I’ve never been able to get a clean first layer with this machine. I have given it a beacon probe and it just made it much faster to get the same problem to happen again. Where am I going wrong?
Your gantry is not in plane. For the rear right corner, you need to perform the following. I suggest you get on the Voron Discord and head to the trident building channel for additional help.
Also what are you doing for beacon probe proximity or contact?
beacon hovers over the bed. it doesnt touch anything.
0.2mm flatness is very flat for a 3D Printer bed, I don’t even know if I achieved that with the Nylock mod on my old Prusa. This simply looks like a bed adhesion problem, clean your bed well with dish soap and water and then rinse well and make sure you don’t toch the surface. Check the first layer height and bed temp as well.
I noticed something while printing. it doesnt have a bed mesh loaded when it prints. it saves the last mesh, but doesnt load it. why?
I think that was a change made a few years ago by the klipper devs. You’ll need to load the bed mesh explicitly in your print start macro before printing
should I have it remesh each time, save, then load the mesh at the start of each print or should I just have it recall a saved mesh?
Thats up to you, if you have a beacon or cartographer, scanning before each print should be pretty fast, though you shouldn’t need to save and load it in that case. The mesh is loaded automatically on finishing the renesh. The change I mentioned from a few years ago is that on restart Klipper doesn’t load a mesh by default, so you’re expected to load it before printing if you don’t remesh.
Try carefully washing the surface with dishwashing soap, rinse and dry before next print. That spot might have something not visible on it that prevents filament from sticking. Grease, dust or residues from some filaments tend to cause that.
I feel you OP. I have a Hypercube Evo that I can’t for the life of me get a leveled bed with. I’ve swapped linear rails with MGN, switched to multiple probes including a beacon that I still need to install but probably won’t help seeing this post.
I think your frame isn’t straight, there’s twist somewhere that combined with an offset from probe to nozzle means it’s impossible to remove with probing as the offset in Z between nozzle and probe changes over X and/or Y. That’s the only thing I can think of at this point.
Where could that twist be?
My guess would be from one side of the x-axis to the other, because then on one side the nozzle to probe offset would be less in Z then on the other. But it could also be anywhere else. Can you check if your frame is square?
I am constantly manually tweaking Z offset at the start of prints so here’s hoping the heatsoak solves that.
You have a lot of possible options here.
- Your bed is ok, mine is 0.6 variation (sad I know)
- dish soap with no additives, rinse and dry it correctly before installing it back
- Verify that it’s doing this in the good order :
- Pre-heat bed
- Home XYZ
- Bed leveling
- Re home Z
- Bed Mesh
- Start print
If everything is ok, as someone else said, it’s maybe a X twist -> https://docs.vorondesign.com/community/troubleshooting/bryansj/twisted_x_extrusion_diagnostics.html
I have reworked the start_print gcode macro to incorporate a heat soak before scanning for a mesh. While its running that 20 minute soak, I notice that the temperature reading on the beacon coil is actual dropping while everything else is increasing during the soak. Is just above the bed, in the center, doors closed, and its readings are getting steadily lower. What could possibly cause that?
EDIT: Here’s where I ended up with my printer.cfg file for my 300 mm^3 voron trident. Do you see anything suspicious? https://pastebin.com/AxxEypK4
Heat soak should not change what you have here. During the heat soak, machine expands a little. You will get a more usefull Z / mesh when the machine gets to equilibrium.
If you still have problems after that, it’s probably a X twisted, see the linked document above. Unfortunately I do not have much time to dig into your configs.
I don’t have any experience with beacon. I use Tap or Klicky. With a twisted X, as tap use the nozzle, it should work even with a twisted X
Do you have a single MGN rail on X or two? It’s easy to end up with twist on dual X rails that can be a bugger to try to straighten out.
Also, are you doing a good heat soak before printing, and do your probe results change at all before and after heat soaks?
If you’ve got a beacon on there, it seems like it would be well worth cranking the probe points way up.
If it’s simply bed adhesion, definitely give the dish soap method a try if you haven’t already.
cranked up the mesh points and switched from lagrange mesh to bicubic. here we are at 35x25 points. weird, I see spikes.
Oh and to answer your question, there is one linear rail on the X axis.
It looked like you have a textured sheet? 0.2 mm variation over the entire built area isn’t huge, might be exaggerating it.
How much of a heat soak? If you’re going to the edge, let that sit for at least an hour, preferably more, look at Ellis’ page on thermal expansion, frame will absolutely expand. I use backers on my 2.4, gantry is giant bimetallic strip, backers do seem to help with that. Klipper does have the ability to correct for this as well, in that link. I do also have a kinematic bed mount (it’s coupled loosely to the frame, basically gives the bed room to expand), which again does seem to help, but I’d personally say heat soak is the first thing to do to achieve consistency.
And to echo others, degrease your bed with dish soap & water (unless your surface can be damaged, Buildtak that’s a no, don’t of that for example). If that’s a textured sheet, may need to give a bit more of a squish, but get it good and clean first. Ellis has some solid 1st layer calibration and troubleshooting guides to go through. For pei, personally I’ve found I needed to rough up the surface a bit with a brass brush, I don’t love pei on my voron, usually use buildtak but have had really good results with the fire resistant version of garolite.
Edit: read up on your probe, eddy current based? Sounds really interesting, my first point is probably moot, though possible you could be picking up the texture or if you have strong magnets it could affect it (my bed has an array of strong round magnets, seen others that are just a magnetic sheet), the do call that out in their FAQ.
I looked at your guide and theres something interesting mentioned for my model, the trident. the actual temp is lower that the measured temp on the top of the bed due to the thickness of the bed! I shoudl raise it by 10 degrees. I’m going to try that in a bit
Yeah, makes sense based on where the sensor is, my heater didn’t have a thermocouple on it so I drilled a hole for a thermistor midplate, it’s super slow to respond is the downside but in theory it should be accurate enough.
I don’t have experience with the bambu, in theory everything will experience thermal expansion but for the voron setup, the bimetallic construction is some of the issue, they have different thermal expansion properties so it can cause deflections. Part quality will vary wildly depending on sourcing too, vorons are very diy and open as the draw, but there’s just so much variability from sourcing, mods, assembly etc.
I notice my open air bambu labs p1p doest have any problems like this and it doesnt need a heat soak. Isnt the voron a better machine with higher quality parts?
its a 25 minute heat soak before each and every print. I wonder if I should break that out of the print start macro and make its own macro.
its a magnetic sheet so that shouldn’t throw anything off. I flipped to the smooth side of the sheet which is much takier but its still having the same issues.
