I’ve been following this thread closely as my machine produces flat roughly by .5 inch on a 18’’ circle, If I follow the calibration.
Upgraded to the latest firmware to make sure I’m not crazy.
I have found a work around for myself, I’ve been able to change the motor spacing from the measured by Maslow 3000.38mm to 2980.5 and get circles within roughly 1.5mm of round. FYI I make the spacing to be 3002mm by tape. Robot around the middle top third. I leave chain sag and all the rest of the Maslow calculated math as it is…I have found for circles the slower I go down to 500mm/min the circle is ever so slightly taller maybe 2mm, so speed plays some part it seems.
I guess I’m asking where have I gone wrong here?
I’ve got a new sled made that has the hole and installed so it can’t get much better than that. I’m using 25mm chunk of steel as a sled weight. IThe feels comparable to a pair of bricks. Should I try more weight? Ideas?
It’s functional for me so I’m not upset I’m just really stumped. I mean, I used to fix aircraft in the military so following written instructions with pictures is usually a breeze lol.
I’m using the latest ring kit batch with a frame loosely based on the new top bar one that’s suggested be built now.
Thanks for your help and ideas. I made sure to use vice grips just soft enough to hold the chain for the machine measurement because it was popping off the odd time and I’ve been doing it alone this far. So it really got a good chance to cinch the chain. 2x4 on edge doesn’t seem to mind.
Oh also interesting, the width is perfectly fine after calibration. I wish I was smart enough to understand the math relationship to the results.
Measuring the distance between motors to be within less than 2mm of what Maslow measures is pretty good. But it sounds like a lower distance yields better results.
How well do you think you were able to measure the distance between calibration cuts? I trust a mechanic to follow directions, but I wonder if this a source of problem (it’s what I’m trying to overcome with an alternative calibration routine). I know I’ve misread a tape measure a time or two. The calibration routine calculates the rotational radius (distance from end of chain to the router bit) and the chain sag factor. If the either or both were calculated incorrectly, then modifying the distance between motors manually can help.
I thought this too but I’ve been setting the 300mm on my tape at the inside of the cut held with a clamp at the edge of the maslow so I can really keep it on the mark.
Also, we are measuring inside to inside right? Last time was something like 1925mm and 1921mm I forget the number 5 measurement.
Just came in from upgrading my firmware/GC and calibrating again. I’m pretty sure it tells you to measure from left side to left side on the vertical cuts
I think the measurements should be from left side to left side or right side to right side. That removes the bit diameter from the horizontal measurements. That’s not possible in the measurement to the top of the sheet, so one half the bit diameter is subtracted form that one in the software.
Hey ArnoldCP…
Is the chart/spreadsheet you posted in #138 available somewhere… I ran the Calibration Benchmark, but do not know what to do with it.
Thanks for your help!
Today I had great success. I tried the calibration again but this time I took the measured number between the motors and divided that by shaft centre to shaft centre. As measured by me was only 2 and a bit mm different from the maslow but slightly longer.
So I took my two numbers and divided my number by the maslow number and came to the conclusion that my chain pitch was off a bit. Setting things up with my measurement and then chain correction I got a 1mm difference in width from one end of the 6ft growth ruler. My best results yet!
My magic number is 6.349mm
Thanks again everyone for all the brain storming in this thread. I think its what spurred me on to experiment.
So, I recalibrated using v1.11 of firmware and GC. I measured the motor spacing using a steel rule and the Maslow chain stretch process. My measurement was 3009.9, whereas the chain stretch registered 3001.9 mm.
I used my measurement, along with rotation radius of 138 per a note I read from @bar . After running through the entire cal routine, including the triangular test cuts, the data summary reported the radius as 140.7.
Doing my circle tests, the are still pretty much dead on in width, but generally shorter in height. Ones in the center and towards the top a little worse that bottom and sides.
The large center circle is 1/8" off vertically, with the smaller one inside about the same.
Your results are very similar to my original results when I started this thread. I have gotten down to about 1/16" vertically on a 10" circle in the center of the sheet.
I think you really want to use your measured value not the chain stretch value. You are more accurate at this point. You might backup your .ini file and adjust your chain pitch. I am calculating your pitch closer to 6.366mm (which seems too long). Ideally, your measured value should match the chain stretch value within a few hundredths.
So, looking at my .ini file, it shows 6.35 for the pitch value, which I assume is the default. In recent posts on this thread, it looks like you used 6.349 as your “magic number”, which is a pretty small adjustment, and smaller than the default. As you note, since my measuremen is longer than the Maslow chain stretch value, the actual pitch also has to be larger. What would you suggest I start with? It seems like a lot of other machine calculations depend on this as well.
