Welcome to the SeeMeCNC Community

First off, a huge thanks to one of our members on the forums here DC42 for his work on the delta least squares calibration code. He's given us permission to modify and adapt his code to more easily implement it with our probe and firmware.

We've modified his online calculator and it is now available on our website. Here's the link straight to the page

https://www.seemecnc.com/pages/delta-calibration-wizard

This is still a work in progress, but we've had a lot of feedback that anyone who has a machine that has a tilted bed/high or low spots on bed/general calibration difficulty has seen a MUCH better calibration after using this wizard to dail in the things that our probe and current firmware/electronics just simply cannot account for and measure/adjust. Reasons being we are out of available memory/cpu/etc... with the RAMBo to run the more advanced calculations that this wizard does. In time we'll have a controller that will be able to do that, but in the mean time, with a few minutes on this wizard, you'll be able to get your machine dialed in better than ever. The inherent variations in manufacturing the kits, from injection molding to laser cutting and then finally assembly variations is the reason some people have different results with the probe.

So, how do you use this wizard?

There are instructions on the web page. Please read them entirely first before starting the calibration. Here's a quick abbreviated instruction set. Download the newest MASTER branch of our firmware. Clear eeprom, upload new firmware, verify you have firmware date of 20161209 or later, run G29 probing on your machine while connected to mattercontrol, open eeprom, on the wizard choose your printer model, enter eeprom values into wizard, download probe gcode file from wizard, run and record probe points, click calculate, enter NEW calculated eeprom values onto your machine, save eeprom, PRINT ALL THE THINGS!


Feedback is always welcome guys!

Is there something special that matter controller does or can I use octoprint instead?

@rocky00717

Well, I don't think we've tested that, but I think it should work too, don't see why not.

We wrote instructions for MatterControl because it's what we recommend and it keeps everyone on the same page if we need to help anyone.

rocky00717 wrote:Is there something special that matter controller does or can I use octoprint instead?

You sure can use octoprint/repetier-host/simplify/ANY host software that has a terminal window that you can read the output from the machine on.

As @partdaddy said, we write our docs based on the software that is officially supported by us. If you're familiar with any host software and it's terminal window no problem. But if you need help, please only use mattercontrol if you need to ask us questions

Thanks guys

Great work!

Remember, this wizard is a first time tool. After using it, you only need to use G29 or the LCD screen to auto calibrate your printer (such as when you clean the glass build plate).

I'm printing on FaBlam in the photo. The HE280 auto calibration makes it way more fun to print on this stuff. It can last 100's of PLA prints. Just use way less bed heat or no bed heat at all. Then careful when removing your print not to cut the FaBlam.

Here's my calibration print:


A couple calibration prints. Note that a skirt loop with a 275mm diameter calibration print is reaching even farther outside of the 10.5" (267mm) diameter working envelope.

http://repables.com/r/800/

I used the wizard twice and now I'm getting excellent results. I was having trouble with my first layer on everything I was printing. Stringy loops and some gaps where there filament didn't extrude. All that is fixed now.

Thanks very much to everyone who had a hand in this!

Thanks Steve and the guys at SeeMeCNC. They ran me thru this wizard, and I couldn't be happier. I had to drop my z-probe max height from -.20 to -.10, and that solved any other little misc problems I was having.

The manual wizard worked great for me!

Wade

What is everyone's experience so far with the calibration tool and scaling? I noticed that the tool changed my arm length by about 3%, which ought to change the XY scaling of the parts that I print... I'm working on a cal cube to quantify the error right now.

Is there an option to run this tool without changing the arm length in the EEPROM?

So, arms changed by about 3%, and the scaling of my printed calibration cube is off by about half of that, so it definitely does affect scaling like changing the arm length would under any other circumstance...

I'm going to give it a try with every value except for the arm length over the weekend. Hopefully that doesn't throw the rest of the math off too badly.

Jimustanguitar wrote:So, arms changed by about 3%, and the scaling of my printed calibration cube is off by about half of that, so it definitely does affect scaling like changing the arm length would under any other circumstance...

I'm going to give it a try with every value except for the arm length over the weekend. Hopefully that doesn't throw the rest of the math off too badly.

So what did you find? While arm lengths are modified, I didn't find too significant of errors.

Wizard makes nice flat 1st layer within .1 mm
BUT part size is off...
275 mm model prints at 270 mm
Fine for vases, frogs, Yoda heads, etc. not acceptable for real functional parts that fit properly.
Wizzer changed arm length from actual 300 to 310


Please make an option to not change arm length.

This is the difference between the six factor and seven factor calibration with dc42's code. The six factor cal doesn't touch the arm length. (My OctoPrint plugin only uses the six factor, as do my Duet-equipped printers.)

g.

geneb wrote:This is the difference between the six factor and seven factor calibration with dc42's code. The six factor cal doesn't touch the arm length. (My OctoPrint plugin only uses the six factor, as do my Duet-equipped printers.)

g.

Okay, so is there a way to use the other method without octoprint?

Is the 6 factor able to calibrate the super flat first layer as well as 7 ?

Brian, simple answer is yes. Most everyone only use 6 factor. It really isn't a good idea to use 7 factor (arm length). The only use case I can think of is to do an initial auto cal as 7 factor. Then print a calibration cube and scale the arms appropriately using the results. Then from then on do 6 factor. I should verify that with David, maybe he'll pipe in here.

bvandiepenbos wrote:

geneb wrote:This is the difference between the six factor and seven factor calibration with dc42's code. The six factor cal doesn't touch the arm length. (My OctoPrint plugin only uses the six factor, as do my Duet-equipped printers.)

g.

Okay, so is there a way to use the other method without octoprint?

Yes, just leave the "Number of factors to calibrate" box in the online wizard set to 6, which is the default. Or if you are using Duet electronics which has the calibration algorithm built into the firmware, use parameter S6 on the last G30 command in the bed.g file, not S7.

Looks like this new page has the factors hard coded to 7, might want to make that an option.

Is the math for the 6 factor calibration different than just omitting the change in arm length from the 7 factor calibration?

Jimustanguitar wrote:Is the math for the 6 factor calibration different than just omitting the change in arm length from the 7 factor calibration?

Yes it is. You get different numbers in the other boxes when you choose 6 or 7 factors.

I'm trying to work through the calibration wizard, all is good until I get to point 19:

"Once it has probed all 10 points and homes out, you're ready to enter these values into the wizard.  Starting in order enter the probe offset numbers into the POINT 0/1/2/etc... Probe Offset boxes below"

Except I see no boxes to enter the numbers into?

I think it should look something like the http://www.escher3d.com/ page but with 10 instead of 7 points?

Here's what I see:


Wizard by Doug Clark, on Flickr

I'm probably being a dumbass what am I missing?

Sorry, worked it out...
If you don't see the boxes select a different printer then go back to the one you want, then the boxes magically appear!

I've now been able to do the full method as described on the wizard page.
The results showing on the first print layers are now awesome, so a big thanks to all involved setting this up.

Well.... I was successfully printing parts, made some changes to the printer and used the 6 factor calibration and dimensions on the part changed. Originally I had used the online wizard which would have been 7 factor. Since the 6 factor was done after the 7 factor I wouldn't imagine that it changed the arm lengths, but I am going back to default arm lenths, redoing a 6 factor calibration and printing calibration objects.

So I get to step 20 and "click on the Calculate button" however the "Calculate new values" box isn't a button and doesn't do anything. (tired firefox & edge). First gcode generation from the upper section worked fine.

Anyone have a reference to what these calculations are? e.g. I would assume I could do the by hand if need be.

back story: have a rostock v2 that upgraded with the He280 kit last October. Initially it printed fine (excellently actually) with the normal calibration, first layers were well fused and worked great. Fast forward to about 10 days ago and the first three layers were starting to string and not fuse well. Played with extrusion temp and extrusion factors to no avail. It appears that even though everything is level it is just starting too high. After it lays down about three layers then the print looks normal. So started looking around for different calibration ideas and found the Wizard which currently appears to be broken.

running seemecnc vers mattercontrol 171204, printing pla currently.
any suggestions?
thanks,
fj..

The "Calculate new values" box is a javascript link. It won't work unless you've filled in all the boxes properly. Also, make sure you don't have an add or script blocker installed on your browser that's preventing the javascript from running.