Domed build platform / bed help...

[Nomasfilms]

Hello all! Long time reader, first time problem poster.I have finished my build of my new Rostock MAX. Everything went REALLY well until I started hot calibration. I have gotten all three of the outside calibration points perfect with the screws, but even though I have added MHackney's nice Aluminum bed plate, my bed is still domed by approximately .20mm. I'm using PolyogonHell's firmware with EEPROM enabled, and have made all the steps per mm adjustment in the EEPROM via Repetier host. I have read on this forum that adjusting (adding or subtracting) to the Offset in the #define DELTA_RADIUS line should make a difference, but I've bumped it up to +.50 with no real visible change. If I could just get this thing level, I could print a test cube. Other than taking my Aluminum bed plate and bending a concave in it, what else can I do? Thanks for any and all input by those who've been there and done that, Chris

[cambo3d]

you have the same problems everyone else has, the onyx has a bow in the middle causing the aluminum plate not to sit flat.

I wonder if seemecnc guys put a slot for the mounting screws instead of a hole in the heat bed. Would allow it to expand and sit flat???????

[mhackney]

Interesting, my aluminum plate is dead flat with the Onyx heated up. The Onyx itself bows awfully when heated. I do have the snowflake sandwiching the Onyx and when I tightened the six screws I used a star pattern to tighten, that might normalize any built in stress. Ovalizing hte holes might work too.

One thing to make sure of though - are you actually measuring the dome with a straight edge or just using the tip of the nozzle? If the later, you may actually be seeing an incorrect DELTA_RADIUS.

@Nomasfilms - you most certainly should see an effect of DELTA_RADIUS. Here is my DELTA_RADIUS directive:

Code: Select all

#define DELTA_RADIUS (PRINTER_RADIUS-END_EFFECTOR_HORIZONTAL_OFFSET-CARRIAGE_HORIZONTAL_OFFSET-1)

.5mm may simply not be enough.

[cambo3d]

with a straight edge after heating to 110 it domes, but cooled down it sits flat.
what is this delta radius. does that compensate for the curve.

[Nomasfilms]

Thanks for posting guys! MHackney, I think something might have happened to my plate in shipping. With TWO different straight edges, I could easily see about .40mm bow in the plate in most directions except directly across the plate starting from the relief you cut for the connections. My guess is the box was under something heavy along the way. I have done a brute strength straightening (using some very technical methods) and now I'm within .18mm or so, and I'm trying to do a calibration cube. Now I have another problem: Hot end gets to about 2 degrees of where you has it to be, but never gets hot enought to run the program and start printing. If I ask for 240, it stops at 237.7. If I ask for 220, I get up to 218.4 or so... That's what up right now...

[foshon]

Thanks for posting guys! MHackney, I think something might have happened to my plate in shipping. With TWO different straight edges, I could easily see about .40mm bow in the plate in most directions except directly across the plate starting from the relief you cut for the connections. My guess is the box was under something heavy along the way. I have done a brute strength straightening (using some very technical methods) and now I'm within .18mm or so, and I'm trying to do a calibration cube. Now I have another problem: Hot end gets to about 2 degrees of where you has it to be, but never gets hot enought to run the program and start printing. If I ask for 240, it stops at 237.7. If I ask for 220, I get up to 218.4 or so... That's what up right now...

You can disable the command to wait for the hotend/bed in your slicer. For a long time I manually removed it from every print I started. Look for "start G-code" in Slic3r or prefix in Kisslicer.

[Nomasfilms]

Thanks for that Foshon! I ended up cheating by preheating manually to above the required temp, then kicking in the START button. Guess what? I printed my first Calibration cube! I'll post a pic soon. MHackney, why do you use any offset if your bed is perfectly flat? Is this just required for a Rostock to work correctly?

[mhackney]

I'm not sure what you mean by offset? Basically, there are two components that must be "flat" and parallel to each other:
1) the print bed itself
2) the motion of the delta head as it moves in the X-Y plane (presumably while printing)

Delta printers have a more complicated motion than Cartesian printers and must be calibrated for planarity and parallelism to the build surface. You can of course, adjust the build surface to be parallel to the head movement, but the software calibration built into the firmware simplifies this to a software solution.

If you are talking about the DELTA_RADIUS I posted above, that is absolutely unique to the way the head on a delta printer like the Rostock moves. If this is incorrect (and it is actually made up of several physical measurements as you see in the equation) the print head moves in an arc - either convex or concave. So, in order to get it planar to the build surface you must remove that arc with the DELTA_RADIUS correction.

It is a bit more conceptually complicated than a Cartesian printer but at the end of the day you need 3 things:

1) a flat (i.e. planar) build surface
2) planar movement of the nozzle in the X-Y plane
3) these two planes to be parallel to each other

--
Note: now that I think about my setup, I do have a borosilicate plate installed. I added it simply because I had it before I made the aluminum plate and rather than put it on the shelf where it will probably get broken, put it on the printer. So, it might be that my Onyx is warping ever so slightly too but the stiff glass is "hiding" that from the nozzle. That said, I did do quite a few prints on the raw aluminum surface before I reinstalled the glass and did not experience a bowing problem.

[rs50pilot]

Mine is bowed with and without the alum plate. I think the bow is in the snoflake or in the spacers. Have not torn it apart yet. Fastening the plate to the onyx makes the plate bow. Then adding the glass to the top makes the glass bow, its a little less but still there. I didn't then the top of the max before after I put it on so it maybe that. Mine is a good .035 out from center to edge.

[Nomasfilms]

So, is the answer here to take MHackney's aluminum plate and "counter" bow it to eliminate the Onyx bow? OK, my head is starting to hurt. I've been able to get my raised center area down to .10 or so... and was finally able to make my first test print tonight: the calibration cube from the instruction manual (THANK YOU GENEB!) I am very happy with the way it turned out. Next, I'm doing the same cube, but sliced in Kisslicer with wipe. I'll post a picture of my first cube when I can figure out how to post pics to this forum...

[rs50pilot]

I think the top of the max table needs to be level as possible then level the onyx. The plate and glass will pull the bow in the onyx down some. I do not think that with what we have any of it will be perfectly flat and parallel. But it will be close enough. The aluminum plate or the glass by themselves are not perfectly flat if you really want to get technical. If you wanted it perfectly flat it would have to be ground flat. I think instead we need to figure out what is flat enough. Is .001 or .002 enough? Is .010 to much. I guess I would rather have that number.

[mhackney]

I think there is an underlying issue with the snowflake/base fit & assembly. If you attach JUST the snowflake without the Onyx to the base, it should be relatively flat as measured with a straightedge. If it is not, you should address this issue first. The snowflake needs to be flat to start with, if it is warped other issues are going to accumulate.

My snowflake mounted on the 6 pillars is about as flat as I could imagine it should be. There is no measurable arc in it as measured with a precision straightedge and feeler gauge. After 100s of hours of use it is still flat. I am not sure what's going on here but I recommend starting with the above suggestion and posting your result on flatness of just the mounted snowflake.

I tested and measured deflection of the aluminum plate on a large hydraulic press. When the perimeter is fixed with the 6 screws (as it is when installed), it takes considerable force to cause the center to deflect .1mm. It is difficult to believe that the Onyx epoxy board could not possibly generate that much force. It certainly doesn't warp in my setup with the aluminum plate on top but it warps badly if just heated alone.

The biggest difference I see with my setup and many others who have no flatness problems is the base and snowflake flatness/alignment. The photo cambo3d posted on how badly warped his base assembly and base top is was a little concerning. If your snowflake is flat, you can and should mount it to maintain that flatness using shims if necessary. You can mount to the base in 3 places, not six. This is simpler and assures planarity without warping. Then mount the Onyx and optional aluminum plate through these same 3 holes to the base. In the other 3 holes, use a screw and washer to affix the Onyx and plate to just the snowflake. Verify flatness when cold. Then heat and verify again. You really do not need to heat to 100°C either. PLA at 60° and ABS at 80-85°C are good targets. Also, as manufactured, the Onyx should be very flat. If it isn't that should be rectified too.

One other comment - when we say "flat" or "perfectly flat", .05mm flatness - meaning no gap greater than .05mm between a precision straight edge and the surface - is very acceptable for this application. Even .1mm is likely not to cause problems printing/bonding of the first layer.

[geneb]

I must have a miracle machine or something - and this actually bothers me. Of all the issues I've seen reported here, the only one I have EVER had was ABS being extruded sideways out of the extruder mechanism and that hasn't happened for a couple of months. I've installed no fixes for it either. The machine is dead square, no convex/concave issues either.

I will say the idea that an FR4 PCB could warp a .125" sheet of aluminum (feels like 6061-T6) is highly unlikely.

[mhackney]

I'm with you Gene! My build went together very quickly and met my quality expectations. My machine is also square and build surface flat, excepting the Onyx's warping at temp - which is a known issue with all PCB heat beds. At the very least, if the snowflake is indeed flat, it can be mounted so that it maintains that flatness. Then, the Onyx and a build surface to keep the Onyx flat can be added. But, start with a warped or wavy snowflake, and the problems begin. If folks are getting warped snowflakes in their kits, they should contact SeeMeCNC and work that out. If the base is not square and flat, it will require shimming to install and maintain the flatness or follow my suggest to use 3 points to mount above.

Hopefully we'll get to the bottom of this soon.

[cambo3d]

did you try heating to 80-110 then test for flex? cold it sits flat hot it bows. I can stick a straight edge on it right now at room temp and it is darn flat.

when using my dial indicator for bed height measurement you can steadily see an increase in the numbers as it heats up and domes.

[rs50pilot]

did you try heating to 80-110 then test for flex? cold it sits flat hot it bows. I can stick a straight edge on it right now at room temp and it is darn flat.

when using my dial indicator for bed height measurement you can steadily see an increase in the numbers as it heats up and domes.

Wondering what you are referring too?
The plate or the bed?

[cambo3d]

both because my onyx is underneath the aluminum plate

[mhackney]

I can say that my build surface does not move as I heat up the bed to 85°C. I have indeed measured it and I print both cold (Nylon) and hot (60° and 80/85°C) and get great first layer adhesion and even thickness.

I have no idea what is going on with these issues, it is very perplexing.

[JakeCShake]

I'm with you Gene! My build went together very quickly and met my quality expectations. My machine is also square and build surface flat, excepting the Onyx's warping at temp - which is a known issue with all PCB heat beds. At the very least, if the snowflake is indeed flat, it can be mounted so that it maintains that flatness. Then, the Onyx and a build surface to keep the Onyx flat can be added. But, start with a warped or wavy snowflake, and the problems begin. If folks are getting warped snowflakes in their kits, they should contact SeeMeCNC and work that out. If the base is not square and flat, it will require shimming to install and maintain the flatness or follow my suggest to use 3 points to mount above.

Hopefully we'll get to the bottom of this soon.

I also agree.....we never checked the flatness but have not had any issues with it. We probably should check it just to stay in this topic. Why do some people need to get to 100C....??? I don't see the neeed but then again...just a noob......

[mhackney]

JakeCShake, there really isn't a need to print ABS at more than 85°C. PLA at 60°C. I am not sure why folks want to go that high unless they are printing other materials or trying to "cure" a first layer adhesion problem. That's better addressed in other ways though in my opinion.

[JakeCShake]

Agree....If Barnett and I had problems, we just used a long "toothpick" to force the first layer down. After that no problems. Thanks for your continued input.....very helpful.

[mhackney]

That indicates an improper Z=0 setting. Getting that 1st layer to stick at first is challenging. Once you've done it and see how it works it seems insignificant.

[JakeCShake]

Yea....we are working on it. Sometimes the first skirt wouldn't immediately hold so we "helped" it stick........not a major concern of ours. We were just happy the thing moved along after our self help......all in all....not our biggest problem. We love seeing this machine in action.....

[jesse]

Hair spray helped to make the first layer stick and added slightly more height.

[Nomasfilms]

Although I've only done one print, the Lexan build plate supplied by SeeMe was amazing. At 75C the ABS stuck like glue, and after cooling, you can remove the Lexan and simply flex slightly. POP and it's off! It's almost too simple. No noticeable residue or damage to the plate that I can see. Buy a few, and you can rotate them...