web based Delta printer least-squares calibration calculator

[DGBK]

I finally got the calculator to work for my Rostock Max V2 running Repetier. I had to do the following:

First, I had to set the program to RepRap. I then converted my offsets to millimeters by dividing my steps by 80. This is because 80 is the default setting for steps per mm for a Rostock Max V2. I always entered mm of offset into the calculator.

Next, I entered all of my rotations as 0. The program doesn't do the exact same calculations if you enter the default rotations of 210, 330, and 90. Again, this is a firmware difference between Repetier and RepRap.

After probing all my points, I converted the millimeters back to steps for my x, y, and z offsets; changed my horizontal radius ("delta radius" in the program); Z-max length ("new homed height" in the program); and converted my rotations angles to angles usable by Repetier. To do this you add if positive and subtract if negative from your default angle number. The angle labeled X corresponds to Alpha (default 210) and the angle labeled as Y corresponds to Beta (default of 330).

You can use several iterations of the program to slowly converge on the proper numbers. Just keep entering the new numbers into your EEPROM and the "Initial" boxes before doing another round of probing. After a while, I was no longer converging. My number ended up with an error of 0.02 pre correction and post correction a couple times in a row, so I figured my calibration was complete.

[Lightpath]

Do you use a PEI Plate too with your FSR setup? No Burnmarks?

PEI is remarkably resilient. I have scraped my head along the PEI surface multiple times with no real ill effects, and after many leveling probes my PEI shows no signs of damage. If something does mark the PEI, I just give it a light sand with wet paper and a sanding block.

My PEI isn't thin, but near as I can tell it is very durable stuff.

[DeltaCon]

After probing all my points, I converted the millimeters back to steps for my x, y, and z offsets

How did you probe? Did you do that using a probe, or could I use the "z-height calibration with paper below the nozzle"-method?
And why do you need to set it to reprep, as it actually has repetier (standard)?

[DGBK]

I have a .04 millimeter feeler gauge. It's the same as reading offsets from paper, but slightly more accurate.

When I posted, the sign correction for repetier was backwards. I suppose I could have multiplied my offsets by -1 instead. David has since updated the tool to correct the end stop calculations for repetier, so you should be able to ignore that.

[JFettig]

I used this calculator yesterday. The repetier version of it didn't seem to work quite right so I switched over to the reprapfirmware version and eventually just had to calibrate the end stops manually to get them close. I did like using it for tower rotations and horizontal radius. I ended up with the center at 0, each tower at -.005" and opposites of the towers at +.005". It seemed to be the best I could do.

[mhackney]

It would be great if folks could provide this feedback to David (the author). I am sure he does not monitor this forum and he would benefit from the feedback.

thanks,
michael

[DGBK]

I tried it again earlier this evening, and the end stop corrections were working fine in repetier. It's interesting that Jfettig and I had very similar results. When I first calibrated, I had an error of 0.12 mm, wich is about the same as your 0.005". I also had lifting between each tower.

The only way I could get to 0.02 or 0.01 was to let it calculate my arm lenght. This of course messed up my x/y dimensions, so I went back and calibrated a smaller area of my bed.

I'd be interested to know what your setup is. I have the new molded carriages, 300mm tricklaser arms, and a rev 7 onyx bed that has distinct doming in the middle.

I'll give David some more feedback in the morning. He may have forgotten the tower angles. The last time we spoke, he said he'd update them with the xyz offsets.

[JFettig]

The weird part about my end stop calibration with it was my Y and Z towers were nearly spot on, then X -.1" or more... it went from bad to worse. Maybe I punched it in backwards, I dunno for sure.

[DGBK]

The weird part about my end stop calibration with it was my Y and Z towers were nearly spot on, then X -.1" or more... it went from bad to worse. Maybe I punched it in backwards, I dunno for sure.

I guess you have two possible sources of error. You could be converting from inches improperly, or you could have your sign reversed.

I just use the height indicator on my lcd. This tells me the height in millimeters. If i get a negative number, that means my 0 was above the bed (enter positive number into the calculator). If my bed height reads a positive number, that means my 0 point was below the bed (enter a negative number into the calculator).

Are you measuring in inches? You keep labeling your numbers with a ". If so, are you sure you are converting to millimeters properly? This could definitely throw off a measurement.

[DeltaCon]

I am trying this method for calibration, but wonder what the "Repetier firmware term" is for rule number 1 in the explanatory text. It says: "If you have tower radius corrections configured, remove them". But in the list of firmware settings I can't find that. Or are "Delta Radius" A, B and C meant here...?

Same goes for most other terms:
Initial Delta Radius - Horizontal Radius ?
Initial homed height - X / Y/ /Z Max length ? (I have 2 different values in these settings...)

[DeltaCon]

Next question: (but don't forget my previous message! )

I am beginning to think that not only my Rostock has a loose screw, but my head too...
Trying to perform the calibration as outlined at the escher website, I am putting some gcode into the terminal.
Always when I send the hotend to height by G1 X0 Y0 Y20 or whatever, MatterControl adds .96 to the Z value.
Just tried Repetier Host and that is not doing the same thing, so I assume it is a software thing and not a firmware setting.
Can anyone explain?

Third question
I was thinking of writing a script (how original!) with the suggested probe point values. Because I have no dial Indicator (yet) I was planning to use the paper-drag-method for now. That means that after each movement I wil have to stop the script to manually dial in and take note of the value when the paper drags. Goolgling through this forum teached me the @pause command, that is supposed to work only from withing Repetier Host. However, the popup for pressing OK to continue appears on screen, but the script completes anyway... Is there a solution for this?

[DGBK]

I am trying this method for calibration, but wonder what the "Repetier firmware term" is for rule number 1 in the explanatory text. It says: "If you have tower radius corrections configured, remove them". But in the list of firmware settings I can't find that. Or are "Delta Radius" A, B and C meant here...?

I believe those are the individual tower radius corrections. They come into play if your carriages aren't a uniform size which causes uneven x/y scaling. By default they are zero in your EEPROM, and this method will not adjust them.

Same goes for most other terms:
Initial Delta Radius - Horizontal Radius ?
Initial homed height - X / Y/ /Z Max length ? (I have 2 different values in these settings...)

Initial Delta Radius is your Horizontal Radius in the EEPROM.
Initial homed height is Z Max length. (if I remember correctly X and Y were set at 250 by default and I never touched these)

You didn't ask but there is also X / Y / Z tower rotation. This is equivalent to our A / B C tower rotations. These are by default 210, 330, and 90. The script uses offsets of 0, so if it tells you -0.14 for tower X, you need to enter 209.86 into your EEPROM for tower A. You will always add or subtract from the default tower rotation. Likewise the x, y, and z offset steps are given relative to zero. He gives you the number to enter rather than asking you to change compared to what you have.

Third question ;-)
I was thinking of writing a script (how original!) with the suggested probe point values. Because I have no dial Indicator (yet) I was planning to use the paper-drag-method for now. That means that after each movement I wil have to stop the script to manually dial in and take note of the value when the paper drags. Goolgling through this forum teached me the @pause command, that is supposed to work only from withing Repetier Host. However, the popup for pressing OK to continue appears on screen, but the script completes anyway... Is there a solution for this?

You can make hot keys instead of a script. I have them set up for P0 through P9. Each one homes before moving to the point (although that probably isn't necessary). All of my hot keys bring the hot end down to 1 mm over the bed. Then I jog it down using the smallest possible move on matter control (0.1mm). Once it looks close or reaches 0, I switch to the dial on the LCD to move the head down by 1 mm/ 80 for each click. I don't see how this would be possible with a script and no dial indicator, as you'd need a dial indicator to avoid the fine tuning.

[DeltaCon]

Thanks DGBK, exactly the info I was looking for!
Now I know what to do this weekend!

I don't see how this would be possible with a script and no dial indicator, as you'd need a dial indicator to avoid the fine tuning.

If it would be possible to pause a script, you could take the time to do the dial in manually, that was the purpose of the pause command, but it just seems to not work. Perhaps that command is not supported in the standard Rostock firmware. Hot keys is something I have not used before, so I will look into that. A separate script for each position is also a possibility ofcourse. Just to avoid typing too many manual gcode

I ordered this dial indicator through ebay, but I am just not patient enough:
[img]http://hennekens.info/rostock/dialindicator-b.png[/img]
It is said to have a data port, but I'm not sure it is usable for Opendact. But if it works good for this manual method, I am more than happy

[DeltaCon]

You can make hot keys instead of a script. I have them set up for P0 through P9.

Can you please elaborate on this? I searched for hotkey setup for both repetier and mattercontrol but can't find anything...

[DGBK]

Just follow the instructions from the calibration portion of the assembly manual and change the macros according to the suggested locations in the web based fields.

[cope413]

Ultimachine does have this guy in the wings...

So perhaps there IS a board change in the works with SeeMe...

Just to make it easy...
No, you can't have it. No, I haven't gotten it up and printing yet. No, I don't know when it's coming out or how much it will be.

But I can let you know as soon as I know

[Eaglezsoar]

Ultimachine does have this guy in the wings...

2016-01-25 16.52.02.jpg

So perhaps there IS a board change in the works with SeeMe...

Just to make it easy...
No, you can't have it. No, I haven't gotten it up and printing yet. No, I don't know when it's coming out or how much it will be.

But I can let you know as soon as I know

A very handsome ARM board, looks like it could make a nice upgrade.

[dmo]

I've been trying to decide which way to go. mHackney seems really fond of the FSRs, but I have read good things about the IR probes. Which is the better rout to go. If I have a led ring will that interfer with the IR probe? Thanks

[Xenocrates]

I've been trying to decide which way to go. mHackney seems really fond of the FSRs, but I have read good things about the IR probes. Which is the better rout to go. If I have a led ring will that interfer with the IR probe? Thanks

I would go the FSR route. With an IR probe, if say you swapped nozzles it it was longer (Prometheus and Volcano for instance) you would need to adjust the IR probe or the mount to change the trigger distance, while with FSR's it doesn't care. IR probes also require some amount of tuning for different bed surfaces, and have some amount of offset too, while with FSR's since you use the nozzle to probe, it will always be right where you are probing, and will always trigger right at the surface.

[kraegar]

I think, if you have a rambo and aren't ready to switch yet, this online calculator + seemecnc's accelerometer board from the eris would be a great way of calibrating a Rostock Max. Until the Archim is out and I upgrade to that.

[teoman]

The pei will get damaged, not instantly but over time.

The center of mine developed a little spot where i did the z height calibration multiple times.

[dmo]

Thanks. I'm switching to the Duet board with the Chimera and Cyclops hot end. I'll give the Fsr a try..

[TN Yankee]

I'm trying this calculator on my V2, with a home-built heated bed, and 0.91 firmware. Sometimes instructions clear to other people aren't so clear to me, and so I have a few questions on the values required.

1) For the initial run, the instructions state to clear any tower rotation values currently in use. However, when another iteration of the calculator is run, the values generated in the first run are retained in f/w, and the offsets are entered as the "initial values," correct?
2) When new values from the 2nd (and subsequent) iterations are run, are they added/subtracted to rotation values from the prior run, or are the new values used as-is?
3) These same questions also apply to the end-stop values.

Essentially, I'm confused as to exactly what values are carried over to future iterations. So far on each run, I've been adding/subtracting to the existing values for both the tower rotation & end-stop values, and have made pretty good progress. It's time-consuming, however, and I want to make sure I'm using the right method, rather than just having dumb luck.

I hope to upgrade to a Duet WiFi in the future, but it will likely be a while.

Thanks!

[dc42]

1. Correct.

2. Used as is.

3. Used as is.

[TN Yankee]

Thanks, David!