Using nozzle as Z-Probe with a rocker mechanism (ARTICLE)

So I decided to develop some articles about things I designed for my own 3d Printer while I was building it. I think this analysis is useful for both the content itself and to show a cool way for exposing knowledge of something even in the internet community.

Enjoy, comment and add ideas. I have been using this mechanism for 3 years now and it has huge success on my printer. It is robust, accurate and precise. The article needs some points to develop, but I'm lazy and I receive no retribution for this so I'll take it easy.

[www.thingiverse.com]

Kisses

Edited 1 time(s). Last edit at 11/05/2017 07:35AM by PurpleSensation.

My delta came with a bed-sensing mechanism similar to this. It wasn't worth the trouble, and I've been happily printing with *no* height sensing at all for years.

A key aspect in your design is the axis/hinge; any play in this hinge will be magnified by the length of the rocker arm when bed-sensing. Similarly, any play in the hinge will be magnified by the distance to the nozzle whilst printing. This should *not* be elastic, but very rigid. I don't think any 3D-printed hinge will be good enough. Probably it needs to be machined to tight tolerances, and rigidly attached to both the nozzle and effector.

Another problem is that you need maybe 0.5mm vertical motion to trigger the microswitch. Somehow the nozzle needs to be held in the "printing" position, except when pressed against the bed when sensing. On my printer, this was done by a spring. Inaccuracies in height sensing and bed-levelling (and other delta geometry) and extrusion rates mean that the spring is sometimes compressed a little whilst printing, resulting in height & lateral inaccuracy. Conversely, inaccuracies in the switch mechanism make it hard to accurately level the bed and correct for delta geometry errors. The strength of this spring is limited by the hardness/strength of the build surface. A very strong spring will result in pressing the nozzle into a soft surface (e.g. painters tape) before the switch is activated. If the bed itself is adjustable on springs, then these springs may compress a little before the height-sense spring compresses.

Ideally, the cross arm will be much longer that the height of the hotend, so that very little motion of the nozzle creates a lot of movement at the switch. That will also minimise the horizontal movement of the nozzle whilst switching. Making the cross-arm asymmetric (hinge closer to the nozzle, switch further away) will also improve this.

So I'm thinking that the hotend should be mounted above the effector (also gaining several cm of free build height ), with the hinge close to the nozzle. Use the height of the hot-end as the rocker arm, with the switch actuated by horizontal movement of the top of the hotend.

I am pretty sure I answered all those issues you talk about on the text. I have no problem with the axis because I designed an "elastic" hole so there's no space between the rod and the rocker, the union is perfectly perpendicular with no vibrations. The rocker itself is rigid. I don't use a spring to fix the rocker because, as you say, it is a bad option, I use a screw easy to turn by hand. If you put the hot end so close to the nozzle you'll have no space for cooling, and if you use a screw you'll have no problems with inertia of vibrations. I have been using this design for years and I print with very good quality with a printer that cost to me to build about 150 bucks. Another thing that you have to realize is that not everybody can forget about calibration, mi printer is made with cheap screwed pieces and my place of living has a lot of thermal fluctuations, so the structure moves a lot and I should calibrate the bed after some printings.

I should upload a video showing how it works.

It's a nice paper, have you considered making it an experimental approach? Take the same machine with another type of sensor, and then the same machine again but this time with an aluminium bed and no sensor. Which gets the best prints?

It's a nice idea. But I'm pretty sure that I have already reached the max quality allowed by my nozzle width, nema motors and delta geometry. The theoretical accuracy I reach is the one provided by those parameters. And the surface quality is probably the best that a PLA printer of 0.4 nozzle can reach, all those pieces showed in the paper were printed by this printed itself (with the previous version of them) and every detail is perfectly visible. I should add a mathematical analysis about the error introduced by the non parallelism between bed and delta head. It is shown that that kind of error (which is the greater one) is minimum for this mechanism because the centralised measuring.

Edited 1 time(s). Last edit at 11/05/2017 07:35AM by PurpleSensation.

Quote
PurpleSensation
It's a nice idea. But I'm pretty sure that I have already reached the max quality allowed by my nozzle width, nema motors and delta geometry. The theoretical accuracy I reach is the one provided by those parameters. And the surface quality is probably the best that a PLA printer of 0.4 nozzle can reach, all those pieces showed in the paper were printed by this printed itself (with the previous version of them) and every detail is perfectly visible. I should add a mathematical analysis about the error introduced by the non parallelism between bed and delta head. It is shown that that kind of error (which is the greater one) is minimum for this mechanism because the centralised measuring.

Sorry, I didn't see any examples of print quality in the paper. If you could, it would be interesting to see a close up image of the layers, particularly the edges to see how they line up. Any play in this system should give uneven layer registration and will clearly be seen in a microscope, or a camera with a good macro. Although I do believe that you have good print quality, everything you say is highly subjective unless you can back it up or prove it.

Heres one of mine --

I think that's the clearest image I actually have, I haven't printed anything for several months and everything I print are mechanical pieces currently in use so is not easy to take a photo from them. That's a 0.25 layer width and It represent pretty well well the overall quality. I can't take images closer without a good camera. I know is subjective, but I can say it tastes really soft when touching the surface. There's no effect of the mechanism on the print because the rocker is screwed while printing. You could find little vibrations on sharpy edges, but that comes from the structural elasticity and the magnetic joints.

Edited 2 time(s). Last edit at 11/05/2017 08:26AM by PurpleSensation.

Interesting.
Some time ago I modified my X Geetech Prusa X carriage to allow print head movement upward for two purposes, avoid crashing when coming across PETG blobs and set the Z Home by opening a contact. It works pretty well.
I don't use a microswith, not fit for the job.
I made my own system that requires a minute head movement. I looked at how Renishaw probes work
I don't use springs but magnets.
Note my system as is, is not fit for a Delta but the concept is. So if I go that route ?

I just want to confirm the system works and is far simpler than using any kind of electronic sensor.
One has to be careful the nozzle and bed are "clean". Heat bed AND nozzle before homing. No big deal.

Edited 2 time(s). Last edit at 11/10/2017 03:56AM by MKSA.

---

"A comical prototype doesn't mean a dumb idea is possible" (Thunderf00t)

Quote
PurpleSensation
... The article needs some points to develop, but I'm lazy and I receive no retribution for this so I'll take it easy.

[www.thingiverse.com]

Kisses

Same for me
Besides, to "share" good ideas when in return you only get magnificent "spool holder" and magic "anti wobble system" is a bit unfair. Not to mention the technical level on Thingiverse is very low.

---

"A comical prototype doesn't mean a dumb idea is possible" (Thunderf00t)

Quote
MKSA
I made my own system that requires a minute head movement. I looked at how Renishaw probes work

Interesting, video of your printer? I have seen some Renishaw videos and It would be very very ok to have some kind of soft contact detector by vibrations or whatever.
Maybe u should write the next paper!