Is openDACT best option for Delta auto-calibration? If not, which one?

1. I've never had luck with the Marlin firmware with delta autocal. I've tried these with both a microswitch Z probe and FSRs under the bed with no success.
2. Pilot626 on the SeeMeCNC forum developed a really nice heuristic autocal for deltas on Smoothieware. I moved a couple of my machines to Smoothieboards to take advantage of it. I've also contributed to the Smoothie source (the TemperatureSwitch module). The issue with this autocal is that the current Smoothieboard does not have enough memory to run the firmware with the autocal, LCD and ethernet support at the same time. So I was constantly reflashing firmware to go back and forth. Once you have autcal working, you'll find that you want to use it often to keep things highly tuned. These machines do change with environmental conditions (temperature and humidity).
3. When I learned of dc42's autocal in RepRapFirmware (his branch) on the Duet, I ported 1 machine to that. It works fantastically. One pass (13 points) on my Rostock or Kossel250 takes less than 30 seconds and the results are excellent. I now run autocall at the start of each print when I'm printing heavily (I manufacture and sell a 3D printed fly fishing reel so near perfection is required). I am now migrating all my printers to Duets and the dc42 fork.
4. Finally, a comment on probes. I've tried the Kossel microswitch and various other early probes. The issue with them is the offset from the nozzle tip. On a delta, I assert that probing at the nozzle tip is essential - or at least gives much better results (as David aka dc42 describes above). When Johann reported his first FSR (force sensitive resistor) usage on a mini Kossel I was hooked. I did a fair amount of experimenting with FSRs with mixed results. But, when the JohnSL board that interfaces the 3 FSRs, provides some sensitivity adjustment, and makes them look like a normal endstop switch to the hardware & firmware became available, that changed. I now have the JohnSL and FSRs on all of my deltas and just completed a 3-point system using them on a Taz 4 Cartesian printer. They all work great. David warned me that there might be a bit of height error on the probe points with FSRs as they require force to trigger. I have not observed this but I think it's because my heated beds are all very rigid and there is "no" mechanical slop in my FSR mounting system (which you can find on the SeeMeCNC forum). I have PEI surface on borosilicate glass on 1/8" (~3mm) aluminum (for heat dissipation) on a 3/8" (~9.5mm) melamine "snowflake" for insulation and stiffness. I have not tried David's IR sensor yet.
5. I've also contributed to and worked with Rollie on the openDACT calibration. It is a work in progress and currently only works with Repetier firmware. I have not been active with it for the last month or two and prior to that I really did not have success with getting it to work. I had done a Mac OSX port of the source too. I think if it can be made to work and output results to various formats to support Repetier, Marlin, Smoothie and even RepRapFirmware it will be a good option for initial calibration. It will not be good for frequent "touch ups" like I do before every print.

Quote
mhackney
By the way, I wrote a Python script to automatically generate the gcode for the bed.g file for the probing points. I posted it on the SeeMeCNC forum. David, I keep forgetting to send you a copy.

Quote
realthor
RepRap Firmware runs only on the Duet, so not yet, I don't have Duet yet. BTW is RepRap Firmware going to support more boards in the near future?

Quote
mhackney
I don't understand how you fried endstops - where they optical endstops?

Quote
Troubleshooting noisy startup of the Overlord Pro
Hi guys, I have traced today the source of the terrible noise at startup/auto-calibration. This is about the same exact noise you see in all owner videos on youtube that show the auto-calibration routine (ex: [www.youtube.com] ... u.be&t=209).

While running the auto-calibration I noticed that the noise came more from one of the towers, didn't know which one was that. I did this over and over until I was certain that the noise is originating in one of the three towers. I turned the printer upside down, got the plastic off the bottom and watched/listened to the motors. The Z motor was the one driving the noisy tower.

Please don't do the following at home, I have burned two steppers in the process. I then disconnected the battery and the power source and moved by hand the Z motor up and down. Terribly noisy. The same noise you hear while it is operating.

I tried to figure if there was something different in the mechanical arrangement of that axis, pulleys, belts, tension. Everything was quite the same. Then I jogged by hand the other axes and that sound wasn't happening. The next thing I disconnected the motor from the board (took out the motor plug/connector - it is very nice to have the plugs/connectors on the motors themselves). I could have also taken out the plug on the board. Surprise surprise the sound was no more there, the carriage was sliding quietly. I then switched between them the connectors of X,Y and Z on the board to rule out stepper drivers and still the Z motor was making the noise.

I think that the problem is somewhere in this motor. There must be friction between the stator and the rotor while it gets powered (by the board) or generates (while manually jogging the carriage) its own power. The sound is worse in reality then you get from youtube videos.

Unfortunately I fried 2 steppers in the process, maybe the power generated while manually jogging the carriages up and down was too much for them to handle.

Interestingly enough, one of the stepper drivers was not the original DFRobot A4988 driver. Apparently the previous owner tried to do some troubleshooting too but he didn't get the right axis and he gave up selling the thing. The two original drivers (DFRobot A4988) are not working anymore. I powered again the board and there was nothing happening, no OLED display turning on, nothing. I almost shit my pants. But after 10minutes of personal cooldown I took it step by step and after taking all the stepper drivers off the think started. Oh what joy. Now I plugged one by one the stepper drivers and yeah, both X and Z were causing the board to remain powered off. I think this must be some feature of the board to prevent the other two axes to move the effector if one motor is not capable of moving its axis.

I find it interesting that the original stepper drivers gave in so quickly while the other stepper (one that looks just like this one: [www.dx.com]) didn't take the nose dive. This speaks maybe about some quality issues of the original drivers?

Ok, last interesting thing is that all these drivers have voltage control pots. I read that it is impossible to modify the current that goes into the motors because the stepper drivers of Overlords don't have pots. It might be that my DMO Pro Black Devil edition could be different from others? I don't know. But it was bad enough that I had to replace a motor, now I have to replace two stepper drivers too

I will make some more tests with another motor (that I already have) and some steppers that a friend will bring over. Hopefully I will not burn anything else and will prove if the motor was faulty after all. If that proves true, i think I will have to replace all motors so all are exactly the same.

Quote
realthor
How do recognize leaning towers?

Quote
dc42

Quote
realthor
How do recognize leaning towers?

Put a carpenter's square on the bed.

Quote
realthor
Morevover, the Rostocks have 1m+ long linear rods. Are these insignificant parameters that can be ignored?