OpenDACT(Delta Automatic Calibration Tool) - For Repetier

[Micael]

No idea where you live but look it up on farnell, rs-components, mouser or Digi-key electronics one of these should be in your country and have what you want.

[RollieRowland]

Version 3.0.1PA - Calibration Fixes

Changes:
-Tower offset calibration now uses horizontal radius to complete step as opposed to delta radii, also is predicted ahead of measurement - reduces number of iterations.
-Changed main control flow of calibration process
-Time wait for printer moves is now calculated based on the set travel speed/distance
-Support for other system languages(not tested, hands on the killswitch)

Not yet finished:
-Iterative mode(might not be reimplemented)
-Heuristic mode(will decrease number of iterations)
-UI

Download - https://drive.google.com/file/d/0B06WCA ... sp=sharing" onclick="window.open(this.href);return false;

Yeah I think that was my problem, the switch I was using before had too much variance in it.
I switched to an inductive probe that has a repeatability of +/- 0.035 mm or less (SICK says its <= 2%), so I set the calculation accuracy to +/- 0.04 mm.

After 62 iterations it completed!

Any issues that cause error in the repeatability of measurements will cause the program to fail at calibration - for obvious reasons - I would measure the heightmap several times to check the repeatability.

Also, make sure your plate diameter is set so the joints of the diagonal rod attached at the effector-plate do not protrude further joints at the carriage(from the center). This will lead to very inaccurate results when measuring heights, i.e. way more iterations.

[Jrjones]

Any issues that cause error in the repeatability of measurements will cause the program to fail at calibration - for obvious reasons - I would measure the heightmap several times to check the repeatability.

Also, make sure your plate diameter is set so the joints of the diagonal rod attached at the effector-plate do not protrude further joints at the carriage(from the center). This will lead to very inaccurate results when measuring heights, i.e. way more iterations.

Hey Rollie, is there any way to adjust the calibration so the steps/mm does not change? I was trying out a print and directly across the Z tower in the -Y coordinate range the extruder path leaves a gap inbetween the lines. I suspect that it is due to the steps/mm being changed from 80 to 81.xxxx.
Pics:

Towards the Y+ the extrusions seem to be closer together. I didn't have this issue before the calibration, so that is what makes me suspect the steps/mm thing.
I can see it as well in the other directions but its not as visible.

[marcoleonardi]

dear Rollie, thanks for your work.
With the version 3.0.0 I've succesfull calibrate my printer to 0.03, using zprobe (set for calibrate with fsr deducted my probe height from global height during calibration, then restored original height for printing)
With the new version (same procedure & parameters) there is no way to calibrate the printer
Do you have the source code of the previous version? I guess the one on github is the PA version, correct?
Also I don't like the step/mm change; in many test everything converge to request result but when the software start to touch step/mm the result are worst.

Thx, Marco.

[RollieRowland]

Hey Rollie, is there any way to adjust the calibration so the steps/mm does not change? I was trying out a print and directly across the Z tower in the -Y coordinate range the extruder path leaves a gap inbetween the lines. I suspect that it is due to the steps/mm being changed from 80 to 81.xxxx.

Towards the Y+ the extrusions seem to be closer together. I didn't have this issue before the calibration, so that is what makes me suspect the steps/mm thing.
I can see it as well in the other directions but its not as visible.

Not at the moment, previous versions accomplished the same height-map via changing the diagonal rod as opposed to the steps per mm but this brought about horrible XY scaling issues. I will look into other issues that may be causing this error, or workarounds to solve it without changing either the diagonal rod or the steps per mm. I know the opposites could be raised/lowered via the alpha rotation, but I feel it may have worse results than changing the steps per mm. If I find anything out then I will make a change to the program and upload a version for testing.

dear Rollie, thanks for your work.
With the version 3.0.0 I've succesfull calibrate my printer to 0.03, using zprobe (set for calibrate with fsr deducted my probe height from global height during calibration, then restored original height for printing)
With the new version (same procedure & parameters) there is no way to calibrate the printer
Do you have the source code of the previous version? I guess the one on github is the PA version, correct?
Also I don't like the step/mm change; in many test everything converge to request result but when the software start to touch step/mm the result are worst.

Thx, Marco.

The source code is on GitHub - including all previous versions of it, but also all of the previous download links will still work properly. The steps per mm calibration still needs some work, but if it is removed then your bed will have high/low spots at the towers and low/high spots inbetween the towers. The only other way to correct this is by altering the diagonal rod length, but I will look into alternatives - ones that are only able to be changed in the firmware.

[marcoleonardi]

thx, I will look @ github.
The adjustment of diagonal rod lenght shall be good, it's the same way that I use in manual calibration.
Tomorrow I will try to understand why not work with version PA

[Jrjones]

Hey Rollie, nevermind my comments about the steps/mm. My problem was based in my slicing settings.
I don't seem to have any scaling issues, except maybe in the Z but that can be fixed easily in the slicer's scale I believe.

[Pyropainter]

Please excuse me if I am asking such an obvious question but, I am trying to get OpenDACT running on my printer. I have searched and have not been able to fine an answer to my problem...My question is as follows... I am supposed to connect the micro switch to the Z axis minimum receptacle on the Rambo board. The receptacle on the board has three pins. Labeled as follows, Posative, Negative and S(Signal? unsure). Which two of the three pins am I attach the micro switch to? I have tried the positive and negative pins and my lcd display shuts off when the switch closes.

Thank You For Your Help
Barry

[geneb]

You want positive and signal I think. Shorting pos & neg together is a Bad Thing(tm).

g.

[mhackney]

NOPE - simple mechanical switches connect to (-) and (S) on RAMBo. You'll short if you connect + and sig.

[RollieRowland]

thx, I will look @ github.
The adjustment of diagonal rod lenght shall be good, it's the same way that I use in manual calibration.
Tomorrow I will try to understand why not work with version PA

Do take note that your scaling will differ significantly in the X and Y dimensions (and some other issues could pop up as well).

Hey Rollie, nevermind my comments about the steps/mm. My problem was based in my slicing settings.
I don't seem to have any scaling issues, except maybe in the Z but that can be fixed easily in the slicer's scale I believe.

Ah, good to know. And I do believe Z scaling is still an issue - although it is easy to compensate for, I would like to fix it anyway.

You want positive and signal I think. Shorting pos & neg together is a Bad Thing(tm).

g.

Just checked and it should be on the negative and signal pin. But yeah definitely bad to use the positive and negative/sig! Nobody wants to have a dead rambo!


Edit: And yes Mhackney is right, you'll short if you hook the positve and signal pin to the endstop!

[Pyropainter]

I don't think I blew anything up yet.

Thank you all for your help.

[Pyropainter]

I finally got to try auto calibrating today. But I am experiencing problems. When running the setup application pressing the Basic Calibration causes the printer to do what I would consider "taking a reading" in several locations while ignoring others. This is what is happening. The probe rapids to each successive location (Center, X, X Opposite, Y) between each location the probe feeds down then back up. The other locations the probe rapids to each location and moves on without probing. Something does not seem right with this. Please inform me of what I should do to remedy the situation.

Also after I get what would be called a successful run thru of the "Setup application" do I then run the OpenDACT application. Please tell me what I am doing by running thru both applications and what to expect so that I am better able to understand what is going on.

Thank you.

[Pyropainter]

If you have read the previous post you are aware of my problem. It pertains to OpenDACT300A, I have since uninstalled it and installed OpenDact3.0.1PA and have had a successful calibrating session. I have no Idea what what was happening with the previous version, I do thank anyone who even gave it a thought.


Thank you so much to anyone who has had any thing to do with this wonderful piece of software.

[RollieRowland]

I finally got to try auto calibrating today. But I am experiencing problems. When running the setup application pressing the Basic Calibration causes the printer to do what I would consider "taking a reading" in several locations while ignoring others. This is what is happening. The probe rapids to each successive location (Center, X, X Opposite, Y) between each location the probe feeds down then back up. The other locations the probe rapids to each location and moves on without probing. Something does not seem right with this. Please inform me of what I should do to remedy the situation.

Also after I get what would be called a successful run thru of the "Setup application" do I then run the OpenDACT application. Please tell me what I am doing by running thru both applications and what to expect so that I am better able to understand what is going on.

Thank you.

This issue is impart from a delay requirement on the program side. The first height measurement is checked assuming that the height is somewhere around the probing height(base is 10mm). So if the z height is initially off, then the program may send commands to quickly, causing this skipping to occur. (This could also happen in 3.0.1)

Also, the setup is really only there to install dependencies - so it only needs to be ran the first time. You may try to run the application without the setup.

[dkaustin]

This issue is impart from a delay requirement on the program side. The first height measurement is checked assuming that the height is somewhere around the probing height(base is 10mm). So if the z height is initially off, then the program may send commands to quickly, causing this skipping to occur. (This could also happen in 3.0.1)

This would be consistent with the problems I am seeing in 3.0.1. It would also explain the various misbehaviors I saw when I tried running the program at different baud rates

[ewm]

This issue is impart from a delay requirement on the program side. The first height measurement is checked assuming that the height is somewhere around the probing height(base is 10mm). So if the z height is initially off, then the program may send commands to quickly, causing this skipping to occur. (This could also happen in 3.0.1)

Also, the setup is really only there to install dependencies - so it only needs to be ran the first time. You may try to run the application without the setup.

I'm also experiencing this same issue with Casper but it's also experiencing other very mysterious and extremely frustrating issues. We are using 3.0.1PA and I've reset everything back to square one made sure our z height is correct for our probe. Tonight upon selecting the Basic Calibration it ran a few iterations and then decided to not tap the bed. It instead gave it an insanely off number of -8.1. At this point I tried to select reset as I know it's not good but it refused to reset. I then turned off the printer and attempted to close out of the program. Upon turning Casper back on he acts possessed and tries to continue the calibration in mid air.

Upon resetting everything again back to square one Casper failed to tap down on the initial first iteration of Basic Calibration. It seems these failures to tap down are sporadic, not consistent at all. The haunting issue after powering down is also extremely strange, it never happened previously.

[Pyropainter]

At first I was able to get OpenDact to operate as expected a couple of times. But recently it has not been operating as expected. I would like to eliminate the variable of operator error so, I will mention my settings at the time of the last failure. Using the Check Current Heights function, my z probe set at zero and the Probing Height (Advanced Settings) set to zero, I observed the probe rapid down to a clearance plane, actuate micro switch, re-home, rapid back down to Z 0.0 (I assume) then slide the probe from Center to X Tower then on to X Tower opposite before getting the printer shut off.

I have also tried the Check Current Heights function with the following settings with failure. Nozzle set to Z zero and Probing Height set to 22.5 (That is the distance from the tip of the nozzle to the point at which the micro switch activates). I do not remember what happened, but it function improperly.

I do appreciate your time in helping me out with this
Looking forward to your help or criticism.
Thank you.

[CaptainMerica]

I believe the probe height parameter is the "start height" for the slower Z move to probe, by having it set to 0, you told the printer to go to Z0.0 and then move Z- until the probe actuates. instead, use like 10 or so for the Z-start. The probe height just tells it what to rapid to and then slow down, so that it has a safe "retract" plane so to speak.

[Mihelich]

Hey guys, is there any way to use this software now manually instead of with the automation like the first few versions?

[Pyropainter]

Good Evening

MIhelich
Although I have not used the earlier versions of OpenDACT I do not believe the current one is able to be used manually. From what I can tell we are not capable of inputting the necessary information by hand.

CaptainMerica

I believe that the probe height is where the software expects to see the probe activate in relation to Z 0.0. Some were at the beginning of this thread the author mentions that if our probe height is above 70mm he would have to change the g-code part of the software. I could be wrong, but from what I observe, when initiating a check current height session or calibrate session my printer rapids to about 70mm the moves at a slower predetermined rate until the probe is activated....

Tonight I decided to play with the software again, I will try to explain my actions as well as my printers respones.

I decided to run the "Check Current Heights Option". This is what I did and what I observed.

1. I Homed Printer
2. Then zeroed my printer to at or near the the activation point of my probes micro switch (I moved the Z axis down until I heard an audible click then zeroed the Z axis of my printer).
3. Re homed Printer
4. Then I made sure the "Z-probe" option was activated in the software.
(not that this is the only change that was made prior to the next action)
3.Activated the Check Current Heights tab.
This is what I saw
1.machine homed itself
2.machine rapids to about 70mm above Z 0.0
3.machine re homes itself
4.machine rapids to about 10mm above Z 0.0 then pauses
5.rapids to x tower then pauses
6.rapids to x tower the pauses

I terminated any further actions because my printer is now moving to each designated probe point and pauses at 10mm above z0.0 without moving towards the heated bed (z 0.0) to confirm its actual position.


I am not sure what I am doing wrong. Please help.
I have once heard of a video explaining how to use this software, does it exist, please tell me where to find it.
If something like this does not exist would someone who has had success be willing to put together a step by step, it would not have to be a video.
The most frustrating thing about this is I was able to make this work the first time that I used it. I do not know what I have done.

Thank you to any one that is willing to spend time on this....

[RollieRowland]

I believe the probe height parameter is the "start height" for the slower Z move to probe, by having it set to 0, you told the printer to go to Z0.0 and then move Z- until the probe actuates. instead, use like 10 or so for the Z-start. The probe height just tells it what to rapid to and then slow down, so that it has a safe "retract" plane so to speak.

Yes, this is correct.

Hey guys, is there any way to use this software now manually instead of with the automation like the first few versions?

Here I added the manual calibration into this version of the program. I am still working on the timing issues, so for now only manual will work here.

Download: https://drive.google.com/file/d/0B06WCA ... sp=sharing" onclick="window.open(this.href);return false;

Tonight I decided to play with the software again, I will try to explain my actions as well as my printers respones.

I decided to run the "Check Current Heights Option". This is what I did and what I observed.

1. I Homed Printer
2. Then zeroed my printer to at or near the the activation point of my probes micro switch (I moved the Z axis down until I heard an audible click then zeroed the Z axis of my printer).
3. Re homed Printer
4. Then I made sure the "Z-probe" option was activated in the software.
(not that this is the only change that was made prior to the next action)
3.Activated the Check Current Heights tab.
This is what I saw
1.machine homed itself
2.machine rapids to about 70mm above Z 0.0
3.machine re homes itself
4.machine rapids to about 10mm above Z 0.0 then pauses
5.rapids to x tower then pauses
6.rapids to x tower the pauses

I terminated any further actions because my printer is now moving to each designated probe point and pauses at 10mm above z0.0 without moving towards the heated bed (z 0.0) to confirm its actual position.


I am not sure what I am doing wrong. Please help.
I have once heard of a video explaining how to use this software, does it exist, please tell me where to find it.
If something like this does not exist would someone who has had success be willing to put together a step by step, it would not have to be a video.
The most frustrating thing about this is I was able to make this work the first time that I used it. I do not know what I have done.

Thank you to any one that is willing to spend time on this....

I have not heard of or seen your issue before. However, the printer may be disabling the probe because it is moving to a lower position than the distance it needs to correctly probe. For instance if your z probe height(distance between effector tip and zprobe switch) is 30mm, then you should increase your z probing height by 30 (putting it at 40mm). Another test you could do is moving your effector down and send G30(you should check your probe with G31 prior to doing so).

Let me know what you find out, that way we can rule out possible causes.

[Pyropainter]

Thank you for the help. I was able to complete a session with OpenDACT (11 iterations). My previous mistake was with the advanced screen and the "Probing Height" I took it to mean the height between the nozzle and the probe actuation point. Your reply is greatly appreciated. Upon completing the session I attached my dial indicator and tested the bed. I found some discrepancies. I will first mention that OpenDACT calibrated my bed assuming a 200mm bed diameter. My macros in Matter Control are set at a 260mm bed diameter. Here is what I found. With the center of the bed set at Zero, each tower and its opposite margin of error ranged from .19mm to .28mm. It might be worth mentioning, that the zero point of the indicator is about 150mm from the nozzle. Could it be that the end effector is rotating enough to show the apparent error.....

Also, what happens if I run OpenDACT using FSR even though I am using a micro switch that is set to Z zero. I may try it, but I have seen too many crashes and thought I would ask first.

Thanks Again

[mirage1207]

Hello Steven,

First, I would like to thank you for your awesome program for calibrating the Rostock V2. So quickly the printbed to get straight is amazing !! Thank you thank you thank you!!!!

Unfortunately, I defy a small but serious problem. my printed parts are too small. if I want to print a cube with an edge length of 50mm, it is 48,99mm x 49.00mm. how can I now "set the correct size" without giving me the Delta radius to destroy again? even if I change the "Delta arm length" my bed is tilted again.

I've run for 3 years one of the last Rostock V1 without problems on and had never secure such problems. I try since ca 2 weeks every day to V2 straight and get dimensionally stable and am slowly despair ....

all Tower are in angles and straight, the arms are made of carbon by Tricklaser, and everything is assembled with great care. I know really slow no longer what to do yet .... please please help me or give me a tip where the error or the problem could lie! Please please

sorry for my poor english!!
Thank you Peter

[RollieRowland]

Thank you for the help. I was able to complete a session with OpenDACT (11 iterations). My previous mistake was with the advanced screen and the "Probing Height" I took it to mean the height between the nozzle and the probe actuation point. Your reply is greatly appreciated. Upon completing the session I attached my dial indicator and tested the bed. I found some discrepancies. I will first mention that OpenDACT calibrated my bed assuming a 200mm bed diameter. My macros in Matter Control are set at a 260mm bed diameter. Here is what I found. With the center of the bed set at Zero, each tower and its opposite margin of error ranged from .19mm to .28mm. It might be worth mentioning, that the zero point of the indicator is about 150mm from the nozzle. Could it be that the end effector is rotating enough to show the apparent error.....

Also, what happens if I run OpenDACT using FSR even though I am using a micro switch that is set to Z zero. I may try it, but I have seen too many crashes and thought I would ask first.

Thanks Again

This error most likely was impart from the change in distance from your new end effector (your dial indicator) and your previous end effector. If the distance from the tip of the hotend to the effector plate and the distance from the tip of the dial indicator to the effector plate is equal then you will most likely not see error. This error is most likely just a minor offset in your machine - nothing significant, given that offset between the dial indicator/hotend is around 100mm.

If you do decide to run with FSR set then I would start with increasing your z probing height from 10 to 10 plus what your z probe offset is (hotend/zprobe height difference). It should run perfectly fine but do take note that you will need to set the z max length after calibration.

Hello Steven,

First, I would like to thank you for your awesome program for calibrating the Rostock V2. So quickly the printbed to get straight is amazing !! Thank you thank you thank you!!!!

Unfortunately, I defy a small but serious problem. my printed parts are too small. if I want to print a cube with an edge length of 50mm, it is 48,99mm x 49.00mm. how can I now "set the correct size" without giving me the Delta radius to destroy again? even if I change the "Delta arm length" my bed is tilted again.

I've run for 3 years one of the last Rostock V1 without problems on and had never secure such problems. I try since ca 2 weeks every day to V2 straight and get dimensionally stable and am slowly despair ....

all Tower are in angles and straight, the arms are made of carbon by Tricklaser, and everything is assembled with great care. I know really slow no longer what to do yet .... please please help me or give me a tip where the error or the problem could lie! Please please

sorry for my poor english!!
Thank you Peter

The scaling issues depend on the version of the program you are using, the newest versions shouldn't have any scaling issues if the calibration runs properly (it is not if the horizontal radius or steps per mm are significantly different). I do have another update in progress - might be up tomorrow - which has a long list of fixes and additions. But one way to temporarily fix your issue is adjust the scale of your prints i.e. 102% scaling in XY dimensions. The newer versions do calibrate the steps per millimeter after calibrating all other parameters, which has helped significantly with scaling - the horizontal radius is more accurate. I used to get values of 81.25 - 82 for my steps per mm, now I get around 80.3 or less.