has anyone with PCB and glass setup successfully dialed in a very flat surface corner to corner?
I spent hours trying to do this meticulously adjusting one corner at a time and then check the other corner. what seems to be an impossible thing to adjust is the fact that I was able to get same nozzle elevation in the x direction at start point and end point but then when I moved the nozzle to the middle point, the nozzle dipped down and touched the surface… this drives me crazy… is this because the pcb flexing the glass?
I’m seriously considering abandoning Mk2A PCB and go with mic6 aluminum and silicone heater…
yes the pcb will never be flat… i have built two heated beds with pcbs ( one a MK2 and one from a bare pcb) . even a 2 mm flat glass attached to it will flex and bend a little. the only way to make it absolutely flat is to use 3mm - 4mm glass. the center of a thin glass will always be either higher or lower than the edges.
I’ve read about another person having similar problems. The belt was over tightend and made the axis rod flex. He thought that the bed was the problem but it was the bent rods which made the extruder go further down in the middle of the bed.
Drilling the screw holes 0.5mm larger in the pcb can help avoid some flexing, depending on how accurate the lower bed assembly is. I use 2-3mm glass, and it is flat enough. I find 4-5mm to be too heavy.
@Richard_Nordstrom that is what I’m afraid of… I could spent time making the bed perfectlg flagt but what is the point if the the extruder dip down in the middle… Although I find it hard to believe that overtighten GT2 belt is capable of bending hardened steel rod…
I level my Mk2b PCB with glass using a dial gauge in place of the extruder.
I can get the corners to better than 0.05mm overall but you cannot measure the bed centre because the dial gauge spring will deform the glass and affect the reading.
Just adjust the corners and rely on the flatness of the glass and that is sufficient.
I can print using a first layer thickness of 0.1mm using this method.
I used a glass from A4 scanner the heated plate is taped to it with kapton and to keep it in contact I put thermal paste between them and cyno glue on the corners. It’s about the closest I’ve had but it’s still not perfect. I tried with 10mm glass but it’s so insanely heavy the motor had trouble lifting it.
I run a Rigidbot Regular (10 x 10 x 10" build volume) and experimented using various glass thicknesses for my bed. With a 6mm glass I was able to get my bed level to within .01 mm with the 4 corner screws (2 1/2 hours later). Any thickness less than 6mm flexes too much, leaving high and low spots.
The second point is that using 4 corner leveling is an inherently bad design. To level a plane, a 3 point system is just as accurate and works much more quickly. I am currently working on a 3 point level system for my RigidBot.
