Here is the tablet’s log
cal-04-19-3-new-tablet.txt (7.2 KB)
2 Likes
I was able to record the process if that helps https://youtu.be/b4sUu8rLJ3c
2 Likes
I’ve been noticing this bug too. I think it happens when you leave the same window open during restarts. I’ve been getting in the habit of opening a new tab every time I restart and closing the old one.
3 Likes
I’ve been playing with the numbers in the calibration tool and found some bugs in my latest release of that. I have a feeling I just missed something in my last merge as I’m getting different answers. I may spend a little time to refactor the main calibration code a bit by adding some events we can hook into and do a PR on the main project to make this easier to merge as things change. Also my new entries for the initial values are ok, but it would be better if we could just copy/paste the whole thing instead of typing them all in. I will see if I can have the main code log them out so they are easier to capture, along with the CLBM.
1 Like
OK, I have a pr out for that and some cleanup I noticed. I’ll hook this into the standalone tool later, I figured out I could just let maslow load with this version, then once its loaded, just paste findMaxFitness([{bl:1697.65, ...,]) into the console to get it to do the calculations there to compare to.
@bar I did notice while testing this that there are times during reload and config changes where the UI seems to lose its configuration (i.e. the maslow settings on fluidNC tab are gone and defaults are there). I think this has happened to me before and doing an esp32 restart fixes it. There must be some race conditions in play here…
2 Likes
I think that seems likely. It feels like we’re keeping the values in too many places at once right now and we should work towards having a single point of truth for them that gets referenced everywhere.
@WFD , Would you be willing to throw these numbers into your excel model and see if it is able to find a solution?
[{bl:1697.65, br:1713.78, tr:1738.63, tl:1768.06},{bl:1627.01, br:1642.88, tr:1812.97, tl:1841.55},{bl:1456.30, br:1833.83, tr:1983.26, tl:1683.02},{bl:1536.39, br:1896.34, tr:1915.30, tl:1602.38},{bl:1618.26, br:1960.78, tr:1851.73, tl:1525.63},{bl:1769.29, br:1783.54, tr:1668.41, tl:1699.00},{bl:1945.64, br:1628.39, tr:1496.38, tl:1883.33},{bl:1877.60, br:1551.22, tr:1574.36, tl:1945.83},{bl:1814.76, br:1472.91, tr:1656.10, tl:2012.83},]
We’ve been running into a bit of a wall here where the JS system keeps failing to produce valid results and it would be awesome to confirm that a valid solution exists that we aren’t finding. It might give us more insight into how to improve.
@ronlawrence3 It occurred to me that I can lean my horizontal frame against the wall in portrait mode and try something somewhat similar so I’m going to do that today and see what I can learn
1 Like
where in git do I find the javascript that is doing the calculations?
David Lang
Bar wrote:
Line 642 here is where the heavy lifting is done:
Edit: Edit to add the link 
…pressed enter too quickly
Bar wrote:
Line 642 here is where the heavy lifting is done:
Calibration-Simulation/index.html at main · BarbourSmith/Calibration-Simulation · GitHub
Thanks
1 Like
I can’t get Excel Solver to converge. I might need a better starting point. Can you tell me your best estimate / measurement of the belt attachment coordinates for the frame corresponding to this data?
Thanks.
2 Likes
My current best guess data for my frame:
copy/paste data from browser
/Maslow_brXBottom right anchor x (normally width in mm) 2504.650
/Maslow_brYBottom right anchor y (normally 0) 0.000
/Maslow_brZBottom right z (normally 117) 117.000
/Maslow_tlXTop left anchor x (normally 0) -53.345
/Maslow_tlYTop left anchor y (normally height in mm) 2364.349
/Maslow_tlZTop left z (normally 144) 144.000
/Maslow_trXTop right anchor x (normally width in mm) 2503.704
/Maslow_trYTop right anchor y (normally height in mm) 2353.292
/Maslow_trZTop right z (normally 97) 97.000
/Maslow_blXBottom left anchor x (normally 0) 0.000
/Maslow_blYBottom left anchor y (normally 0) 0.000
/Maslow_blZBottom left z (normally 75) 75.000
this is for CLBM:
CLBM:[{bl:1697.65, br:1713.78, tr:1738.63, tl:1768.06},{bl:1627.01, br:1642.88, tr:1812.97, tl:1841.55},{bl:1456.30, br:1833.83, tr:1983.26, tl:1683.02},{bl:1536.39, br:1896.34, tr:1915.30, tl:1602.38},{bl:1618.26, br:1960.78, tr:1851.73, tl:1525.63},{bl:1769.29, br:1783.54, tr:1668.41, tl:1699.00},{bl:1945.64, br:1628.39, tr:1496.38, tl:1883.33},{bl:1877.60, br:1551.22, tr:1574.36, tl:1945.83},{bl:1814.76, br:1472.91, tr:1656.10, tl:2012.83},{bl:1747.50, br:1402.68, tr:1745.70, tl:2078.44},{bl:1553.87, br:1581.70, tr:1895.80, tl:1914.16},{bl:1378.64, br:1779.57, tr:2059.70, tl:1763.39},{bl:1224.03, br:1991.60, tr:2234.34, tl:1629.82},{bl:1305.44, br:2043.23, tr:2168.71, tl:1538.13},{bl:1392.80, br:2099.30, tr:2106.82, tl:1449.25},{bl:1483.16, br:2159.56, tr:2049.18, tl:1363.75},{bl:1574.94, br:2224.08, tr:1996.17, tl:1282.19},{bl:1695.26, br:2034.28, tr:1798.41, tl:1448.10},{bl:1837.61, br:1861.88, tr:1608.21, tl:1628.36},{bl:2006.49, br:1715.61, tr:1428.07, tl:1818.53},{bl:2193.66, br:1595.08, tr:1262.51, tl:2016.00},{bl:2125.04, br:1504.53, tr:1344.94, tl:2068.83},{bl:2063.33, br:1415.84, tr:1431.20, tl:2126.19},{bl:2005.54, br:1332.28, tr:1520.78, tl:2187.67},{bl:1952.55, br:1248.62, tr:1612.97, tl:2253.20},]
1 Like
Also here for 0.69: ESP3D-WEBUI/www/js/calculatesCalibrationStuff.js at Maslow-Main · BarbourSmith/ESP3D-WEBUI · GitHub
1 Like
It turns out that I had made a copy and paste error with the CLBM values, which is probably why Excel Solver wouldn’t converge - the numbers made no sense. When I corrected these and used the starting point that @ronlawrence3 gave me, I ended up with:
| Excel | ||
|---|---|---|
| Maslow_tlX | -49.00 | |
| Maslow_tlY | 2424.06 | |
| Maslow_trX | 2450.42 | |
| Maslow_trY | 2408.20 | |
| Maslow_blX | 0 | |
| Maslow_blY | 0 | |
| Maslow_brX | 2449.24 | |
| Maslow_brY | 0 | |
| Max Error: | 0.97 | Relative to centroids |
| Avg. Error: | 0.50 | Relative to centroids |
| SSE: | 10.76 | |
| ‘Fitness’: | 1.310002688 |
…which is pretty good.
I notice that in @ronlawrence3 's message the CLBM values have a lot more terms than the ones in @bar 's message, although they started out with the same values. I used the shorter set in @bar 's message - was that what you all intended? Let me know.
1 Like
…and note that the ‘fitness’ number is incorrect, probably lower than it should be. If someone could give me a more detailed description of how it is calculated then I can fix that.
1 Like
I will plug these in and jog a bit to see how they are. at first glance they are reasonable looking in terms of where my anchor points are in relation to one another. I don’t have anything more accurate than a tape to measure with.
1 Like
I put in WFD’s numbers, restarted (made sure the maslow config looked right) retract/extend/ then I homed and moved the sled around in a 1m square using the jog buttons. It behaved well. I noticed the lower belts (right more than left) would be a little bit slack, but not too much. Certainly better than I’ve seen so far. Also, my “skipping” when going up is no longer happening, but that may be the hardboard I attached on top of my OSB, which has much less friction.
I’m tempted to try a cut, but I am a little weary still.
1 Like
That sounds like some excellent progress!
If you can jog around comfortably then you should be good to cut at least within the area that you have jogged around in.
1 Like
trying a dry run of the maslow logo with Z up and no router going… looking good so far. I may cut it after I get through this.
1 Like
Got this far and the power plug somehow fell out of the bottom of the maslow board. It looked like it was going well. Unfortunately I’ll never find “home” again, so likely this will turn into a pretty weird logo when I try it again.
1 Like
WOW so it remembered where home was but now it got up to top of the logo and stopped. All I see in the console is MSG:ERR Emergency stop triggered over and over and I can’t stop it to select anything to paste here…
OK, powered down and now can copy messages:
[MSG:INFO: Measured waypoint 0]
[MSG:INFO: Center point deviation: TL: 0.218 TR: 0.355 BL: -1.177 BR: -2.827]
[MSG:INFO: Center point deviation: TL: 0.218 TR: 0.355 BL: -1.177 BR: -2.827]
[MSG:INFO: Center point deviation within 15.000mm, your coordinate system is accurate]
[G54:-100.000,-130.000,-22.000]
[G54:-100.000,-130.000,-22.000]
[MSG:DBG: /SD/Maslow Logo.nc found]
[MSG:DBG: path /SD/Maslow Logo.nc CT text/html hash ]
[GC:G0 G54 G17 G20 G90 G94 M5 M9 T0 F0 S0]
[MSG:INFO: Tool No: 0]
[MSG:INFO: Using spindle NoSpindle]
[GC:G0 G54 G17 G20 G90 G94 M3 M9 T0 F0 S0]
[GC:G1 G54 G17 G20 G90 G94 M3 M9 T0 F762 S0]
[GC:G3 G54 G17 G20 G90 G94 M3 M9 T0 F762 S0]
[GC:G2 G54 G17 G20 G90 G94 M3 M9 T0 F762 S0]
[GC:G1 G54 G17 G20 G90 G94 M3 M9 T0 F762 S0]
[GC:G3 G54 G17 G20 G90 G94 M3 M9 T0 F762 S0]
[GC:G1 G54 G17 G20 G90 G94 M3 M9 T0 F762 S0]
[GC:G3 G54 G17 G20 G90 G94 M3 M9 T0 F762 S0]
[GC:G1 G54 G17 G20 G90 G94 M3 M9 T0 F762 S0]
[GC:G3 G54 G17 G20 G90 G94 M3 M9 T0 F762 S0]
[GC:G2 G54 G17 G20 G90 G94 M3 M9 T0 F762 S0]
[GC:G3 G54 G17 G20 G90 G94 M3 M9 T0 F762 S0]
[GC:G1 G54 G17 G20 G90 G94 M3 M9 T0 F762 S0]
[GC:G3 G54 G17 G20 G90 G94 M3 M9 T0 F762 S0]
[MSG:ERR: PANIC! Stopping all motors]
[MSG:ERR: Emergency stop. Update function not being called enough.1002ms since last call]
[MSG:ERR: Position error on Bottom Right axis exceeded 15mm while running. Halting. Error is -21.939mm]
[MSG:ERR: Emergency stop! Stopping all motors]
[MSG:WARN: The machine will not respond until turned off and back on again]
[MSG:ERR: Emergency stop triggered.]
[MSG:ERR: Emergency stop triggered.]
[MSG:ERR: Emergency stop triggered.]
[MSG:ERR: Emergency stop triggered.]
[MSG:ERR: Emergency stop triggered.]
[MSG:ERR: Emergency stop triggered.]
[MSG:ERR: Emergency stop triggered.]
[MSG:ERR: Emergency stop triggered.]
...