Maslow Home Maslow Community Garden

Sled balance (possibly custom sled weight issue)

calibration

#1

Using the default frame (https://github.com/MaslowCNC/Mechanics/wiki/Choose-A-Frame-Design#the-default-frame) with non standard weights, on the ring sled, with the the recommended Ridgid R22002 router, with Z-axis kit. Latest fimware and GroundControl (1.2.4).
My sled uses 5 Lbs diving weights which have been angularly aligned and positioned to where the Brick’s CG was. Wish to understand how position of weights and their weight influences movement.
The ‘Ring’ on my sled is set at ~3" above the top of sled on Z axis. Had figuring out this height as I could not figure out clear directions. So maybe got it wrong. The clearest instructions were found under Step 4 in “Maslow Ring System Sled” (http://maslowcommunitygarden.org/Maslow-Ring-System.html?instructions=true) it says:
“Step 4: Attach the ring and bricks” - “Note that the ‘L’ brackets are designed to let you move the ring up and down to adjust for your router’s center of gravity. When supported by the ring the router should hang vertically. You can move the ring in and out to account for your router’s weight distribution.”
I assumed I should balance the sled using the top most (z-axis) point of the ring but this caused problems with the sled tipping when aproaching the edges of the work area. The closest cuts to the edge were 2".
The angle of sled changes depending on the position it is in on the cutting area. Here is a quick crude diagram.
sled-balance-rough

This is a side view (z-axis) from cutting surface.
⦁ RED is cutting surface.
⦁ BLUE is sled position (at top and bottom) of cutting surface with tilting effect.
⦁ GREEN is dimensional center of sled (Not CG of sled).
⦁ ORANGE is contact position of chain to sled ring.
⦁ YELLOW is mount point of motors and the chains.
⦁ BLACK is current sled angle when supported by rope from top of ring.
⦁ Diagram is not to scale and exageratted to emphasize points.

PROBLEMS:

  1. When sled (BLUE) moves to the top of the work area then the chain contact point to the sled (ORANGE) moves below the the physical center of the sled (GREEN). The top of the chain at the motors is at least 4" further out, away from work area, than the sled ring and causes the sled to start tipping over the top of the work area (RED). Conversely when the sled is at the bottom of work area the chain is almost vertical, and the contact point to the sled is high on the ring, which maybe pulling the top of the sled away from the work area.
  2. Another factor is the CG of the sled which is below the physical center. So when the sled approaches the bottom edge of work surface the CG moves of the board and cause top of sled to tip away from work area.
  3. I lose accuracy (badly) when cutting within 1’ of left and right edge of 8x4 board. The pieces shown below should look the same left to right (mirror image) but the left side is horibbly disfigured.
    P1000534b

To fix these I adjusted the sled to NOT hand straight vertically when supported by rope at the top of the ring. It is more like 5-10 degrees (BLACK). This seems to be the optimal angle to keep the sled from tipping. I can now move the sled so that the router bit is directly on the edges of work area, top and bottom, without the sled tipping. It also helped to stop a slight tip on left and right edges of work area.

QUESTIONS:

  1. Have I balance the sled properly?
  2. What is the minimum/maximum for sled weights?
  3. Should I change the sled balance setting in Ground control? This uses a measurement from center of router bit to CG? How does this setting affect the calculations?
  4. Should I move the motor mount beam closer to the edge of work are to reduce the angle of chain to the sled?
  5. Could moving the sled weight closer/further to center, OR up/down influnce accuracy when working on left/right extremes of 8x4 board?

Sled tilt - Wrong frame dimensions?
#2

Using the default frame
(https://github.com/MaslowCNC/Mechanics/wiki/Choose-A-Frame-Design#the-default-frame)
with non standard weights, on the ring sled, with the the recommended Ridgid
R22002 router, with Z-axis kit. Latest fimware and GroundControl (1.2.4).

My sled uses 5 Lbs diving weights which have been angularly aligned and
positioned to where the Brick’s CG was. Wish to understand how position of
weights and their weight influences movement.

There are two directions that the position of the weight matters.

first, as you look down on the sled from the Z direction, you want the weights
to be offset towards the bottom of the sled so that it doesn’t rotate under the
motor torque. The reason for this is that if the sled rotates up to where the
chains hit something, it throws off the distance calculations

The ‘Ring’ on my sled is set at ~3" above the top of sled on Z axis. Had
figuring out this height as I could not figure out clear directions. So maybe
got it wrong. The clearest instructions were found under Step 4 in “Maslow
Ring System Sled”
(http://maslowcommunitygarden.org/Maslow-Ring-System.html?instructions=true)
it says:

  1. Have I balance the sled properly?

No, you want to position the ring so that if you hang the sled by the chains, it
is balanced or tilts the top of the sled slightly forward (the chains attaching
at or just above the balance point)

you then want to adjust the frame so that the chains are level to the workpiece.
You can do this by either moving the top beam back or by adding additional
thickness under the workpiece with additional wasteboard material (consider
using foam insulation panels, they are cheaper than plywood)

  1. What is the minimum/maximum for sled weights?

we don’t know the minimum, we know that if the sled is around 30 pounds, it is
heavy enough to cause problems with the current motors and stock frame
dimensions.

  1. Should I change the sled balance setting in Ground control? This uses a
    measurement from center of router bit to CG? How does this setting affect the
    calculations?

That setting does not apply when you are using the ring, that was for the
original L bracket setup.

  1. Should I move the motor mount beam closer to the edge of work are to reduce
    the angle of chain to the sled?

yes, see above

  1. Could moving the sled weight closer/further to center, OR up/down influnce
    accuracy when working on left/right extremes of 8x4 board?

yes, but only a bit. If the chain angles are bad, and/or the attachment is too
high/low on the sled, the sled does not move as smoothly, and may end up being
tippier than it should be at the edges.

David Lang


#3

Thanks David (@dlang) for all the info. It is extremely helpful to know.

  1. Sled balance. OK My sled is now tilting towards work surface at the top of the sled at about half the angle of the work surface. I tested more cuts today and the sled remained flush with the work surface even when cutting 1" away from edge. So very happy about that.

  2. Not sure about sled weight but it’s all stock apart from using 2x5lbs weights instead of bricks. The reason I asked was because I saw slack in one chain when cutting in extreme bottom left/right of work surface which I think caused the disfigured cuts and wobbles in straight cuts seen in the second image.

  3. Sled balance ground control - OK good to know that is one more setting I can rule out.

  4. Chain angle to work surface - Will certainly adjust the chain angle probably using foam backing material as you suggested. Great idea.

  5. Sled weight position - The ring is perfectly centered in relation to the router bit. I did a test of manually rotating the sled at a fixed point which resulted in a perfectly round hole that was the size of the bit. I do see that the pulleys on the ring do not always move smoothly and have slight jerks which introduces slight wobbles in straight lines. For the most part of the cutting area, this is very slight, but when the chains are very long it seems to introduce a resonance to the chain which makes it worse. Was thinking this might be binding between sled and cutting surface. Sled is sanded and poorly waxed where as the 3/4 ply being cut certainly has rough imperfections. Thought I might need to sand the ply board before cutting. Does anyone do this?

Thanks again.


#4
  1. Sled balance. OK My sled is now tilting towards work surface at the top of
    the sled at about half the angle of the work surface. I tested more cuts today
    and the sled remained flush with the work surface even when cutting 1" away
    from edge. So very happy about that.

That’s a good data point for us to have.

  1. Not sure about sled weight but it’s all stock apart from using 2x5lbs
    weights instead of bricks. The reason I asked was because I saw slack in one
    chain when cutting in extreme bottom left/right of work surface which I think
    caused the disfigured cuts and wobbles in straight cuts seen in the second
    image.

the long chains do get slack in the bottom corners, but wobbles in straight cuts
(when cutting down I assume) are caused by the sled sticking. sometimes
sanding/waxing the bottom ofthe sled helps, sometimes going slower helps.

the other problem you can see is that movement to the outside to the corner may
not happen and corners can be rounded. going slower, decreasing friction, or
going to a 12’ top beam can help here.

  1. Sled weight position - The ring is perfectly centered in relation to the
    router bit. I did a test of manually rotating the sled at a fixed point which
    resulted in a perfectly round hole that was the size of the bit. I do see that
    the pulleys on the ring do not always move smoothly and have slight jerks
    which introduces slight wobbles in straight lines.

Some people are sanding the inside of the ring to try and improve this.

For the most part of the
cutting area, this is very slight, but when the chains are very long it seems
to introduce a resonance to the chain which makes it worse. Was thinking this
might be binding between sled and cutting surface. Sled is sanded and poorly
waxed where as the 3/4 ply being cut certainly has rough imperfections.
Thought I might need to sand the ply board before cutting. Does anyone do
this?

I haven’t heard of this.

David Lang


#5

Thanks again David for all the insight.

the long chains do get slack in the bottom corners, but wobbles in straight cuts
(when cutting down I assume) are caused by the sled sticking. sometimes
sanding/waxing the bottom ofthe sled helps, sometimes going slower helps.

The wobbles in cuts are actually when cutting horizontally. Vertical cuts going up are probably the best. Going down are acceptable. I am cutting using 1/4" bit at 1/4" depth at full speed (1000mm/m), based on suggested speed in Fusion 360, so I could try slowing that down, and/or reducing depth.

Sled is well sanded with 60, 100, 200 grit with all edges rounded over but think I should do better waxing. Any suggestions on wax?

going to a 12’ top beam can help here.

I already thought about going to 12’ top beam as the majority of my problems are within last 1’ left and right. Would I have to get longer chains?

Some people are sanding the inside of the ring to try and improve this.

Again very useful suggestion. Will try this out. I was surprised that with the great design and dual bearings that the movement on the ring was not silky smooth.


#6

you would need longer chains (12.5-13’ ) for a 12’ top beam

can you get a video of it as it’s running into problems cutting?

the chain should never be dropping below the midpoint of the sled.


#7

Ran more cuts and got much better results.

Changes made:

  • Sanded the sled base with 300 grit and added more feathering to the 1/4" round over bit used on both inner and outer sled edges.
  • Added more wax. Using liquid car wax which is probably not a good product as it is too thin. Let it set for one hour at room temp before wiping excess off.
  • Sanded the Ring on inside edge, and top/bottom surfaces within 1/18 of inside edge.
  • Used better grade 3/4 Ply that had pre-sanded surfaces ($10 more per 4x8’ sheet)
  • In Fusion360 Post Process set ‘Disable Arc’ to ‘True’ and used ‘Feed Optimization’ with default settings.
  • Reworked z-axis with bronze bushing attached using screws. (Other issue but may influence current issues as Z cut depth was more consistent)
  • Wiped EEPROM (other issue) and performed full re-calibration, with test cuts. Most probably little effect as the new values were very close to original ones.
  • Moved chain end mount points, not where motor attaches to beam, but where end of chain is attached to top beam with nail. Moved it from ~6" from end of top beam, to 12", as close to top edge of top beam as as possible. Noticed before that when cutting close to top cutting surface the chain would rub on the nail and move unevenly as it rolled over it.
  • Slowed down cuts to max 30 ipm and reduced cut depth to 0.2"

Results:

  • The movement of chain supports around the Ring were a marked improvement after sanding the Ring. No jerking movements were visible.
  • The sled movement over work surface was also much smoother which I think is due to the additional sled sanding, edge round over feathering, and wax.
  • The sled remained in parallel contact to the cutting surface almost all of the time. Which I think is due to sled balance and possibly due to improved z-axis cutting depth accuracy.
  • Massive improvement in accuracy when sled changed cutting direction. This was surely due to Fusion360 adjusting feed rate. Square corner cuts are now less than 1 degree off from 90 degrees. The feed rates were adjusted by Fusion360 based on the radius of the arc and angle changed with smaller radius cuts slowed down more. 1/2" radius arcs were 15 ipm. Square corners were 7 ipm.
  • Marked improvement in horizontal cuts close to the top edges of 4x8’ area. Noticed that the chain no longer rubbed against the ‘end of chain’ nail.

Existing issues:

  • Still have big accuracy issues in bottom left/right corners within 1’ of edge with the longer chain going slack. Think I am going to extend the top beam as suggested, with longer chains, as I really want to be able to cut a full board accurately. Measured sled weight at 23.8Lbs and am thinking about adding another 5lbs to increase the tension on the chains.
  • Still have wobble only in horizontal cuts going left.

#8

Existing issues:

  • Still have big accuracy issues in bottom left/right corners within 1’ of edge with the longer chain going slack. Think I am going to extend the top beam as suggested, with longer chains, as I really want to be able to cut a full board accurately. Measured sled weight at 23.8Lbs and am thinking about adding another 5lbs to increase the tension on the chains.

once you add weight, test horizontal cuts going along the top edge. When you get
too heavy for the motors (speed plays a factor hear as well), you will see the
cut curve down and then hook up at the end as the pulling motor can’t keep up

  • Still have wobble only in horizontal cuts going left.

try swapping your motors (but keep the cables in the same place) to see if
you’ve got a bad motor.

David Lang


#9

Are you using a stock Maslow board, or one of the TLE5206 versions? On the stock board the left motor and the z motor are both driven by the same chip, which must divide the current it supplies between them.


#10

Stock Maslow board & have Z-Axis kit hooked up so this could be an issue.

Most, if not all, of my cuts set the Z depth first and then only move X & Y. No Ramp cuts. Would the current division still affect this?

Wondering if Fusion 360 -> Sideways compensation -> Left (Climb) or Right (Conventional) might influence this. So far I have been using Left (Climb) [default] which results in clockwise path cuts. Have not tried Right (Conventional) yet which would result in counter clockwise cuts.


#11

In theory it should not. When the motor is inactive the ‘enable’ pin is set false and the motor leads are set one high the other low. Having the enable shut off should prevent current from reaching the motor.
You could test by unplugging the z axis to see if the wobble still happens during a run with no bit.
‘Wobble’ on a horizontal cut is more of a problem in the upper third of the workarea, and at higher feed rates. Moving ‘Home’ to run the pattern lower, or running it at a lower feed rate would be other things to try.