Precision Piezo Z-probe Now available.

Sweet results 7 microns accuracy and 2 standard deviation - can anyone beat that?

Please add an image of your setup if you get chance Stef, it would put some context on it.

I hope your pleased with those results!

---

Simon Khoury

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

Yes Simon, very happy with the results. Will post pictures as soon as I find my camera that has the build on it. I need to take the bed apart to install the 41mm piezo to check it out. I came to your site out of sheer frustration with the BLT leveling thing that sometimes worked and is now in the scrap box. If you remove the s from the preceding, you will get the message. Looking forward for your next digital incantation.

Stef

Yeah I keep meaning to buy a genuine BL touch and a clone so I can test them and really see how they work.

---

Simon Khoury

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

My thots, save your money get Simon's V2. just love it. Forget the clones. No more offsets, no more poor registration of probe points. Just exact posistioning .
Stef

So while i am still trying to Frankenstein my way to a somewhat working hotend mount on my Wanhao I3 i made some progress recently which made me hopefull that there is light at the end of the tunnel however there are a few issues which i can't seem to pinpoint why they keep happening seemingly randomly.

So i have made a contraption to have a single piezo being triggered around my hotend carriage. I won't post pictures because a) its a work in progress and horrible to look at and b) its not yet working 100%.

I had some promising results with homing registering correctly so i thought i'd do a mesh. On a 10x10 mesh i have about 8-10 points where the probe either falsely triggers even tough no contact has been made with the bed or the head has already touched the bed but it didn't register so the z motion continues pushing down onto the bed further until eventually it registers and moves on. Now for the second part of the problem i think i have that solved mostly by using Double probing setting and also by increasing somewhat the velocity of the Z motion to have a more firm contact.

What i can't figure out is why i still get the odd false positive in mid air. I have tryed and failed with the following adjustments: disabled Z jerk, set z accelaration to 30 for the probing, increased or decreased general z axis movement speeds, increased wait between z probe movement from 400ms up to as high as 1200 ms, tryed to decrease sensitivity ever so slightly on the piezo board by adjusting the two screws (negative side effect was that with somewhat decreased sensitivity more contacts with the bed did not register correctly which is not good).

The weirdest part of the whole thing is still this: The false triggering of the probe seemed to occure exaxctly at the same probe points somehwat in the first 3rd of the mesh run. Now this is perplexing to me as if the sensitivity might be too high then i would assume false positives to happen randomly troughout the mesh probing but i did some repeat tests and sure enough about 5 out of 5 runs the false positive in mid air registered at basically the same probe points. It can't be ringing because that would have an effect at other points as well so i just don't get it.

So after the lengthy info do you still have some advice to further try to eliminate the false (mid air) positives?

Also for the life of me i do not know exactly what the second screw on the board adjusts exactly? The first is the sensitivity which is clear and i figure the second is somewhat the same???

Cheers

Okay so the vr1 is gross sensitivity if set to around 0.4Mohm it gives standard performance which I use in the ready made Piezo20 module. 0.6 makes it more sensitive 0.2 less etc..

VR2 sets the threshold, the point at which the unit triggers, the closer to the threshold point you are the greater the fine sensitivity. So I start with vr1 at 0.4Mohm then tune vr2 close to the threshold point then test, if it's too sensitive move vr2 away from the threshold point just a 1/8 turn maybe less and try again. If still no joy then set vr1 to a lower value then retune vr2.

If your tuning window - defined as the range of useful performance where false mid air triggers don't occur and the contact with the bed is light and gives an accurate trigger, is narrow it might be your tuning on the pcb but it's more likely to be your setup.

Make sure you use a pause before probing (you can do this in duet/RRF, marlin 1.1.x onwards, smoothie and repetier. Most mid air false triggers with a well tuned system are due to mechanical vibration/ringing in the frame which takes half a second to subside when the head stops over a probing point.

Ensure polarity of piezo disc is correct by doing the push upwards test, if it triggers on push its correct, if it triggers on release its the wrong way around.

Edited 4 time(s). Last edit at 12/13/2017 11:49AM by DjDemonD.

---

Simon Khoury

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

Here are some pictures of my rig early version. Please note that I have used shielded cable from the piezo and grounded at the electronics.. Image on right shows the insulating blanket underside of bed. Although I have the 20mm and27mm (green) only 1 piezo is active for testing.

Stef

Can you give us a picture of it assembled? I'm trying to work out whether you have anything more rigid than the foam between the bed and the piezo.

Idris

---

{Precision Piezo} Accurate, repeatable, versatile z-probe plus piezo discs, endstop cables, pt100, 50w heaters.

Quote
CaptainHero
So while i am still trying to Frankenstein my way to a somewhat working hotend mount on my Wanhao I3 i made some progress recently which made me hopefull that there is light at the end of the tunnel however there are a few issues which i can't seem to pinpoint why they keep happening seemingly randomly.

So i have made a contraption to have a single piezo being triggered around my hotend carriage. I won't post pictures because a) its a work in progress and horrible to look at and b) its not yet working 100%.

I had some promising results with homing registering correctly so i thought i'd do a mesh. On a 10x10 mesh i have about 8-10 points where the probe either falsely triggers even tough no contact has been made with the bed or the head has already touched the bed but it didn't register so the z motion continues pushing down onto the bed further until eventually it registers and moves on. Now for the second part of the problem i think i have that solved mostly by using Double probing setting and also by increasing somewhat the velocity of the Z motion to have a more firm contact.

What i can't figure out is why i still get the odd false positive in mid air. I have tryed and failed with the following adjustments: disabled Z jerk, set z accelaration to 30 for the probing, increased or decreased general z axis movement speeds, increased wait between z probe movement from 400ms up to as high as 1200 ms, tryed to decrease sensitivity ever so slightly on the piezo board by adjusting the two screws (negative side effect was that with somewhat decreased sensitivity more contacts with the bed did not register correctly which is not good).

The weirdest part of the whole thing is still this: The false triggering of the probe seemed to occure exaxctly at the same probe points somehwat in the first 3rd of the mesh run. Now this is perplexing to me as if the sensitivity might be too high then i would assume false positives to happen randomly troughout the mesh probing but i did some repeat tests and sure enough about 5 out of 5 runs the false positive in mid air registered at basically the same probe points. It can't be ringing because that would have an effect at other points as well so i just don't get it.

So after the lengthy info do you still have some advice to further try to eliminate the false (mid air) positives?

Also for the life of me i do not know exactly what the second screw on the board adjusts exactly? The first is the sensitivity which is clear and i figure the second is somewhat the same???

Cheers

Are you using a bowden setup? I've found that when my piezo triggers mid-air always at the same points is because the PTFE tube somehow in that position vibrates or something and as is touching the piezo, it makes a false contact. You can easily check it by taking the PTFE tube in your hand and putting it straight when the probe is in those points that fails, and if it works, you will know that's the problem.

To fix it, I use a zip tie to try to avoid PTFE movements during probing, or moving it a bit so it doesn't touch the piezo inside the groovemount... I'm having that issue usually at the right side of my CoreXY, where the extruder motor is.

Regards

Edited 1 time(s). Last edit at 12/13/2017 04:04PM by okercho.

Quote
CaptainHero
So while i am still trying to Frankenstein my way to a somewhat working hotend mount on my Wanhao I3 i made some progress recently which made me hopefull that there is light at the end of the tunnel however there are a few issues which i can't seem to pinpoint why they keep happening seemingly randomly.

So i have made a contraption to have a single piezo being triggered around my hotend carriage. I won't post pictures because a) its a work in progress and horrible to look at and b) its not yet working 100%.

I had some promising results with homing registering correctly so i thought i'd do a mesh. On a 10x10 mesh i have about 8-10 points where the probe either falsely triggers even tough no contact has been made with the bed or the head has already touched the bed but it didn't register so the z motion continues pushing down onto the bed further until eventually it registers and moves on. Now for the second part of the problem i think i have that solved mostly by using Double probing setting and also by increasing somewhat the velocity of the Z motion to have a more firm contact.

What i can't figure out is why i still get the odd false positive in mid air. I have tryed and failed with the following adjustments: disabled Z jerk, set z accelaration to 30 for the probing, increased or decreased general z axis movement speeds, increased wait between z probe movement from 400ms up to as high as 1200 ms, tryed to decrease sensitivity ever so slightly on the piezo board by adjusting the two screws (negative side effect was that with somewhat decreased sensitivity more contacts with the bed did not register correctly which is not good).

The weirdest part of the whole thing is still this: The false triggering of the probe seemed to occure exaxctly at the same probe points somehwat in the first 3rd of the mesh run. Now this is perplexing to me as if the sensitivity might be too high then i would assume false positives to happen randomly troughout the mesh probing but i did some repeat tests and sure enough about 5 out of 5 runs the false positive in mid air registered at basically the same probe points. It can't be ringing because that would have an effect at other points as well so i just don't get it.

So after the lengthy info do you still have some advice to further try to eliminate the false (mid air) positives?

Also for the life of me i do not know exactly what the second screw on the board adjusts exactly? The first is the sensitivity which is clear and i figure the second is somewhat the same???

Cheers

I'm in the process of designing a mount for printers like yours (and mine)

Take a peek: [www.thingiverse.com]

Edited 1 time(s). Last edit at 12/13/2017 06:40PM by pyr0ball.

Okercho (and others with bowden issues) - there is a sure fire solution. Design a place above the piezo sensor unit, to screw your bowden coupler into, then run a PTFE tube from the bowden coupler down through the piezo sensor unit into the hotend. This decouples the bowden tube mechanically from the piezo disc. I accept it might not always be easy or neat to do this, but if the bowden tube is pulling on the sensor and triggering it, perhaps it is too tight to begin with (just a thought).

Pyroball - that looks good, be keen to follow how well it is working.

In other news I am working with someone on a cr10 idea - and have a cr10 on order to play with, as I know how popular these are right now. Watch this space, or let me know if you want to try it (before me).

---

Simon Khoury

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

Just to say I have made a correction to the online manuals for both the universal kit and the Piezo20.

Where I stated that to measure VR1 resitance it was between centre of VR1 pot and piezo negative, this should read positive.

If you measure Vr1 to piezo negative you get 10Kohms constant. To positive you get the actual resistance of VR1.

Apologies for my mistake.

---

Simon Khoury

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

I used the piezo mount from Simon before this but it was not 100% reliable (did get false triggers once in a while).
Since I wanted to regain some Z-height, I designed this mount for my P3Steel printer: [www.thingiverse.com]

Basically, I sandwiched the piezo between the x-carriage and the Titan mount.
I have repeatable sensitivity around 10 micron which is plenty since my bed is no way flat

Some data from my probe macro, I have a Duet Wifi (so no M42 command):

11:57:22 AMStopped at height -0.175 mm
11:57:19 AMStopped at height -0.165 mm
11:57:16 AMStopped at height -0.175 mm
11:57:13 AMStopped at height -0.172 mm
11:57:09 AMStopped at height -0.172 mm
11:57:06 AMStopped at height -0.175 mm
11:57:03 AMStopped at height -0.175 mm
11:57:00 AMStopped at height -0.175 mm
11:56:56 AMStopped at height -0.170 mm
11:56:53 AMStopped at height -0.170 mm
11:56:50 AMStopped at height -0.168 mm
11:56:47 AMStopped at height -0.170 mm
11:56:44 AMStopped at height -0.172 mm
11:56:41 AMStopped at height -0.172 mm
11:56:37 AMStopped at height -0.170 mm
11:56:34 AMStopped at height -0.162 mm
11:56:31 AMStopped at height -0.170 mm
11:56:28 AMStopped at height -0.170 mm
11:56:25 AMStopped at height -0.168 mm

based on this data: I have mean dev of 0,087186958 mm, stdev is 0,00353181 mm

Edited 3 time(s). Last edit at 12/14/2017 06:15AM by whosrdaddy.

Quote
DjDemonD
Okercho (and others with bowden issues) - there is a sure fire solution. Design a place above the piezo sensor unit, to screw your bowden coupler into, then run a PTFE tube from the bowden coupler down through the piezo sensor unit into the hotend. This decouples the bowden tube mechanically from the piezo disc. I accept it might not always be easy or neat to do this, but if the bowden tube is pulling on the sensor and triggering it, perhaps it is too tight to begin with (just a thought).

Pyroball - that looks good, be keen to follow how well it is working.

In other news I am working with someone on a cr10 idea - and have a cr10 on order to play with, as I know how popular these are right now. Watch this space, or let me know if you want to try it (before me).

That's a good idea, I will think how I can implement it in my setup

Thanks!

Quote
DjDemonD
Sweet results 7 microns accuracy and 2 standard deviation - can anyone beat that?

Please add an image of your setup if you get chance Stef, it would put some context on it.

I hope your pleased with those results!

I can

Mean: 0.101750 Min: 0.98 Max: 0.102 Range: 0.005
Standard Deviation: 0.001601

Video of the setup here: Please forgive the mouthbreathing ._. [www.youtube.com]

Edited 1 time(s). Last edit at 12/14/2017 09:57AM by aMpeX.

Quote
whosrdaddy
11:57:22 AMStopped at height -0.175 mm
11:57:19 AMStopped at height -0.165 mm
11:57:16 AMStopped at height -0.175 mm
11:57:13 AMStopped at height -0.172 mm
11:57:09 AMStopped at height -0.172 mm
11:57:06 AMStopped at height -0.175 mm
11:57:03 AMStopped at height -0.175 mm
11:57:00 AMStopped at height -0.175 mm
11:56:56 AMStopped at height -0.170 mm
11:56:53 AMStopped at height -0.170 mm
11:56:50 AMStopped at height -0.168 mm
11:56:47 AMStopped at height -0.170 mm
11:56:44 AMStopped at height -0.172 mm
11:56:41 AMStopped at height -0.172 mm
11:56:37 AMStopped at height -0.170 mm
11:56:34 AMStopped at height -0.162 mm
11:56:31 AMStopped at height -0.170 mm
11:56:28 AMStopped at height -0.170 mm
11:56:25 AMStopped at height -0.168 mm

based on this data: I have mean dev of 0,087186958 mm, stdev is 0,00353181 mm

So range (accuracy) is 13 microns, dev 4 microns.

Great results from a bracket type system which by its nature is indirect.

If you get chance see what effect probing speed has between 240-1000 mm/min

Also worth trying the new M558 P8 mode which removes filtering to give an even more precise result.

---

Simon Khoury

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

Quote
aMpeX

Quote
DjDemonD
Sweet results 7 microns accuracy and 2 standard deviation - can anyone beat that?

Please add an image of your setup if you get chance Stef, it would put some context on it.

I hope your pleased with those results!

I can

Mean: 0.101750 Min: 0.98 Max: 0.102 Range: 0.005
Standard Deviation: 0.001601

Video of the setup here: Please forgive the mouthbreathing ._. [www.youtube.com]

Well I think you currently hold the record, and its not fair, you're getting way better results than me! Do you want a job?

Excellent, Mike/Leadinglights is quite convinced with enough trial and error, and big enough piezos, and accurate enough firmware we could go to 1 micron accuracy, with probing force 1g! Something to aim for although worth saying this is physics now not engineering.

---

Simon Khoury

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

Quote
DjDemonD

Well I think you currently hold the record, and its not fair, you're getting way better results than me! Do you want a job?

Excellent, Mike/Leadinglights is quite convinced with enough trial and error, and big enough piezos, and accurate enough firmware we could go to 1 micron accuracy, with probing force 1g! Something to aim for although worth saying this is physics now not engineering.

I'm pretty sure I just got lucky, and the VORON is an impeccable machine so some credit has to go to the creator

The first image shows placement of the Piezos under the insulating blanket and the high density foam, the kind that the Tevo Black Widow was packaged in. I don't know what the insulating foam is but it has to handle the bed temps and is covered with a heavy duty aluminium foil so the contact signal is damped by the foil. The second image shows the relative diameters of the various piezo disks and specifications. Note that the 27mm disk has twisted pair wires rather than shielded something to try to reduce inductive/capacitive effects. Also the piezos are hot glued to the bed carriage.

Stef

Edited 1 time(s). Last edit at 12/14/2017 11:04AM by Chowa.

Quote
DjDemonD
So range (accuracy) is 13 microns, dev 4 microns.

Great results from a bracket type system which by its nature is indirect.

If you get chance see what effect probing speed has between 240-1000 mm/min

Also worth trying the new M558 P8 mode which removes filtering to give an even more precise result.

Well this is already tuned at 1000 mm/min in P8 mode
I ordered some V2 boards to see if I can make it even more precise (pots on V1 board are almost dead, so I leave it like is now)

Edited 1 time(s). Last edit at 12/15/2017 03:36AM by whosrdaddy.

First print layer using piezo

Quote
tekkydave
First print layer using piezo

[attachment 100519 2017-12-1511.50.09.jpg]

#First Layer Pr0n

Lovely. It's interesting with this technology you get quite used to every first layer being perfect, +/- one babystep perhaps. Got to love a bit of first layer porn.

---

Simon Khoury

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

Made some under bed mounts. So far working well.
[www.thingiverse.com]

Looking good nice video on there too.

---

Simon Khoury

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

Apologies if orders are going out a little later than usual Christmas is bearing down on us. Around 2-3 days to ship from receipt of order. Christmas post now making pre-christmas delivery now unlikely wherever you are based (it will still arrive sometime after though).

---

Simon Khoury

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

Hi. I Have a prusa i3, with a E3dv6 hotend mounted on a gregs extruder
Will your kit work? I jave a MK3 aluminium heated bed, wich somehow warps like 0.05 in the center.
So my estios are 2: will your system work in my machine? question 2: I still dont understand much about autolevelling. I have repetier 0.92 version, will work with it? Also a question about autolevelling: the system probes all around the heated bed... then auto ccorects Z values in the first layer only or it autoccrects all Z values through the home Z layers? My questios is because I was thinking: If I hav a tall square... if all Z values are corrected, for example in my bed I have a depression of 0.05 mm in the center. That would mean that the final object will not really be flat on the top surface but haveing a 0.05 depresion in its center?

Looking for you answer and hopefully put an order

Should work fine for an i3 mk1/1.5 type, just fit the groovemount Piezo20 unit in between your extruder and the hotend.

It's not usually your bed which has a "hill" in the middle it's your x axis rods which are sagging. They deflect downwards especially if they are 400mm+ in length. I proved this on my corexy swapping from 500mm smooth rods to a linear rail, which is much stiffer and the "hill" disappeared.

Use the later versions of repetier with our probe as they have the pause before probing option you'll need to enable and use around 500ms delay to allow mechanical noise to settle before probing, otherwise false triggers can occur. I'm not sure 0.92 has this, see repetier on github.com to find out more.

If you autolevel which strictly means compensate for a bed which is flat but tilted, by using a plane to define the z height at any given xy position, then objects will technically be slightly tilted all the way through.

If you chose grid (also called mesh) levelling then the bed is divided into virtual squares, you probe at the corner of each square (often square size or number of points is user defineable) and a map or mesh is made which can achieve an even first layer especially on your bed with its hill in the middle, but objects will still have unevenness (and possibly tilt) throughout, this will compensate for a bed that is uneven or tilted or both. Often the artefact is too small to notice,but it gets your first layer down so it a step forward if this is a problem.

I don't know if repetier can do this but RRF can fade out the compensation after x mm, which is user defined 10mm usually does the trick for me. This is where the real gains are, now you can print a straight, even object onto a bed that looks like the hymalayas as long as you have an accurate probe, and that's where we come in.

There is a long running debate between those who want to use these software compensations to correct for cheap printers which have uneven tilted beds, axes that are not orthogonal and those who think that printers should be expensive and heavy but have flat beds and square axes. I am all I'm good with both approaches. It was having a cheap printer that wasn't square that got me interested in this in the first place, now I have an expensive printer which is square and use the probe to check it before printing in case anything has shifted/broken.

Edited 4 time(s). Last edit at 12/20/2017 08:09AM by DjDemonD.

---

Simon Khoury

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

What is RRF?

RepRap Firmware