Tired of unsuccessfully calibrating this thing - please help

[miglo]

On another note, adjusting the enstops via the offset in eeprom for some reason no longer works. No matter what values I put in, it's completely ignored. Strange and frustrating. I may do what stonewater suggested and just reflash the firmware.

I have never had to adjust endstop offset in eeprom.

Tom C

Its another way of calibrating without having to adjust the endstop screws. Its pretty accurate actually.

[miglo]

So more hours of fun filled calibration today.

Reflashed my firmware to standard .91 with correct voltages for the kysan motors. Entered extruder PID info in. Removed my "paper shims" at the ZX and ZY zones. Did initial Z adjustment.

I got my digital angle gauge, all towers are 89.9-90degrees with respect to the print bed center. Y tower belt tension was readjusted so all 3 "seemed" pretty close and taut with respect to each other. All print bed screws adjusted so the print bed is 13.00-13.05mm off the melamine box at each location. Used a metal feeler gauge (.127mm thick) to zero all towers and center of bed. Horiz radius is 130.75.

The result:
[img]https://lh3.googleusercontent.com/-RIjL ... -04-18.jpg[/img]

The hotend is practically 1.5-2mm off the bed at ZX and ZY zones and mildly mashes the bed at the XY zone. I'm at a lost for words.

[Dale Eason]

I suppose you have already checked this but just in case... Put a straight edge across the build plate while it is on the heated bed. Check to make sure it is flat. You should be able to see a 1.5mm gap if it exists there. If it is flat then the problem has to be elsewhere.
Are you using the G code short cuts that were in the build manual to move to each tower calibration position? If so make sure the numbers in them are correct. I would especially check the one for the Z tower position.

Dale

[miglo]

I suppose you have already checked this but just in case... Put a straight edge across the build plate while it is on the heated bed. Check to make sure it is flat. You should be able to see a 1.5mm gap if it exists there. If it is flat then the problem has to be elsewhere.
Are you using the G code short cuts that were in the build manual to move to each tower calibration position? If so make sure the numbers in them are correct. I would especially check the one for the Z tower position.

Dale

Macros are straight from the manual. Last time I use a straight edge on the heated bed it was pretty straight. I also actually rotated the bed 180 degrees to see if it's curvature was the problem. It's not. Exact same problem.

In running out of ideas to try. My next step is actually rotating the arms to see if maybe one of tl's is off or something. But i doubt it.

[TFMike]

Are you using gene's 2nd edition manual and not re-checking the end stop adjustments until you have gone through the rigamarole of Z Y X Bed Center when adjusting the horizontal radius?

EDIT: Page 238 -

"When you're done, click the Close button on the Macro Editor. The CONTROLS screen
should now have a macro display that looks like the figure below.
Each one of the listed macros is a button you can click on. DO NOT CLICK ONE YET.

In order to make sure that each axis is higher than the destinations of the four macros you just
created, I want you to position the machine using the terminal window. Enter these commands:

G28
G0 Z5 X0 Y90 F3500

Perform the same check on the other axes by issuing G28 followed by G0 Z5 X77.94 Y-45
F3500 and G0 Z5 X-77.94 Y-45 F3500.

You're not after accuracy at this point, you just want to get the
nozzle from smashing into the build plate.

Once you're confident you can go to the heated bed without striking it, you can begin to
precisely adjust the end stop screws.

Click the “Z Tower Adjustment” macro button. This will send the g-code you entered
previously to the Rostock MAX. Make sure you've got your sheet of paper under where the hot end
will “land”. Having these in macro form makes the repeating task of setting the end stops much easier.
You adjust the height of each axis by turning the end stop adjustment screw to the right to raise
the platform and to the left to lower the platform.

Each time you make an adjustment, click the “Z
Tower Adjustment” macro button.

Repeat this process until you're getting the same amount of “grab”
on the paper as you did when setting the initial Z height.

When you're satisfied, move on to the “Y
Tower Adjustment” and “X Tower Adjustment” macros.

Lay your sheet of notebook paper on the center of the build platform and click on the “Bed
Center” macro button.

The nozzle tip is going to end up in one of three positions. It's going to be above your paper by
a visible amount, it's going to pin the paper firmly to the bed, or if you're incredibly lucky, it will be
“gripping” the paper the same amount as the tower base calibration steps. If it IS, I strongly
recommend you go buy a lottery ticket. Your luck is just that good. "

Page 239

"If you're merely mortal like the rest of us poor suckers, you're going to have to make an
additional adjustment. Delta configuration printers like the Rostock MAX v2 have a very interesting
geometry that will result in the hot end traveling in a non-flat path if it's not perfectly calibrated. This
tiny error will express itself as a “virtual” convexity or concavity in what it thinks the bed shape is.

If your hot end is pinning the paper to the build surface, the error is expressing itself as a concavity – the
firmware thinks that it is moving flat, but the path of the hot end is actually concave and that's why it
pins the paper to the build surface – the center is actually lower than it should be. The reverse is also
true – if the hot end is not touching the paper at all, it thinks that the bed is dome shaped (convex).

The concave/convex shape of the bed is controlled by the EEPROM table entry labeled
“Horizontal radius [mm]”.

What you're going to do is change that figure by 0.5 until the nozzle is touching the paper just
the same as it was when you calibrated at the base of each tower.

In order to lower the nozzle, you'll need to increase the Horizontal Radius value.
In order to raise the nozzle, you'll need to decrease the Horizontal Radius value.

Each time you change the Horizontal Radius, you must re-calibrate the base of each tower as
you did in the previous steps using all four macros in the Z, Y, X, Bed Center order.. It may take a
number of iterations to get the center nozzle height nailed down, but it IS worth the hassle. Your first
layer quality and plastic adhesion require that the nozzle track across the entire bed as perfectly flat as
it can. Please make sure you click the Save To EEPROM button each time you make a change,
otherwise the new Horizontal Radius value will not take effect!"

[miglo]

@TFmike, yup, followed the exact thing to the letter. I've calibrated this from 127-131 over so many hours I've lost count over the span of the past 2 weeks since putting this thing together. I've also use the method of setting endstop offsets too, instead of adjusting the endstop screws. This method is actually more accurate, but the end result is still the same.

At Horiz 130.25 (currently), all towers are approximately .157mm (using a feeler gauge and not a piece of paper) from the print bed. Center is just a tad over that, but good enough.

But the same problem happens. Heading close to and past the virtual line of the X-Y tower and the head will smash the bed. Head close to or past the virtual ZX and ZY lines, the head is lifting almost 2mm off the bed. Moving the head close to the Z,X,Y tower will have it close to the calibration (.157mm off the bed).

It looks like this (pic not to scale obviously, but the info is the same, getting near those virtual tower lines results in badness):
[img]https://lh3.googleusercontent.com/-5QiK ... obueno.jpg[/img]

[craftymethod]

I had a similiar problem, I could only ever calibrate to one small area in the middle.

I thought it was my bed which I got a steel ruler and found it curved a little, so i put some brass washers to hold the bed up a little, recalibrated but for some damn reason my test prints would swing and push into the glass even back doing my smaller prints.

Had me really confused, so I took apart the arms and cheched for a little grab where they meet the carriages (there was a little) Oiled things up and re adjusted my belts and then while I was watching the machine home over and over (good way to check for a machine seemingly de calibrating itself)

And I noticed a wire for one of my fans was catching on one of the Z arms at the back.... so slighty and I think this was my issue.

Bottom line, check for and weird snagging and catching when the machine is coming down from its home command.

Maybe even limit switch bounce/recoil.

Best of luck!

[miglo]

@craftymethod - Ya. I disassembled all arms and hotend again. Rewired some stuff to make sure theres no snags. Theres not much else to do at the moment. I'm just out of ideas. I may have to just low-tech shim the two bad areas with 2-3 sheets of paper again just to get things going.

[Mac The Knife]

Nobody has brought it up, or maybe I missed it, but have you tried adjusting your tower rotation settings? by default, the z tower is at 90 degrees, the x tower is 210 degreees, and the y tower is at 330 degrees. Leaving the z tower setting alone, change the x tower from 210, to 211, and the y tower from 330, to 329. After that change, check your tower calibrations, z height, and radius.

[KAS]

Back to your digital angle finder. How did to zero and measure each axis?


Also, check the hot plate to see if you torqued the bed down to tight or not equally at all 6 mounting points. Use the depth gauge of the digital caliper and measure from the melamine to the flat spot of each screw to set the height equal all the way around. You'll have to check calibration afterwards.

[TFMike]

Can we get a pic of the arms and your bowden and wiring setup ?

EDIT: Also maybe you should refrain from playing with the endstops in eeprom and do it all manually until you get better dialed in.

[stonewater]

I think your tower arms are of unequal length looking from the top of your diagram-

z tower they are equal
x tower the right arm is longer
y tower the left arm is longer

either that or the upper arm brackets are not parallel with the print surface on the trolley, please check with a level.

Tom C

[miglo]

Well changing the tower angles ala mac's advice significantly improved things. Didn't get a chance to run it through all the calibration tests but initial tests shows a big improvement. It still has areas where the head rises a little more than others but it's more of a 0.5 rise than a bit 1.5+ rise, and in shorter lengths too. Its also doing this at different areas, but it's more manageable now. Maybe with more tweaking I can get it even better.

Im away to conference for a week but will test this further and read up on it to see why this is so.

[Earthbound]

My suggestion would be to take the entire assembly of hot end, effector, and arms and rotate it 120 degrees. Disconnect the arms at the cheapskates, and reconnect the whole thing rotated 120 degrees. See if the pattern rotates with it. If so, then this is an arm or effector issue. If it doesn't change, then it is a bed, tower, or firmware issue.

[craftymethod]

Those black arms, did you notice they seem to be different on one side than the other? Look for one fork having a weird bump different to the other side.

I did some things not sure if it mattered but your OCD will be ok after that if you have any that are odd side up.

[miglo]

@earthbound - I've done everything u said aside from disassembling cheapskate. I may do that eventually if there's no other improvement.

@craftymethod - I'm using tricklaser's carbon fiber arms

[stonewater]

my trick lazer arms were exactly the same, but considering the way they are assembled an error is always possible. I do doubt it, but at this point we are moving thru possible issues one at a time. my instinct is this is mechanical. a difference of a .5 millimeters could cause the problems you are having.

Tom C

[KAS]

It's mechanical for sure, either the extended extrusion towers, the fly-n-strude, trick arms, or a combo of everything. But I'm banking on the left/right lean of each tower. How did you use the digital angle finder to correct that?

I'd start over because you'll never get it figured out with all these micro changes in the software, along with mechanical adjustments. If just the arms caused the issue then his print would change to match the arm/effector rotation Earthbound mentioned above.

[miglo]

It's mechanical for sure, either the extended extrusion towers, the fly-n-strude, trick arms, or a combo of everything. But I'm banking on the left/right lean of each tower. How did you use the digital angle finder to correct that?

I'd start over because you'll never get it figured out with all these micro changes in the software, along with mechanical adjustments. If just the arms caused the issue then his print would change to match the arm/effector rotation Earthbound mentioned above.

I took out the fly-n-strude and it still demonstrates the same issue. As for the angles of the towers, I zeroed it by having the gauge in the center but oriented parallel to the tower I'm after. I also tried it where I zeroed it at the edge of the bed near that tower in the same orientation. The only thing I'm not sure about is the over all lean of the towers and machine. But basing it on the bed, all towers are 89.9-90 degrees.

I also used a depth gauge at each screw location in the bed and made sure it's 13mm +/-. 05mm at each location.

The only thing that has significantly helped me so far is changing the angles in the eeprom. I'm going to try tweaking this more after my medical conference in a week. Worst case I'll disassemble the skates and towers if I can't get a decent 200mm print area.

[Mac The Knife]

Anybody care to chime in about how changes in humidity may effect the calibration of our printers? It's the main reason I painted the edges of the melamine.

[KAS]

In regards to having aftermarket extruded towers, are they 1" or 25mm?