OpenDACT(Delta Automatic Calibration Tool) - For Repetier

[RollieRowland]

This is the curvature that i was speaking about:

[img]http://boim.com/img/towerShiftDistortion.svg[/img]

from http://wp.boim.com/?p=67

Well.. Does this distortion only occur when the towers are off? Or does this also occur when they are calibrated properly?

According to that website, that specific example was because the z tower (I believe) was 10mm off. So, that is quite exaggerated.

I'm going to look into this a bit more, I do have matlab, so I will play around a bit. If I find anything out I'll let everyone know!

[RollieRowland]

Thanks for all the work on this Rollie, and a video to boot!

On a side note, I do need to update the video to show how the calibration can be done in one iteration. Or even just using the program in its current state. It has changed significantly since the video.

You're welcome by the way!

[enggmaug]

I have one question to address to the calibration experts. In order to use 2 or more nozzles, a machine shall not tilt the head when moving in the XY plane.

I believe that the only way to reach that, is to have a perfectly square machine. I don't think calibration can correct this kind of problem over a machine that has its towers not perfectly square.
So I don't believe this program can help me much regarding this aspect.
Am I correct ?

[teoman]

Yes you are.

[DGBK]

I'm pretty stumped here. I have gone through this program twice (v1.02 and v1.03), and I cannot get a good printing calibration. Both times I had all of the macro points within 0.01mm of each other. When I start a print, the head is all over the place. It is way too low in some places, and then too high to even contact the bed in others.

I noticed that after a few iterations my hot end zero position moves towards the y tower when I hit my bed center macro. The final position is maybe half a center from the real center of the bed. Could this be causing the trouble? Is there a way to reset center? I have only seen a way to reset z=0. I was really looking forward to this working. Any help would be appreciated.

[RollieRowland]

I have one question to address to the calibration experts. In order to use 2 or more nozzles, a machine shall not tilt the head when moving in the XY plane.

I believe that the only way to reach that, is to have a perfectly square machine. I don't think calibration can correct this kind of problem over a machine that has its towers not perfectly square.
So I don't believe this program can help me much regarding this aspect.
Am I correct ?

If the head is tilting as it moves around the build plate, i am sure you can correct this via delta radius. Or you can at least lower the effects. I can not currently tell you a method to correct delta radius... Although, KAS posted a trick to measure the offset by seeing where the nozzle hit the plate in center then at each tower. You would want to have distances between where the effector centers itself on the plate and the three points against the towers. For the points against the towers, you would want to make sure that the arms are close to being parallel to the extrusions. Then if you measure the distances between the center and all 3 outer points you may have some idea what to set the delta radius, to correct this offset.

Now this is not a tested method, but it is worth a shot.


Also, DGBK, the offset from the center is caused by changing the XYZ offset. This also occurs when you change your endstops.

But since this is calibrating while your dial gauge/calipers are mounted and not while you are printing, it may be due to where you mounted them. Where are they placed at when you take measurements? It should be as close to the centers of the effector, as the effector plateay shift around the bed, as stated above. I would recommend trying this first and if it doesn't work, we will go from there.

Tomorrow I will try to find a way to manually correct delta radius, and this should help this problem.

[DGBK]

Also, DGBK, the offset from the center is caused by changing the XYZ offset. This also occurs when you change your endstops.

But since this is calibrating while your dial gauge/calipers are mounted and not while you are printing, it may be due to where you mounted them. Where are they placed at when you take measurements? It should be as close to the centers of the effector, as the effector plateay shift around the bed, as stated above. I would recommend trying this first and if it doesn't work, we will go from there.

Tomorrow I will try to find a way to manually correct delta radius, and this should help this problem.

I think I figured out a major part of my problem with the odd movement. At some point I cut a little bit off of my bowden tube after I got it kinked. That missing 2-ish inches combined with my E3D being mounted below the effector was pulling quite a bit on the effector plate. I'm pretty sure my company uses teflon tubing for groundwater sampling, so I can probably get some more tubing pretty easily. I will have to grab some of that before giving this another go. I'm just glad it wasn't your program.

Once again, this was user error on my part.

[RollieRowland]

Also, DGBK, the offset from the center is caused by changing the XYZ offset. This also occurs when you change your endstops.

But since this is calibrating while your dial gauge/calipers are mounted and not while you are printing, it may be due to where you mounted them. Where are they placed at when you take measurements? It should be as close to the centers of the effector, as the effector plateay shift around the bed, as stated above. I would recommend trying this first and if it doesn't work, we will go from there.

Tomorrow I will try to find a way to manually correct delta radius, and this should help this problem.

I think I figured out a major part of my problem with the odd movement. At some point I cut a little bit off of my bowden tube after I got it kinked. That missing 2-ish inches combined with my E3D being mounted below the effector was pulling quite a bit on the effector plate. I'm pretty sure my company uses teflon tubing for groundwater sampling, so I can probably get some more tubing pretty easily. I will have to grab some of that before giving this another go. I'm just glad it wasn't your program.

Once again, this was user error on my part.

Ahh, yes, I have had that happen as well... Glad you figured it out!

[techstorage]

After figuring out I should use the Enter Defaults button to get the program to work...

I used your program with just a dial indicator and adjusted everything within .01mm. Then I installed my hot end with the wires and tube, the opposite tower ends were still out but a lot closer then anything I have been able to get in the past. I believe adding the cables and tube showed me my extruder slop even with bands on the arms.

I did more passes with a paper gauge to check hot end gap and was able to get it very level. Thanks for your program, it has allowed me to use my full printing surface because of getting accurate calibration numbers.

[geneb]

I haven't looked at the program yet, but if you're interested in turning it into a desktop application, let me know.

g.

[RollieRowland]

After figuring out I should use the Enter Defaults button to get the program to work...

I used your program with just a dial indicator and adjusted everything within .01mm. Then I installed my hot end with the wires and tube, the opposite tower ends were still out but a lot closer then anything I have been able to get in the past. I believe adding the cables and tube showed me my extruder slop even with bands on the arms.

I did more passes with a paper gauge to check hot end gap and was able to get it very level. Thanks for your program, it has allowed me to use my full printing surface because of getting accurate calibration numbers.

Glad it worked for you!

Also, geneb, that would be fantastic! My goal is to turn this into an actual auto calibration program for deltas with z-probes. I know this would definitely not be an easy task but even just switching to a desktop application is a step in the right direction.

[geneb]

The problem with integrating a Z probe is getting data back from the printer that a program can read. It would be something like:

<rapid to 5mm above the bed>
<move a tiny amount>
Are we there yet?
NO!
<move a tiny amount>
Are we there yet?
NO!
.
.
.
<move a tiny amount>
Are we there yet?
YES! Now quit asking me dammit! BTW, the z height is <value>. Now go away!
<record value and go to next position>

Send me a PM about it tomorrow and I'll grab the code from the repo and take a look at what it'll take to "appify" it.

g.

[U.S. Water Rockets]

I have hacked a pair of those cheap Chinese calipers in the past to output data serially. I don't know if the circuit inside is still the same as it was 5 years ago, but I added a small connector and was able to read data out with a UART on a microcontroller.

It's totally within the realm of possibility to connect a USB to Serial cable to the calipers and read them directly with a PC.

Make a modified extruder platform with a bracket for the calipers, and a spring or rubber band to provide the tension on the calipers, and you could do all the readings directly on the desktop PC.

If there is a spare UART on the RAMBO board, you could read directly with the printer electronics. You could also make a standalone system with an Arduino or MSP430 LaunchPad to read the calipers and send commands to the RAMBO.

Would any of this be of interest to you?

[RollieRowland]

The problem with integrating a Z probe is getting data back from the printer that a program can read. It would be something like:

<rapid to 5mm above the bed>
<move a tiny amount>
Are we there yet?
NO!
<move a tiny amount>
Are we there yet?
NO!
.
.
.
<move a tiny amount>
Are we there yet?
YES! Now quit asking me dammit! BTW, the z height is <value>. Now go away!
<record value and go to next position>

Send me a PM about it tomorrow and I'll grab the code from the repo and take a look at what it'll take to "appify" it.

g.

My main issue is I know absolutely nothing about connecting to COM ports/arduinos. Which would be the most convenient way for auto calibration, as it would connect and could pull required info from the EEPROM, while also able to send gcode. I mean every host program for printers can do it! (repetier/matter control etc.) Why couldn't we?

[RollieRowland]

I have hacked a pair of those cheap Chinese calipers in the past to output data serially. I don't know if the circuit inside is still the same as it was 5 years ago, but I added a small connector and was able to read data out with a UART on a microcontroller.

It's totally within the realm of possibility to connect a USB to Serial cable to the calipers and read them directly with a PC.

Make a modified extruder platform with a bracket for the calipers, and a spring or rubber band to provide the tension on the calipers, and you could do all the readings directly on the desktop PC.

If there is a spare UART on the RAMBO board, you could read directly with the printer electronics. You could also make a standalone system with an Arduino or MSP430 LaunchPad to read the calipers and send commands to the RAMBO.

Would any of this be of interest to you?

This is also a possibilty! The only issue with this would be speed, as you would have to dissassemble your hot end, then mount a bracket for the calipers. This would still be more efficient than the current method, however, I figured it can still be possible through even mounting an endstop switch to the hot end. Similar to the retractable z probes for some Kossel printers. Then the software would check if the switch clicked before or after it was supposed to and by how much. This could be used as the displacement values to insert into the program.

[geneb]

It could be done interactively while using a feeler gauge. Very similar to the method used by MatterControl in the bed level wizard. No extra hardware would be required.

g.

[RollieRowland]

It could be done interactively while using a feeler gauge. Very similar to the method used by MatterControl in the bed level wizard. No extra hardware would be required.

g.

This would be the easiest way, probably the best route for now. I have not yet developed a gui for a desktop app but I'll look into it tomorrow or later tonight.

[geneb]

That's what I do for a day job.

g.

[RollieRowland]

That's what I do for a day job.

g.

Fantastic! What languages do you know? Or would prefer to use?

[Jimustanguitar]

Silly question... Can Javascript reach the COM ports and control USB devices? I know it has limitations for security reasons, but I don't know what they are.

How cool would it be to have a printer calibration website that could send the GCode to your machine without a separate host program? All in the same window, no switching back and forth... You could even make it a step by step tutorial kind of a thing instead of a single page with all of the values.

Here's the million dollar idea part, You could add support for other models over time and be the one stop shop for printer calibration. That's one of the biggest gripes that we all have, you could corner that market. Imagine every printer in the world being able to level and calibrate simply by visiting a webpage and following the guided prompts?

Stepping off soapbox... Happy Friday.

[geneb]

Jimustanguitar, I don't know of any scripting language that runs in the browser that would permit (or even provide!) access to the serial ports on the host machine - you'd have to write something in Java and that would be wrong.

I'm thinking C#. I haven't gotten it installed yet, but Visual Studio 2015 seems to support .Net applications on Android, iOS and MacOS(?). I have no way to test on non-Windows platforms though.

I should have it installed later on today.

g.

[KAS]

Chrome has API's for both serial and USB access from apps. I'm not sure if that's the Chrome browser or OS, maybe both?

https://developer.chrome.com/apps/usb
https://developer.chrome.com/apps/serial

[RollieRowland]

Jimustanguitar, I don't know of any scripting language that runs in the browser that would permit (or even provide!) access to the serial ports on the host machine - you'd have to write something in Java and that would be wrong.

I'm thinking C#. I haven't gotten it installed yet, but Visual Studio 2015 seems to support .Net applications on Android, iOS and MacOS(?). I have no way to test on non-Windows platforms though.

I should have it installed later on today.

g.

Yeah, there is no way to access com ports or nearly anything on the pc with javascript. Ok, I'll get Visual Studio as well. I'll start converting the calculator portion.

Also, KAS, it would be easier to add support for auto calibration in a desktop application later. And since that is an API, it would most likely require a php or node.js environment to operate properly.

[Jrjones]

C# is a good way to make a simple GUI. I've used it before for some simple test programs. No idea if it works on non Windows machines.

[geneb]

All the .Net stuff is being open sourced by Microsoft. Via Mono you can build .Net apps for Linux and MacOS. That's how MatterControl is done - it's all C#.

I'm having problems getting VS2k15 to behave. It won't ask me for a license key, so I'm having to re-install. (Ent. Ed. vi msdn...)

g.