Bed Stability: 3-point leveling should be in fact 2-point leveling and a fixed one

Quote
MechaBits
Lol...Just put X arm on a hinge...then you might only need to level bed in one direction.

I don't get it... what X arm?
And really, I don't know if 10mm rods will flex before heatbrake gives away ...

---

RepRap Lander concept on Concept Forge
RepRap Lander concept on RepRap Forums
My Things, mostly experimental stuff

They might bend if you are fortunate enough to have your head crash in the centre of their travel, but near the supports at the ends of the travel - probably not a great deal at all.

Sorry to be a little puzzled at this topic but head crashes are occasional unfortunate events, which generally don't cause much harm as they usually only happen when calibrating/modifying the machine and something is not set right, often during these events you are watching closely and can pull the plug before anything snaps. Electronics failures happen too, but with NC sensors, a failed sensor (or broken wire/connector) results in no z movement, and in my experience this is the most likely and common electronic failure leading to a head crash. Whilst a system that can prevent any head crashes would be nice if it were simple and reliable aren't we trying to force a solution where there is little or no problem to begin with?

I have a belt driven z axis so if I get my motor torque just right, then the z motor skips on a head crash instead of the heatbreak (or any other part of the print head) breaking/bending. Less easy to do with leadscrews/ballscrews I agree due to their gearing but still possible.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

Quote
DjDemonD
They might bend if you are fortunate enough to have your head crash in the centre of their travel, but near the supports at the ends of the travel - probably not a great deal at all.

Sorry to be a little puzzled at this topic but head crashes are occasional unfortunate events, which generally don't cause much harm as they usually only happen when calibrating/modifying the machine and something is not set right, often during these events you are watching closely and can pull the plug before anything snaps. Electronics failures happen too, but with NC sensors, a failed sensor (or broken wire/connector) results in no z movement, and in my experience this is the most likely and common electronic failure leading to a head crash. Whilst a system that can prevent any head crashes would be nice if it were simple and reliable aren't we trying to force a solution where there is little or no problem to begin with?

I have a belt driven z axis so if I get my motor torque just right, then the z motor skips on a head crash instead of the heatbreak (or any other part of the print head) breaking/bending. Less easy to do with leadscrews/ballscrews I agree due to their gearing but still possible.

Yeah, this topic started as something else and it drifted off. Now, while I was working on some design issue, I remembered about this discussion and posted. I am not too concerned about a head crash into the bed...
I knwo that the nozzle must be in the middle to have the maximum amount of flex and because I imagine that such a crash could only happen during homing, positioning the head in the middle of the XRods is just a software configuration, so easily achievable.

This discussion might be quite trivial but getting rid of the flimsy spring-loaded bolts that secure most beds nowadays is quite an achievement if enough failsafes can be managed to make a head crash non-damaging.

---

RepRap Lander concept on Concept Forge
RepRap Lander concept on RepRap Forums
My Things, mostly experimental stuff

Have you seen this method, which I am using on my corexy at the moment:


I ordered a 300x200mm aluminium bed from Leap3D in the states and it came with what appears to be just thick silicone tubing instead of springs. Very pleased with it, gets rid of all that lateral movement and wobbliness of springs and allows for a reasonable range of adjustment, and its simple and cheap. Won't really give very much if you have a head crash but then springs wont help that much either they'll just compress until they are compressed then if something is going to snap/bend it will do so anyway.

Edited 1 time(s). Last edit at 06/12/2016 05:19PM by DjDemonD.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

X rods X arm, potato potaatto, flexing your rods as a feature just sounds like your going off in the wrong direction, anyway I thought you'd skipped the 10mm rods, which is a good idea, unless your tuning the weight of things, and seeing what you can reduce, I am yet to notice any flex in the 8mm rods I have, can swap them out for 12mm easily, but if I tear it down I might as well add the linear rail...

I'd second that, my 8mm rods are not causing the "hill in the middle of the bed" syndrome that I've seen on i3's, even with a vertical stacked arrangement. My x bed dimension is 300mm. And I'd also agree that if they were considered too bendy, linear rails would be the logical next step.

Edited 1 time(s). Last edit at 06/12/2016 06:19PM by DjDemonD.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

A mechanism for the hotend to move would be a better solution, would be cool if the auto leveling could work in conjunction with hotend rise n fall, but more weight when z already can do the trick.

Nah, linear rail is still too expensive for justifying the price. My latest design is using 10mm with SF-1 or sintered bronze bushings (will probably go for SF-1 for now) some 60-70mm apart for good stability, so the bearings' weight is saved here (they're incredibly heavy compared to the LM8UUs).

The hotend to move is to make a Renishaw-probe arrangement where the hotend itself is the probe.

Edit: oh and i've seen a lot of movement in G+ towards silicone instead of the springs... thought of it as a good idea.

Edited 5 time(s). Last edit at 06/13/2016 04:03AM by realthor.

---

RepRap Lander concept on Concept Forge
RepRap Lander concept on RepRap Forums
My Things, mostly experimental stuff

Hi, I am new to 3d printing, I bought a Monoprice i3 (wanhao i3 clone) about a week ago. I have done some milling and leveled various milling and other manufacturing machines in the past, so the concepts of xyz accuracy for manufacturing are not unfamiliar to me. I have noticed it was hard to level the 4 point system on the i3 variant I have. I have looked at google images for bed leveling and read a few posts and have a few questions, being a total noob and all to additive manufacturing.

Is 4 point leveling possibly a 'borrowed practice' from leveling machines with solid cast iron frames that need 4 or more contacts to support the weight without a fairly massive 3 point hinge system? It seems likely to me that someone looked at how a mill or lathe was leveled and figured 'lets just do it like that and move on to the next problem' which is understandable, but not the best design for the problem.

So, wanting a 3 point leveling system I have started to think about it. Before printing something fancier though, I modified my off the shelf i3 with a quick 3 point leveling hack I wanted to ask about, ask some experienced people if I am going to run into some unforeseen problem.

I simply removed a spring and the adjustment thumb screw below the carriage on the back right corner. It creates an axis between the front right and back left springs/screws, I adjust the roll for the plane of the heat bed and glass print surface balanced on this axis with the front left spring and thumb screw.


How I level with this setup:
level the extruder first to the back left with both the front corners lower. Move the extruder to the front right avoiding the rear right of the heat beds possibly higher triangle, and get it close but not exactly dialed in. Then move the extruder to the front left and dial it to the back left so I can move the extruder (actually moving the Y carriage) front to back leveled, then with extruder again at the front of the table I move it right and dial the right side in. The back right corner is free hanging and upon checking it is level and the whole bed is possibly perfectly level

It is with a glass bed upgrade printing to .02 corner to corner on some pillow blocks made as a 4 part set, so spanning at least 100mm of the center of the heat bed. I was planning on making a 3 point mount, but barring some unforeseen issue this 2 diagonal 'main' level points and a 3rd adjustment of roll at a corner seems pretty good at providing a level plane without a 4th point pulling it out of alignment into a concave or convex surface.

That is unless there is some unforeseen problem with setting up my heat bed this way? Anyone?

So the next question/problem is:
With the new glass bed, as well as a new thicker 5mm Y carriage, the Y Carriage/heat bed/print surface mass is wobbling and prints got slightly worse at higher print speeds. Ripples on areas after path direction changes, especially in the Y to X axis as the Heat Bed and glass plate wobble front to back. A tighter Y belt will be done. I was thinking about isolating the issue to just the belt which would at least solve some of the problem, and the springs are an obvious source of lateral instability so I was going to eliminate that first. Since lead screws can solve the belt backlash issues completely, but would be sadly ineffective with this springy wobbly mess on top still.

So has anyone made a vertical brace for the edges of a heat bed that allows it to travel vertically but not horizontally in relation to the Y carriage? I am printing tests of that, but are there any known issues there? Like 'oh man everyone did that 3 years ago but it causes X horrible problem!' type info is sorta what I am after here, as well as any and all good advice.

I am preliminarily thinking about bracing on the corners I have conected as a pivot axis (front right back left). Bracing the front, sides, and back of the heat bed with metal plates that allow the heatbed to slide vertical if the extruder head crashes into the bed, and allowing the rolling of the entire build plane across that diagonal axis for leveling, I hope to eliminate the wobbling table as much as possible while retaining the springs function of allowing the z home to be on a simple switch and bottom out slightly without damaging the extruder.

Hope that all makes some sense, I am just learning the dialogue and terminology of 3d printing.

Thanks

Re. brace for the heat bed. I use the method shown in the picture below: Smooth stainless steel pillars connected to the heatbed slide in snug fitting 3mm thick PTFE bushes.

This is fairly effective although I am still looking for something better.

Mike

@Milton - Your 3 point levelling system is absolutely fine, most people use 3 point levelling if at all possible for ease/speed of levelling and to prevent bending the bed. Many better heat beds have two holes at the back corners and one in the middle at the front, but having 3 corners fixed works too. As Realthor points out above it really only needs to be 2 point levelling, one of the three points could be a non-adjustable fastening, if you can offset your z axis this is even easier. I have had a lot of success using thick silicone rubber tubing rather than springs, the bed is far more stable and it can be compressed a fair way allowing a reasonable range of movement. See image above.

As for bracing the bed, have a look online at bed mounts for delta printers, which you might be able to adapt. These often need to support a round bed, albeit a fixed immobile one, allowing some vertical movement but no lateral movement.

Here are some examples: [www.thingiverse.com]

The issue really is that the i3 and its derivatives suffer from a small loss in print quality because the y axis and therefore the bed moves. Deltas achieve a fixed bed, at the cost of considerably harder calibration and precision required when building, corexy/hbot type machines manage to only move the bed in z direction which is preferable as these are very slight movements.

Edited 1 time(s). Last edit at 07/04/2016 06:07AM by DjDemonD.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

You can use a switch to detect crash, as long as it is not the same as the Z endstop. Allowing the head to tilt back and break a contact (open a switch) has been done before for bed leveling, there is a commercial printer that does that, if only I could remember which.

An option is to use a head tilt switch to cut the main power supply, this way you don't rely on the already misbehaving firmware to save your printer.

Obviously for safety reasons the tilt switch should drive a relay, not mains directly.

P.s. Loved the piezo auto leveling. Where can I find more details on the hardware/firmware?

Quote
Milton
...
So has anyone made a vertical brace for the edges of a heat bed that allows it to travel vertically but not horizontally in relation to the Y carriage? I am printing tests of that, but are there any known issues there? Like 'oh man everyone did that 3 years ago but it causes X horrible problem!' type info is sorta what I am after here, as well as any and all good advice.
...

Instead of trying to brace the corners laterally to prevent bed plate wobble, try bracing the leveling screws so they don't allow the bed plate to wobble.

I use three teflon blocks with holes for the leveling screws, firmly screwed to the undercarriage plate. The leveling screws go into the blocks at least 12 mm and are held very firmly so they can wobble. The leveling screws have flat heads and the bed plate, 1/4" cast aluminum, holes are countersunk so the screws sit below the surface of the plate.

In my printer the alignment screws at at the center of the bed plate in the the X direction, controlling pitch in the Y axis, and at the left side of the bed, controlling roll about the Y axis. You can see how the bed leveling is done by clicking the link in my sig, below. I recently converted a Taz 3 printer to a cast tooling plate bed with 3 point leveling. You can see details of that here: [www.instructables.com]

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Quote
lhartmann
...
P.s. Loved the piezo auto leveling. Where can I find more details on the hardware/firmware?

Thingiverse at [www.thingiverse.com]
Discussion on [forums.reprap.org]

I use standard Repetier firmware although it would be improved by having some code that introduced a short delay before the Z down movement.

Mike

THANKS! I appreciate the advice and ideas to try out. Tubing! Going to the shop to get some now! Ah crap its a holiday, going to the shop tomorrow! Happy 4th!

Oh wait... the old water cooling gear... I have tubing!