You are correct. I’m sorry, I was being rhetorical, it wasn’t actually a question . The clamping works quite well; when I drilled the bars for my tests I drilled all 6 at once (clamped with C-clamps). I was using a drill press so it remained accurate. With a hand drill it will be harder to keep the holes perfectly vertical through that thickness. I’m sure it can be done but drilling them in well-marked pairs might prove better for hand drilling…
That’s only 44.09 pounds of force though…
EDIT: I just saw your correction for 30Kg/cm! Yeah, that comes out to 66.14 lbs.
For anyone curious, I believe the equation looks like this:
Force = Torque / (Length * sin (Angle))
The Length is the radius of the sprocket. Also, since the sprocket is basically a pulley the Angle will always be 90˚ and sin(90) is 1 so you can just ignore that part If someone wants me to show my work I’d be happy to
Anyway, I see that the motors are capable of putting 66 pounds of tension on each chain, that’s pretty awesome!
Thanks @dlang!
So, lots of thoughts of how to build this from wood, etc., and also some of doing custom CNC/laser parts.
Have we considered standard hardware store metal strips? Wondered by the local Home Depot display tonight and they had a couple of sizes of this stuff:
That would solve a lot of the issue with hole accuracy, I expect…
BUT, the piece I played with wasn’t stiff enough, I believe (although I’m not sure on the quad-hole mounts how much is in tension and how much in compression… tension no problem.
You could probably double a layer or two (or find a source with slightly thicker metal).
one issue is how to attach the chains to this. The examples I posted have the bar shaped so that you can attach the chain with a master link. If you don’t do that, you have to figure another way to attach it so that the pivot point is at the hole in the metal, and you can swing through the required range (± 40 degrees for one approach, 10-80 degrees with the other approach)
Unfortunately, the bars aren’t just under tension, some are mostly tension, but some have to handle side loads and not buckle. with my proposals above, I suggest using 3/16" thick metal, and if that’s not strong enough, it works really well to use angle metal that’s 3/16 thick.
The remaining problem is that the bolts need to be a very snug fit to the holes, the error from a sloppy fit can add up quickly to noticeable amounts.
Seems like a shackle or splitring (keyring) would do that… through the appropriate hole and through the end of the chain. You just have to show the chain as “longer” than it really is to cover the extra connection liengh.
Where does one find these tiny shackles? Mine all have big pins, up to an inch or so, but nothing under about 3/8".
Of course they’re all for farm equipment…
I found something interesting, Tsubaki end bolts, but the smallest is the RS40EB for #40 roller chain. Nothing small enough for #25, although you could make one with a bolt, grinder, and drill press.
I think that’s iRoc999. Here they try to move inside about this time of year. Mrs. Moose doesn’t have a welcoming attitude to little furry critters so those potential food sources can be hazardous
remember that whatever mount you make has to hold 66 pounds of force from each
chain without flexing noticably.
we need to have the pivot in a known place. The stock design doesn’t have this,
some of the pivot of the chain is the chain moving in the hole it’s put through,
and then when it hits the limit there, the pivot point moves to the first link
of the chain outside the mounting hole.
If you want your connection to be perpendicular to the workpiece (so that it’s
your pivot), than you have to use the pin holes in the chain.
the pins are .091" in diameters. 3/32 is 0.0975, so I’m calling the holes 3/32"
the strongest way to use these holes is with a master link. This will also let
you use something that’s ~3/16" thick to anchor to rather than just the 1/8"
thick that a sprocket can be.
If you are going to put something through the chain parallel to the workpiece
(like the large pins the stock design uses) you have .126" between the side
plates of the small links and .120 between the rollers
a 1/8" bolt/link/whatever will not fit through the chain in this direction, you
need something just a bit smaller (3mm is just enough smaller than 1/8 that it
would probably work)
Note that the 66lbs of force that @dlang mentions is not a working load. 66 pounds is the max (tensile) force those motors can impart in a chain when fitted with a 1cm sprocket. Under static circumstances with a 20 pound sled there is about 33 pounds of force on each chain (when the sled is at the top center of the work area). Under working conditions that likely increases some due to friction of the sled and drag from the bit and inertial loads, however I would be extremely surprised if it ever neared 66 pounds.
I am not saying things shouldn’t be over built, it’s much safer to build for 66lbs. But it’s unlikely either chain will ever see that much force unless something malfunctions (in which case having a weak link might be a great idea to save the motors and frame and sled…)
Just wondering… Would any of the benefits of these designs theoretically act to improve accuracy (and how much), or is this purely to free up cpu and improve the calibration/setup process? Find the discussion fascinating but my primary concern before trying any of the options would be to improve cutting accuracy. My initial project had numerous circle cuts, and its pretty obvious by sight that they aren’t perfect.
Yes! It should absolutely improve accuracy! Not only is the math much simpler but the sled should be more stable as well since it’s now a triangle instead of a trapezoid the sled can’t tip away from the calculated cutting point. Hopefully there will be some difinitive tests soon!
We actually had someone check on the power fed to the motors, and they found
that most of the time the motors never had more than 50% of their rated power
fed to them, but some additional testing was suggested and he saw conditions
where the motors (and therefor the torque produced) went up over 90%
Also consider situation like the sled moving close to vertically down at max
speed, and all of the sudden the motor stops (zero voltage, zero strain on the
motor), the worm gear, chains, and sled mountings will be trying to decelerate
the sled close to instantly
it’s one thing to have a weak point that will give under stress, but if you have
a weak point that just bends (and worse, bends in a way that isn’t visually
obvious) that will cause no end of frustration.
for what it’s worth, the chain (and the sprockets) are rated at something around
700-900 pounds of force.
They increase accuracy by removing several sources of error.
you should take a look at the math used to position the normal sled, it’s very
ugly, and you have the folloing dimensions that all produce error, even if only
off by a mm or so
space beteen the motors
length of chains
accuracy of measuring where the chains are mounted to the sled
where the chains pivot
space between the chain mounts on the sled
distance from the line connecting the chain pivot points down to the bit
distance from the bit to the center of gravity of the sled
rotation of the sled caused by cables/dust collection hoses
the approaches being discussed eliminate #5-8 entirely and could eliminate #4
and almost eliminate #3
We try to measure #1 by stretching the chain. This assumes there is virtually no
give in the chain, no significant droop in the chain as it’s being measured, and
that the person doing the measurement really gets a sprocket straight up (It may
be easier to be precise either setting a tooth straight donw, or having a hole
in the bracket that you can screw something into as a positioning marker
for measuring the length of the chains, we calculate how many links we have let
out based on how far we’ve turned the sprocket, assuming there is no stretch or
droop in the chain.
The current calibration cuts take what we think is correct about the machine
dimensions, make some cuts that we know what the distances should be, and then
change #5 in the list above and try again until we get something that ‘seems
right’. However, we know that errors in other measurements can cause errors that
look the same in our tests, so it’s very possible that the results will still be
wrong elsewhere in the work area.
The triangular math approach supported by these types of sled attachments
eliminate so many of these error sources that the remaining error sources become
much easier to test for.
. The clamping works quite well; when I drilled the bars for my tests I drilled all 6 at once (clamped with C-clamps). I was using a drill press so it remained accurate. With a hand drill it will be harder to keep the holes perfectly vertical through that thickness. I’m sure it can be done but drilling them in well-marked pairs might prove better for hand drilling…
If someone wants me to show my work I’d be happy to 
