Good… So from a GC/Firmware standpoint, the sled is being moved where they think it should be… the problem is where they think it should be isn’t producing the results you expect.
So here’s the thing that’s been bothering me. During calibration of triangular kinematics, we don’t measure the vertical distance between cuts. That means that to me that with triangular kinematics once you align the horizontal distance you’ve aligned the vertical distance. You’ve seemed to have dialed in the horizontal distance but the vertical distance is off. That means that either the assumption regarding triangular kinematics is wrong or there’s something mechanically wrong (that’s why I kept asking about the linkage kit).
This will be a pain, but consider doing a series of measurements to help define the specific problem. Start at 0,0 and move left and right in 6 inch increments and measure the error across the board and then repeat the process up and down from 0,0. If inclined, maybe do some up and to the left, up and to the right, down and to the left, down and to the right… just to produce a matrix. Maybe the simulator can be used to see if we can match the errors your are measuring.
Alright, some results. I performed the test that @madgrizzle suggested to verify my encoders steps were working correctly. No issues there, the motor started and stopped with the sprockets straight up.
A. Verify my linkage lengths
My linkage lengths are all 9" (229mm)
Backup my ini file again,
Done. Started clean.
Set my known measured motor spacing to 3010mm,
I actually measured this again using a board to space the tape measure at the right height so there was no sag in my tape measure and I came up with 3008.5mm. I used this as my motor spacing value. For another data point on chain pitch I used the machine to measure the motor distance and again ended up shy of my measurement at 3005.04mm. That pulls my chain pitch back to about 6.356
Set my ideal rotation radius to 260mm,
Took the time to set the sled upside down and clamped a piece of 1/4" plywood to the clevis and measured to that and then determined the overlap of the chain into the clevis. With that I had my rotation radius measurement of 268.3 mm.
Test the up-down accuracy of 6" movements before doing a calibration.
Using my manually input numbers I completed the upper left quadrant of the sheet in 6" increments. Measured with a caliper my horizontal moves ranged from 5.99" to 6.02" and my vertical moves ranged from 5.95" to 5.98". I did notice that using a level there is a slight sag moving from the outside of the board to the middle which is not unexpected.
I may actually adjust my chain pitch and do the upper right half of the board and see how the numbers compare to the left side. Might be an interesting comparison.
That’s a classic symptom of flexure of the frame. You’ve posted an excellent picture of your frame at the start of this thread. It’s hard to tell form the angle, is the top beam a 10 foot 1x6 or 2x6?
You’ve already got a detailed measurement test going and that could reveal what we’re looking for, but if not, would you consider checking whether the top beam bows during the motor measurement? A measurement from the chain to the top beam in the center, or between any two points at the ends of the beam, before and during tension would do. If the tape measure value was around 3010mm and the calibration measurement was 3005mm, doing the math that would suggest a bow of almost 3.5 inches. Maybe I misunderstood the values, though.
Keep digging! We’ve going to get this solved in time for that project you’ve got planned!
My top beam is a solid 10’ 2x4 on edge so it is better able to resist the bending moment. (Oriented and fastened in the same way that the new stock frame top beam is done). If it is flexing I would be absolutely amazed but I will absolutely check it.
I did actually buy a second 2x4. If it is indeed flexing I can screw/glue that to the top of the existing beam to provide even more stiffness.
When you manually measured the distance between motors, was it while the system was under tension (i.e., while the machine was measuring the distance)?
I think that is very important to do since the system will be under tension under normal operating conditions. I’ve done my measurements during the calibration routine, but you might consider hooking the sled up, moving it up and down while horizontally centered, and see what the distance between motors is while the sled is at different heights.
I’d believe so, but you might want to consider doing it there as well doing it with the sled attached at various heights. I don’t know if anyone has done that yet and I’d be interested in seeing if you have any variations. That could help figure out what’s going on with your setup.
Yes, measuring during the motor measurement should reveal any flexing at its maximum.If you like, while the chain is taught, give it a bit of a downward tug, just a little, and see if things move. Be careful, though, you would be standing in the middle of ‘the slingshot’ and flying chain could really mess you up!
What If during the process of zeroing the motors and ensuring the motors are at 12 o’clock they’re off by a degree or more. I’ve looked at those things until I’m cross-eyed and get paranoid about their position. But i would imagine with the longer chains that 1 deg or so off at the top would translate more significantly as the chain got longer. Which actually brings up any recomendations for how to accurately ensure they’re at 12 o’clock…
But I know @madgrizzle had run the macro to spool out and check to ensure return to the 12 o’clock position, but if you’re off to begin with the problem will likely persist. Just a shot in the dark as everything else seems to be run into the ground…
That is slick…I COULD cut that with a 1/8" flat end mill on the Maslow…if it was behaving itself. I unfortunately do not have access to a 3D printer but I have been measuring to a framing square on both sides to confirm alignment.
Note that the files on Thingverse are different, the creator changed to a more cylindrical part. The sprocket outline is in that file, though, so you could make your own shape if you choose.