Print Bed Leveling

I’m new to this and don’t have a Taz yet. Planning to order a Taz5 at the end of April. So excuse this rookie input if it’s already been covered.
Since bed must be level relative to the XY plane, why not let the head XY axis print the bed leveling supports?
With no glass print base or corner blocks in place, print four corner “donut” leveling spacers directly on the base frame itself. Print enough layer thickness or number of layers to make the four corners level. Then place glass plate corner supports on those printed donut leveling spacers and secure with studs/bolts/whatever.
See attachment. The corner blocks would need to insulate the printed spacer donuts from any plate heat. Seems maybe a quick and easy enough initial step if it works.

There are socket cap screws on the plate below the silicone heat pad. They attach the linear bearings. As long as everything is square in general your idea could work. Keep in mind you’ll have to square everything up to the bed… which could change as you make adjustments… So it could be a much lengthier process than adjusting the bed to the build imperfections.

Leveling the bed isn’t that difficult. Just need a piece of card stock (business card) at the minimum or metal feeler gauge set. A dial depth gauge would be most quantitatively accurate. Then:

  • Heat up the nozzle and bed, home the head, adjust z-axis using the feeler / card / dial gauge.
  • Then just move the head to the three corners (in theory only two other orthogonal points, the third should line up if bed is flat) and adjust the corner screw to achieve the same nozzle height according to the feeler / dial gauge. Tip here, leave room for the Allen wrench to reach the adjustment screw.
  • Check a few points in the middle, adjust or reevaluate the corners.
  • Print the calibration print. Check for adhesion and consistency of the filament bead (ie. oval shape means a good initial squash)
  • If the print doesn’t adhere well, or adheres too well, redo the above. Alternatively fine tune the z-axis through the g-code offset variable in your slicer software… As long as the bead of filament is consistent.

Once the bed is level, it should be checked on every few weeks or if you move the machine.

Each corner pad has two screws. The first attaches the pad to the aluminum plate. The second screw is offset from the first and has a tension spring under the pad. By adjusting the screw you compress or decompress the spring thereby lowering or raising (respectively) that corner of the bed.

I used a digital gauge and I will tell you that the bed on mine is extremely flat. I tested 17 points (4x4 grid plus center) and they all were ± half a 1/1000th of an inch. That works out to 0.0127mm max deviation. Thats 1/6th the thickness of a piece of paper. I don’t think you really have to worry about surface flatness.

Leveling, well thats a different animal. I don’t know what the tolerance SHOULD be. But I made a jig (well donloaded one from thingiverse and then modified it for my dial) that sits in the head mount so I can run the corners easily. I find it drifts about 4/1000 or so. So I tweak the screws to get it back.

Oh, and do this hot if you plan to print hot. I didn’t measure it but I suspect there could be uneven heating as a possibility which could change the plate levelness. So I run the corners with it hot (110c) to account for any changes due to heat.

Its likely that all this is overkill. Just call me Capt’n Overkill. :wink:

That is the first time I have ever heard a suggestion of doing the leveling process while hot. Interesting!

Well, my thinking is…

Everything changes size to some degree when it changes temperature. Its why they put grooves across bridges on the freeway to allow for expansion on hot days. Its why they put expansion joints in oven chassis so they can expand and contract.

So, if the bed heated absolutely perfectly even, then leveling cold or hot would make no difference. But, if it does not heat unevenly then some of the bed will be larger than other portions. Heating (uneven) could cause the bed to warp slightly.

You print with it hot right? Level it at operating temperature and you account for any uneven heating or warping. Whether it makes any difference at all, I don’t know. I just figure it can’t hurt doing it that way, so I do. I am a noob so what do I know :slight_smile: I just applied logic to it, right or wrong.

I try to level it as accurately as possible. 1/1000th of an inch is 0.0254mm. If I am printing layers at .05mm and my bed is off on one side by a mere 1/1000th of an inch, thats HALF a layer height error. Does it moatter in prints? I don’t know. But I don’t want to waste filament finding out either so I make it as right as I can to start with. Nearly every guide and faq you read all will tell you that the first layer is the most important layer. Logic says they are right. All other layers are built on that layer. If its not right the others are not right either. I took it one step lower. If the bed isn’t right, the the first layer isn’t going to be right either. A sheet of paper is around 0.004" thick just for reference (varies by paper make and bond weight).

We are talking degrees here though. .2mm layers and .025mm error isn’t that much of a problem. At .05mm layers it is.

I’ve only checked my bed level cold, but I may have to check it hot to see it there is a measurable difference. The bed is made of borosilicate glass because it has a VERY low coefficient of expansion. They also make some telescope mirrors out of it for the same reason, so heating or cooling during the viewing session doesn’t change the figure of the mirror. The reason I think it may be useful for us to check, is that the plate the bed is mounted to and the entire frame of the machine is made of aluminum, which has a VERY high coefficient of expansion. So it may be possible that heat from the bed is radiated to the aluminum bed plate or frame causing thermal warpage. Thanks for bringing this up, I’ll have to check. :slight_smile:

Mark -

After my test I can report less than .ooo5" change in Z height at home position. I measure with a feeler gauge so it is a little subjective. I checked Z height at home when cold at .004". I preheated the HIPS profile for a TAZ5, which is a bed temp of 110C, and a nozzle temp of 240C. I then ran a one hour print, cleaned the nozzle of residual filament, and re-homed. Checked again and the .004" feeler still slid under the nozzle easily, but with a very slightly greater drag. So either Z home changed (which is somewhat suspect), or there was residue on the nozzle, or it was due to conducted or radiated thermal expansion. I’m pretty satisfied that there is very little change, at least on a short print like that. Maybe I’ll check again, and do a full bed level check this time on my next 10+ hour print, and report back.

Mark -