Bed Levelling Tolerances
Recently encountered some 'issues' with first layers with seemingly poor extrusion / inconsistent extrusion in certain regions on the bed. A levelling map viewed in Optoprint shows a 0.7mm deviation 'high' in the rear right corner vs the front left 'home' position of the bed.
Just how obsessive should I be over this because nothing that I do seems to improve things (including a full on 100% rebuild) and there is zero mechanical adjustment available.
All the tools that I have say the Pinda probe is wrong (a pretty accurate DTI, digital level and a micrometer) - could this be down to poor magnet placement or magnetic variation in the heated bed ? I've also had occasions where point 2 on the mesh bed levelling fails to trigger which is seriously *******g annoying since there's nothing wrong that I can see and the LED on the Pinda says it is triggering, this doesn't happen all the time.
I don't see why mesh bed levelling is needed all the time since if the printer doesn't move then what's going to change? unless you get dirt under the build plate,so I'm tempted to turn it off.
If the Pinda is reporting erroneously then the bed levelling is also compensating erroneously and this would explain a number of issues I'm seeing. It would be so nice if the Pinda values could be dynamically displayed then I could verify with feeler gauges (oh and I use these to set the Pinda exactly 1mm higher than the nozzle tip at the bed centre).
So has anyone come up with a good way to mechanically level the bed - yes I know this is supposed to be automatic but it makes more sense to me as an engineer to rely on this 'automatic' as little as possible and to my mind 0.7mm is far too large a deviation - I almost wish I could level it manually like the old days. I tried to shim the rods at the front and lift the low end, that's assuming that the back of the bed lowers as the front raises in relation to the Extruder axis, this didn't behave as expected and the discrepancy remained so I don't think that the problem is in the rods / bearings but the bed assembly. The machine is printing OK 'mostly' but my OCD can't let this alone.
Any ideas anyone ?
Re: Bed Levelling Tolerances
https://github.com/PrusaOwners/prusaowners/wiki/Bed_Leveling_without_Wave_Springs
https://github.com/PrusaOwners/prusaowners/wiki/Bed_Leveling_with_Wave_Springs
neither tried by me the pinda seems to be ok on mine
Re: Bed Levelling Tolerances
The problem is that I think my bed is level - the Pinda disagree's but I have no way to verify the data it is using - i.e. I lack a method to measure the magnetism at the mesh measuring points at specific heights. It is the matrix data from the Pinda that says the bed has a 0.7 deviation front to back - my other tools (especially the DTI) disagree - but there is no way to tell the Pinda it has got it wrong.
That's an interesting strategy though, I think I'll may reconfigure for that, I had ordered some springs 8mm high and 5mm deep but couldn't figure a way to stop things rocking although they're pretty stiff springs.
I'm tempted to buy a new heatbed and see what numbers I get off that.
Do you have access to your Pinda levelling matrix ? would be interesting. There were a few posts I found from earlier in the year that had a consistent high front right deviation.
Re: Bed Levelling Tolerances
nope adding an octoPi is a future project...
I assembled my bed using a torque driver set at 10 inch pounds so I have confidence that it is done up evenly. I also Used a micrometer on the spacers and found them to be within 0.02mm the same hight all but one which was 0.04 higher than the mean I brought it down to the mean size on a lapping plate in my machine shop....
Re: Bed Levelling Tolerances
My spacers were all over the place 0.1 variance - did similar on a glass plate with 800 grit wet & dry which took forever.
If you enter the gcodes G80 and G81 you can get a matrix up in Pronterface or Simplify3d - for whatever reason Prusa decided to deviate from the G29 used by Marlin .
There is a website where you can post the numbers and it will show you the results - can't remember the URL currently
Re: Bed Levelling Tolerances
I also started with about 700mils of variation across the bed. Try as I might, I couldn't get a reliable first layer. No matter what I did with live Z and calibration correction. I did two things. First, I printed replacement Y rod mounts that surround the entire rod rather than using zip ties. This dropped my variance to about 300mils. I replaced the stock spacers with 5mm spaces with a 2.5mm wave spring underneath. With careful tweaking, I got down to ~15mils of variation. At that point I just seemed to be chasing the high point and I expect I hit the repeatability of PINDA measurements limit.
No calibration corrections needed.
spacers: https://www.mcmaster.com/94669a007
wave sprints: https://www.amazon.com/gp/product/B07B45Y37D/ref=od_aui_detailpages00?ie=UTF8&psc=1
Y rod holder: https://www.thingiverse.com/thing:2800776
Re: Bed Levelling Tolerances
Thanks,
I'll redraw the rod holders as you have because that's a good idea. Where did you get replacement spacers made up ? Did / are you having any issues with the wave washers losing their spring ?
I really wish there were decent parametric drawings available I really can't get motivated to learn scad - could do this in fusion 360 in seconds
Re: Bed Levelling Tolerances
Thanks,
I'll redraw the rod holders as you have because that's a good idea. Where did you get replacement spacers made up ? Did / are you having any issues with the wave washers losing their spring ?
I really wish there were decent parametric drawings available I really can't get motivated to learn scad - could do this in fusion 360 in seconds
Links to where I ordered the spaces and springs from is in my post. As to loosing their spring, I haven't had an issue yet, but I've only had them in place for a couple of days. My printer is only about two weeks old. When I installed them, I did compress all of them fully, to find the high spot on the bed and then loosed the low ones. They sprung back just fine from that, but whether they will have issues over time? That I can't say. I will probably order another set or two since it is pretty fiddly trying to get them installed and I'm sure I'll loose one next time I have to take the bed off. 🙂
Re: Bed Levelling Tolerances
Doh ,,, hadn't spotted the links - ta muchly ... I'll be ordering some today. I believe from other posts that it is heat cycling that causes them to weaken over time so a spare set won't hurt.
Edit : Happen I won't be ordering some today - they're in the US, I'm in the UK and usually shipping from US to UK is brutally expensive (if they will even ship)
Edit : I won't be ordering from those links - won't ship to UK, now searching similar. I've found the spacers on Mouser, may find other stuff there too.
Re: Bed Levelling Tolerances
I found that mesh leveling was actually making my first layers worse. I still don't have a good solution on how to disable mesh leveling, but retain the correct first layer height, as the stored live Z adjust value seems to be ignored when mesh leveling is disabled.
Re: Bed Levelling Tolerances
I wouldn't bother with the wave springs. Get some shim stock, or even use paper to make shims and mechanically level the bed that way.
Re: Bed Levelling Tolerances
I tried that, calculating and making shims to suit potentially 9 locations was a major league chore and didn't add much benefit for the effort it took.
The printer is starting to drive me bonkers, it has printed fine, it's hardly had a day off since January but just lately it is getting to the point where I don't feel able to leave it alone, which is affecting what I print, it is making me not want to start large / long prints because I just end up adding to the plastic waste mountain. I make 3D parts for scale models but the printer hasn't produced a print that I can use since July. Whether it be pillowing, perimeters separating, holes where there shouldn't be, straight lines coming out curved ....
The printer started by messing up layers, not bottom layers but infill of top layers, all the symptoms of too much heat - there wasn't - all the symptoms of too much speed - there wasn't - then the pinda started to fail levelling on trigger point 2 when I watched the LED flicker. I never use anything but Rigid Ink PETG except for the odd PLA but that too is Rigid Ink so I have the printer well dialed in for their material - I have the prints to prove it but just lately it is driving me up the wall.
On the otherhand build plans are in progress for a dual Z 400x400 Hypercube Evolution driven by a Duet 2 Ethernet so I need to get this layer issue sorted so I can print the bits ... the hypercube will then become my large item printer and the MK3 will become my 'detail' printer.
Re: Bed Levelling Tolerances
I wonder if a https://en.wikipedia.org/wiki/Belleville_washer Belleville washer stack might be a better choice for this hack?
they are not prone to fatigue if used within their specification
Re: Bed Levelling Tolerances
May be worth getting some to try I guess it would depend how 'magnetic' they were and what impact they'd have on the Pinda but this applies to pretty much any spring. I may replace the Pinda and the heated bed just to 'eliminate' one or the other from the equation - at least I'd end up with spares if they aren't the culprit, maybe I'll pull the Pinda and test it via an Arduino and measuring station - something similar to what Tom Sanladerer built.
The biggest issue with anything under the bed is simply accessing it, fat fingers really aren't an asset under there.
The odd thing is that when I was playing with the shims the levelling impact wasn't always as expected. If you lift the front left, assuming a pivot in the centre then you'd expect the rear right to lower, and to some extent it does, but once you go past a certain point the pivoting stops.
I'm just printing out some new y rod holders, then I'll start over, tired of putting tie wraps in the trash.
Re: Bed Levelling Tolerances
Belleville washers are made in just about any material you can think of they are a very elegant engineering solution much used in the aerospace industries
Look at my post on handy extra tools for builders a Kelly clamp is a tool that makes placing the spacers a snap
Zip ties used as assembly fasteners does feel sort of McGuvier... as I have a machine shop in my basement I am considering if some parts might be better replaced with 6061-6 parts made on my Bridgeport Mill and Schaublin collet lathe... the lathe is an instrument and tool makers lathe type and I can make parts on it to ± 0.0001 all day if I need to.
Re: Bed Levelling Tolerances
Your basement sounds like my dream cave ...... I just don't have the space unfortunately. I couldn't agree more on the tie wraps the way the linear bearings are mounted in the Y axis carriage has bugged me from day one also, not only the mounting but also the lack of symmetry, it rotates much too readily around the Z axis especially at the extremes of travel. I noticed that the filament and plastic sleeving were causing the rear of the bed to twist and lift so I've not had those in from day 2.
Re: Bed Levelling Tolerances
it must be such a hard biz to succeed in.... as soon as you ship a product the dozens of outfits in China are rushing a clone to market undercutting your price and leaching off of all your R&D. Reading some of the stuff on these forums from the spring i would bet money that some of them are from cloners pretending to be customers....
The Y axis linear bearing mounts are another area where I am considering making some proper Aluminum mounts...
When I bolted up the frame I had the use of my Granite Starrett AAA surface plate to be sure it was all squared up and the use of the Torque hex driver to be sure that everything was equally done up.... I have toyed with the idea of using precision dowel pins as aid if I end up turning this printer into a project but I hope not to.
I have my big Delta Rostock Max Delta printer that has been a fun and sometimes frustrating ongoing project for the last 4 years... I plan to do a total rebuild on it over the winter replacing the Rambo with a Duet WiFi board the steppers with .9 deg per step type oh lots of things and I expect this MK 3 to be a workhorse support machine for making printed parts that I might need.
In that tools thread I mention that very serviceable surface plates (and /or Machine bases though for that option I would laminate several sink cutouts to get more thickness) can be had for some quitting time beer at your local kitchen countertop fabricator... a lot of scrap precision ground stone can be a byproduct of such fabrication.
My first surface plate (used with my first Lathe in the 60s ( a Sears 12 X 36 engine lathe with quick change threading gearbox) that I inherited from a Great Uncle) I actually got from the local undertaker it was a granite "Rock Of Ages" brand headstone that had developed a frost crack and i used it for years... it would have graded about ABB or ABC as a surface plate they are made to a high degree of flatness that most do not suspect... a friend of mine who is a hobby gunsmith is still using it.
If the stone carver has made a mistake in the inscription they may have a few rejected stones out back to be used for practice and training new hires to use the sandblasting system that they carve with these days... the problem with headstones is that they are very very heavy and much thicker than a surface plate needs to be.
if you look at the Video on youtube when Prusa took Tom Sanladerer on a tour of the factory after the Prague makers fair you will see one of his development printers built on a surface plate.... though I accept that the weight and hence shipping penalty would not work on a consumer product.
Re: Bed Levelling Tolerances
a little tech data on the Belleville washer disc springs
https://www.belleflex.com/techtips_6.html
What is a Belleville disc spring?
A Belleville disc spring is a conical-shaped spring with an open center. It is shaped much like a washer and is typically smaller than a coiled spring. Belleville disc springs are ideal for cushioning heavy loads with short motion. The load/deflection ratio can be changed by using more than one Belleville disc spring, stacked either in series or in parallel.
A Belleville disc spring is characterized by four dimensions: outside diameter (De), inside diameter (Di), material thickness (t), and deflection-to-flat (ho).
Belleville disc springs provide the following advantages:
• Conserve space;
• Offer a long service life;
• Can be used in conjunction with coiled springs;
• Offer great versatility when stacked in series or in parallel;
• Increase the reliability of machinery;
• Prevent bolt failures;
• Maintain the positioning accuracy of ball bearings;
• Minimize thermal expansion;
• Are self-damping;
• Distribute loads evenly;
• Absorb shocks.
Our engineers can help you determine the springs needed to meet the requirements of your specific application. Well-designed springs contribute to increased efficiency and prolonged service life for your equipment.