Piezoelectric disks for Z contact detect and bed levelling

That's a big improvement.

---

Simon Khoury

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

V2, think V5 might work.....

Needs dropping down a bit as it binds against my titan mount plate, captive nut holes inserting instead of using 4mm rods.

You're certainly on the right lines. The nuts might work okay as long as you tighten them evenly.
Might be worth borrowing Lykle's clamp [www.thingiverse.com] as its much better than the one I did and remixing it.

I am right now remixing it into a 20mm version, no idea if it will work as the top surface of the clamp which presses the piezo will be 12mm diameter, unless I make it quite elaborate, so this covers almost all of the piezo material on a 20mm disc. But worth a try.

Heres a taster:


Edited 1 time(s). Last edit at 03/03/2017 03:55PM by DjDemonD.

---

Simon Khoury

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

Okay so I have a 20mm piezo hotend probe put together, based on Lykles Clamp, but smaller, with 4x M3 screws and no rods, for simplicity.




So definitely getting a signal, these are taps on the heatsink.

I will give it a try in a machine tomorrow.
MkIdraft1.stl

---

Simon Khoury

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

Connected to piezo board, seems to work just fine:

[www.youtube.com]
(sorry about the background noise it's my son's toy truck wheels - and yes I've greased them they still make that noise).

Now to borrow from Phytone's idea above and put a groovemount on it so it can be attached to a titan.
Should only add about 10mm to the hotend's length.

Its very small, very simple and if the LED on the piezo board is anything to go by, it works. Annoyingly I don't have a direct drive machine with a titan to try it out, but I can rig one up.

---

Simon Khoury

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

Just thinking out loud but maybe it can be made into a snap together unit, get rid of some of the hardware. Plus I was thinking with a slight tweak to the flange on top of the clamp, it could be possible to slide the hot end out without disassembling the unit, which we would need if we made it snap together.

---

Simon Khoury

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

Quote
DjDemonD


Now to borrow from Phytone's idea above and put a groovemount on it so it can be attached to a titan.
Should only add about 10mm to the hotend's length.

+1 to that.... I;m tied up insulating the roof before I have chance to tweak. But definately if there is a way to simplify it would be great.
It needs about 15mm under the collar to be about right. The one in my picture was 11mm under the collar, and didnt fit right.

Repository of 20mm draft versions if anyone fancies making one or remixing one.

Titan version added - totally untested, but I took Phytones idea, but instead used a dummy of an e3d v6 hotend to make the groovemount, and made it a two piece print for bonding so the top plate can be printed more precisely, and to perhaps allow a tiny bit of leeway if the fit into the titan is not quite right.

Edit would help if I included the link...

[www.thingiverse.com]

Edited 2 time(s). Last edit at 03/05/2017 09:12AM by DjDemonD.

---

Simon Khoury

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

So Titan version fits well. Anyone with a groove mount direct drive machine want to test it?


Edited 1 time(s). Last edit at 03/06/2017 03:52AM by DjDemonD.

---

Simon Khoury

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

Off topic but related, does anyone know where I can get 10mm M4 female threaded steel balls that are stocked in the UK? Loads on ebay but I don't want to wait for the slow boat from China

Idris

Edit, nevermind, I'll order some from china. What I'm planning involves a complete rebuild of the printer so I'll have to order other stuff too.

Edited 1 time(s). Last edit at 03/05/2017 03:25PM by Moriquendi.

Holy crap - these things are great!

I think my board has an LED problem (email to Idris on this), but it senses fine...

I mixed DJD titan file for a 27mm version and made the titan offset longer (but its not needed with sunk bolts - so will change) and made it an all in one print.

For testing purposes:



This is what I got with DC42's sensor AFTER spraying my aluminium heat bed with black matt paint and probing my Printbite (the sticky tape makes it a bit patchy as in the photo - is this causing the problem?):



A quick tune by guesswork as the LED wasn't flickering:



Looks a world different!

Cool it works then,you've tested the first one of those.

They are a really good sensing system. All positives as far as I can see. Plus the 20mm version can potentially be a full assembled unit. I might look at getting some pcbs made which can be fitted to it with just standard endstop wiring. If we are going to get more people into this we want a drop in system.

Edited 1 time(s). Last edit at 03/06/2017 03:55PM by DjDemonD.

---

Simon Khoury

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

I had to enlarge the collar for the head of the heat-sink, for my chinese clone by a mm - but its a great snug fit. Still needs some nuts adding on the collar to keep it nice and tight.
The piezo, wasn't drilled centrally, its a little off - but works.
Not tried printing yet, but sensing - oh yeh.

Great work guys

Thanks James. Means a lot.

Phytone - just thinking looking at your version make me realise there is no reason why the groovemount ought to be a separate part, I think I had it that way as I wanted to print lots to get the spacing just right.
Will try a one piece one now, see if it fit perfectly. I've made the groovemount collar itself about 0.2mm wider to ensure a snug fit.

Slicing with the new prusa edition slic3r - much better supports than the master version.

Edited 2 time(s). Last edit at 03/07/2017 02:58AM by DjDemonD.

---

Simon Khoury

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

Quote
Phytone
Holy crap - these things are great!

I think my board has an LED problem (email to Idris on this), but it senses fine...

I've replied, I'm happy to replace the board, sorry you've had a problem but very glad it's (mostly) working well for you.

Idris

Quote
Moriquendi

Quote
Phytone
Holy crap - these things are great!

I think my board has an LED problem (email to Idris on this), but it senses fine...

I've replied, I'm happy to replace the board, sorry you've had a problem but very glad it's (mostly) working well for you.

Idris

Confirm the above, stellar support from Idris in this. I've decided to decline the offer of a replacement, as well, the unit works and it's just an LED. Idris also provided a suggestion on using the seperate led ports on the amp board to see if that works, if so, I could rig up a seperate led. Nonetheless,the offer of a replacement stands.

DJD also provided access to his tinkercad stl files for the prior version, as I was having problems in A123D and the stl. Before I got round to tinkering it, the new smaller alpha stl came out. Still had to play with the file, as I ordered 27mm piezo.

Great support here from the team driving this development.

I'll post some more pictures later, and my experience on seeing up as an early adopter.

Duplicate post - deleted...

Edited 1 time(s). Last edit at 03/07/2017 02:27AM by Phytone.

On another forum [forum.seemecnc.com] a forum user wrote that drilling the piezo disks was tricky and that he had damaged several of them. Having had some odd results with a couple of 20mm and 27mm piezo disks I made up a jig to hold the disk in a lathe while I milled the hole with a rotary Proxxon tool and a small carbide end mill. Although I haven't tested any yet, they seem to be free of the deformation on the edge of the drilled holes that may have caused the strange behavior. They also a lot less stressful to make than drilling and even allow a nice 0.5mm relief to be cut on the edge.


The holder was made from a length of 1" nylon and the clamp was the neck and lid of a 500ml Listerine bottle.


Mike

So, every time I run a probe, it shows the bed as being under the z0 axis:



When I go to print, the nozzle drops below Z0 and does a number on my printbite - hit stop quickly.

I've gone to bed center, done the G92 Z0 to define that height as Z=0, probe, print and exactly the same - the nozzle is below Z0.

Suggestions?

Mike that looks good. Its neat, to be fair I'm not that far off with my spur point bit. If I had a pillar drill (might get one) it would be fairly neat, especially if I could clamp the piezo properly (maybe to a metal backing to heatsink it while drilling).

Phytone - you're using RRF lets speak over on the duet forum as whatever the solution is is likely to be of most benefit to duet users.

---

Simon Khoury

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

I've been thinking about effector mounting of piezo sensors as opposed to the under bed sensors that I'm using at the moment, here's what I've come up with.

For a long time now I've been mounting my hot end so that the nozzle is as close in height to the arm mounting as possible. I've been doing this for a couple of reasons:

I hate losing build height, I regularly print things that scrape the top of my Kossel Minis 230mm build height, with an underslung hot end I'd probably lose 50mm or more.

I think that a smaller distance between the arms and the nozzle will minimize effector tilt.

With that in mind, here is the first prototype of my integrated piezo effector.





I've soldered brass M3 nuts to the back of the piezo disks which avoids having to drill holes in them. Unfortunately I can't test this on my printer at the moment, my printer is set up with metal balls on the arms and cups on the carriages and effector, I need to swap them around but I'm waiting on more steel balls.

Idris

Be interesting to see if the 3 Piezo design offers something extra for the slight added complexity. Lykle's effector allows a raised hotend just need 3x carbon tubes and some long m3 bolts. It's well executed though.

I'd like to go raised too more to minimise effect of tilt on the nozzle at the extremities.

Edited 1 time(s). Last edit at 03/08/2017 10:43AM by DjDemonD.

---

Simon Khoury

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

Oh my, that sure is an attractive approach, and solves the problem of needing to drill a hole through the sensor. You'd probably want to mount the electronics for that on the head so that you can run a single digital signal back to the motherboard instead of 3 pairs of analog wires.

Hi Idris,
Neat idea, and I see that you did something similar before:- [forums.reprap.org]
Just a thought, because mechanical noise seems to be my most common problem, if the three discs were mounted at an angle - like on three faces of a tetrahedron, the outputs for X and Y axes would cancel so that mechanical noise would only be what is seen in the Z axis.

Mike

The main problem I see with this design is that it's not a plug and play solution, not something you can add to an existing printer without significant changes. This one has 52mm arm separation, anything larger than that should work but things would get awfully cramped if you tried to reduce the arm spacing.

The piezo disks are mounted at an angle though it's only 12deg to vertical, is a tetrahedral arrangement the best geometrically? I can see that the electrical signals could cancel but is it the stiffest mechanically?

One potential problem might lateral force exerted by the bowden tube and wiring, I won't know about that till I try it in a printer.

Idris

Quote
Moriquendi
.......
One potential problem might lateral force exerted by the bowden tube and wiring, I won't know about that till I try it in a printer.

Idris

My guess is that your nearly flat arrangement will cancel out the pull caused by the bowden tube and wiring. The tetrahedron may even make it worse, particularly as the bowden/wiring may swing for some time after the completion of an X or Y movement.
I will follow this with interest.

Mike

I thought about that but here's my thinking. Piezo disks are very stiff in the radial direction and most compliant in the axial direction (push on the centre, hold the rim). My concern is that with the attachment of the bowden at the top and the piezos on a plane (or their centre points on a plane) a lateral force by the bowden tube is translated into axial force on the piezo elements. Hopefully the signals from this kind of force will cancel out if the disks are connected in parallel but we'll have to see. I suspect that there is an ideal angle for the disks that maximises rigidity against unwanted deflections but maintains sensitivity for z-probing, I'll give some thought to how I might calculate that, otherwise it's good old fashioned trial and error.

Idris

So I made a sensor test rig, shamelessly copying Thomas Sandladerer's rig he used in his recent sensor tests, which unfortunately didn't include a piezo, and had a few factual inaccuracies in them, but thats another discussion.


Mine is a bit less elegant, but it does the job. Its some printed parts to create a motion stage, nema 17, 5mm threaded rod (from my old deceased I3), 8mm smooth rods, an arduino/ramps running marlin rc8bugfix. There is a big piece of elastic from the motor bracket to the carriage to remove the horrible backlash the m5 nut/rod combo has, I forgot how terrible these were (saved in the i3 only by the weight of the z carriages and x assembly resting on them).

I tested the rig in the same way Tom did using direct electrical contact between a steel rod and the "print surface" aluminium sheet, via the z_min_endstop connector and despite this being only attached with a bulldog/binder clip, it was accurate to 0.01mm.
Here's a video of it working [www.youtube.com]

My plan is to attach piezo hot-end sensors, via the various signal conditioning boards Moriquendi is supplying, and ensure that the various iterations of sensor module designs and boards are up to the job. I might build into it the ability to add some varying print surfaces to probe against.

---

Simon Khoury

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

Good job! It'll be excellent to get some definitive numbers on accuracy and repeatability.

Idris