Jump to content

How flat is flat? (bed deviations)


Maurici

Recommended Posts

SO... I may already have explained that  on another of my ramblings, but I do have a severe OCD with BED flatness...

I´m guessing it all started thanks to my firs ender, that the bed was an absolute mess (I still have it somewhere), and got as task shimming every single spot till I became pretty good with it.. so were my first layers.

Back then marling would not do mesh leveling, so I took some tools from my workshop and a dial gauge and I did it all the old school way.

BL touch, manual meshes and inductive probes, have made my life much easier since then, however my OCD has not diminished. I find very little discussion in internet about how flat a bed must be...

I allow myself a total range of 1 tenth, this is 0.1mm for a 200x200 bed, and work from there. I know the stratasys machines at work, will have the bed serviced when deviation is greater than 0.1  (and why I took this value).

Currently this is the status of my beds (recently measured).

Switchwire: I3 style knock off PCB DC bed 230x230mm.

-Measured on 36points at 110 degrees  (its max operating temperature for Polycarbonate)

- PEI spring sheet on top.

-0.091mm of range (+0.028/-0.063). Omron Chinese knock off. probe accuracy brings back no deviation up to the 5th decimal number. Klipper.

image.thumb.png.49cf7ac06501b1828d7126df385f466c.png

My big cartesian. Non machined 3mm aluminium sheet 400x400mm with a silicone AC heater.

-Amazon mirror on top.

-Measured on 81 points at 85 degrees (that is the maximum temperature it will ever operate)

-0.15mm of range (+0.075/-0.08) with BL touch. probe accuracy brings back no deviation up to the 5th decimal number. Reprap

image.thumb.png.2c34d165966f39f650829298deff9df1.png

 

My V0.1. 120x120mm fysect machined bed.

-PEI sheet on top

-Measured on 9 points at 100 degrees for ABS printing.

-0.05 max range. -0/+0.05. Manual mesh. Klipper. (is basically concave with an small dip in the middle when hot, fully flat when cold...).

image.thumb.png.730fb7f16d028b244326bb354f9f559e.png

 

 

So... far from trying to measure swords lengths here... how are your beds? what are your acceptable limits and how is your first layer once mesh compensated?

This could build up as a nice database for people to keep as reference, and maybe avoid someone my own OCD trying to achieve the impossible with Chinese materials as is maybe not necessary

Edited by Maurici
Link to comment
Share on other sites

12 minutes ago, Maurici said:

proven that nobody gives a damn flying thing about bed health

Not an expert on this and since doing an adaptive bed mesh, have not looked at the mesh for some time.

Sorry did not mean to ignore you, just don't have enough expertise in this to sensibly comment.

  • Like 1
Link to comment
Share on other sites

13 minutes ago, mvdveer said:

Not an expert on this and since doing an adaptive bed mesh,

Not all will be lost... I never heard of that before and I find it an absolute brilliant idea. saved to my bookmarks for tomorrow.

 

2 years ago I made the great expense of moving to reprap firmware in my (by then) workhorse... as was the one to have. back then was THE BEST thing. Now I´m regretting more and more having done the switch, as literally with a creality STD board with tmcs2209, a dirty soldering to put them in UART and a raspberry b3 (of witch I do happen to have an small army of them along with their screens) you can literally do ANYTHNIG. What should I do with my RRF machine? should I bin it? I guess I could throw Klipper in the duet, but then... what a friking waste it was!

Edited by Maurici
  • Like 1
Link to comment
Share on other sites

Not being religious, I'm still minded of the serenity prayer (Serenity to accept things I can't change, courage to change what I can and wisdom to know the difference). I can empathise with the obsessive part, being in IT I've often chased some nirvana-like state of parallelism or massive storage or resiliency (whatever the goal is) before eventually realising/remembering that good-enough is,well, good enough and I'll never gather (or even afford) the disproportionately vast levels of resouce required to build my own Petabyte storage array with 99.99999% durability and 1M (sustainable) IOPs and multi-gigabit thru-put.

I've watched (and marvelled at) my dad tramming his milling machine still in imperial measures (the UK has been metric for half a century) which is essentially an enormous lump of hundreds of kilos of cast-iron and steel in a reasonably temperature stable environment of a workshop. I then contrast that with the effort of tramming a 3d-printer which in comparison to the milling machine really is a rinky-dinky little hot-glue gun made from paper thin materials, but at the same time the task is way more complex because there's more variables to deal with: pathetically thin work-surface; all those [flexiblle] joints holding the tool-head above the bed; flying grantrys; belts; stepper motors (Dad's mill is manual no CNC here); all those insubstantial aluminium extrusions; electronics.....oh, and we're also adding a whole bunch of heat into the system so it's not even in a particularly controlled environment. The heat (or rather all those thermal gradients we create) will just change all the variables even more (electronics are affected as resistance changes; metal, joints, belts all change and move)....heat-soaking only "helps".....I can testify that my dad, an engineer, curses uses engineering language much more when he's tramming his 3d-printer than a "proper" piece of kit like a mill or lathe (although that may also have something to do with the fact he has hands like bunches of bananas). My dad taught me the value of "good enough".

I think the ranges you're getting on your beds are good, and I wouldn't torture yourself about it. Certainly wouldn't compare to the Stratasys which are industrial machines orders of magntitude more expensive,etc.

I would also recommend the KAMP stuff, I only found that recently (thanks to this brilliant forum!) and think it is fantastic.

  • Like 3
Link to comment
Share on other sites

32 minutes ago, smirk said:

realising/remembering that good-enough is,well, good enough

this is something that, I struggle sometimes, but just in some aspects as I´m normally perfectly happy with functional bodges. However, when it happens (normally racecars, or hobby like things) it gets to a point it can be considered a mild mental condition as I get extremely obsessive and anxious. Thankfully, with the printers my thing is the bed, and not ultimate astetically pleasing prints, as it can be twice as difficult to tune than the bed... I normally settle for structurally sound, as I normally print functional stuff (being said so... functional dirty stuff with a perfect first layer).

Is not that I´m comparing to stratasys. but ultimately I use stratasys PC for my posh parts, so I aim to copy that. The only thing that isn´t achievable at home, is the 145 "oven" temperature.

Other than this... they are pretty run of the mill 3d printers nowadays, and many DIY stuff under 2k surpases them in everything but chamber temperature. So for kinematics and overall structure they are realistic stuff to look to as a result goal.

 

What is KAMP?

Edited by Maurici
Link to comment
Share on other sites

2 hours ago, Maurici said:

it gets to a point it can be considered a mild mental condition as I get extremely obsessive and anxious.

You're definitely not alone 🙂

2 hours ago, Maurici said:

What is KAMP?

Ah! Sorry, being imprecise,  that's the "Klipper Adaptive Meshing and Purging" that @mvdveer had linked it, as I say I think it's an excellent feature.

Link to comment
Share on other sites

That KAMP library is awesome. They did all the hard work and made it super simple for the rest of us. When I found it, I prostelyzed it here a bit. I had a different adaptive mesh that was much more difficult to get set up, and my own too-complex purge line routine.

With running a mesh on just what you need and a reasonably flat bed (whouch yours is), you should be getting great first layers every time.

Link to comment
Share on other sites

The important thing is, how do your prints look? Does the variation even affect them in a negative way? We can obsess over getting the flatness of our beds to 0.001 mm, but does that even mean anything?

I have never really used bed-mesh calibration. I'm not knocking in in any way. I just haven't felt the need to do it.

When I print something,  look at the first layer - does it look right? There are tests you can do to print little squares over areas of the bed then see if you get even bed adhesion on the bottom. In my personal experience, the tolerances to print successfully are quite forgiving. If you are getting elephant's foot in one area of the bed, and first layer barely sticking in another area, then I would say you may have a problem, which then can be corrected by bed mesh compensation. Otherwise, use your printer and have fun making stuff. 

  • Like 1
Link to comment
Share on other sites

I gotta say that having struggled with a crappy Geeetech printer to start with I don't worry much about the Voron bed flatness. If it is close enough that the mesh can "fix" it then it is fine for me.  I have other things to do with my time.  I may want a nice flat bed but I'm not willing to spend the $ or time to get beyond "good enough" 😉

That said, I will check out this KAMP

 

Link to comment
Share on other sites

I have a Fysetc rolled (?) aluminum heat spreader that's far from flat.  I used 121-point probing on my magnetic spring steel sheet to identify the low spots, then shimmed on top of the fridge magnet with layers of kapton tape rolled firmly with a brayer (mine adds about 0.06 mm per layer) until the mesh probing had < 0.06 mm deviation from high to low spots.  Once finished with physical leveling, I changed back to a sane 25-point probing for actual printing.

So far it has continued to give me a flatter actual surface at both PLA and ABS temperatures and does make a significant difference on first-layer adhesion, with mesh leveling in use both before and after, over the unmodified platform.  The flatter surface means not only less work for the mesh leveling but less "wild approximation" of the mesh leveling surface.

Link to comment
Share on other sites

11 hours ago, neufeld said:

I have a Fysetc rolled (?) aluminum heat spreader that's far from flat.  I used 121-point probing on my magnetic spring steel sheet to identify the low spots, then shimmed on top of the fridge magnet with layers of kapton tape rolled firmly with a brayer (mine adds about 0.06 mm per layer) until the mesh probing had < 0.06 mm deviation from high to low spots.  Once finished with physical leveling, I changed back to a sane 25-point probing for actual printing.

So far it has continued to give me a flatter actual surface at both PLA and ABS temperatures and does make a significant difference on first-layer adhesion, with mesh leveling in use both before and after, over the unmodified platform.  The flatter surface means not only less work for the mesh leveling but less "wild approximation" of the mesh leveling surface.

Yeeeeey! I´m not alone! 🙂

I do see a significant improvement in consistency too when the mesh has to work as little as possible.I´ve also found that in the areas of big low or high spots, It will easily approximate it the wrong way.

Having said that, I do have KAMP working now. A bit of fiddling with the voron purge thing to avoid massive blobs, and I find it to be a pretty well though thing.

So, having the bed as flat as possible, and a much more dense cloud of points for the mesh to approximate, makes a significant impact on consistency once your Z-offset is correctly calibrated.

Now I should do a macro to vary the Z offset depending on the bed temperature to be able to work on something else than PC and ABS, but as starting point is a MEGA thing.

However, I do insist... the flatter the starting point and the least the mesh has to correct, the better.

  • Like 2
Link to comment
Share on other sites

7 minutes ago, Maurici said:

Now I should do a macro to vary the Z offset depending on the bed temperature to be able to work on something else than PC and ABS,

I looked into this a little while ago and found it wasn't too difficult. You can add this to Prusa/Super Slicer custom gcode:

MATERIAL=[filament_type]

And then I have this at the top of my _PRINT_START macro:

gcode:
    {% set MATERIAL = params.MATERIAL|default('ABS')|string %}
    {% if MATERIAL == "PLA" %}
        # SET_PRESSURE_ADVANCE ADVANCE=0.035 SMOOTH_TIME=0.040
        SET_GCODE_OFFSET Z=-0.035
    {% elif MATERIAL == "PET" %}
        # SET_PRESSURE_ADVANCE ADVANCE=0.035 SMOOTH_TIME=0.040
        SET_GCODE_OFFSET Z=0.0
    {% elif MATERIAL == "ABS" %}
        # SET_PRESSURE_ADVANCE ADVANCE=0.035 SMOOTH_TIME=0.040
        SET_GCODE_OFFSET Z=0.0
    {% else %}
        # SET_PRESSURE_ADVANCE ADVANCE=0.035 SMOOTH_TIME=0.040
        SET_GCODE_OFFSET Z=0.0
    {% endif %}

The hard part is dialing in good offset numbers for each different filament.

  • Voron FTW! 1
Link to comment
Share on other sites

On 2/12/2023 at 7:52 PM, Maurici said:

proven that nobody gives a damn flying thing about bed health and I´m the only one mental wit this then! XD

Hi @Maurici

I don't think you're wrong in being anal about your bed flatness. I think what might help you get over it is to visually see what 0.091mm really is. In inches it's .0036". Grab a caliper or micrometer and dial in .091mm, hold it up to the light to see the gap and then forget about your bed being off because in reality... that distance isn't enough to even worry about. Especially if you do a bed mesh beit full or adaptive.

Here's a picture of my Ratrig V-Minion, first layer, 90% of the entire bed and guess what? on the X axis... the side to side variation (X0.0 - X180.0) is 0.40mm. And I'm using a crappy pinda probe, not even the super pinda.

228118167_1stLayer.thumb.jpg.6fd31bd1662123be556b118b65394d20.jpg

 

  • Like 3
Link to comment
Share on other sites

4 hours ago, Penatr8tor said:

Hi @Maurici

I don't think you're wrong in being anal about your bed flatness. I think what might help you get over it is to visually see what 0.091mm really is. In inches it's .0036". Grab a caliper or micrometer and dial in .091mm, hold it up to the light to see the gap and then forget about your bed being off because in reality... that distance isn't enough to even worry about. Especially if you do a bed mesh beit full or adaptive.

Here's a picture of my Ratrig V-Minion, first layer, 90% of the entire bed and guess what? on the X axis... the side to side variation (X0.0 - X180.0) is 0.40mm. And I'm using a crappy pinda probe, not even the super pinda.

228118167_1stLayer.thumb.jpg.6fd31bd1662123be556b118b65394d20.jpg

I´çm very aware of what a tenth (well, .9 of a tenth) is... however with my background (high performance engine design and integration) an tenth happens to be A LOT. is also 50% of the first layer thickness,  and about the same ammount you may babystep to fine tune your Z offset. I don´t think the ammount itself is irrelevant, however, i do agree that on this values, with a decent mesh compensation, should be perfectly enough.

 

Nice colour on the RatRig BTW... All my vorons have that same ABS in some of its components!

 

  • Like 1
Link to comment
Share on other sites

LOL. Love the comment "glorified glue gun". That's exactly what a 3D FDM printer is. Even if your first layer is a little off, by the time you get to the 3rd or 4th layer, the FDM process itself has straghtened out any minor discrepancies. We are not making turbine blades for jet engines - we are squishing plastic out of a nozzle as precisely as possible to acheive a desired result in our creation.

Edited by SuperBoppy
  • Like 3
Link to comment
Share on other sites

32 minutes ago, SuperBoppy said:

LOL. Love the comment "glorified glue gun". That's exactly what a 3D FDM printer is. Even if your first layer is a little off, by the time you get to the 3rd or 4th layer, the FDM process itself has straghtened out any minor discrepancies. We are not making turbine blades for jet engines - we are squishing plastic out of a nozzle as precisely as possible to acheive a desired result in our creation.

The worry here isn't really the final precision. If all it takes is a tenth off in every direction coz the bed has a dip, I´m happy with it.

First layer tho, this is the thing. and getting consistency with all materials. when you have a 0.09 of deviation each way, this is a 50% of your first layer squish, the mesh has to work harder and improvise and gess between points really.

The  flatter it is the less you leave to the algorithm chances.

However, I´m seeing that with KAMP all the above isn´t much relevant anymore unless you print something flat as big as your bed.

 

  • Like 1
Link to comment
Share on other sites

1 hour ago, SuperBoppy said:

LOL. Love the comment "glorified glue gun". That's exactly what a 3D FDM printer is. Even if your first layer is a little off, by the time you get to the 3rd or 4th layer, the FDM process itself has straghtened out any minor discrepancies. We are not making turbine blades for jet engines - we are squishing plastic out of a nozzle as precisely as possible to acheive a desired result in our creation.

I'll just leave this here... 😜

  • Like 1
  • Haha 2
Link to comment
Share on other sites

Well. Now that all my printers are running KAMP (even the duet board, yes, I did it!) I have to say that some of my OCD has disapeard...

Now that I´m only probing the printing area, my first layers have been AMAZING without any babysitting.

What a game changer this is!

How long has this been in existence? and why nobody talks about it?

  • Like 1
  • Voron FTW! 1
Link to comment
Share on other sites

6 minutes ago, Maurici said:

nobody talks about it?

Shhhh.. don't tell anyone - TeamFDM secret 😄

Adaptive meshing has been around for some time but was complicated to set up. KAMP has made this so easy. Looking at the Github page, it seems this has been released about 3 months ago

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...