Capacitive Auto Bed Leveling

Why doesn’t Lulzbot use capacitive auto bed leveling like Ultimaker?

Ultimaker’s capacitive auto bed leveling actually taps the bed numerous times with the hotend with great accuracy.

Hmm… a capacitance sensor wouldn’t need to touch the bed. Could be FSR or BLTouch which require touching the bed.

All these work of the idea that the print plane is determined by 3 points (basic Geometry). Additional points are typically a placebo effect… :slight_smile:

Lulzbot’s auto-level is innovative, but flawed by the use of the nozzle tip for conduction. Ooze or filament build-up seems be the drawback.

The capacitive auto leveling tapping the print surface is described here:

https://community.ultimaker.com/topic/15687-inside-the-ultimaker-3-day-6-active-leveling/

(It’s also described a little bit mid page here: https://www.matterhackers.com/news/tech-breakdown-ultimaker-3)

Touching the print surface, if done accurately like the Ultimaker’s, is better than tapping the corners like the TAZ’s.

  1. Glass is not perfectly flat so it’s better to probe its surface.

  2. The aluminum bed mounting plate on the TAZ is probably less flat than the glass and clamping the glass to it probably introduces slight deflections / bows in the glass.

  3. When the bed travels back and forth there may be a slight tilt introduced unless it’s dialed in perfectly.

There’s probably some other variables that can cause less than planar results.

The cumulative effect of these non-planar variables can give less than optimal first layer results, which, depending on what your printing, might make it impossible to get good and consistent first layer results when trying to use the whole print surface.

I’ll have to read that post on the ultimaker forum again… it didn’t describe the capacitance sensors. Here’s a video: https://www.youtube.com/watch?v=x0WelRvsU5I

Seems like it also uses the 3 points… but you’re right it touches the glass bed and not the frame.

Overall not much different from the conductive method utilized by Lulzbot. Author also acknowledges the flaw of this method regarding filament residue… but minimizes the effect of the thin layer of plastic.

Overall, I agree with your flaws about the LB system. Glass is pretty flat… so I’ll have to disagree with you there. :slight_smile: I guess it could potentially be warped under duress, but I’d be more concerned with shattering and breaking.

Anyway, I really like the PB system which probes with an induction sensor. I’m surprised folks haven’t tried the sonar sensors… they are a bit bulky.

I think the Ultimaker Capacitive system is alot different than the conductive method used by LB. If you read the first link it mentions how the capacitive sensor works and the part number of the Capacitive sensor that Ultimaker uses.

The way LB suspends the glass on the Ninjaflex washers still introduces flexion in the glass from the aluminum bed mounting plate if the plate is not as flat as the glass (and its not as flat). I notice on one of my bed corners that I can slip a piece of paper between the washer and the glass with only slight friction, but the other 3 corners are more snug. That introduces another error in the auto bed leveling because the glass is closer to the washers on 3 corners than 1 corner.

There’s alot of noise in the LB System that’s causing me problems getting precise first layers for small parts distributed throughout the build plate.

The new Ultimaker S5 measures the bed at more than 3 points.

If the system is based on capacitance, it wouldn’t need to touch the bed. Which would be a step up from having to touch the bed. I never found the capacitance sensor part that’s built into the toolhead… but did find the reference in the “Electronics” section of the guide.

Couldn’t find a reference in the S5 pages I skimmed through for additional touch points… which would be a waste on a flat glass plate. I did notice that the glass plate sits on metal plate similar to LBs modular bed… an inductive sensor would be more reliable.


Regardless, I think there have been community members who added a capacitance sensor and FSR to perform bed leveling on the TAZ5. It would be a good mod to tackle for the TAZ6… probably just an adjustment of the probe coordinates in firmware. Probing the PEI would be susceptible to bubbles and could be detrimental to the surface if the nozzle is at extrusion temp.

I’d be interested to hear the results if you do pursue. It would probably mitigate a lot of issues with parts stuck to the bed from bad bed leveling.

Not to derail the conversation, but I have absolutely no problems with the TAZ 6 touch probe leveling.

I tuned in my softening temp and then added a slightly cooler pre-wiping temp (which I believe Lulzbot also has in more recent Cura startup scripts). That prevents any oozing during probing.

Since doing that, I have easily done over 300 prints without failing the probe sequence once. And I also haven’t had to replace the felt pad either.

A few other thoughts…glass can be considerably stiffer (higher Young’s Modulus) than aluminum. It depends on the type of glass. And, thickness matters when talking about which one wins the deflection battle when clamped in the corners. It’s a cubic function, so any additional thickness makes a huge difference in stiffness. For a thin rectangular plate:

Flexural Rigidity = ((Young’s Modulus)(Thickness)^3) / (12(1-(Poisson’s ratio)^2))

So, instead of guessing about which not-perfectly-flat plate wins when the corners are clamped, it can all be calculated…that value can then be used for further calculations where the boundary conditions are defined to mimic the corners being clamped, and then the stresses, strains, deflections at points within the plate can also all be determined.

I don’t have any fails in auto bed leveling, my nozzles are always clean and I get good first layers when printing larger objects with the TAZ 6. It’s when printing smaller objects across the build plate where the bottoms of these parts all need to have the same squish so they all have the same nice shiny flat bottoms without any gaps showing between adjacent extrusions and without stringing from one part to the next that I see the variance in the first layer on the TAZ 6.

If you smoosh too much on the first layer you get stringing and if you don’t smoosh enough you get gaps between agent extrusions.

Finding the sweet spot on one part of the TAZ 6 build plate does not produce the same excellent results across the entire build plate with the TAZ 6 in my experience.

Interesting… Does LB provide guidance about the leveling discs? It sounds as though they should be fairly tight against the glass/PEI bed for the proper leveling effect. I’m guessing the actual Z-offset is based on the thickness of the disc where it meets the glass. If you can slip a piece of paper under the disc and washer, this could affect the offset. That disc probably needs to be tightened.

To prevent ooze, adding the wipe to the end script while the nozzle is cooling down might be an idea. Then adjust the temp of the probe for thermal expansion of the nozzle material. Then wait at home for final extrusion heat up.

I think the wipe at the end of print is a great idea. The only time I have had a failed probe it was due to not removing the oozed NGEN filament at the end of the previous print. As long as I remove any ooze at the end of the print it has been perfect.

I wonder if you could take the wipe sequence and post it in Cure LE at the end and have it execute…

That might work as long as you can guarantee that the tool head won’t crash into the finished print…

I also find the way the Taz (and my Mini) do auto leveling problematic. It starts out fine, but eventually works poorly enough you can watch the print bed deform during probing. On the days I’ve printed after a number of oozy-PETG runs, I really have to be on top of how good the level actually is. It’s easily off by a mm sometimes if there’s some caked on junk you didn’t catch beforehand.

I bought a 3" round makeup mirror, an LED flashlight, and a brass brush. With those three tools, keeping the nozzle clean is pretty easy. You do have to watch out for the heater cartridge wires and if they are exposed, you’ll may need a fourth item, some high temp silicone. I just put a small dab where the wires are exposed and now I can brass brush without worrying about shorting out the heater cartridge (alternate plan B is to turn off the power before brushing but if you aren’t quick enough, you have to power up and reheat).

Would there be an issue with the print head colliding with the freshly printed item still on the bed ?
Maybe a remote wiper strip at a higher level on the assembly ?