I’ve calibrated and find that the sled sits about an inch lower than board centre when finished. I will probably recal later, but want to sound you out on what may have caused it before starting over. I’ve used the latest version of CG, and have used the calib process to measure the frame and chains. My question is about the linkage of the end of the chain to the ring. I have a bearing block at the end of each chain into which the chain locks with a split pin. I’ve measured the internal diameter of the sled ring and have used this as the number for radius. However, I’ve not included the distance between the end of the chain and the internal diameter of the ring. This is probably about 3/4" on each block, so I guess I should be adding that measurement to the ring dimension, right? This would likely be the cause of the discrepency! I think I’ve just answered my own question! Anyone out there with a ring sled able to elaborate on the linkage and measurement? My linakage can kinda be seen from the pic:
Out of curiosity, did you complete the triangular kinematics test cut pattern, the last step of the calibration process? That step is designed to determine your effective rotation radius, including any extra links and mounting brackets.
Your sled sitting below center could be an inaccurate rotation radius, as well as an inaccurate motor Y offset or even motor spacing.
What parameters did the calibration process end with?
Yes the measurements are from chain end to bit center! Add it in👏🏻
Yes, obvious really, but I figured it was probably something simple, so I didn’t progress until I’d sorted that. As such, I’ve yet to perform the test pattern, will probably try and find the time tonight now that we’ve figured out that discrepency.
Is it ‘chain end’ to spindle centre, or ‘centre of last chain hole’ through which the pin goes?
Cheers,
Miles
I have it at chain end.
Yeah, I figured that’d probably be best, but it did make me wonder if there is an exact answer as although it’s only likely to be a 1-2mm difference, it could have a noticeable bearing on cut accuracy at the periphery of an 8x4 board. Any thoughts?
I have cut a simple circle or square and changed the value in the settings to see the difference. It is kind of a fine tune. Just cut /or draw a known circle and measure with calipers. If it’s oval adjust mm differences until it’s round. Since you have the pen you could change colors!!
There’s a bit of uncertainty about this, but the automatic calibration for triangular linkages does a great job of finding the correct value with three tape-measure values and some calculations.
I’ve played around with it, and it looks to me like the value will be from the center of the bit to the point on the chain which was at 12 o’clock top-dead-center when you did the chain ‘zero’ step.
until you do the calibration, and ‘return to center’, the position of the router
doesn’t mean anything
k
This does not work. Your circle could be accurate becouse your different sources
of error cancel at that point in the work area, but then as you move away from
there the errors increase/decrease at different rates and the result will not be
consistant in different areas.
This approach is what the calibration used to be. The current calibration
actually figures out what the multiple numbers should be rather than just adding
error in one to try and cancel out errors in several others.
This is exactly what it is.
So, Blurfl, you mention three tape measure values. Did you prefer physically measuring with a tape, over letting GC figure it with the chain process? I’ve measured with a tape, quite accurately, and it differs significantly (~3-4mm) from the GC measurements.
The way I figuerd it, if I can set the motors perfectly centred above the board a known distance from the centre of the board, AND, everything is true and square, then I’ve got all the measurements I need from basic geometry. I got to this point, realised there was catenary error from chain sag and went into a head spin for a few days!
The issue I’m trying to resolve in my head is whether GC is best for making those measurements as the software calculates everything from that, but can’t help thinking my accurate tape measurements are correct.
Was going to run the calibration program and return to centre this evening, but had to pop out and it’s late, so I’ll have a crack at it in the morning. Will probably run it based on GC measurements first, then plug in my measurements before running it again, to discern any difference. Figure if I do it with the sharpie, in differing colours the difference should be apparent. However, if like most of this project it comes together first time, I might actually mount the router and cut something up! Probably got a bit of sled balancing to do before that though.
I’m sorry, I must have been unclear. starting with v1.03, the final step of the calibration sequence, if you’ve chosen the triangular Kinematics, makes four cuts and asks you to take three measurements from them. Those are the measurements I meant.
I use a pen to check calibration too 
. I use 0.8mm pens, and enter that value for the bit diameter in the final cal. Panel. Remember that the measurement to the top edge in that group adds 50% of the bit diameter to calculate motor height. The horizontal measurements are right edge to right edge, so the diameter doesn’t matter there. I’ve observed that being accurate about the thickness of the pen stroke makes an important difference in the calculations.
Do run the auto cal and compare it to your calculations, and let us know what you find. We all want more precision!
Thanks for the nib width tip.
The way I figuerd it, if I can set the motors perfectly centred above the
board a known distance from the centre of the board, AND, everything is true
and square, then I’ve got all the measurements I need from basic geometry. I
got to this point, realised there was catenary error from chain sag and went
into a head spin for a few days!
how much error does this actually cause? When this was first reported, I did
some calculations based on the chain length and the amount of sag and calculated
it at some tiny value (IIRC, something like 0.1mm error). If I was wrong and we
need to include a correction factor at this stage, it would be really good to
know.
The issue I’m trying to resolve in my head is whether GC is best for making
those measurements as the software calculates everything from that, but can’t
help thinking my accurate tape measurements are correct.
your tape measure is correct, the question is why the software calculation is
incorrect.
3-4mm error on the motor distance is going to cascade down the line.
you can use a tape measure to measure motor distance and rotation radius, it’s
hard, but possible to use it to figure out Yoffset, but you cannot use it to
figure out the effects of chain sag and machine angle.
so let’s see about spending some time figuring out why the motor distance
measurement is wrong. If there’s some error there, it will cause all the rest of
the calculations to be wrong.
Yep, that’s what I’m working on tmrw. I think an afternoon on the forum has provided me with plenty to be getting on with.
My current motor mounts are a bit shonky, but sturdy. CG measures the distance between them as 3073.3 and I make it 3077. This is likely due to my poor measurements on squareness of motors to frame (not easy to explain what I’m going on about) I’ve sourced a couple of the proper motor mounts and have a plan going forward as to how to use these to get better accuracy.
It has also occurred to me that the last calib of chain length I did I may only have ‘tightened chains’ once, whereas a few goes may actually have pulled the chain tighter (?). There was still a noticeable sag following the tightening.
I suspect it’s all down to my working late in the garage and missing something obvious. Will report back following the next session.
This could also be the result of flex. see if you can stand something up next to
each of the motors, and then while GC is measuring, see if they’ve moved a
couple of mm
error caused by chain sag would cause GC to think the distance was longer than
what you actually measure
Does everyone usually hit the ‘tighten chains’ button several times?
i don’t know, but if the frame is rigid enough, it shouldn’t matter 
I don’t remember your frame design, do you have a solid top beam? one with the motors offset?
yes, and I give the chain a very gentle ‘lift’, but these don’t seem to matter much beyond making me feel like I’m helping 
.