZ: 34.000000
brZ: 78.000000
the fact that you have yours at:
tlZ: 53.500000
trZ: 41.500000
blZ: 19.500000
brZ: 63.500000
says that you have raised the anchors a fair bit. Which probably means you
should set your spoilboard thickness to zero
Frame is made of 2x4 fully covered with 5/8 ply
The rising piece is about 12” x 12” (minus the arc cut in it) adding 12.5mm in height
The steel triangle anchor (2 Steel corners welded together) is screwed in with 3 x 3 inch screws through the 2x4 and ply. It does not move
The spoil board is 18.4mm thick on top
this is confusing, 18mm on top of the surface that the rising piece is attached
to? or is the rising piece rising up from the height of the top of the
spoilboard?
an overall picture would help us visualize this (the pictures you post clarify
the anchors, but not how they relate to the spoilboard)
David Lang
Are all your anchors raised by the same amount? I think David may have got it right, you are compensating for anchor height twice, meaning the maths thinks the anchors are higher than they are.
or doing and undoing this. if the z offset for an arm was reduced by 13mm but
you say that you have a 18mm spoilboard, that would mean that the anchor is 5mm
below the level of the sled.
David Lang
Thank you for keeping the conversation going to find a solution.
I’m not ignoring you, as I mentioned, l m away for a week, and I will get the drawings and picture when getting back.
As for the different anchor heights, the TL anchor is higher then the rest. (To be closer in height to the highest arm on the cnc.)
Back in the days when building the frame, I did some testing while jogging toward the TLAnchor and it felt that raising that anchor made the sled a little less tippy
Will be reporting back on details when I get back
Thanks
1 Like
Hi,
I’m back with some sketches for the anchor heights,
First, for clarity, this is my understanding of the set up when using the Default Z
tlZ: 100.000000
trZ: 56.000000
blZ: 34.000000
brZ: 78.000000
With “Maslow_spoilboardThickness” field = 0 and the machine at the very bottom
Now this is my setup
For TLZ
So TLZ Value, would be 74mm-18.5mm (Spoilboard) = 55.5mm and
“Maslow_spoilboardThickness” field = 18.5mm
For TRZ
So TRZ Value, would be 60mm-18.5mm (Spoilboard) = 41.5mm and
“Maslow_spoilboardThickness” field = 18.5mm
For BLZ
So BLZ Value, would be 38mm-18.5mm (Spoilboard) = 19.5mm and
“Maslow_spoilboardThickness” field = 18.5mm
For BRZ
So BRZ Value, would be 82mm-18.5mm (Spoilboard) = 63.5mm and
“Maslow_spoilboardThickness” field = 18.5mm
so the config items would be adjusted to this:
tlZ: 55.5
trZ: 41.5
blZ: 19.5
brZ: 63.5
and set the “Maslow_spoilboardThickness” field = 18.5mm
This seem in line with the default Z offset of 22mm difference between heights (except TLZ at -8, But it is raise 30mm more so It’s OK)
This is consistant to what I had. (just the TLZ was missing 2mm, I probably forgot the steel corner height and the 0.1mm on the spoilboard). In any case I will adjust the 2mm for TLZ
And retry calibration tomorrow.
I just want to make sure that I have this right
Thanks
On a side note, when you roll over the blue information circle in the “config items” page for TLZ TRZ BLZ and BrZ, The info provided is not the same as the default Z offsets provided in the clean Yaml file
should be 100
Should be 56
Same goes for the other 2
I’m back with some sketches for the anchor heights,
First, for clarity, this is my understanding of the set up when using the Default Z
tlZ: 100.000000
trZ: 56.000000
blZ: 34.000000
brZ: 78.000000
With “Maslow_spoilboardThickness” field = 0 and the machine at the very bottom
correct
Now this is my setup
For TLZ
So TLZ Value, would be 74mm-18.5mm (Spoilboard) = 55.5mm and
“Maslow_spoilboardThickness” field = 18.5mm
you could do that, or you could leave spoilboard at 0 and just set the tlZ value
at 74. internally the maslow just adds them together. So if you think there is a
possibility that you would use different thicknesses of spoilboard, you can keep
it separate. If you plan to always use 18mm stock, you can put it all in the tlZ
value
I just want to make sure that I have this right
yep, you have the right idea. As a note, the flatter the belts are, the less any
Z errors matter (you just have to make sure the anchors and frame are not going
to flex from the leverage of the belts pulling higher up). Many of us raise the
anchors up so the Z offsets are near zero.
On a side note, when you roll over the blue information circle in the “config
items” page for TLZ TRZ BLZ and BrZ, The info provided is not the same as the
default Z offsets provided in the clean Yaml file
@bar this seems like a bug in the UI
David Lang
1 Like
Good catch, I assigned it to the AI here Fix hover text · Issue #928 · MaslowCNC/Maslow_4 · GitHub
So I didn’t understand this before. I haven’t tried raised anchors yet. The default config has z heights already that match the height of the arm in the machine? So if I did perfect raised anchors that were straight out sideways what would they look like?
1 Like
Excellent drawings
1 Like
That’s a great way to think of it. If you raise the anchor points so that all of the belts go straight out, you would set all of the vertical offsets to 0.
One thing to note is that 0 is defined when the z-axis is lowered all the way down.
1 Like
wouldchuck wrote:
So I didn’t understand this before. I haven’t tried raised anchors yet. The
default config has z heights already that match the height of the arm in the
machine? So if I did perfect raised anchors that were straight out sideways
what would they look like?
if you take the drawings and increase the thickness of the blocks until they are
straight out, the Z offset would be 0
the closer to flat they are, the less the error matters.
on a stock frame, the default Z height for the top arm as 100mm, when you are
closest to the corner to a stock frame, you are ~600mm away (convienient numbers
to use that are a plausible worst case)
you put a bit in the router that sticks out 2 in (50mm), so your expected height
is now 150mm. the arm and belt end are not supposed to flex (they do, which adds
error), and are ~150mm long, so there is ~450mm of belt to handle the 150mm of
height difference. This means the belt length is
l^2 = 450^2 + 150^2
l=474mm
rearranging this equasion to show the orizontal length
450 = sqrt(474^2 - 150^2) (actual result is 449.6 close enough for the example)
but if you add a 18mm spoilboard, now the 150 is really 168 and the resulting
horizontal distance is 443mm, 7mm shorter than expected (or for the same
position, the belt is longer than expected, and calibration calculations are
thrown off)
now, let’s say the belts close to horizontal, 10mm instead of 100mm
the correct calculation is
l^2 = 450^2 + 60^2
l= 454
450 = sqrt(464^2 - 60^2)
add the same 18mm spoilboard so it’s now 78mm and the result is 447mm, 3mm off
and if you had a shorter bit (20mm of stickout)
it would be
L = 451 and with the 18mm error the horizontal error would be 449.4mm
Besides just taller blocks like I mentioned above, see my bolt down anchor
thread for a 3d model of anchors that have different heights (and are
configurable so you can set them for what’s right for your exact setup,
including overcomplication like the ability to correct for an uneven floor in
the prints rather than in the Z offsets, which I probably should remove 
)
David Lang
2 Likes
Ok,
After doing everything I could to avoid disassembly and opening the arms, I finally did. Not being able to retract with a force of 800 to 900 on all arms convinced me!
Here is what I found, 3 of the rings ( I changed the 4th one not so long ago because it had 2 broken teeth) once “ungear” were really hard to turn by hand and full of black plastic dust/powder. Once clean, they were freely rotating. What I also found, was that 1 of those gears was about ready to break some teeth, and 1 had a belt stretch.
Safe to say that I probably had to much tension at some point in the past! Luckily, I had the rings in stock!
From there, I recalibrated with free turning rings and retraction force of 900. It worked well with 0.79 fitness, but after running test files,jogging around, releasing and applying tension, It gave me the Center point deviation over 12.000mm and was stuck in extended mode. So I recalibrated at 800 retraction force and got a 0.89 fitness and so far so good after running files, jogging around, releasing and applying tension. Hopefully, the over 12mm deviation will not come back!
Thanks David, Bar & Ian for your insights
Here is one of those project
Now I have to go install it !
4 Likes