Any guesses as to discrepancy in arm length vs. measured arm length?
-
Today I received my 360mm Haydn Rods. After a few rounds of 8-factor calibration I ended up with this:
Calibrated 8 factors using 16 points, deviation before 0.013 after 0.008
So it is definitely the Tricklaser Rods that are causing the calibration issues, I will ask the tricklaser guys if they have an idea why… (Or did anybody already contact them?)
-
@MiR:
Today I received my 360mm Haydn Rods. After a few rounds of 8-factor calibration I ended up with this:
Calibrated 8 factors using 16 points, deviation before 0.013 after 0.008
So it is definitely the Tricklaser Rods that are causing the calibration issues, I will ask the tricklaser guys if they have an idea why… (Or did anybody already contact them?)
What radius did you probe at?
What arm length did it auto calibrate to, does it match real rod length?
Does it print parts at true scale now?What did you change to go from ball-cup pivot balls to mag-ball pivots?
-
I did not notice if anyone is using the "h" parameter in their G30 probing?
This value is a height correction for probe trigger height.
What I understand is that your probe triggers at different heights depending on location of bed.
So, unless you measure actual trigger height at each probe location (and put those values in bed.g) I do not believe you are getting true calibrations, especially when tweaking down to such small deviations like you all are chasing after.
Furthermore, I suspect those trigger values change when you change arms, joints, or any other part of your mechanical linkage.If you think you need to get closer than what auto calibration probing does, why not use mesh bed height compensation on top of the auto calibration??
If you know your real arm length, I don't think you should let auto calibration change that value.!
Also, probing at extremes of travel (140 mm radius) does not help, I feel the geometry out there is to unstable to be reliable (and worse the shorter the arms are) I get more predictable results at 125 mm radius.
-
I probed with 120mm radius, forgot to re-enable the larger 130mm radius (I was using shorter 300mm Haydn Rods before the new ones arrived, and for those rods a larger radius was a little extreme).
I did only S8 calibration, dc42 does advise against using S7
I cannot yet say if the parts are true scale with the 360mm arms as I ran into another issue with the hotend cooling fan (more on this in another thread), but with the 300mm Haydn rods Calibration cubes were close to perfect in size.
To change away from tricklaser I printed a small adapter for my carriages that has the same hole size/distance as the tricklaser ball bar-bells, and I also printed a new effector that uses Haydn's spheres instead of the bar-bells.
I will replace the adapters soon, but so far they do their job.On h-parameter:
I used then in the beginning, when doing all the testing I did not use them anymore because the effort was simply too much as I was trying a lot of probing patterns/changed probing points. Now that I have my 'final' configuration I will create proper h-values again for all my measure-points.
With tricklaser rods the best I could ever get was a deviation of 0.087, that's factor 10 worse than what I see today, now that I have proof that they are really the parts to blame for the issues I can move on finetuning the printer.
As far as I know the correction factors will not be used with mesh calibration so I am not sure if the height map I get out of it is really the heightmap of my bed or just a map of sensitivity of my FSR's…. I have finished piezo-probe kit from DJDemonD ready for installation but I did not want to change too many things before the Haydn rods arrived to be able to compare the results.
-
That is very interesting info. Thanks for sharing.
I think there is more going on with the printer than simply changing between a mag-ball joint and ball-cup+springs, essentially they both should be the same spherical joint.
Would you have the time to gather some data on probe trigger distances? comparing the two joints, probing a the same locations the auto-calibration uses.
I am very curious if probe trigger offsets are the same. -
That is very interesting info. Thanks for sharing.
I think there is more going on with the printer than simply changing between a mag-ball joint and ball-cup+springs, essentially they both should be the same spherical joint.
Would you have the time to gather some data on probe trigger distances? comparing the two joints, probing a the same locations the auto-calibration uses.
I am very curious if probe trigger offsets are the same.If memory serves, they were similar on the probe points. I never really found any points that differed more than ~0.1mm in probe distance and I never could find a pattern of different probe distances at the towers vs between them (where I'd previously been adding up to ~0.3mm to ~0.5mm).
Right now I have everything working nicely though and I've been printing continuously, so it may take me some time to check further.
I'll go ahead and order some custom length 350mm ball joint arms from you though and we can finally solve the question of what exactly is causing the calibration weirdness. Then I can run a 300mm test and a 350mm test with the same arm style and 350mm should be plenty of length to reach the 300mm edges of the bed.
I feel like the 300mm ball joint arms were pretty good up to around 250mm in diameter for printing, which probably works for 90% of people using them.
-
So I finally got around to installing some ball-cup joint arms that are 350mm in length and they look pretty similar to the 325mm arms that were the standard M3 traxxis metal joints in how they are calibrating with the FSR sensors.
My old 325mm arms config-override:
[c]M665 L325.000 R133.101 H332.907 B150.0 X-0.208 Y0.178 Z0.000
M666 X-0.608 Y0.878 Z-0.269 A0.00 B0.00[/c]The new 350mm ball-cup arms config-override:
[c]M665 L350.000 R141.775 H311.606 B150.0 X0.150 Y-0.031 Z0.000
M666 X-0.043 Y0.130 Z-0.086 A0.00 B0.00[/c]Calibration on the 350mm arms looks like:
[c]Calibrated 6 factors using 19 points, deviation before 0.081 after 0.031[/c]My 300mm ball-cup arms simply wouldn't calibrate properly for prints over ~200mm in length, but these longer 350mm arms appear to be working much better.
I will let this print run tonight (it takes about 5 hours and is 268mm in length so it's using a lot of the print diameter) and once it is finished I should be able tell how well the FSR sensors and the 350mm ball-cup arms work together. Initially, I'm not seeing the variation on the first layer outer perimeter from the 300mm ball-cup arms, so this should isolate if the ball-cup arms and FSR sensor setup simply needs longer arms to calibrate properly.
I can also post more calibration output later after this print.
-
Lots of work here. Mind posting some pictures of the different rods you have used with the results you obtained?
-
I'll try to gather the arm sets together and post pictures later. For reference the three sets I tested on a Max Metal frame were:
300.15mm Trick Laser Ball-Cup Carbon Fiber Arms (these gave me lots of calibration and print issues for larger items)
325mm Trick Laser Traxxis Joint Arms M3 screw variety (these print quite well and solved my print issues)
350mm Trick Laser Ball-Cup Carbon Fiber Arms - Custom Length (these print as well or better than the 325mm arms and also resolve or rule out issues related to the ball-cup design)
The first print on the 350mm arms had an issue that shifted layers in the last hour (this happens to me maybe once in 100 prints - but it doesn't look related to the arms, I think I had X/Y movement too fast. I will update if it becomes a pattern).
My second print came out excellent though. I'd say on par with the 325mm arm quality if not better. In short, unless you need the height, the 350mm arms are a good trade-off for getting a full ~280mm diameter and beyond print quality for the first layer. It could be that the ball-cup joint artificially lowers the arm length since the ball-cups are larger diameter (I'm not entirely sure).
-
Images as requested. The 325mm arms are the standard traxxis joints.
The 300.15mm and the 350.00mm arms are the ball-cup joints used on the Rostock Max v3.
Not pictured is the older Trick Laser carriage that pairs with the M3 Traxxis joint arms. It's the same as the ball-cup carriage, only larger with larger wheels.
-
My delta, from a kickstarter project, has haydn arms.
They are measured at 304mm.
It has custom metal carriages with 2020 V slot style extrusion.Using S8 I get
; Delta parameters
M665 L335.697 R163.858 H245.141 B115.0 X-0.423 Y-0.941 Z0.000
M666 X-0.443 Y-0.225 Z0.668 A0.38 B0.45-Calibrated 8 factors using 10 points, deviation before 0.015 after 0.002
Using S6 I get
; Delta parameters
M665 L304.200 R154.758 H245.057 B115.0 X-0.252 Y-0.735 Z0.000
M666 X0.330 Y-0.387 Z0.057 A0.00 B0.00- Calibrated 6 factors using 10 points, deviation before 0.309 after 0.308
I have no springs, all magball. All metal construction,
So how can the arm length be a measured 304 and a calculated 335 ?I guess its recommended i keep to the S6? I've also used mesh bed compensation, which as others state looks like a ski slope with S6 and a flat plate with S8
-
You have some large mechanical errors going on. Either your bed is out of level, one of your towers is tilted, or a good bit of effector tilt. Possibly all of the above.
You need to go through the mechanical aspects of your printer and find what's out of square.
-
Neither S6 and S8 would modify your arm length, are you sure your calibration routine is correct?
Do you have other M665 in your config.g or config-override.g? -
no other M665 commands.
I usually move any config override values to the main config file and delete the override.My delta may be out of square somewhere but not massively. All towers are square to the bed when using an 8" precision square.
The end stops are within .3 of each other I believe. I can run though this again.
The effector may have some tilt but not visual to the eye and no room for a level. Haydn carbon arms, all measured to 304mm, delrin ends, magball connections. All 6 pairs of magballs measured in width (of each pair set) to be within 0.1mm or so with calipers.Its not terrible if I use S6, I just find it odd that when I use S8 (or S9) it calculates and changes the L length of the rods to be something so far off.
-
One idea on your s8 values:
my guess is that you did s9 calibration, put the values to config and then switched to s8, forgetting to replace the arm length with the correct length of 304.2
As said above, s6 and s8 do not guess the arm length, you have to provide it as input.
Also, run s8 (or s6) several times, it can take a number of runs until everything is dialed in, you will see at some point that the average deviation does not change much anymore. Then do M500 and copy the values over from config.override to config -
@MIR yes you may be correct, I was using S9 at one point.
I have been running 2 or 3 auto calibrations till the deviation amount converges.
Thanks for the tips. -
Sorry to dig up an old thread here, but I am fighting the same issues with my Rostock with all of the trick laser components.
I eliminated the "ridges and valleys" in the bed mesh by upping my stepper current and lubricating the joints of the 300.15mm Trick Laser arms with some 50K diff lube for RC trucks. When I got that result, I was initially very excited, but with the ridges gone I could see the same three point high/low shape described by many other users across this forum.
I have the components to build some new arms to whatever length I choose, but I am having a hard time deciding what length to go with as a compromise between a) getting a good bed leveling result and calibration via FSR by eliminating the effector tilt and b) having the arms become dangerously near vertical when getting near to the edge of the bed near a tower.
Can anyone report further on the performance of their Rostock with longer arms? Have you had any trouble with being too close to vertical on the arms when near a tower?
Thanks ahead of time!
-
Having the arms go vertical or even beyond when printing close to a tower isn't a problem. What you need to avoid is arms going below about 20 degrees to the horizontal.
-
Perfect, that's exactly what I needed to know. Thanks!
-
As an interesting test before making the longer arms, I just reduced the calibration radius to 87.5mm and reduced the probing radius for the bed mesh correspondingly, and the results were pretty astounding. Previously, the best calibration I could get was roughly 0.18 deviation, and with no physical change the the printer, the deviation went down to 0.021!
The bed mesh in the smaller radius was also greatly improved. Linked below are the results of the bed mesh at a 140mm radius followed by the 87.5mm radius.
I will post my results when I change the rod lengths as well.
