Differential IR height sensor, Attemptin' Clone....ain't sure of success...now stage 10

I could just move the BuildTek I suppose. That might be easier. Won't help me if I need to use the heated bed though so maybe I should keep messing with the grid values until I get it right instead.

Oh it should be positive? I've never changing it to be honest. Do I even need it anymore?

Also, you changed things in the Marlin_main.cpp. Should I change them too? What difference did they make to your setup versus just changing the Configuration.h?

 

Yes you still need to set the z offset with M565 and a positive value for Z. Think of it this way. M565 is always opposite what you think it should be. Stock it is negative, and I had to think of it as the nozzle being the probe and hitting the plate is Z=0 and I want to back it up 1mm for example so Z-1.0. Now with the IR the IR triggers above the plate as Z=0. So you need to advance it to get the correct offset. Hence using Z 1.5.

Yes you should make the changes I did in Marlin_main.cpp. That keeps from the printer trying to out think you and force your z offset to be negative.

 

Ok I will make the changes. Does your printer home the z-axis on command? Mine doesn't.

 

It does. Just stops when the IR triggers.

 

Ah ok. Mine moves up and down about 5mm regardless of height. Doesn't seem to be triggered by the IR. If I move it up to 100mm for example, it does it's auto-level bounce 9 times in mid air.

 

Ok I got it to do an autolevel where all the points it measured were on the bed. Here is the result:

READ: Bed x: 15.00 y: 5.00 z: 0.01
READ: Bed x: 110.00 y: 5.00 z: 2.72
READ: Bed x: 205.00 y: 5.00 z: 4.78
READ: Bed x: 205.00 y: 110.00 z: 5.03
READ: Bed x: 110.00 y: 110.00 z: 3.15
READ: Bed x: 15.00 y: 110.00 z: 0.02
READ: Bed x: 15.00 y: 215.00 z: -0.01
READ: Bed x: 110.00 y: 215.00 z: 2.75
READ: Bed x: 205.00 y: 215.00 z: 4.84

Is that normal? The z-values seem a little odd but you guys know better than I do.

 

looks like your right side is a lot higher than your right side of the bed. Can dial that in a bit by hand adjusting the right leadscrews. Mine are pretty close across the probe. I think my calculated diff is like .15 or something the last print I did.

 

Visually it's perfectly flat. There's definitely not 3 to 5mm in the difference. I'll do a few more levels and tweak the rods though.

 

bed is flat...ok...but what is the x axis level compared to the bed is what it is saying. So your x axis is just high. Put a small level on the bed and again on one of the x axis smooth rods. You should see an angle difference in comparison to one another. That is what the probing is telling you when z is 0 and the left and +5 and the right.

 

I did it and noticed a difference left to right rather than right to left. I tried deliberately making an obvious slant so that the left side was visibly higher than the right side, but the right side values are still larger. Thanks very much for sticking with me through this by the way. I appreciate it!

 

ok...turn off your printer and get out a sheet of paper. put the paper on the bed and manually put the nozzle down to the paper so you can feel it tug a bit when you pull it. Move the printer head from one side to the other and manually turn the leadscrews until you get both sides pretty close to level where you get about the same tug on the paper when you pull it. That should get it pretty level and then turn it all on and do your G28 then G29 and see what you get.

 

Good call. I was doing that as you typed! Seems to have produced a better result:

READ: Bed x: 35.00 y: 20.00 z: -0.01
READ: Bed x: 120.00 y: 20.00 z: 0.29
READ: Bed x: 205.00 y: 20.00 z: 0.03
READ: Bed x: 205.00 y: 125.00 z: -0.02
READ: Bed x: 120.00 y: 125.00 z: 0.45
READ: Bed x: 35.00 y: 125.00 z: -0.01
READ: Bed x: 35.00 y: 230.00 z: 0.00
READ: Bed x: 120.00 y: 230.00 z: 0.29
READ: Bed x: 205.00 y: 230.00 z: -0.02

 

Boom! Now let her rip on a cal cube so you can get the z offset dialed in and you will be all set man.

 

The printers old Autolevel technique never helped me at all. This is some revolutionary stuff! Jim really built something outstanding. Thanks very much to you for helping me along. I'll do the test print now. These are the values I just got from another autolevel. They are quite different from the ones before but I guess I'll see how this setup compensates versus how it used to.

READ: Bed x: 35.00 y: 50.00 z: 0.36
READ: Bed x: 120.00 y: 50.00 z: 0.44
READ: Bed x: 205.00 y: 50.00 z: 0.06
READ: Bed x: 205.00 y: 140.00 z: 0.28
READ: Bed x: 120.00 y: 140.00 z: 0.52
READ: Bed x: 35.00 y: 140.00 z: 0.34
READ: Bed x: 35.00 y: 230.00 z: 0.43
READ: Bed x: 120.00 y: 230.00 z: 0.63
READ: Bed x: 205.00 y: 230.00 z: 0.50

 

hmmm...mine is pretty consistent. does the buildtech have white areas on it or anything? There have been some reports of the IR not triggering correctly over white areas.

 

Nope. I'm using the black side.

Getting different z values every autolevel shouldn't make a difference should it? Wouldn't my z-offset screw it up if I keep getting different z values?

I've always noticed that every autolevel produces different values. With the endstop autolevels too.

 

no...you should be ok...z offset is more off of where your homing of z comes from. the ABL readings just help calculate the tilt of the bed and will keep the nozzle x amount of mm away from the plate across the entire plane...in theory.

 

Ok. It's pretty late here now so tomorrow I'm going to put a test cube in each corner and test to see if the first layers are all pretty even. Thanks so much for your help!

 

no problem...sleep is important and you got pretty far with it. I'll be around tomorrow if you have any further questions. Take care.

 

So I did the test with a cube in each corner. Had to turn on the heated bed to get the top right cube to stick down. I posted pictures of the result. Seems too far from the bed at the top and too close at the bottom.

I know my BuildTek surface is pretty roughed up. Not much sticks to it, but whatever does sticks like glue.

Z-offset is set in Simplify3D at -1.85mm.