Precision Piezo Z-probe Now available.

Quote
Suggy
-core xyz Marlin 1.1.0 RC6
-300x200mm bed
-round holder fixed to bed carrage, piezo inside, lid then plunger (https://www.thingiverse.com/thing:2466725)
-4 point fixing around 40mm in from each end both back and front
-using DjDemonD 20mm underbed mounts
-3.6 Mohm VR1 (lowest I could get it to go) VR2 tuened until LED went out then about 1/8 turn
-z accel 3000 mm/s/s Jerk 0.4 mm/s (I think)
-probing speed ?mm/s travel ?mm/s (don't know where to find these)

Hey Suggy, can you please post a picture of your setup? I'm curious about how you setup everything. Are you happy with the accuracy? How is it working?

Cheers!

There are 4 censors, 1 next to all 4 bed leveling springs it work realy well but it did take me around 2 weeks to get it working right. So far I'm happy with it.
This is on fb of it working, very slight pause when probing, not sure what it is
[www.facebook.com]

ps dident have to glue bottom and middle parts together, piezo is a press if as is the middle section

Edited 2 time(s). Last edit at 08/03/2017 05:50PM by Suggy.

Hi Ian,

Thanks for posting, probing speed is usually "z homing feedrate" on marlin, I think it does the first probe at that speed then you can configure a 1/2 speed verification probe. "Travel speed between probing points" is the other one, or something phrased like that. But I am a little out of date with marlin.

Just to clarify having just spoken to Idris/Moriquendi the LEAST sensitive setting for VR1 is the LOWEST resistance measured from pot adjuster to Piezo negative pin. Obviously the inverse is true if you want the most sensitive setting, but this will probably trigger if a mouse sneezes 10 feet away, so its not very useful. Worth saying I have not managed to get reasonable behaviour with VR1 set to (as close as possible to) 0 ohms, you need a little resistance to make it work.

For hotend setups aim for 0.4Mohm on VR1, go up to 0.5 or 0.6 if not sensitive enough, or down to 0.2-0.3 if too sensitive
Be aware a little adjustment of sensitivity is also possible from VR2, by setting it as close to the point where the LED changes state as possible for MOST sensitive, and a little further away for LEAST sensitive.

For underbed, it depends on the mount and mass of the bed, and the smoothness of its motion. We are currently looking into a "cut-trace" to get 10x less sensitive option on the kit board, which would have a solder bridge that could restore the board back to "normal" if desired. This would effectively make VR1 go from around 0.1Mohm down to 0ohm, instead of 10Mohm to 0 ohm, thereby allowing more precise control of sensitivity for heavy beds.

Any thoughts on this idea?

Edited 3 time(s). Last edit at 08/03/2017 04:47PM by DjDemonD.

---

Simon Khoury

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

Hi Simon
I think it is a great idea but may I suggest a tiny micro switch, there are loads of people out there that don't have or know how to solder

Cannot find the settings in the config.h file do you mind if I pm it you for you to have a look?

PS forgot to say the case on this machine is acrylic witch enhances any vibration caused by movement of any kind. Still to test with a print, waiting for an extruder feeder as the one I orded is unusable. Only had a left, need a right

Edited 1 time(s). Last edit at 08/03/2017 05:54PM by Suggy.

The problem with switches is that they're expensive and take up a lot of space on the board. I think that most people that are homebrewing their piezo system will have some soldering experience and the only people who would need to solder the jumper would be people who had gone from low sensitivity to high sensitivity (by cutting a trace) and then decided to go back to low sensitivity. A jumper is another possibility but again, there isn't much free space on the board, especially if I add mounting holes.

Idris

---

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

I take your point Ian, and we can think about it for future releases. But I do agree with Idris that a blob of solder isn't a problem for most, if they can't do it themselves it would be easy to find someone who can put a blob of solder across two pads.

As for the probe speed/travel speed, just let me know which marlin version you are using (is it the now 1.1.x bugfix with our modification?) I'll just look at it on github.

line 671-678

// X and Y axis travel speed (mm/m) between probes
#define XY_PROBE_SPEED 8000

// Speed for the first approach when double-probing (with PROBE_DOUBLE_TOUCH)
#define Z_PROBE_SPEED_FAST HOMING_FEEDRATE_Z

// Speed for the "accurate" probe of each point
#define Z_PROBE_SPEED_SLOW (Z_PROBE_SPEED_FAST / 2)

So your travel speed between probes is whatever you defined for xy probe speed, and your probing speed is whatever your homing_feedrate_z is divided by 2.

Its worth aiming to get your probing speed to be around 5-7mm/s. So what I did is change
#define Z_PROBE_SPEED_SLOW (Z_PROBE_SPEED_FAST / 2)
to
#define Z_PROBE_SPEED_SLOW (Z_PROBE_SPEED_FAST / 1.5)

and then define HOMING_FEEDRATE_Z as 10.

---

Simon Khoury

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

How about a 2-pin 0.1 header so you can just use a jumper.

Quote
Moriquendi
The problem with switches is that they're expensive and take up a lot of space on the board. I think that most people that are homebrewing their piezo system will have some soldering experience and the only people who would need to solder the jumper would be people who had gone from low sensitivity to high sensitivity (by cutting a trace) and then decided to go back to low sensitivity. A jumper is another possibility but again, there isn't much free space on the board, especially if I add mounting holes.

Idris

That might be possible in a later design.

---

Simon Khoury

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

So still getting nowhere with the underbed sensors on the corexy. Currently oscillating between a height map with some unexpectedly low points on it, as it doesn't trigger well enough in the parts of the bed which are not well supported (I'm using 3 point mounting, my z carriage isnt large enough for 4 point - plus it works fine with hotend mounted sensor, titan bracket sensor in this case), and a height map with some unbelievably high points in it. So its on a knife edge between the Himalayas, and the Mariana Trench. So I'm going to try some redesigned mounts which will use a full 20mm plunger to compress the piezo, hoping this will mean mechanically less sensitive, leaving me the ability to use the PCB to get the best tuning.

A few thing have to do be done differently, the piezo has to be the other way up, as the plunger has to rotate, as its also a manual levelling screw, there needs to be a cut out for the solder pads in the lower part, and the piezo will need to be lightly bonded in, so it doesn't get twisted and break the leads off when the thumbwheel upper part is turned.

But then this is what 3d printers are for isn't it - prototyping things, funny that its always a part for a 3D printer.

---

Simon Khoury

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

When I first put underbed piezos on my Delta printer I had a head-banger type of problem, it seems to sudeenly get far too sensitive and even stayed in a "contact detected" mode while it should only give a fairly short pulse (50ms or the time from nozzle touch to nozzle lift - whichever is shorter). The problem turned out to be that the piezos gave a DC output of 50mV for the whole of the time the heated bed took to get to temperature and an extra 5 minutes just for kicks. Solution may be to reduce the gain of the amplifier by making R3=R2, somewhere about 10k to 22k

Mike

Thanks Mike but I haven't even turned the bed on yet. I'm printing new mounts now will see if they change the dynamic at all if they don't it's not mechanical.

---

Simon Khoury

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

Do you have tensioner springs? I had to tighten the front ones 1/4 turn more that the back, I think it's something to do with more flex at the front than the back ( XYZ setup)

No springs but thanks for the suggestion.

---

Simon Khoury

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

Can you look directly at the output of the piezos with an oscilloscope ( times 10 probe if you have one ) This should show up any odd mechanical movements.

Mike

I can but I've not got a proper scope with storage, just a pocket scope, might have to get one if I'm going to figure this out. Changed mounts to much more mechanically dampened ones and yet it's still on a knife edge.

---

Simon Khoury

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

I think what's most puzzling with this machine is the hot end solution works perfectly, and is the same bed and hotend moving. So it's either my mounts, which work on a stationary bed machine, or something in the electronic tuning. I might switch over to my prototype arduino Nano, at least I can serial monitor the adc values from the piezo and see what's going on a bit more quantitatively.

---

Simon Khoury

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

Try putting some pressure on the bed replicating bed tensioner springs and see if the changes anything?

That's worth a try the springs might act as dampeners however... I've concluded it's mechanical noise in the bed.

I did hook it up via the arduino Nano so I could see the adc value and its triggering on almost nothing when the bed moves, and I'm moving it as gently as possible. It's triggering at adc 3 (of 1023) with 1Mohm resistor in parallel to the 3 Piezos. As a result it either doesnt trigger or has to be set too sensitive that it triggers whenever it moves.

I think it's my z axis which is too mechanically noisy. It's belt driven and the belts ride over pulleys on the toothed side (608 bearings). The trap I've fallen into is presuming that the mechanicals are the same from hotend to underbed, but they aren't as with hotend sensing, the x and y axis are mechanically noisy and cause triggers during movement, but once they stop and pause, then probe they are stationary. The effect of a noisy bed is irrelevant as the hotend is insulated from the vibration to a sufficient degree, and the bed only has to raise and tap the nozzle.

When Piezos are under the bed, on the z carriage, the mechanism is noisy and has to move to probe and so there just isn't enough signal to noise ratio to give me a usable range to tune.

So I think it's leadscrews and linear rails or stick with hotend sensing which is working great.

Otherwise I might be able to twist some belts and use higher microstepping on the z motor.

Edited 3 time(s). Last edit at 08/05/2017 07:17AM by DjDemonD.

---

Simon Khoury

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

Hello All

I have been playing with a 20 mm piezo and have noticed that when I place my hand close to the sensor input wiring the blue led flashes uncontrollably.Suspect it could be a static charge from my body. Am not using shielded wiring from piezo to the input of the board. Since I do not have a scope to see what noise is being generated. The circuit is very sensitive and responds to a light tap. I'm using the under bed holder. The bed weighs in at 960g and is 400mm W X 250mm D. I plan to use 4 -20mm piezos in parallel. Will post once I rewire with shielded cable. I have not adjusted the pots as yet settings are as straight out of box.

Steve

Cool. Yeah the capacitance in your finger can make a difference. Try to keep the wires between the piezo and the PCB short if possible. Let me know if shielding them helps, as I've not tried it. Can't see how it would hurt.

---

Simon Khoury

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

Well, good news, I connected the piezo with 2 wires shielded a length of 25cm the shield connected to the DC -. No capacitive effect Now to get more shielded wire and hook up the other three. Could also use ferrite beads to choke off any other noise. Something to try later.

Good work.

---

Simon Khoury

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

NEW STOCK ANNOUNCED.

So I have 12 more kits available. [www.precisionpiezo.co.uk]

However due to not receiving new stock when I need it they will NOT ship with an endstop cable (its a standard dupont 3 pin female-female 70cm) but please swap the SIG and GND leads around at the piezo PCB end if you use one you have already). I have lowered the price to reflect this.

Simon.

---

Simon Khoury

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

Printed these mounts for underbed on my delta to replace the FSR's. Black parts are CF PETG and white are 40 hardness TPEE. Should I install the piezo's as shown or with the solid metal side against the plunger or does it matter? Don't have time tonight to install them and test, but will report back when I do.



Edit: mockup showed I designed the plungers a little too tall so have to reprint them.

Edited 1 time(s). Last edit at 08/07/2017 10:16PM by Alexander Mundy.

They look good.

I'm not sure it makes any difference. In my latest bed mount design which has a thumb wheel plunger so I can use it to manually level, I installed the piezo with the ceramic face down as I figured I'd get less wear on the brass side of the disc from the rotating plunger, but I'm sure it would be quite minimal.

The piezo when used in squeeze-configuration can be either way around. If using it in bend configuration where you support it on the extremities and press in the middle polarity can come into play, but swapping the connector around is easy enough to do.

Edited 2 time(s). Last edit at 08/08/2017 01:59AM by DjDemonD.

---

Simon Khoury

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

Quote
DjDemonD
They look good.

I'm not sure it makes any difference. In my latest bed mount design which has a thumb wheel plunger so I can use it to manually level, I installed the piezo with the ceramic face down as I figured I'd get less wear on the brass side of the disc from the rotating plunger, but I'm sure it would be quite minimal.

The piezo when used in squeeze-configuration can be either way around. If using it in bend configuration where you support it on the extremities and press in the middle polarity can come into play, but swapping the connector around is easy enough to do.

Dito to that I hade to change the polarity on 2 out of 4 on my setup

Quote
Alexander Mundy
Printed these mounts for underbed on my delta to replace the FSR's. Black parts are CF PETG and white are 40 hardness TPEE. Should I install the piezo's as shown or with the solid metal side against the plunger or does it matter? Don't have time tonight to install them and test, but will report back when I do.

N92GuLb.jpg

Edit: mockup showed I designed the plungers a little too tall so have to reprint them.

look very much like what I have done the last two days. I try to put some pictures online the next days. did a first G33 yesterday with std dev 0.048 over the whole bed. Not bad - but I want more (or less? )

The solder connections to the ceramic are very delicate, whichever way up you have to disks you'll want to make sure that there's a cutout for the solder and wire. I also suggest fixing the wires to something close to the piezo, any movement of the wires will pull the wire off the ceramic eventually and give rise to problems that are really hard to pin down (ask me how I know...).

I'm curious to hear how you get on with your soft plungers.

Idris

---

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

Dancer - can you test your sensor setup? If using Marlin send M48 at bed centre this gives you the sensor deviation, or if using RRF you can download a test gcode (this page on page one has the test gcodes [www.duet3d.com]) - which does a bunch of G30's and then pools the results. If this gives really high accuracy i.e. std deviation (probing repeatedly over one point not the full levelling process) of 0.007mm or better which is the most common value we came up with in testing then your sensor is as accurate as it needs to be, you are then better off looking at other aspects of the machine to get it down lower, bed rigidity, carriage rigidity etc.. .

Its tricky to isolate whether it is the sensing method needing improvement, or whether its the mechanics of the machine.

This is where the ready made sensor module comes in, in that its tried and tested to a certain level of accuracy and repeatability, so it can be relied upon, whereas the kit depends on how you implement it.

---

Simon Khoury

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

std. deviation in Marlin with M48 0.0089

Especially, as my heatbed is not bolted anywhere. Just Cups with the Piezo put in it, and Inserts (for the Cups) bolted to the heatbed. I can easily take away the whole heatbed without any screw.
I'll try to take some pictures tonight and maybe upload openscad-Files for my parts.

What do you use as Probe Offset? X0 Y0 Z0.2 is slightly to low on my machine - think it will need about 0.25-0.3 in Z

I used this circuit diagramm - the first one with the 4 OPAMPS, but inversed the imputs of the 4th one. Thingiverse
Disabled the input-pullup in Marlin, so if circuit is not connected, or not available it reads triggered. As I did not want to give additional load to the Ardino 5V regulator, there is it's own 78L05 on the board. And the possibility that none of the four piezos triggers - I'd say it goes to Zero, so should be pretty safe against failing probe.