I was measuring my maslow today, and I think the default Z offsets are wildly wrong.
I measure 36mm from the bottom of the sled to the bottom of the first arm. The anchor is just over a mm higher.
but the lowest default Z offset is 34mm, which is not only slightly short (but it could have been my measurements) but it also leaves no room for a wasteboard or workpiece.
with the thinking being that the wasteboard is 3/4" and the workpiece is 3/4" that’s another 36mm to add, so the lower arm should be 72-75mm which would put the top arm pretty close to 140mm, not 100.
am I missing something obvious? or have we been strugling with bad offsets all along?
I don’t know. I think the offsets are inconsistent. I’ve been keeping track of my Z offsets because they seem to cause tertiary problems. I’m doing a test run of an 8" Yggdrasil relief with the router turned off and the z-stop set WAY above the project board. I’m 62.5mm between the bottom of the first router clamp and the top of a z-axis motor. My cut will move the z-axis down to 52.0mm(same measurement). I haven’t cut anything with my M4 yet. However, I’ve done some dry runs. I noticed that if I don’t have my zhome/zstop set at a fair height, I can run into a 15mm error. 15mm Position Error While Moving Investigation - #19 by bar
So I loosened my router chuck and dropped the bit a cm.
But yes. I have to be careful to set the Z-axis carefully.
Unrelated, Thank you for all of your development work! I’ve been lurking for a long time. I might actually have some small PRs in an age or two. Perhaps something to send telemetry and errors to a central log server for QA? It would have to be secure and not an invasion of privacy, so it could be better that this take a longer time to develop. CISSP with 3yrs of Criminal Justice, hackspace owner for 15yrs. Coding since 90’s. I’ve changed legislation to support privacy. If I do this… I’ll probably still screw it up… So I welcome advice.
It’s true we aren’t accounting for the thickness of the wood currently. I think that we should do some testing to see how big of a difference it makes but it’s a great point.
It’s true we aren’t accounting for the thickness of the wood currently. I
think that we should do some testing to see how big of a difference it makes
but it’s a great point.
I knew that we didn’t account for the workpiece size, but I thought we accounted
for a wasteboard size.
we have had some people post that raising the anchors and setting the Z offsets
to zero make a huge difference, it could be that just the Z offsets being
correct could account for the huge difference.
for the top left arm, the 40mm error results in a difference of ~8mm in the
effective length of the arm when it’s near the top left corner (I use 600mm of
belt length for checking this sort of thing as a nice round number)
I would say that the best fix for this will be to add in separate wasteboard and
workpiece thickness variables rather than changing the offset numbers (set them
through the popup that does the retract/extend/tension/etc)
I would like to see some testing to see that it matters before we add things to the UI just because of how confusing folks already find the settings that we have
I would like to see some testing to see that it matters before we add things
to the UI just because of how confusing folks already find the settings that
we have
I cam come up with tests that will show it’s a huge problem (using a 2" foam
board for your wasteboard, and a thick workpiece)
but what sort of testing are you thinking of?
do we want to talk someone who is struggling with calibration through changing
the offsets and see if it helps?
do we want to have someone who’s machine is working well change their offsets to
introduce deliberate errors?
do we want someone who’s machine is working well to save their yaml file and
re-do their offsets and do a new calibration?
I would like to see some testing to see that it matters before we add things
to the UI just because of how confusing folks already find the settings that
we have
how about adding the variables to the code, but not add them to the UI yet (and
have them default to 0 if not set in the yaml). we could direct people to the
fluidnc advanced config tab to set them while we are testing to see how much of
a difference they make. (it would be easier and safer than changing the 4 z
offset values in that same tab)
I would make it two variables, not one. While one can work (tell people to add the wassteboard and the workpiece thicknesses), I think people will be more likely to make mistakes if it is one variable rather than two.
also, I measure the arms at 21mm thick, so if the lowest one is 34mm, the next should be 55mm (current 56), the next 76mm (current 78), and the final one 97mm (current100). I don’t think being off single digit mm in height matters much, but being the ocd type, if we make any changes to the defaults, we should make them correct.
I have noticed a big difference by changing my work piece from 15mm to 9mm thick, the difference is noticeable in the clearance of the lower belts on my vertical frame, in my case it is useful to be able to adjust the height of each arm independently, since the anchors are not at the same level between them, when I change those few millimeters in height the clearance is corrected and the precision increases
I have noticed a big difference by changing my work piece from 15mm to 9mm
thick, the difference is noticeable in the clearance of the lower belts on my
vertical frame, in my case it is useful to be able to adjust the height of
each arm independently, since the anchors are not at the same level between
them, when I change those few millimeters in height the clearance is corrected
and the precision increases
the Z offset is the way to adjust an individual arm, the proposal to add a
workpiece and wasteboard offset should help accuracy, but I don’t understand how
clearance would change (are you changing the anchor locations???)
I may not be understanding correctly, due to the translator. My anchor points aren’t aligned with each other or with the waste table. Some are lower than others, so I have to individually change their programmed height. Later, when I change my work material, I increase or decrease them all equally according to their thickness. My point was that those millimeters of error are reflected in the clearance and precision.
I may not be understanding correctly, due to the translator. My anchor points
aren’t aligned with each other or with the waste table. Some are lower than
others, so I have to individually change their programmed height. Later, when
I change my work material, I increase or decrease them all equally according
to their thickness. My point was that those millimeters of error are reflected
in the clearance and precision.
the idea of the proposed change is that you won’t have to change each of the
four individually, you will set them once for what’s correct with one wasteboard
and then as you change your workpiece, there will only be one setting to change.
but I’m still not understanding how those changes affect clearance (precision I
understand, but not clearance)
Oh, I’m so sorry, it’s definitely a translation issue. The term “clearance” referred to “loose belts”. And I do agree that it would be wonderful to adjust a single measurement without having to always go back to the yamln file.
A couple of thoughts on this and shims from while i’ve been pondering other stuff.
It might be the bits i’ve been using, but I have to raise the Z-assembly 30+mm from z-stop to Z-home. Given I really only need a range of about 20mm (cutting 12mm ply most of the time, a few mm cut clearance and 5mm travel clearance), I could alter Z-stop by shimming the z-motors 10mm+ (if that makes sense). It’s relevant because that would mean you could/should raise the anchor points 10mm to minimise belt angles in normal use.
I set my anchor points based on the spoil board, and then shim them with the depth of the material i’m going to cut. Depending on the users setup that’s either pretty easy, or a pain, but it again is a good way of minimising belt angle without a software value (though I do thing a software value is useful).
