MAX V3 calibration woes or am I an idiot? HA

[zakfarias]

So, I decided to finally add a delta to my printer collection. I have been avoiding building one like the plague due to the bad reputation they’ve developed over the years for calibration woes.
However, seeing the HE280 in action, I figured I would give it a try and worst-case scenario I could always switch to a duet Wi-Fi setup and run their smart effector when it becomes an official release.
A Rostock MAX V3 kit was ordered.
Assembly of the machine went relatively smoothly. Followed the newest instructions as closely as possible and verified towers were square to bed with a machinist square.
I went through the initial setup process and started a 75mm disk test print, high and low spots everywhere. Ran the online delta calibration tool 3 times until it claimed an error of 0.02mm. Reprinted, there were improvements but still large differences in high and low points as well as part scaling.
Reset the firmware and went through normal calibration stages again, except, this time I ran the octoprint calibration plugin. Best results deviation 0.14mm; still have low spots in bed going towards x tower and high spots between Y and Z.
Reset firmware again.
Again, went through basic calibration. G28, G29(x3). Endstop screws were adjusted until offsets were 0, 1, 1 (within 4 as suggest by mhackney on a previous post).
This time I really paid attention to the probing points for the calibration wizard.
I very consistently get points of -.02 to 0.02 at the towers
In between every tower, the probe point is 0.39 to 0.42
Middle of the bed is usually -0.05 or so
This would mean the bed is identically high between all towers right?
I have tried multiple combinations of bed clips to see what differences there are. The results are always relatively the same.
Rod length in eeprom is stock 291.06
Horizontal radius comes out to about 144.40
Another side note is that the heat bed does not go over 82C even with power supply set to 13.44V and a ½” of cork sitting on top of it. Resistance is 1.2 ohms, which is within spec. In addition, the hotend temperature does not seem right. On my main workhorse triple C-Bot I run PETG at 250-255. On the Rostock (even after three rounds of PID tuning hotend) it seems to be extruding too hot even when set to 230.
SeeMe is sending a new heat bed and mounts after a previous conversation with them, but I want to make sure I am not overlooking something or having a “duh” moment here since I am a novice when it comes to deltas. All my other machines I have built are C/D bot variants or Cartesian
Reflective note: On extrusion, the closer to the X tower the farther the nozzle is from the bed. The farther from the X tower @ a diagonal 60 degrees or so, the closer the nozzle is to the bed until it is too thin. Saw this when printing a single layer disk w/ concentric infill that took up the entire bed. Oh and S3D for slicing. The width of the area affected by this looks like a hourglass >-< like this with the affected area being widest at the points closest and further away from the X tower. I have loosened and tightened everything on the base multiples times now so that the machine can relax into the appropriate position, re-verified squareness to bed.

[Xenocrates]

The high spots between towers seems to be semi-solvable through minor adjustments of the axis steps per mm. The symptoms specific to a tower point to a problem in the seating of the tower (So check your measurement instruments, as well as the angles there). There's some others working the problem here:
http://forum.seemecnc.com/viewtopic.php?f=112&t=12025

Best of luck resolving it, and welcome to the forums (I don't think you're an idiot until you make the same mistake more than twice for minor ones). I hope we get your printer working properly, and you learn to love the interesting sight of a delta.

[dc42]

There is a long thread on this topic here https://www.duet3d.com/forum/thread.php?id=2002&p=1. It appears that over-strong springs between the pairs of arms make this problem worse.

[Gdot]

Just checking to see if you've made any progress on the outer perimeter sine wave calibration issue you have. I have the exact same issue, I see where the previous post made mention of spring tension contributing to it. I had the issue with super loose stock arms and springs and with the TrickLaser carbon fiber arms with tight springs so I dont know.

[zakfarias]

So far I am still waiting on a new bed and bed mounts from SeemeCNC. However, I did Remove the plastic springs and just used a heavy rubber band around the outside of the arms at the top and bottom. It didn't make much of a difference but the Octopi delta calibration plugin seemed to move smoother. After some AHA moments I noticed that the WHIP and bowden would slightly touch the "X" tower arms when running calibration. I shortened the whip a couple inches shorter than recommended in the build instructions. It helped some but still was hitting the arms at the "X" tower. So, for fun, I reran the calibration but held the cable whip away from the arms at all times. Calibration mean deviation went from .16 average to .10mm. Did this a couple more times and always got 0.1mm I then reinstalled the plastic springs and tried it, deviation consistently became 0.13mm so springs definitely are influencing the issue.
So, part of my issue at least is Effector tilt from the cable whip towards the "x"
However, I haven't run the test while watching the probe offsets to see if the sine wave issue still persists. I'm sure it hasn't gone away but at least it calibrates a bit better and allows me to print on one half of the bed pretty consistently I did a print 8"x1" with the part slightly offset from center towards the Y tower and first layer was great...If only I can now get it to not have vertical and horizontal ripples/dimples down the entire length of the part (yes both kinds).

[Gdot]

I noticed a similar issue with the Octopi calibration plug-in. It seems to travel too fast. It seems to travel so quickly that it does not travel at a steady speed from point to point. I never noticed this issue when running any other gcode.

[Gdot]

Just checking in to see if you ever had any luck with your calibration issue?