Interesting New Delta Printer with Autoleveling

[jetpad]

This one is on kickstarter but here is a video showing how it autolevels the bed with a little flip-down probe.

http://www.youtube.com/watch?v=X1iHC4VARQI

[geneb]

That feature is in Marlin and I think a beta of Repetier. Johann developed it about a month ago.

g.

[Eaglezsoar]

That feature is in Marlin and I think a beta of Repetier. Johann developed it about a month ago.

g.

I sure would like to see an implementation of this for our Max. That is really neat and one heck of a time saver.

[Gimbal]

Hi I found this a few days ago:
https://github.com/Extent421/Repetier-Firmware
, seams to have some new futures, didn't hade the time to test it yet

[626Pilot]

I tried to order the PCB for that but they don't have shipping to the US in their order form. Specs are here: http://www.reprap.org/wiki/CrashProbe

I think they will ship, but you have to contact them directly. I was just going to see if I could order the parts myself and put them on some perforated silicon board.

There is another mod which just uses a switch (same type used in the endstops) to detect the bed height in the center, and maybe 10 lines of G-code to tell the controller "this height is 0." You'd put it in the custom G-code section of your slicer. I found it on Thingiverse a few days ago but I'm having trouble finding it again. I developed some tower clamps based on someone else's idea and I found that since I re-tightened everything with them in place, the bed is *perfectly* level, to where I can print at 80 microns and it's perfect. So, all I really need is to know the height at the center. That's the one thing that always bugs me, having to go through three or four cycles of trying to print something and adjusting the tower heights by .02 of a millimeter. Should be a huge time saver.

[kbob]

626pilot, my dial indicator mod made that process a whole lot easier and more repeatable. It's not as sexy as a fully automatic solution, but it can be ready to use in the time it takes you to print a bracket and drive to Harbor Freight. My bed is dialed in to about 0.0006 inches (0.015 mm).

http://forum.seemecnc.com/viewtopic.php ... =90#p16474

Unfortunately, I shorted out my thermistor before I printed anything, and now I'm waiting for parts.

[geneb]

The auto bed probe builds a "map" of the bed, it's not just setting a zero point.

To give you an idea of what this means - if you were to change the height of all three end stop screws (so they're ALL different), running the auto-level routine before a print would result in a perfectly accurate first layer, every. single. time.

g.

[626Pilot]

626pilot, my dial indicator mod made that process a whole lot easier and more repeatable. It's not as sexy as a fully automatic solution, but it can be ready to use in the time it takes you to print a bracket and drive to Harbor Freight. My bed is dialed in to about 0.0006 inches (0.015 mm).

http://forum.seemecnc.com/viewtopic.php ... =90#p16474

Unfortunately, I shorted out my thermistor before I printed anything, and now I'm waiting for parts.

That... would have saved me about $30 buying the SeeMe pen holder! I use that and a set of four scripts to do the leveling. I get it accurate to within 1 thou. I don't think it's even necessary but when I was learning how to do it I wanted to get it as good as possible. I did a realignment today with the dial gauge and the bed not heated, and when it WAS heated things were off, so I printed the first layer of something at .08 and if it looked too smooshed or too round near one tower I would just tweak the adjustment screw. Works well enough.

The auto bed probe builds a "map" of the bed, it's not just setting a zero point.

To give you an idea of what this means - if you were to change the height of all three end stop screws (so they're ALL different), running the auto-level routine before a print would result in a perfectly accurate first layer, every. single. time.

g.

That's why I am DYING to do this.

What would be even better (although I don't know if it's possible with RAMBo) would be to get the printer to just do that WITHOUT a depth probe. The RAMBo uses A4982 stepper drivers, which appear to have current sense capability. If they do, and if the right lines are wired to the CPU, it should be possible to do so with a firmware change and nothing else.

[626Pilot]

https://github.com/repetier/Repetier-Fi ... issues/120

The Repetier dev guy says the dev version has some new features, like "endstop leveling." There isn't much new code referencing it but maybe it'll turn into something decent.

[dgirard]

I'm working on adapting this to our rostock max to see where I can get with it...

http://www.thingiverse.com/thing:111397

I've got a different hall effect switch than this guy used...and I've got some magnets on the way...If I can figure out how to fix the cad model it seems pretty straightforward. I'm going to start looking at the firmware to see if I can make sense and merge the auto-leveling code in...

This could really make setup a breeze...after this we just need a camera or some type of other sensor to tell when we've moved from "printing" mode to "making a big pile of plastic spaghetti" mode and auto-cancel the print job... <g>

[Flateric]

I'm running the autoleveling version or I believe it's "alpha" repetier firmware now. It has a number of nice features and seems to be running fairly well on my machine.

However, I am uncertain on how to wire up the switch needed for the probe to my old school Rambo 1.0. What input would I use for this? Just a spare endstop switch referenced to the variable.

Polygonhell, you know everything firmware related that there is to know on the planet (atleast it seems like it to my mere mortal limited knowledge anyways), suggestions guidance? Tips? Mocking abuse? LOL

[Nylocke]

I plan on printing off the auto level probe for my Rostock Mini I built for myself, I'll post how it goes so I might help some guys around here, though I'm using Marlin with my printer, and if I have time ill try and put one on the RMAX I built for my work and post how that goes as well. Good luck to whoever tries this (the Kossel creator has been able to print PLA direct to glass, no heat involved because of this probe, I think thats pretty sweet)

[Eaglezsoar]

I plan on printing off the auto level probe for my Rostock Mini I built for myself, I'll post how it goes so I might help some guys around here, though I'm using Marlin with my printer, and if I have time ill try and put one on the RMAX I built for my work and post how that goes as well. Good luck to whoever tries this (the Kossel creator has been able to print PLA direct to glass, no heat involved because of this probe, I think thats pretty sweet)

That would be great if you could help us implement this on our RMax. The majority of us use Repetier firmware hopefully the one you have at work also uses Repetier and you can get
it to work with your new probe.

[626Pilot]

I'm working on adapting this to our rostock max to see where I can get with it...

http://www.thingiverse.com/thing:111397

I've got a different hall effect switch than this guy used...and I've got some magnets on the way...If I can figure out how to fix the cad model it seems pretty straightforward. I'm going to start looking at the firmware to see if I can make sense and merge the auto-leveling code in...

This could really make setup a breeze...after this we just need a camera or some type of other sensor to tell when we've moved from "printing" mode to "making a big pile of plastic spaghetti" mode and auto-cancel the print job... <g>

Do you have a source for the Hall effect sensor board that's in the US or will ship to it?

[dgirard]

I bought a few of these:

http://www.ebay.com/itm/261186753702?ss ... 1439.l2649

They look to be pretty similar (although this is new territory for me), but the form-factor of the board is different. FWIW the seller shipped them out fast, and they were well packaged.

I'm planning on getting them hooked up in the next week or so. I'll keep everyone posted if they work.

David

[Nylocke]

Unfortunately I prefer Marlin over Repetier (I built the one at work and loaded Marlin onto it), and I'm having to leave my job because it was only for the summer, but I'll download and modify a version of Repetier for my Mini and experiment with that until I get it and share it with the lot of you guys if I can figure it out

[JohnStack]

I believe it is all still very immature and lacks sophistication. Hall effect sensors are subject to way too many different conditions. Dial indicators rely on adapting another physical device onto the whole thing, which is also a calibration step.

I've been doing napkin cads with lasers and mirrors; where four and/or five mirrors are used. My goal is to have a receiver mirror with a bulls-eye etched onto it (or with software, a screen), where, like a marksman, you dial in your end stops and EEProm settings.

Last month, I tried doing a laser trigger (photoresistor)!with an LED turning off when the hot end crossed the plane...it is still a valid test; however my implementation was sketchy at best. Two challenges, precision of the trigger (height) and futzing with the imaginary line between the center and an axis. I asked Steve to silkscreen on some lines on the hot bed, not sure if he has, will know later today. Whee!

[Eaglezsoar]

I believe it is all still very immature and lacks sophistication. Hall effect sensors are subject to way too many different conditions. Dial indicators rely on adapting another physical device onto the whole thing, which is also a calibration step.

I've been doing napkin cads with lasers and mirrors; where four and/or five mirrors are used. My goal is to have a receiver mirror with a bulls-eye etched onto it (or with software, a screen), where, like a marksman, you dial in your end stops and EEProm settings.

Last month, I tried doing a laser trigger (photoresistor)!with an LED turning off when the hot end crossed the plane...it is still a valid test; however my implementation was sketchy at best. Two challenges, precision of the trigger (height) and futzing with the imaginary line between the center and an axis. I asked Steve to silkscreen on some lines on the hot bed, not sure if he has, will know later today. Whee!

Thanks for your continued interest in this which is sure to benefit us all eventually. Rather than relying on silkscreened lines on the hotbed which would only benefit those with such a
hotbed or be willing to buy one, I believe we need a solution that will work with ALL Rostocks including the hundreds that are presently built. I would not mind purchasing a small probe
of some sort but a new hot bed? I saw the one You Tube solution that used a common switch similar to an end stop switch that dropped down on command and used the switch closings
to calibrate the bed. Now that is a cool, inexpensive solution. This is the link http://www.youtube.com/watch?v=X1iHC4VARQI Note: This is the same link JetPad used at the beginning of this thread.

[JohnStack]

So, now for my ignorance...

How does calibration work beyond the imaginary lines between the center and the three axis? What is the equation that translates for 7 or more measurements beyond averaging that translates into EEPROM settings (the five that we know?)

#hatemath

[626Pilot]

So, now for my ignorance...

How does calibration work beyond the imaginary lines between the center and the three axis? What is the equation that translates for 7 or more measurements beyond averaging that translates into EEPROM settings (the five that we know?)

If you sample the height at the center of the build surface and a bunch of points around it, you can calculate the angles between these points and the center, and use these to calculate the plane defined by the build surface. More samples provide more data points, which provides a better average and compensates for whatever vagaries exist in the sensor system.

[Eaglezsoar]

So, now for my ignorance...

How does calibration work beyond the imaginary lines between the center and the three axis? What is the equation that translates for 7 or more measurements beyond averaging that translates into EEPROM settings (the five that we know?)

#hatemath

I don't know the math behind it but that video shows the simplest form of Auto Bed Leveling that I have seen. Now if we can find someone with a PHD to design the switch to fit the RMAX
and modify Repetier to work with it. What are the chances? I would say pretty good if enough interest is there. This leveling procedure is the most time consuming part of a 3D printer and has to be accurate. I am surprised that someone has not developed an easy to implement solution such as what is in the video.

[JohnStack]

So, steps before the momentary switch is turned flipped.

An idea:

A separate, non-board solution...not requiring software...only an arduino and a display and potentiometer (to set steps per on the separate arduino)

1. Do your Z0 center measurement manually.
2. Have a circuit wired (isolated from Rambo/Arduino but using the same switch) in conjunction with momentary switches at the bottom each axis.
3. Values collected are the time it takes to move from end to start and then translated back to a measurement.

It might be easier to do the laser/mirror thingy. I just can't figure out which to reflect where. What I was thinking:

You put one at the center of the bed, one on the hot end, and then for each axis, somewhere on the cheapskate. I tried a few variants of that but got nowhere. I'm sure someone has done it for something.

[daftscience]

http://www.adafruit.com/products/1362

I am really interested in getting that working as a way to sense the bed. The idea would be to wire it to the hot end.

Pros: No need to mount a prob, just hook up one wire.
Cons: Need a capacitive material for the bed.

[mrblade]

Where to connect in the Rambo Z probe?

[geneb]

Yeah, that's not how that technology works. Capacitive touch sensors pretty much require a finger to work. (Note that I didn't watch the video, so I don't know if they discuss printer applications - I'd still call it fishy if they did)

g.