Will this probe work for the auto-levelling?

[insta]

http://www.amazon.com/gp/product/B008DG ... AZDOWP300U

'bout to order 4 up in this bizzy

My plan is to use clean glass with braked 22ga cold-rolled carbon steel sheets bulldogged to it, with BuildTak on the steel. I can get a zillion sheets of that steel stamped at a local metal shop for nothing and the BuildTak isn't terribly expensive either. The auto-level should compensate for any slight bulge the steel gets, and I can flex it like a poor-man's Zebra Plate.

[willnewton]

This is the one I have used. Same thing, 2$ cheaper without Prime shipping, but is still free shipping.

http://www.amazon.com/Detecting-Distanc ... 4H16S3YEXZ

You will need to make a voltage divider with a couple of resistors to get the output signal down to 5v. Maybe. I have not done this on my M2, but on my RAMPS 12v printer, I used two 10k resistors and took power straight from the 12v lines. The end stop 5v supply would not trigger the probe correctly.

I am not sure about the Rambo board hookup protocol.

PS- I also ordered the 8mm detecting probe. It is really large, probably double the diameter. Just stick with this one.

[willnewton]

Also, autolevel does not compensate for bulge. In fact if you have a bulge it can make your bed totally off.

The leveler builds a flat plane averaging all the points. A high center measurement will cause one of the corners to print lower than the others.

Your bed can be unlevel as you like, but it still has to be dead flat.

That is why it is better to call it tramming than leveling.

[insta]

I was planning to use 9-point mesh probing.

I'll see how it works on 5v straight from the endstop harness before going to 12/24v on the power supply pins. Thanks for the feedback!

[willnewton]

I was planning to use 9-point mesh probing.

I use 9 points too, but it doesn't make it better. It takes those nine points and builds a flat plane. It does not make a mesh with hills and valleys to compensate for an uneven surface.

I was having a hard time finding proof of this to show you via an internet link and was only 99% sure myself, as I am pretty sure my piece of 18x24x3/8" piece of Mic6 is flat, but I have never directly tested for "point specific" leveling.

I do know on that machine, my rails dip a bit in the middle due to the wide bed. The autolevel has never compensated for this and causes one of my corners to always be off.

So to 100% verify this, I placed a metal file under one of the nine bed points as my auto level probed and guess what? The print starts 3mm high, it does not dynamically lift over a single probe point. It builds a level plane only and does not compensate for uneven spots.

Again, this is why the "leveling" part is inaccurate and tramming is the better term. The software assumes a flat bed and as best as it is able, builds a plane parallel to it, not conforming to it.

Hell, I could still be wrong, but not from my direct experience.

[insta]

I have a large (18x18 inch) machine that does 16 point probing, and it definitely does build an internal grid of points. The Z axis will move up and down as the head tracks around. That firmware may be using "mesh levelling" instead of just 4-point autoprobing, but it's definitely something Marlin supports.

[Bratag]

I have a large (18x18 inch) machine that does 16 point probing, and it definitely does build an internal grid of points. The Z axis will move up and down as the head tracks around. That firmware may be using "mesh levelling" instead of just 4-point autoprobing, but it's definitely something Marlin supports.

I am using the 9 point mesh with the probe on my M2 and have to agree. The z axis does move up and down as the head tracks around which would compensate for an uneven bed as well as being useful for an unlevel bed. The only down side is you will need to adjust the z height in the software/firmware because as a rule the trigger point for the probe is not ideal for printing correctly.

[willnewton]

Hell, I could still be wrong, but not from my direct experience.

Well crap. So basically this means my firmware or something is not working right.

I have been pretty ticked about how crappy the leveling feature is!

[Tim]

Well crap. So basically this means my firmware or something is not working right.

I don't think that the grid-based leveling has been in Marlin for very long, so you might want to compare firmware versions. I've mentioned before that I got all this working under Smoothieware, which had grid-based correction implemented a bit earlier, so I can't be much help to you with the Marlin firmware.

Also make sure you're issuing the right commands for the grid probing. It would not surprise me at all to find that if you issue the wrong command, it will probe lots of points and then reduce it to a plane.

[willnewton]

I don't think that the grid-based leveling has been in Marlin for very long, so you might want to compare firmware versions.

Aaaaaaaand there's the answer. Thanks Tim.

I am using the latest stable Marlin release, 1.0.2, which has multipoint-based leveling projected as a plane.

I checked the Marlin Github release notes for the upcoming 1.1.0 version and saw the dynamic mesh bed leveling feature was not implemented until 1.1.0-Release candidate 2

Sorry for the confusion Insta, but I was happy to see that I was not totally crazy.

Tomorrow, new firmware install is high on the priority list!