Auto-Bedleveling

ekah · April 1, 2017, 5:20pm

Hello Cetus-Team

Whats about the possibility to install an Auto-Bedleveling Sensor?

If i am looking through the menu of nozzle calibration there is a (greyed) Auto-Button.

Are we having luck?

system · April 2, 2017, 11:23am

Hello?

Current hardware setup is not designed to be compatible with any auto leveling function. Base on my experience, manual leveling is still better than auto leveling in most desktop 3D printers.

jringloff · July 17, 2018, 2:14pm

I machined a kinematic-contact type bed leveler in 6066-T6 aluminum, and anodized it (well, had it anodized) and delrin and a few printed parts, and a few screws nuts and spring washers.I am finding once I tram (level) the bed using a dial test indicator (made mount that allows me to attach that) that it stays trammed unless I upset the printer in some way. I have two work plates and the heated bed, and mount the plates via registration pins I epoxied to the heated bed, and milled pockets in the work plates to register to the pins. So the plates go on only in one orientation. I haven’t trammed the printer since I last pulled the z-axis rail, I won’t again until after I move it in to an enclosure.I lost ~27mm of z-height (on the extended mk2 so no big deal,) by adding the bed lever and the heated bed, and the non-headed workplate on top of that (which is heated by the heated bed). But the bed is parallel to the X and Y axes, and I get square prints. The balls are delrin, and the ball nuts are also delrin epoxied to m3 screws. The vee channels are slanted 1.5° which provides just enough travel. Missing from this early drawing (sorry no photos, and not pulling the printer apart to take any) there are two screws with spring washers and nuts. One goes through both plates and the corner ball, the other the opposite corner about 10mm further inboard than the ball for asymmetry of compression load, (no holes in this drawing screen shot), to pre-load the system. The dual hole patterns are for the cetus carriage and heat bed, and an alternate heated bed I tried and didn’t like. I made the top and bottom parts identical/symmetric for faster milling setup. The plate/rail-carriage holes are all threaded. The end-adjusters brackets and float bars that hold the m3 screws are 3d printed and tapped after printing, with a undersize through tap for a snug fit that won’t rotate without being intentionally turned. The ball is a bead in shape, with a large enough through hole that it can float while a screw is through it.I also checked the work plates and found one of them them not exactly “flat” compared to the other and it appeared “used” as well, and helped them toward more flat (don’t try that at home unless you have a pile of experience flattening plates and measuring the flatness and thickness of things, I am a machinist, if you aren’t, you are more likely to make it less flat and cause your plate to have uneven thickness as well, consider your plate flat enough, the factory plates are good enough, really, ±0.01mm is plenty good enough given the variation in thickness of the coating, and the resolution of the filament and repeatability of the belts) checking on a surface plate and dychem which I am finding more than “flat enough” since I tend to print the first layer heavy (150% of extruder diameter), and if the first layer is critical, on a raft. Becuase I use the two workplates interchangeably, I didn’t want to have to re-level the bed each time I changed plates which is why I made them flat enough to be interchangeable without re-tramming or adjusting z. This also involved fly cutting the underside of one of the plates by on a slope of just 0.0007in/in across one axis over the length of the plate.I haven’t done a software based bed level compensation routine since installing this, and I use the printer exclusively with G-code.I tram it (wrote a tramming routine in g-code,) adjust it, and run the tramming routine with the indicator in place until its good on the X axis, then I do the same on Y, (I trammed the Z manually using other methods,) both against 6" machinists parallel to average any surface irregularity in the surface coating, and set my z-height using a tool height indicator gage (typically used in milling machines to set tool height in the tool holder) against the extruder nozzle over the center of the print area/over the carriage, and I am done. I don’t touch it again unless I knock something out of adjustment by moving the printer or disassembly (although I occasionally check tramming and tool height), or rough handling (which I don’t do). As the leveler plate is kinematic, it should be and so far has proven to be entirely repeatable within the resolution of my 0.0005" test dial indicator. I don’t expect that to hold that forever with delrin ballend nuts and balls forever, as the round surfaces will develop flat-spots through compression eventually. I will probably switch to aluminum balls and ball-end nuts at some point without anodizing (so they are softer and wear rather than the plates that take more effort to fabricate). This design isn’t particularly easy to install or service, as such I consider it a V1, and will likely improve it (particularly the geometry, it has terrible geometry for good load distribution of the carried mass) in the future, but I also want it light and compact and flatter, so it doesn’t eat too much z-height, or add to much moving mass to the system. The next one I will likely skeletonize to further reduce moving mass as well while retaining reasonable rigidity. More moving mass equals reduced maximum (clean) printing speeds. And I am finding the cetus is quite capable of printing small things really fast given some tuning. My next mod is likely to go move the filament feeder off the print head (bowden style) to reduce moving mass up there. I use slic3r and I my max print speed limit setting is currently 160mm/s which I don’t think it’s hitting because of extrusion pressure limitations (how fast it can extrude).Sometimes a well-crafted (or in this case, a v1 proof of concept,) mechanical solution is better than a boatload of fancy-pants software.Love the printer, hate the software/firmware. I wish it were capable of running like most open source machines with Octoprint front-end pi over USB, and had a larger, richer g-code vocabulary. Oh, and fan speed control, a second fan channel, and g-code heat bed temp sensing and on/off/PID temp control.I bought it because it was a system that used ball rails (and well-above-decent IKO’s at that), and was reasonably sound in design, I knew if I needed to make it stiffer I could, and if I needed to replace the controller I could, but I didn’t want to scratch build a linear rail printer for “my first 3d printer”. I got a better printer than I expected, with a much worse software/firmware/support group than I expected. Still very happy, and this little machine prints circles around MANY $1500 printers I have used or are owned by friends. It just needs better software/firmware, Cetus would be wise to open source the software, and full mechanical drawings, they could compete directly against Makerbot and Lulzbot if they did, because the cetus prints better with fewer hassles out of the box than I have previously witnessed from any 3d filament printer to date.