OpenDACT(Delta Automatic Calibration Tool) - For Repetier

[Jimustanguitar]

Sorry if this is obvious to everyone else, but it's just not clicking for me. What are the updated directions for using the webpage tool in expert mode? Using the basic calibration, I can dance around .25mm variation, but can't get any closer. When I got started using the expert fields like adding 1 to my diagonal rod and taking those measurements, clicking on calibrate didn't seem to change any of the fields. What is the correct procedure for using the expert fields? Are the different percentages supposed to auto-populate from the measurements that I enter, or am I supposed to calculate those numbers?

Also, I'm supposed to be using the second set of GCode commands that you posted, right? Perhaps adding a text document to the GitHub repository with the commands that match the current iteration of the html page would be best.

Thanks for your work on this awesome project, you're a lifesaver!

Jim

[geneb]

Well personally I don't need a fully-automatic calibration system - primarily because it's a "do it once" task until you change some physical aspect of the machine that would require doing it again. Unlike our unfortunate cartesian brethren, we DON'T have to recalibrate at the start of every print day.

g.

[RollieRowland]

Sorry if this is obvious to everyone else, but it's just not clicking for me. What are the updated directions for using the webpage tool in expert mode? Using the basic calibration, I can dance around .25mm variation, but can't get any closer. When I got started using the expert fields like adding 1 to my diagonal rod and taking those measurements, clicking on calibrate didn't seem to change any of the fields. What is the correct procedure for using the expert fields? Are the different percentages supposed to auto-populate from the measurements that I enter, or am I supposed to calculate those numbers?

Also, I'm supposed to be using the second set of GCode commands that you posted, right? Perhaps adding a text document to the GitHub repository with the commands that match the current iteration of the html page would be best.

Thanks for your work on this awesome project, you're a lifesaver!

Jim

Let me start with saying that yes, I do need to update my documentation.
For using the expert settings you need to start by increasing that value by 1 unit- 80 for XYZ offset- then take the measurement, and add the heightmap to the corresponding boxes. For higher accuracy, you need to repeat this for diagonal rod, horizontal radius, and XYZ offset. Then once all of those heightmaps are filled in you can press calibrate. Once you press calibrate, it will autofill the percentages, so no calculation on your part.

I am adding an advanced calibration to the automatic version which will do this for you, and also allow you to manually change settings.

[RollieRowland]

Version 2.0.1 - Advanced Calibration

This update has a lot of new features, and has not been tested...

I finished what I wanted for this update, but since I am away for a week, I am not able to test any of the features that I have implemented. These include:
-support for FSR and Z-Probe
-control over z-probe speed
-less lag/thread sleeping, now depends on how long your printer takes to communicate
-live height-map
-accuracy control
-advanced calibration to learn specifics about Your printer - similar to expert mode in V1.0.3
-maximum iterations
-control over time pausing for printer communication (sends script and waits designated time)
-FSR plate offset control (raises the print height slightly so you have a first layer)
-determine z-probe height, this requires you to set your z-height first
-and most likely other features that I have already forgotten...

I will say it again, I haven't had a chance to test this yet. However, I have gone over the new code 3 times, so it should work - i.e. finger on the kill-switch.

Also, I still have only tested this on Windows 8.1 and 10.

https://drive.google.com/file/d/0B06WCA ... sp=sharing

[duvdev]

Cool you are doing a great work for us all .
I am installing my z probe now. Do I need to connect it to Z min?

If the z probe I just under the hot end but lower from the hot end let's say by 50 mm how do I take that as an input? Or is there something I am missing ?

[RollieRowland]

Cool you are doing a great work for us all .
I am installing my z probe now. Do I need to connect it to Z min?

If the z probe I just under the hot end but lower from the hot end let's say by 50 mm how do I take that as an input? Or is there something I am missing ?

You can use any minimum endstop, you will have to add it in your firmware though. I have made a couple of posts on this if you wonder what you need to change. Yes, the difference between the z-probe and hotend is the z-probe height in the EEProm settings.

[Jimustanguitar]

I did the expert calibration with the webpage tool's expert settings, and after 4 time through the cycle, I'm between -.01 and .01 at all 6 point for the first time (I think) ever! Awesome, tool, thank you!

[HumanLiberty]

I did the expert calibration with the webpage tool's expert settings, and after 4 time through the cycle, I'm between -.01 and .01 at all 6 point for the first time (I think) ever! Awesome, tool, thank you!

Congrats on those outstanding results!
What kind of Probe did you use, and how did you attach it?

I just ordered the below Depth Probe, and have been wondering how I'm gonna mount it, when it arrives:
http://www.amazon.com/dp/B00OACN5XI/ref ... epth+gauge

Planning to take out the whole hot end and try to somehow mount this in it's place.

I'm sure a printable mount for the effector platform would be pretty simple for a person with CAD skills to make - unfortunately
I'm not one of those. If anyone out there who is likes this approach and feels like helping, hopefully the community would benefit...

[RollieRowland]

I did the expert calibration with the webpage tool's expert settings, and after 4 time through the cycle, I'm between -.01 and .01 at all 6 point for the first time (I think) ever! Awesome, tool, thank you!

It's fantastic to hear that it worked for you!

My next step for the program now will be to add support for more than repetier, just need to read up on Marlin, Smoothieware, etc. Once I get back in front of my printer to test the current auto-cal, make sure it works first, I will start to work on support for other firmware and operating systems.

[Jimustanguitar]

What kind of Probe did you use, and how did you attach it?

I used the HTML version with a dial indicator. It's a Pittsburgh Tools digital one from Harbor Freight (same thing, different sticker http://www.harborfreight.com/1-inch-sae ... 93295.html ) mounted in the Trick Laser pen holder ( http://tricklaser.com/Pen-Tool-holder-f ... RM-PTH.htm )

My next step for the program now will be to add support for more than repetier, just need to read up on Marlin, Smoothieware, etc. Once I get back in front of my printer to test the current auto-cal, make sure it works first, I will start to work on support for other firmware and operating systems.

YAY!

[cyber.shifter]

I did the expert calibration with the webpage tool's expert settings, and after 4 time through the cycle, I'm between -.01 and .01 at all 6 point for the first time (I think) ever! Awesome, tool, thank you!

Congrats on those outstanding results!
What kind of Probe did you use, and how did you attach it?

I just ordered the below Depth Probe, and have been wondering how I'm gonna mount it, when it arrives:
http://www.amazon.com/dp/B00OACN5XI/ref ... epth+gauge

Planning to take out the whole hot end and try to somehow mount this in it's place.

I'm sure a printable mount for the effector platform would be pretty simple for a person with CAD skills to make - unfortunately
I'm not one of those. If anyone out there who is likes this approach and feels like helping, hopefully the community would benefit...

Repables has this: http://repables.com/r/344/

I was looking at it because I was trying to figure out the same thing. My question is, if I take off the hot end to do the calibration, won't I have to make adjustments to something later unless the probe end is in the same spot that the hotend will be? I mean, the end of the probe will be lower than the nozzle would be, but I'm not sure how this will affect the process, or if it matters. Maybe it's just as simple as resetting the Z-height once the hot end is back installed?

[bvandiepenbos]

I did the expert calibration with the webpage tool's expert settings, and after 4 time through the cycle, I'm between -.01 and .01 at all 6 point for the first time (I think) ever! Awesome, tool, thank you!

woohoo!

is that .01 MM ?

[Jimustanguitar]

Yes, millimeters.

[cyber.shifter]

What kind of Probe did you use, and how did you attach it?

I used the HTML version with a dial indicator. It's a Pittsburgh Tools digital one from Harbor Freight (same thing, different sticker http://www.harborfreight.com/1-inch-sae ... 93295.html ) mounted in the Trick Laser pen holder ( http://tricklaser.com/Pen-Tool-holder-f ... RM-PTH.htm )

My next step for the program now will be to add support for more than repetier, just need to read up on Marlin, Smoothieware, etc. Once I get back in front of my printer to test the current auto-cal, make sure it works first, I will start to work on support for other firmware and operating systems.

YAY!

How did you account for the difference in heighte between where the probe mounts and where the extruder mounts? Sorry if that's a seemingly simple question, but I'm still on the steep part of the learning curve.

[HumanLiberty]

What kind of Probe did you use, and how did you attach it?

I used the HTML version with a dial indicator. It's a Pittsburgh Tools digital one from Harbor Freight (same thing, different sticker http://www.harborfreight.com/1-inch-sae ... 93295.html ) mounted in the Trick Laser pen holder ( http://tricklaser.com/Pen-Tool-holder-f ... RM-PTH.htm )

My next step for the program now will be to add support for more than repetier, just need to read up on Marlin, Smoothieware, etc. Once I get back in front of my printer to test the current auto-cal, make sure it works first, I will start to work on support for other firmware and operating systems.

YAY!

How did you account for the difference in height between where the probe mounts and where the extruder mounts? Sorry if that's a seemingly simple question, but I'm still on the steep part of the learning curve.

First of Thanks for commenting+Info Mustang and Cyber.

Cyber: I'm going on the expectation that the small difference in height between the the probe and where the hotend will be can be will be worked out by simply re-adjusting z-height once the nozzle is back in. However, I'm a noob and totally winging this - it might not work at all for all I know.

That gives me an idea though - If I knew how many mm the nozzle extends below the effector plate on a standard build, I could calibrate that as my "zero" on this gauge, which could hopefully nip any problems in the bud... Does anyone know that measurement?
If not, I'll call JJ in SMCNC tech support tomorrow.

If this still doesn't work for me, my plan B is to buy the FSR kit and a borrow a windows 10 machine to run the auto-calibration.
Here are pics of how I did the mount - got lucky and my son's rc truck was delivered at the same time with perfect twist ties or I don't think
I would have gotten it on... I'll let you know how it works out...

BTW - It wasn't mentioned on the Amazon spec sheet, but turns out this probe has some sort of USB output (non-standard shape) -
I'm going to find out if/how it can be used to directly load data into EEPROM for Rolly's auto-calibration; calling the company tomorrow - stay tuned...

[DGBK]

That gives me an idea though - If I knew how many mm the nozzle extends below the effector plate on a standard build, I could calibrate that as my "zero" on this gauge, which could hopefully nip any problems in the bud... Does anyone know that measurement?

I would think that before removing the hot end to put on your measuring rig, you should make sure your nozzle height is properly zeroed at bed center. Then, once you have your assembly strapped on, hit your hotkey z=0 button (i think it is G28 z0 f3500). This will put your setup at what was zero for your nozzle.

I'm assuming that once your reattach your hot end, you'll still end up a little off the mark due to torquing everything back on, but you'll be pretty close.

[HumanLiberty]

That gives me an idea though - If I knew how many mm the nozzle extends below the effector plate on a standard build, I could calibrate that as my "zero" on this gauge, which could hopefully nip any problems in the bud... Does anyone know that measurement?

I would think that before removing the hot end to put on your measuring rig, you should make sure your nozzle height is properly zeroed at bed center. Then, once you have your assembly strapped on, hit your hotkey z=0 button (i think it is G28 z0 f3500). This will put your setup at what was zero for your nozzle.

I'm assuming that once your reattach your hot end, you'll still end up a little off the mark due to torquing everything back on, but you'll be pretty close.

Thanks for the advice - I already made the swap, so will see what happens, and if results are not good I'll try again with the method you describe.
Unfortunately, the .html system is freezing up, so no results yet. Will keep trying tomorrow.

[RollieRowland]

Unfortunately, the .html system is freezing up, so no results yet. Will keep trying tomorrow.

If you're getting lag, try lowering the accuracy from 0.001 to 0.005(or more). This takes longer/lags because the code is in a loop and will run until it has reached the desired accuracy, which can be a problem if the height-map values are significantly off.

[Jrjones]

Unfortunately, the .html system is freezing up, so no results yet. Will keep trying tomorrow.

If you're getting lag, try lowering the accuracy from 0.001 to 0.005(or more). This takes longer/lags because the code is in a loop and will run until it has reached the desired accuracy, which can be a problem if the height-map values are significantly off.

Ah, so thats why it appeared to crash on me when I tried setting the accuracy to 0.0001, haha. I ran out of filament so I am trying it with with my analog gauge that has 0.0001" resolution.

You had previously asked for what my measurements were, so here they are:
Results with the digital dial indicator (the same one that Jimustanguitar has) (units are in mm):

Results with the analog dial indicator (still going with this, maybe I need to try the expert settings) (units are in mm):

Do those results look reasonable to you Rollie?

I also noticed that while printing before that the corner of a larger piece didn't stick too well near the Y tower, and then warped. The rest of the print stayed flat.

[HumanLiberty]

Unfortunately, the .html system is freezing up, so no results yet. Will keep trying tomorrow.

If you're getting lag, try lowering the accuracy from 0.001 to 0.005(or more). This takes longer/lags because the code is in a loop and will run until it has reached the desired accuracy, which can be a problem if the height-map values are significantly off.

Thanks Rollie - turned out in my case it was due to a failure to press "Enter Defaults"
which I learned from the below post from earlier in this thread.

It also address Cyber's concerns about accuracy surviving transition from probe back to nozzle, which apparently it did, in techstorage's case:

"Re: Delta Calibration Program for Very Accurate Bed Leveling
Postby techstorage » Thu Jul 23, 2015 2:01 pm

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."

[RollieRowland]

Unfortunately, the .html system is freezing up, so no results yet. Will keep trying tomorrow.

If you're getting lag, try lowering the accuracy from 0.001 to 0.005(or more). This takes longer/lags because the code is in a loop and will run until it has reached the desired accuracy, which can be a problem if the height-map values are significantly off.

Ah, so thats why it appeared to crash on me when I tried setting the accuracy to 0.0001, haha. I ran out of filament so I am trying it with with my analog gauge that has 0.0001" resolution.

You had previously asked for what my measurements were, so here they are:
Results with the digital dial indicator (the same one that Jimustanguitar has) (units are in mm):

Digital indicator cal results.JPG

Results with the analog dial indicator (still going with this, maybe I need to try the expert settings) (units are in mm):

Analog indicator cal results.JPG

Do those results look reasonable to you Rollie?

I also noticed that while printing before that the corner of a larger piece didn't stick too well near the Y tower, and then warped. The rest of the print stayed flat.

Yes, these do look reasonable, although it seems your frame is a bit warped. This won't necessarily be a problem, but it will affect XY scaling of printed parts. You shouldn't see anything more than 0.25 to 0.75 mm change over the whole size of your build plate, from what I see. This issue can be countered with modifying the Delta Radii, but at the moment there isn't an efficient way to do so. I am looking into a fix for this though.(Or rebuilding your frame to make it square.)

Thanks Rollie - turned out in my case it was due to a failure to press "Enter Defaults"
which I learned from the below post from earlier in this thread.

It also address Cyber's concerns about accuracy surviving transition from probe back to nozzle, which apparently it did, in techstorage's case:

Also, forgot to mention that pressing Enter Defaults is a necessity, unless you are using expert settings! But you seemed to find that, good thing... This causes an error because it will not be able to compute the XYZ/diagonal rod offset percentages (without providing the 4 height-maps or pressing enter defaults they are zero), then it will do some fun dividing/multiplying by zero, messing up the calculation, causing an infinite loop too(until it is closed), etc.

For the "accuracy surviving transition" statement, I have also noticed a small error when calibrating with a probe/calipers that protrude beneath the nozzle. That is actually the main reason as to why I decided to upgrade to FSRs. That distance is then zero, and of course you don't have to recalibrate the z-height.

[geneb]

I've been pondering making a z-probe out of an alligator clip and a feeler gauge. The alligator clip is soldered to one switch lead and grips the wrench faces of the nozzle. The feeler gauge gets the other wire. You just side the feeler gauge under the nozzle each time it probes. As long as you've cleaned your nozzle, this should work without a problem.... You'd make the Z-probe height whatever the feeler gauge thickness is.

This is the method I use when setting my Z-zero point on my ShopBot. It has a "z-zero plate' that's around 1/8" thick. I install a cutter and run the probe process - it comes down and touches the plate quickly, backs off and comes at it more slowly, then sets the zero point by adding the plate thickness to where ever it found itself and then zeroes the current Z position.
(on a side note, I also do this for a "3 axis zero" that can be seen here: https://www.youtube.com/watch?v=ENoIdEcxeFc)

g.

[mhackney]

I did exactly that when I was prototyping my delta compensation gcode preprocessor. It works very well indeed. .001" feeler is .0254mm. To make a more permanent setup, run one wire up the Bowden tube and fasten to the hotend (clip or a screw depending on your hotend, etc) and then you'll be good to go anytime.

[HumanLiberty]

For the "accuracy surviving transition" statement, I have also noticed a small error when calibrating with a probe/calipers that protrude beneath the nozzle. That is actually the main reason as to why I decided to upgrade to FSRs. That distance is then zero, and of course you don't have to recalibrate the z-height.

[/quote][/quote]

So I did the .html calibration and it looks like the results are a huge improvement - 10k x Thanks!

So here's the question:
Tricklaser's custom 325mm CF arms are en rout to me. Am I going to have to re-run all the calibrations after installing?
I'd imagine so, and if so, will wait for them before swapping the probe out for the hot end.

I suppose I'll have to change the Diagonal Rod Length starting point to 325 in EEPROM...?


Thanks all!

[Jimustanguitar]

Yes, I'd run the calibrations again. I'd set everything back to stock, change the arm length accordingly, and see where you're starting from.

I'm very curious how much it will change what you're experiencing without correction.