what is the slop in the bearings and between the bolts and the bearings?
another thought that just hit me.
the current calibration routine tinkers with the spacing between the chains on the sled to make the results come out even
I wonder if triangular kinematics could end up with something similar in two stages.
-
tinkering with the space between the motors (very similar to the existing routine)
-
tinkering with the chain length error by cutting at the top an bottom of the workpiece and comparing the heights of the two pieces (if the chains are not the right length, the error will be far more significant at the top than at the bottom)
This won’t be able to tell the difference between constant error (error in the pantograph/chain mounts) vs percentage error (slop in the chain itself) unless we then do some tests at the side of the workpiece…
if we can do this, that may make the 45-balanced pantograph better because it will have a more consistant error, which may be able to be detected.
Probably too much with a 5/16 bolt, With a 5/16 2" bolt, and the hold and wiggle with a ruler technique, maybe a 16th at the bolt end, but it looks like it’s all slop in the 8mm hole. Good enough for a wine bottle cutter or 3D printer idler but not this. Couldn’t find an 8mm bolt (know I have some…) to test. Could always tighten them up with some loctite but that might put it off center. Bronze bushings might be tighter, not like they’re going to get any appreciable wear.
My bolt holes are less than a 64th, drilled at 1/4 and needed a bit of hand reaming with the bit to get them from jam them on to turning in the hole, didn’t break out the adjustable reamer. Of course, anything made out of wood will change dimensions anyway.
Wonder what waterjet or laser cut steel would cost for a sled and arms if made in quantity? Prove (or disprove) the design in wood, do a pre-buy for metal parts. Not feasible here, don’t know of a job shop that has one, and CNC Bridgeport is likely spendy.
How do you fiddle the calibration numbers for the 3 (really 4) arm balanced sled?
Got do do some more floor breaking and wood cutting first, Mrs. Moose is coming back home in the next day or two and need to show some more woodpile progress before having fun. 7C down, 35 to go, although 10 will make it through the winter unless it’s back in the -20s F.
Thanks for a good, thorough review.
I’ve wondered about the hoop approach and the 80 degree limit with regard to the ring top mounting point. My napkin calculation suggests that at 80 degrees on a 6” hoop, the chain needs to reach about 1” from full vertical. That would mean that the sheave radius plus half the width of the top mount would need to be less than 1”. Does that sound right? This wouldn’t be an issue with the single-sheave and ring approach you suggested, even if the full ring were used, as the upper mount would only need to clear one direction.
Moving the motors farther outboard, or accepting a narrower work area would be options too, I guess - but who likes compromises?
a couple ways
-
connect the fixed chain at the same distance from the bit as the moving one
-
lie to the machine about the chain position
since each link is 1/4", if your fixed chain is 1" further out than the moving
one, move it 4 links further out when you start paying it out.
1 Like
For us visual learners:
In an ideal world this appears to work just fine even at the extremes of the current Maslow set up. This is based on motor measurements provided by several of you, and a work surface of 4’x8’.
Note, the distance between the two anchor points does NOT need to equal the distance from the lower anchor point and the center of the router. As has been mentioned before (by myself, @dlang, and likely others) the distances between the top two pivot points on the vertical members must equal the distance between the two anchor points and the distance between the lower pivot point and the chain attachment point on the vertical members must equal the distance between the lower anchor point and the router center. The length of the horizontal members is relatively arbitrary, though in this drawing they are about 5.36" between pivots (I forgot to label that one). Edit: Sled is 18", I forgot to mark that too…
Personally I think I’d be quite happy with this or the 45˚ design.
-Logan
1 Like
Here’s a symmetrical 45˚ design that should have no problems:
Here’s a slightly more complicated version but it’s more compact by a couple inches if that’s what you’re after:
3 Likes
Not sure if I can describe this, but given the other posts about the difficulty of “side loading” the pivots and having different heights on the chains (above the sled)…
Could you have some one of the upper arms from the router to the pivot be “single” and it’s opposite be “double” so that they interleave?
Have one arm:
And the other:
With the single layer arm going in the gap of the double. If you alternated single and double between the sides, you would get around some of the side loading on the pivots, and allow for them to slide between each other.
Maybe - I’m going to have to find some sticks and build a mockup before anybody can understand this one, I expect
2 Likes
I think @krkeegan has a perfect example of what you are talking about here: Throwing my hat in the sled modification ring
If not then maybe I misunderstood your question…
1 Like
Right you are… I’m going to give up trying to have good ideas, all you smart guys beat me to them.
Sorry @krkeegan - I wasn’t trying to steal your thoughts 
Well, it does seem to work!
Now I need to build one…
2 Likes
Ha! Don’t apologize for having good ideas! We are all working on this thing together, brainstorming like this only works if there are multiple brains in the storm.
5 Likes
One big benefit to the 45˚ version over the top mounted version is that the chain shackle for mounting the chain would not need to be anything special. You can see in the Top Mounted version that the shackle would need to be pretty long in order to clear. You could easily make a home-made shackle with two pieces of wood and two pins (basically another bar)… but the simplicity of being able to use a pre-made $.75 part might be nice.
1 Like
In this 3-Bar 45 degree design, the slop of the pivots like the slop of the chain links can be counted as part of the chain measurement, because these slop points are always under tension. The tension varies, it’s true, but I think the difference would be less than the 1/64” Target. Am I missing something?
2 Likes
minor note, the angle can get down to just under 10 degrees from horizontal when you are at the far top.
The router base is 6" diameter, but the motor is only 4" in diameter (but you do have the Z axis, handles dust collector adding to this on the sides and bottom)
the chain has a minimum length of ~2 ft from the sprocket to the center of the bit, so call it about a foot to the arms.
roller chain can handle a fair bit of angle (think about how it angles on a 10 speed bike), so a little bit won’t bother anything.
If this is done with 3/16 metal, I wouldn’t bother trying to do anything about the different heights. Doing this with 3/4" thick wood, you still have to figure out how to attach the chain to the wood, attach it to the top on one side and the bottom on the other side and the resulting difference is almost nothing.
actually, both approaches need care for the chain. It’s easier on the 45 version as the chain goes from +45 to -45 instead of from 10 to 90 (relative to the bar it’s attached to), see the onshape pictures I posted about 2 days ago.
The first one is harder to see in the post, it’s got an interesting shape in the center where the chain attaches.
yes it is under tension, so it should be fairly consistant, but it will be fairly consistant at 2x the slop (once for the holes in the mount, once for the hole where the bars join)
and interesting ‘feature’ of the top mount is that the error is based on the ratio between the upper and lower portion of the vertical bars. With the upper portion 2x the lower portion, the error is the same (I don’t know how much it varies based on different positions), if you were to make the upper portion 4x the lower, the error would actually be half as much.
1 Like
So that slop amount can be added to the length of chain as a static measurement. It will be included when one measures from the bit to the sprocket while under tension. The same is true of the 3-bar balanced version as well.
Once the slop is accounted for in the measurements it shouldn’t affect the kniematics.
1 Like