Homing and Initial Position: How does it work?

1

Good Morning everybody, I wanted to start a discussion about how Maslows homing function works and how it knows where the initial position of the unit will be after calibration.

Here is my current experience: after homing the machine I ran a test of the profile I wanted to cut prior to actually cutting something. Things were going well until it got close to the upper edge of the board and then it stopped moving vertically upward, the belts began to form excess slack, then eventually the machine just stopped all together. Initially I thought this may have been a connection issue (as I did disconnect from Maslow also) but after sorting out a registry issue in my computers network settings, I ran the sequence again and had the same error but this time with no disconnection.

The next thought then was that the machine was hitting a maximum height prior to finishing the profile. So I decided to lower Maslow by a few inches from its preset starting position initially and after doing that, it ran the profile without a problem.

This is fine, however, at the moment it means that I cannot fully utilize the entirety of a 4’x8’ cutting area if the machine starts from its initial position set from calibration, which then raised the question: How is this initial position determined, e.g. what set of variables/parameters is this based on? I assume it’s based on the parameters input on frame size and the anchor point location, and I also assume that Maslow assumes your work area is directly “centered” to the frame parameters also so that you may fully utilize the entirety of your work space. I will do a visual check today to see where the bit center is located relative to the center of the board and that may shed more light on this, but for future cuts, assuming the position of Maslow is not centered to the board, will part of my start up sequence be to adjust the offset initially and rehoming prior to starting a project, or is there a way to set this in the parameters on start up. Thank you!

2

what is your frame and workpiece size? try putting them into the tool I have at
http://lang.hm/maslow/maslow4_frame.html and see if the area you are having
problems with is in the green, white, or red areas.

what are the heights of your anchors (Z direction)

when you get into the area that it is pulling away, is the sled hanging over the
edge of your workpiece?

can you show a video of the problem?

with the frame too near vertical, I would not expect to run into problems with
calibration, but would expect to have problems cutting in that as you plunge
into the workpiece, the maslow would push away from it

0,0 in the maslow internal coordinates are the bottom left anchor, 0,x is the
bottom right anchor.

when you finish hanging, you will be close to the center of the material.

David Lang

3

There shouldn’t be any constraints on where the machine will move other than physical limitations (ie it hits something or falls of the edge).

You should be able to define the XY home to be in any location on the sheet and then it will work there.

4

David, thanks for the information. Frame size is 2958mm x 1472mm, work piece is a standard 4’ x 8’ piece of 3/4" plywood. I assume Green is good, white is caution, and red is no go? My question on this would be why would there be a no-go section on this machine at all? @bar can you comment on this?

Z-Height of the anchors is 27mm. This is measured from the surface of the frame that the anchor is attached to, to where the belt end sits.

I’m not really experiencing any pulling away of the sled anymore ever since I fixed the frame and my angle is now 25 degrees measured from vertical.

No video to show, but what I can tell you is that the G-Code file I generated is based on the home position of the Maslow being dead center to a 4’ x 8’ sheet of plywood (I also generated the .svg file to show my project having the origin at the center of a 4’ x 8’ sheet of plywood). When I upload the g-code and run the file, things run fine until it gets towards the top of the workpiece, then it stops moving upwards vertically but you can clearly see slack forming within the belt dials, particularly in from the belt anchored to the bottom left (presumably the bottom right also). The machine then shuts down and waits for me to basically perform a hard reset. My assumption here is that, although I have designed the .svg and the g-code in such a way that everything should be centered around the middle of a 4’ x 8’ piece of plywood (and therefore everything was positioned within this boundary accordingly) the physical setup does not reflect what has been designed on the computer, and the Maslow is more vertically upward initially than being actually centered to my workpiece, thereby going too “high” and running our of vertical “room” if this makes sense. By moving the initial position of the Maslow down approximately 5", I was able to run a test profile no problem in its entirety without the issue presenting itself, lending to my belief that the Maslow is going to high relative to the physical setup, coming to a conclusion that my Maslow may not be centered to the workpiece but IS centered to the frame.

This was the reasoning for my asking about how is “home” defined after calibration, as it sounds like it is centered to the frame and not to the workpiece, since the workpiece parameters are not input anywhere that the software and the Maslow would know is too far.

5

I wouldn’t rely too heavily on that sheet. It’s technically true that the arms hitting the supports will lead to a small amount of inaccuracy, but it certainly isn’t a no go zone…it’s more of a “slightly increased inaccuracy zone”. There are other sources of inaccuracy that are more significant that we are working to fix.

I think that this sounds like a spot on description of what is probably happening.

When you see the gcode preview in the user interface, what do you see? Note that clicking on the preview will toggle through several camera angles

6

Mark Thomas wrote:

David, thanks for the information. Frame size is 2958mm x 1472mm, work piece
is a standard 4’ x 8’ piece of 3/4" plywood. I assume Green is good, white is
caution, and red is no go? My question on this would be why would there be a
no-go section on this machine at all? @bar can you comment on this?

the different colors are just different types of errors.

Red is the area where the two adjacent arms are both hitting the verticals of
the sled (top center, both arms are jammed against the side brackets as they
try to get more horizontal)

white is where the two adjacent corners are both hitting the verticals (towards
teh top left, it means that the top right and bottom left will both be hitting
the verticals, when they really want to be pointing more towards the bottom
right corner)

as you get into these areas, the arms that are hitting the verticals become
shorter than the machine thinks they are. Instead of the belt + arm being
directly in line with each other, they are at an angle, and the more they are
angled, the shorter the distance from the center of the bit to the anchor.

you can go a little bit into the white/red area, but it’s not a linear
relationship, going in say 100mm may produce 1mm error in belt length, but going
in 200mm may produce 10mm error in belt length (faked up numbers as an example)

if you are limited in your height, play around with that page setting your frame
to be narrower, that will get you a larger usable curring area

Z-Height of the anchors is 27mm. This is measured from the surface of the frame that the anchor is attached to, to where the belt end sits.

what matters is the distance from the center of the belt at the anchor to the
center of the belt at the arm. The default Maslow_*Z values in the .yaml file
assume that you have the anchors directly on the frame with a 19mm wasteboard
and a 19mm workpiece.

since you are moving the anchors up on the Z axis by 27mm, you will want to
reduce all four of these numbers by 27mm from stock

that said, I would not expect that to be the cause of pulling away.

I’m not really experiencing any pulling away of the sled anymore ever since I
fixed the frame and my angle is now 25 degrees measured from vertical.

the weight is holding it flatter.

No video to show, but what I can tell you is that the G-Code file I generated
is based on the home position of the Maslow being dead center to a 4’ x 8’
sheet of plywood (I also generated the .svg file to show my project having the
origin at the center of a 4’ x 8’ sheet of plywood). When I upload the g-code
and run the file, things run fine until it gets towards the top of the
workpiece, then it stops moving upwards vertically but you can clearly see
slack forming within the belt dials, particularly in from the belt anchored to
the bottom left (presumably the bottom right also). The machine then shuts
down and waits for me to basically perform a hard reset. My assumption here is
that, although I have designed the .svg and the g-code in such a way that
everything should be centered around the middle of a 4’ x 8’ piece of plywood
(and therefore everything was positioned within this boundary accordingly) the
physical setup does not reflect what has been designed on the computer, and
the Maslow is more vertically upward initially than being actually centered to
my workpiece, thereby going too “high” and running our of vertical “room” if
this makes sense. By moving the initial position of the Maslow down
approximately 5", I was able to run a test profile no problem in its entirety
without the issue presenting itself, lending to my belief that the Maslow is
going to high relative to the physical setup, coming to a conclusion that my
Maslow may not be centered to the workpiece but IS centered to the frame.

excess slack probably is caused by the Z values not being correct. I would
expect that you would be running into current limits towards the top center.

This was the reasoning for my asking about how is “home” defined after
calibration, as it sounds like it is centered to the frame and not to the
workpiece, since the workpiece parameters are not input anywhere that the
software and the Maslow would know is too far.

the maslow does not have any software ‘you are moving off where I think the
workpiece is, shutdown’ limits, just mechanical "I can’t move any further, I’m
pulling but not moving’ limits.

David Lang

7

Bar wrote:

I wouldn’t rely too heavily on that sheet. It’s technically true that the arms hitting the supports will lead to a small amount of inaccuracy, but it certainly isn’t a no go zone…it’s more of a “slightly increased inaccuracy zone”. There are other sources of inaccuracy that are more significant that we are working to fix.

that depends how heavily you go into those zones. With his frame size, the area
before the arms hit is only 200m up from the center.

I agree there is more going on than just this (see my longer reply)

David Lang

8

Thanks for reaching out and getting back so quickly @bar

When I toggle through the camera options, I see my design come up, and I see a black outline, my design and then the location of the Maslow in space. This black outline: Is this frame or is this the workpiece of a standard 4’ x 8’ piece of plywood?

1 Like
9

Mark Thomas wrote:

When I toggle through the camera options, I see my design come up, and I see a
black outline, my design and then the location of the Maslow in space. This
black outline: Is this frame or is this the workpiece of a standard 4’ x 8’
piece of plywood?

the centered black outline is a 4x8 sheet of plywood

David Lang

1 Like
10

I guess I am confused, why not measure where you want the home position to be on the board you are cutting (in this case dead center), jog the sled to that location, and then set the home there? Not sure I understand why the calibration home is all that important.

I wouldn’t depend on something I couldn’t personally see and know, but maybe I am old school.

1 Like