Piezoelectric disks for Z contact detect and bed levelling

Quote
DjDemonD
Yes I have been commenting about polarity a lot as it seems to now make sense based on what you've said to be quite random. I presumed it was because I was using the inverted logic instead, one way around they seem to trigger but the voltage rise (maybe Mike you've seen this on your scope?) is not fast enough to trigger quickly enough and you get a crash, the other way around they are amazingly precise.

So something to be wary of with our experimental beta units, if its performing poorly, reverse the piezo connector. For pre-built units I will make sure they are the right way round, by testing them on the rig, before they go out of the door.

That is almost something I have seen before, but not quite. The first batch of no-name 27mm piezos were all checked and recorded and they were all within a 10% spread and all had the same polarity. A batch of Murata 27mm piezos were purchased but they were without leads so had to have leads soldered on. I again classified all of these although I lost some to hamfisted soldering; they all had the opposite polarity to the no-name ones and a much worse spread of peak voltages - within a 25% spread or so. A second batch of no-name discs was purchased from a new source on ebay and some of them were checked and were identical to the first no-name batch. No results were recorded on this batch as I had gone into "same old, same old" mode. One 27mm disc was drilled and installed in one of my rigs so that it could be checked as if it were in Simon's hotend. Output was good in that mode, however when it was installed as if it was an undrilled one, under a pressure pad, it gave a worrying double hump in th voltage output. I think this was because the drilling had given a local buckle to the brass.
A single drilled 20mm disc was checked along with the 27mm disc and gave a useful output. 10 further 20mm discs were modified by having a hole milled in them, this was done to eliminate the possible buckling from the force of drilling. On testing this last batch, a much wider spread of peak outputs was obtained although voltage rise for the first few ms appears to be in the Goldilocks range (livable)
I have seen one mention before of a reversed piezo, I think it may have been on a maze runner robot forum.
I am not sure how the PZT is polarized in these discs as it is normally done by putting a high voltage across the ceramic, possibly while it cools through the Curie temperature. Since thse discs are radially polled, I imagine that the ceramic is in naked form while being polarized and is then mounted to the disc and the top contact plated on.
I have a further batch of 20mm disc and will record them before and after drilling - this may take a while though as I have managed to sprain my wrist*

Mike

* Nothing to do with printers, I was turning over a couple of tons of compost at the allotment.

Is there an easy, no oscilloscope needed method of determining the correct polarity before installation as it doesn't seem obvious just watching the led on the piezo board?

---

Simon Khoury

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

Here's what I was referring to by a "vertical" mount:

[imgur.com]

The piezo in the photo is 27mm, so there's plenty of room in there.

g.

Okay so the idea is one part of the disc is attached to the hotend the other to the mount?

---

Simon Khoury

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

Challenge accepted. Stay tuned for my version of an effector. Started it today.

Quote
DjDemonD
Okay so the idea is one part of the disc is attached to the hotend the other to the mount?

Not really. The idea is to suspend the piezo between the two spacers that hold the hot end assembly to the effector platform. If the piezo by itself isn't sensitive enough to record the hit, a #6-32 nut could be glued to the ceramic face in order to add mass to that side of the device.

g.

So you're hoping to get a trigger just with the piezo stuck on the side of it? I'd be dubious if it's that sensitive but I'd be pleased if it is. Keen to see what happens.

Clearlynotstefan looking forward to it.

Edited 1 time(s). Last edit at 03/17/2017 12:07PM by DjDemonD.

---

Simon Khoury

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

Yes, that's the idea. Based on a tap test I did on Wednesday, it should work. I just need to get one of the boards soldered up so I can test it "for real".

I won't be offended if someone beats me to it.

g.

Much thinking just done, see edit below:

Quote
DjDemonD
Is there an easy, no oscilloscope needed method of determining the correct polarity before installation as it doesn't seem obvious just watching the led on the piezo board?

A simple way is to make or print a little plastic ring that will support the edge of the piezo and connect a digital multimeter (generally an analogue one won't do) Red lead to red wire and black to black. Push in the middle with the pen and then release suddenly. Pushing should give a signal one way and releasing the other. If the multimeter is a good one you should even get an idea of how much of a signal but sampling interval may make this difficult


On this latest batch I did get two which were reversed but his is the first time I have classified by radial sensitivity and it is possible for a disk to be positive on radial but negative on axial. Needs thinking about.

Edit:Ahhh!! I believe I understand. On the rig probe goes down and releases pressure on disc; half a second later probe goes up and pressure is re-established. Poor potty cnc training means that going up is G0 Z1 and going down is G1 F60 Z-0.2, G0 and G1 are at different speeds but acceleration is the same. Oscilloscope is only triggered on the positive going pulse whether it is going down or going up, speed once it gets in its stride depends on down or up but initial speed is always the same because of acceleration so initial voltage good but peak may suck big time.

It seems likely that, for whatever reason, the 27mm discs are all one way but the 20mm discs are all over the place.

Mike

Edited 1 time(s). Last edit at 03/17/2017 02:44PM by leadinglights.

Quote
JamesK
For my intended use case of three point under bed mounts I wanted to be able to cope with the bed moving in opposite directions at the sensors for some probing points, so my circuit rectifies the signal before passing it to the op-amp, and uses three separate channels with the output combined digitally at the comparator output. It's not particularly expensive using quad op-amp, quad comparators.

Could you share your schematic with us please. In particular how do you rectify the signal? Thanks

Edited 1 time(s). Last edit at 03/19/2017 08:37PM by Quazzer.

Quote
Quazzer

Could you share your schematic withus please. In particular how do you rectify the signal? Thanks

Sorry, I don't have a schematic done up. Rectification was just the typical diamond arrangement using four discrete diodes. One side of the resulting output is tied to ground, the other goes to the op-amp. I'm using the LM324 which copes with running on a single supply and is cheap on ebay. The output is fed to a LM339 quad comparator which tests the output against an adjustable reference voltage (generated with a cermet and buffered by the spare 4th op-amp on the LM324).

I should 'fess up that I only tested a single channel and not on a printer, so the benefits (or even the feasibility) of the circuit are unproven.

Quote
JamesK
Sorry, I don't have a schematic done up. Rectification was just the typical diamond arrangement using four discrete diodes. One side of the resulting output is tied to ground, the other goes to the op-amp. I'm using the LM324 which copes with running on a single supply and is cheap on ebay. The output is fed to a LM339 quad comparator which tests the output against an adjustable reference voltage (generated with a cermet and buffered by the spare 4th op-amp on the LM324).

I should 'fess up that I only tested a single channel and not on a printer, so the benefits (or even the feasibility) of the circuit are unproven.

Thanks JamesK. I guess you mean a bridge rectifier as used in this circuit.

Edited 1 time(s). Last edit at 03/20/2017 07:06AM by Quazzer.

Quote
Quazzer
Thanks JamesK. I guess you mean a bridge rectifier as used in this circuit.

Oh, good find. Yes, very much like that, but with a comparator in place of the second op-amp. I also used a zener to clamp the input to the op-amp within safe limits as the piezo's are potentially capable of generating voltages above the supply voltage.

I see that somebody has put a video of a piezo z probe on YouTube [www.youtube.com]
There seems to be very little additional information other than that the video was posted by Celtek Imports. A Google search has not identified what company (?) this is.

Mike

Looks very familiar. Be interested to hear if they have refined it, and how? Lets hope its not a closed source project that would be a shame.

It might be worth also highlighting this module [www.thingiverse.com] which seems to pre-date our efforts, or perhaps arose around the same time.

Edited 1 time(s). Last edit at 03/23/2017 06:56AM by DjDemonD.

---

Simon Khoury

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

Quote
DjDemonD
Looks very familiar. Be interested to hear if they have refined it, and how? Lets hope its not a closed source project that would be a shame.

It might be worth also highlighting this module [www.thingiverse.com] which seems to pre-date our efforts, or perhaps arose around the same time.

That is Njål Brekke who was the first to try piezo discs back in 2015. It is amazing quite how many people mentioned the idea without apparently trying it even before him.

Mike

I left our man a comment on his video above.

Apparently he's marketing it already. He gave me a product link which was dead, and a link to a second video:

[www.youtube.com]

Which is remarkably similar to this one
[www.youtube.com]

Oh well I suppose I ought to be flattered.

---

Simon Khoury

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

The product link does seem to be dead but I found that the main site in Brazil is sort of ebay-ish (without auctions) [www.mercadolivre.com.br]

I fear that in the RepRap field the flattery of imitation is as good a reward as it gets. As far as I know, nobody in the community has yet got a gold plated Rolls-Royce.
You may have noticed that no matter how clear a case for piezos you make, the most common response is something like "Sounds interesting but I will go with Capacitative sensors (or even microswitch_on_a_stick_on_a_servo)". Perhaps any exposure will be good exposure.

Mike

I'm not bothered but it would be great if they are willing to share any improvements they've made or testing they've done. I'd like to sell some of these here, Im sure I still will, postage/VAT/Duty from Brazil will be way to high to make it competitive for them to sell them here.

---

Simon Khoury

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

That site is basically eBay for South America. The gentleman that did the MPCNC design has had problems with people printing parts and selling them on that site, using his and other folks pictures. More on it here: [www.vicious1.com]

I did get my boards in and got one assembled - hand-soldering 0603 parts was not fun. Unfortunately, I ran out of spare time before I could test it. Hopefully next week...

g.

Very interested in trying this out on my delta. Do you guys already have boards for wider testing? What are these costing to produce?
I found the brazillian version, but £30 + delivery + duty + etc will be a bit expensive
[produto.mercadolivre.com.br]

Hi there are boards available from Moriquendi. You can print the sensor housing and a piezo disc is just 40p. We are planning on making a ready made unit it will be around £25 when it's ready (3-4 weeks).

---

Simon Khoury

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

I've got five or six boards ready to go and another batch on the way. £10 including UK delivery or £10 + shipping for the rest of the world.

Drop me a PM if you're interested.

Idris

Transferred from the "Proximity Sensor for Glass Sheet" thread.

Quote
frankvdh
I'm not pro/anti Piezo... in fact I hadn't heard of them until a couple of weeks ago. Just trying to see what is practically available.

Quote
leadinglights
[*] Claim: Piezos become insensitive at higher temperatures
Truth: This varies with the make but all that have been tried are satisfactory up to at least 60°C\

But surely that kills their usefulness right there? For PLA, I have my glass bed heated to 70C, and for ABS 100C. How satisfactory are Piezos at those temperatures?

There have been no thermal problems with piezo discs that have not been easily overcome. Using Piezo discs directy under a heated bed without insulation in the manner that FSRs are used is unlikely to work although Moriquendi comes very close to doing exactly this - details and photos [reprap.org] and also [forums.reprap.org] showing that just a thin layer of cork will do the job.

Although I have tested a number of discs in a range of sizes and from different manufacturers for mechanical response and sensitivity, to date I have only done temperature characterization on one Murata 27mm disc and one "no-name" 27mm disc - these showed the Murata disc working to >100°C and the "no-name" disc working to only about 60°C. I also did an informal "put a disc on a heated bed and thwack it with a screwdriver" test on another "no-name" disc which showed it working well at 110°C

So there you have it, more work needs to be done but not to show that piezo discs are as good as anything else but to improve on this, to find out precisely which make to use, how to drill them and mount them and anything else needed to get the very best from them.

Mike

Edited 1 time(s). Last edit at 03/31/2017 12:11PM by leadinglights.

How do you handle the bit of plastic ooze on the nozzle? Do you probe hot, or snip off the filament before probing?

Either way works, but I prefer to probe at 130 nozzle temp (irrespective of filament) as its below ooze temperature and then if there is a need to factor thermal expansion (I did work this out to be less than 0.02mm from 130 deg C to 250 deg C) its from the same baseline. Bed is at whatever temp I need for adhesion.

---

Simon Khoury

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

Quote
nebbian
How do you handle the bit of plastic ooze on the nozzle? Do you probe hot, or snip off the filament before probing?

What method of setting the nozzle height does not need the nozzle to be cleaned?

Mike

Well you might not clip the loose filament from the nozzle if using a non contact sensor and you know your sensor offset. But I suspect most of us do either out of habit or because we want a new print to start cleanly.

---

Simon Khoury

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

Okay so Idris and I are getting closer to having a drop in module ready. More or less settled on the design, just a bit of tweaking. It's done several thousand probes on the rig with no issues. Very stiff as it turns out you don't need any noticeable vertical compliance to get a trigger. Idris is making a custom pcb for it.

Should be a case of remove hot end, insert module between hot end and extruder/carriage and then plug in your endstop wire. It's for groove mount but could be easily adapted with a few bolt holes for screw mounting.

It's small around 30x20x18mm. Should illuminate on trigger (but will also make a solid colour version also).

Aiming for around £25 delivered (UK) for final version. The initial units will be considered beta units and be £19.99 delivered (UK) on the assumption we will get some feedback on improving the product.

Obviously we will ship internationally but shipping quotes on an individual basis.

Edited 1 time(s). Last edit at 04/02/2017 12:32PM by DjDemonD.

---

Simon Khoury

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

So the idea is this is a module which will sit between the groovemount in your effector/carriage/extruder such as a titan, and your hotend. It will have just a 3 wire endstop connector on it. Its pretuned and tested on my rig. If you are using one in a chamber or your ambient temp is 35 deg C+ then let us know we will tune one for higher temp environment, or leave the PCB adjuster with no threadlock on it so you can tune it yourself.

If using bowden, you'd just feed your bowden tube through the middle, there will be space to attach and tension the bowden tube on the hot end, as well as remove the hot end without dismantling the module. If using direct I will supply a short length of bowden tube to use as a filament guide from extruder down into the hotend.



This is what it looks like in my test rig. If you don't want groovemount I'll do a version with captive nuts and bolt holes for direct mounting to underside of a carriage or effector. Almost no compliance of the nozzle in this design, the piezo is pre-tensioned quite firmly.

The new custom PCB will be attached to the back of the module for the transparent version and it will light up the module when powered and hopefully change colour when triggered. If preferred a solid colour version will also be available with the pcb on the front (you can insert the module into your printer whichever way around you want) so it will show two LED's one for power, one for triggered.

Its an open source hardware project so I will push all the files to github as soon as I'm convinced it will be reliable and effective, you are very welcome to make your own, Idris and I can supply PCB's at some point over the next 3 weeks. I will start a new thread here, or possibly utilise a free forum for user support.

This module completed 2000 probes on the rig with no issues. I intend to do a test run of 36800 probes which will simulate a full year's use in a heavy-duty-cycle environment, based on doing 16 probes on a delta to autocalibrate each print, and doing 10 prints per day for 46 weeks per year, 5 days a week. This is likely to represent 3+ years of use in a slightly more normal environment.

---

Simon Khoury

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