When is it OK to give up on calibration and use Maslow_Scale_X and Maslow_Scale_Y to "fix" scale errors

Troubleshooting
,
1

I’ve been struggling to get my maslow to draw squares - it is giving me rectangles

Rather than burn through wood I have a pen holder fitted and am drawing on cardboard with it

I have tried the automatic calibration, and a manual measurement calibration and I get the same result.

Drawing a 500x500mm square results in a 502x489mm drawing
Nice right angles, but short on the Y axis by 2.25%

This is fairly consistent between the auto and manual calibration

I have a horizontal setup approx 3700x2700mm between anchors
This is the results from @dlang’s fantastic calulator showing the dimensions and a 1mm measurement error in the manual diaganol measurements

image
image502×779 26.4 KB

For the automatic calibration I get numbers that look something like this

Flex measurement: TLBR: 2.514 TRBL: 2.848
...
Calibration values:
Fitness: 0.8983546724010442
Maslow_tlX: 2.8
Maslow_tlY: 2640.5
Maslow_trX: 3720.0
Maslow_trY: 2648.4
Maslow_blX: 0.0
Maslow_blY: 0.0
Maslow_brX: 3756.8
Maslow_brY: 0.0

So, not good, in that they don’t match the measured values at all well but they do give me the same X and Y scale errors when I plot my test squares.

So is it time for me to give up on getting the perfect calibration and just set Maslow_Scale_Y = 1.0225?

It feels like cheating to fix the issue this way without understanding the cause

Any advice is welcome

-Andrew

3 Likes
2

if it works for you it works for you, period.

keep an eye on it, the error may not be linear, so the scaling may need to
change (either due to position, say near the edge, or due to improvements in
future firmware upgrades)

the x/y scaling is a hack, but that just means it works for at least some
people, even if not the theoretical ‘right’ answer.

David Lang

4 Likes
Multiple user consistent Y axis calibration error
3

I was wondering if anyone has found the x/y scale isn’t working with FW1.11?

I tried to route 4 of 300 x 300mm squares and adjusted the x/y scale after each time before the next route and each time no change to the error, it was exactly the same error x=296 and y=303

1 Like
4

Keltyinnty wrote:

I was wondering if anyone has found the x/y scale isn¢t working with FW1.11?

yes, there was a report on that I saw yesterday

I tried to route 4 of 300 x 300mm squares and adjusted the x/y scale after each time before the next route and each time no change to the error, it was exactly the same error x=296 and y=303

note that the maslow error is unlikely to be the same in different places on the
worksheet.

David Lang

1 Like
5

I can confirm that we did break those and I should have a fix by the end of today. I just need to test it. Great catch!

6

hey @bar are you still expecting to release new firmware that allows x/y scaling? Per my understanding above, scaling doesn’t work in 1.11?

I think I have my router all set up, I’ve been drawing 100mm squares and found that mine is 100.88% bigger that is should be on X axis and 95.77% of what is should be on Y.

In 3d printer land, we’d generally fix this sort of thing by applying a factor to the mm / step for the stepper on a given axis. It can also be done in post processing (slicing) by applying a factor to the generated g-code. It seems most elegant to allow a simple scaling feature as you have done.

Thanks again.. Excited to start cutting!

1 Like
7

Yes! I should have a version of that out today!

I meant to do it on Friday, but then I found another but and got deep into fixing that one :grinning_face:

8

Hi, where can I find the link to that calculator?

image
image502×779 26.4 KB

9

http://lang.hm/maslow/maslow4_manual_calibration_simple.html

1 Like
10

Thanks for the quick reply.

If I use the calculated anchor points, is it still necessary to fine-tune the axis scale?

11

If your measurements and calculation are 100% right you do not need to adjust the x scale or y scale.

If you Maslow has any issues, like stiff spools or misaligned magnets then it still won’t work.

I had to sand and polish my belt spools to get them to wind and unwind freely before I could get a cut that was close to correct.

A

1 Like
12

JURA23 wrote:

If I use the calculated anchor points, is it still necessary to fine-tune the axis scale?

it depends on how accurate you need the machine to be and how far off it is.

in theory, no. In practice, possibly.

David Lang

13

You may find that the answer is in the height differences between the anchors and the arms on the machine. I changed the order of the arms by mistake which changed the heights of the arms as recorded in the .yaml file. I started cutting ovals instead of circles. After manually measuring the height differences and changing my .yaml file appropriately the x/y scaling was much better.

2 Likes