Thanks David.
How do I change the Z offset values to match the mounts? I have not tried that.
You could try a smaller grid size. I recently started having a similar issue. I had calibrated before and had good values and my frame hasn’t really changed, so I decided to try running the M4 with the lower values. It worked really well, but wasn’t perfect. Had a few cuts that were a bit wobbly (though that might have been sled friction), and, when it did another pass on some detail cuts, the two cuts (which should have lined up perfectly) had a tiny bit of deviation between them (almost like the material had shifted like 1/32" between the passes though I’m pretty sure that didn’t happen and most likely the machine had gotten slightly lost). The final product still looked really good even with the slight errors, but there definitely were some errors.
I’ve seen similar passing comments made about changing these offsets, but also not sure what settings these are or even what is recommended. I’m pretty sure I can find the settings to adjust, but what should we be measuring exactly (top of spoil to middle of belt end, floor to top of belt end?), and what is the baseline that we are offsetting/measuring against.
The z offset does make a difference, I have mine set all at zero as I put the arm height on my anchor points so my belts run parallel to waste board.
My first cal I had a fitness of 0.56 which I was a bit disappointed with so rechecked my anchor points and noticed that one was 5 mm low, fixed the error recalibrated and now have a fitness of 0.67 so it helps
Thanks guys. Where do I find the place to enter those offsets? What is the best way to measure those? While the belts are mounted or when retracted?
So grid size doesn’t determine the size you cut? Just the accuracy? Like on a 3d printer when you do a larger grid vs smaller grid. Larger grid is more accurate?
Good to know. So to confirm my understanding, we’re looking for the difference in height between the machine resting on the spoilboard and the center of the belt ends when inserted into the anchors (in other words, height of the center of the mounted belt ends minus the thickness of your spoilboard)? Since the M4 is resting on top of the material and the material is on top of the spoilboard, should these values also be adjusted between projects to compensate for the additional height of the cutting material (updated equation: height of center of the mounted belt ends minus (thickness of spoilboard plus thickness of cutting material))?
Lastly, what is positive and what is negative in terms of this offset? E.g., if the mounted belt ends are higher than the spoilboard top, is that a positive offset value or negative? Or, if the mounted belt ends are lower than the top of the spoilboard, is that a positive or negative offset?
TimmyZ wrote:
How do I change the Z offset values to match the mounts? I have not tried that.
first a note that it may be good to put down a wasteboard rather than
calibrating on concrete, the sled will probably slide better on the wasteboard.
The Z offset values are in the maslow.yaml file (available also through the
advanced settings tab in the same window you go to to upload new firmware)
These values are the Z axis distance from where your anchors are to the arm on
the sled. I don’t know how tall your anchors are, but they look tall enough to
be significant compared to the default (which would be if they were against the
floor)
David Lang
Andith wrote:
I’ve seen similar passing comments made about changing these offsets, but also
not sure what settings these are or even what is recommended. I’m pretty sure
I can find the settings to adjust, but what should we be measuring exactly
(top of spoil to middle of belt end, floor to top of belt end?), and what is
the baseline that we are offsetting/measuring against.
This is how much the belt is offset, so top of belt to top of belt, or bottom of
belt to bottom of belt, either one works.
The default values assume 3/4" wasteboard and 3/4" workpiece, with the bottom of
the anchors being even with the bottom of the wasteboard. on a full size frame,
working towards the center, you can be off an inch or so without it making a
noticable difference. But the more you are off, or the closer to the anchors you
get, the more it will matter.
David Lang
TimmyZ wrote:
Thanks guys. Where do I find the place to enter those offsets? What is the best way to measure those? While the belts are mounted or when retracted?
go to the window where you would upload the firmware and look at the tabs on
that popup, over to the right you have settings or advanced (don’t remember off
the top of my head), when you select that you will see lots of values, you will
be looking for Maslow_*Z values, one for each arm.
So grid size doesn¢t determine the size you cut? Just the accuracy? Like on a 3d printer when you do a larger grid vs smaller grid. Larger grid is more accurate?
No, the grid size is just telling the machine how big a grid to move in while
trying to calculate where the anchor points are. We think that the larger the
area the grid covers, the more accurate the calibration will be, UNLESS the grid
is large enough that two of the arms hit the uprights at the same time and
the belts and arms are no longer a straight line (i.e. staying inside the green
area of my frame size visualization http://lang.hm/maslow/maslow4_frame.html )
We think this because the wider the area you use for the calibration, the larger
any errors are compared to the total belt lengths, so the easier it is to find
and correct for them.
David Lang
As always, thanks, @dlang! You’re a wealth of knowledge and help! For my mounts, I should be able to raise each arm’s anchor point individually. I have the belt-pins go into a coupler on a threaded rod that goes down into concrete anchors, and I can increase where the coupler and pins sit by using a longer length of threaded rod (as well as nuts or washers as spacer beneath to lock the coupler in place).
Instead of fiddling with the offset values, do you think it would be better to try (or at least would achieve the same results) raising each of my couplers and pins so that each belt is parallel to the ground (and thus a 0 z-distance from the anchored-point of the belt to the respective arm)?
Andith wrote:
Good to know. So to confirm my understanding, we’re looking for the
difference in height between the machine resting on the spoilboard and the
center of the belt ends when inserted into the anchors (in other words, height
of the center of the mounted belt ends minus the thickness of your
spoilboard)? Since the M4 is resting on top of the material and the material
is on top of the spoilboard, should these values also be adjusted between
projects to compensate for the additional height of the cutting material
(updated equation: height of center of the mounted belt ends minus (thickness
of spoilboard plus thickness of cutting material))?
Ideally, it should be the difference in height between the center(or top or
bottom) of the belt at the anchor and the same point of the belt on the arm with
the Z axis all the way down, sitting on top of your spoilboard and workpiece.
Lastly, what is positive and what is negative in terms of this offset? E.g.,
if the mounted belt ends are higher than the spoilboard top, is that a
positive offset value or negative? Or, if the mounted belt ends are lower
than the top of the spoilboard, is that a positive or negative offset?
positive is the anchor lower than the arm.
you don’t want the anchor higher than the arm because that will tend to pull the
sled away from the workpiece.
David Lang
1 Like
I have A spoilboard i’m calibrating on. Here are some pics.
So does the offset need to be the length of the mount to make up for the gap if the belt end was just flush against the concrete floor?
Andith wrote:
As always, thanks, @dlang! You’re a wealth of knowledge and help! For my
mounts, I should be able to raise each arm’s anchor point individually. I
have the belt-pins go into a coupler on a threaded rod that goes down into
concrete anchors, and I can increase where the coupler and pins sit by using a
longer length of threaded rod (as well as nuts or washers as spacer beneath to
lock the coupler in place).Instead of fiddling with the offset values, do you think it would be better to
try (or at least would achieve the same results) raising each of my couplers
and pins so that each belt is parallel to the ground (and thus a 0 z-distance
from the anchored-point of the belt to the respective arm)?
you would still need to change the offset values from the defaults to 0
There hasn’t been enough testing yet, but I believe that having the anchors at
the height of the arms (assuming no flex in the anchors) will give the best
results, but I have no proof of this yet.
David Lang
1 Like
That just saved me a lot of typing and a better explanation than I would of gave.
I should note the reason I changed mine was due to my frame being on the small side just over 2 metres square, I’ve only just started playing around with it and not cut anything as yet as I’m waiting on a spindle to arrive plus I want to hard wire the encoders to control pcb and maybe move its location to the sled as opposed to the top of router
TimmyZ wrote:
So does the offset need to be the length of the mount to make up for the gap if the belt end was just flush against the concrete floor?
you need to reduce the current offset values by the length of the mount.
David Lang
Oh man. @Andith I was thinking the same thing about making each anchor point the same height as the belt. Should we be the testers? maybe I should get a proper calibration first? lol
Thanks David. So I’m subtracting 68mm the length of the mount from the existing setting?
For example, if the setting is 30 then I make it -38?
TimmyZ wrote:
Oh man. @Andith I was thinking the same thing about making each anchor point
the same height as the belt. Should we be the testers? maybe I should get a
proper calibration first? lol
to be a proper tester, you need to be able to compare the two options, so get
something working first (remember to save your maslow.yaml file as that contains
the calibration results)
with your anchors, how confident are you that you can replace them in the exact
same spot when you remove them? an error of 2 mm would make the theoretical best
possible calibration only have an fitness of 0.5
David Lang
TimmyZ wrote:
Thanks David. So I’m subtracting 68mm the length of the mount from the existing setting?
For example, if the setting is 30 then I make it -38?
Yes, but that would mean that as that belt is tight, it will be pulling the sled
UP away from the workpiece, which is probably not a good idea.
David Lang
I made the changes for the Z settings, and attempting calibration now. What’s the threshhold for a good fitness? Between what numbers?
TimmyZ wrote:
I made the changes for the Z settings, and attempting calibration now.
What’s the threshhold for a good fitness? Between what numbers?
fitness is 1/error so the higher the better. The default threshold is 0.5, so an
average error of up to 2mm in the anchor location.
David Lang