Precision Piezo Z-probe Now available.

Hi interested to see how your signal conditioner works, we have played around with some micro controllers to signal condition but haven't really settled on the best way yet. But they are cheap and versatile.

I think mounting a piezo in the bracket like our titan bracket can work, but the tuning needs to be very precise as the force transmits to the piezo much more indirectly than when the piezo is above the hotend. In that case it is very direct and therefore tolerant using our analogue circuit of a much wider range of tuning.

As for drilling a hole, get a spur point wood drill bit. Mark the centre of the piezo disc.place the piezo disc with ceramic side upwards, on a disc of blutack which dissipates the heat and stops it spinning around. Press down with the point of the drill bit on the centre mark, it will puncture the upper metal layer, thereby remaining centred. Now use medium RPM and medium pressure to cut out the rest of the thickness. As long as it does not go hotter than you can hold in your hand you wont damage the ceramic. Discard the dust carefully it contains lead(Pb).

Edited 1 time(s). Last edit at 07/14/2017 03:45AM by DjDemonD.

---

Simon Khoury

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

Quote
interested to see how your signal conditioner works, we have played around with some micro controllers to signal condition but haven't really settled on the best way yet. But they are cheap and versatile.

Very simple for now, I just use the DAC that can be configured as a Vref source for the comparator, you can adjust the threshold voltage in 32 steps. I use an interrupt service routine that will detect an edge, then determine the low to high edge, then turn off the comparator interrupt enable. Start a timer for say 100 to 500ms that ignores any other signals. During this period I pull an output pin low.
Then when the timer enters the ISR turn the comparator back on and the detection is armed once again, pull the output pin high again.

I've found a bigger piezo disk in my junk bin, 35mm in diameter (35mm brass, ceramic disk is about 26mm). These can be found in alarms. I've found mine in a telephone which is the ring tone transducer. The bigger size may be more suitable for 3mm bowden setups.

It would seem that ultrasonic transducers contain ring shaped elements, this might also be an interesting source of piezo ceramic elements. I also found square ones but were very expensive.

Do you have a picture of the Titan bracket in an actual setup? I'm having trouble imagining how it would work. Thanks!

I've been thinking of making an L shaped bracket for the x carriage as a test. Then another L shape bracket for the hotend holder. Between these the piezo disks (I plan on using two on both sides) can be placed so that they face downward. If this works, I can integrate the L bracket directly into the x-carriage. Some slots for the bowden tubes and access to the connector. This could also help reduce the added thickness and reduced build space in the z axis.

What if I could add two more disks along the x and y axis, or an accelerometer, then compare the signal strength between them to decide from which angle the change came, eliminating the triggering of the sensor when the force came from the x or y axis? I'd need to use the ADC to capture the 3 signals and process it. It would probably take a dsPIC or PIC32 to process the data quickly enough for a low latency between action and detection/reaction.

The thing with homing is, the motors stop suddenly, I think the acceleration setting is ignored in that case. I've lowered the homing speed. I'll test later if this works sufficiently to get at leas to the probing stage of the bed. I can imagine if the Z axis triggers during printing, for example if there's a jerk motion on the x axis, the auto bed leveling may fail to work.

Thanks for the drilling tip! I've tried drilling using steel drill bits and clamping it down with a sheet of wood with an 8mm hole in it. The drill bits weren't sharp enough, just pushed the metal downwards, creating horrible burrs. Would a stepdrill be better in this case? I also used a drill bit like the one you said would work, but it pulled the metal upward along the flutes. I will try adding a piece of metal under the disk, drilling through both and see if the result is cleaner.

---

--
Kind regards
Imqqmi

NFAN CoreXY printer:
[reprap.org]

This is very interesting, let me answer some of your points as directly as I can:
- The titan bracket, so I do have a photo [www.precisionpiezo.co.uk] the piezo sits in the base, which is attached to your carriage, the hotend (or extruder in this case) is mounted on the vertical face. When the bracket moves upwards the piezo is forced to bend.
-Homing etc.. I wonder why you are having problems like this. When I home X and Y no piezo is involved and my controller is not looking for a trigger so its irrelevant if the piezo signal conditioner pcb sends triggers.What firmware/printer controller are you using?

When I home Z, I move to the bed centre, pause 0.2 seconds to bring mechanical noise down to zero and then dive to probe. It works well on my core xy as the bed moves up not the head, but it works on the delta too with the effector moving. There is no doubt on a flimsy printer with poor linear motion these sensors are going to have a harder time seeing the signal above the noise.

As far as firmware setup goes now were settling on having Z acceleration set permanently at 200mm/s³. And a pause before homing of 0.2sec, see above in this thread, repetier can do this, we now have a fork of marlin that can do this (untested), RRF can do this, smoothie might be able to but they are so far the only major firmware team not to have responded to my request on github. Probing sped I find wants to be minimum 4mm/s but works better at 5-8mm/s.

I will try to do a video for drilling piezos also with my method.

Edited 1 time(s). Last edit at 07/14/2017 07:09AM by DjDemonD.

---

Simon Khoury

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

Quote
DjDemonD
As for drilling a hole, get a spur point wood drill bit. Mark the centre of the piezo disc.place the piezo disc with ceramic side upwards, on a disc of blutack which dissipates the heat and stops it spinning around. Press down with the point of the drill bit on the centre mark, it will puncture the upper metal layer, thereby remaining centred.

If you place the piezo disc with the ceramic side upward, how is the point of the drill bit going to puncture the metal layer to keep it centered?

There is a metal layer on top of the ceramic also its just quite thin. Puncture it with the point of the spur point drill that's why they work so well.

---

Simon Khoury

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

Looks like there is video editing in my near future.

---

Simon Khoury

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

Okay so the last few are still uploading but

videos are now online starting with drilling a piezo and continuing through how to build a piezo 20 and finishing with how to tune the PCB.

[goo.gl]

---

Simon Khoury

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

Nice video. Thanks for posting them

No worries I felt like Tim the toolman.

---

Simon Khoury

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

lol that made me giggle

Hey @DjDemonD, what's your DUET configuration on your delta with the piezo installed ? especially M566, M203, M201 and M906

We've got similar build, maybe I'll find a better tuning to perform fine print.

These are my current settings

 1st machine - Kossel XL delta, 27mm Lykle design piezo effector, v1.20 board (early kit board).
Motor current M906 X1100 Y1100 Z1100 E800:800 I60
Accelerations M201 X3000 Y3000 Z200 E200
Max feedrates M203 X15000 Y15000 Z15000 E3600	
Jerk M566 X1200 Y1200 Z1200 E100	

M558 P5 I0 R0.2 F300 X0 Y0 Z0 ;digital piezo sensor, output rises on contact, probing speed, not used to home axes
G31 X0 Y0 Z-0.14 P10	;sensor is nozzle and debounce value.

But the e settings are for flex3drive so use whatever works best for your extruder.

Honestly it seems now with the R parameter in M558 all you need is z acceleration at 200 and M558 R0.2 (can increase it if still getting false triggers). Everything else to do with axis movement is not relevant to probing.

But if you're after general settings there you go. I have fairly conservative settings it can go a lot faster.

Hope this helps.

Edited 2 time(s). Last edit at 07/15/2017 05:58PM by DjDemonD.

---

Simon Khoury

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

Quote
Hergonoway
Hey @DjDemonD, what's your DUET configuration on your delta with the piezo installed ? especially M566, M203, M201 and M906

We've got similar build, maybe I'll find a better tuning to perform fine print.

I am using the identical settings for mine that DJDemonD uses with great success. I do have a couple minor differences in my M558 command...

M558 P1 X0 Y0 Z0 F300 T3000 H2 R0.2

P1 is an analog probe using the probe connection on the Duet. About the only difference is where the probe connects to the board.
T3000 sets the speed the head moves from one probe point to the next.
H2 controls how far the probe will be moved before a no trigger event occurs. If the piezo doesn't trigger, this is the maximum distance it will move in mm.

Using a smaller H value requires the H value in the M665 command be fairly close to the actual height. If the H value in M665 is too high, you will get no trigger messages. If it is too low, it will push the head into the bed before trying to probe. But having a smaller M558 H command can speed up the probing.

thanks !

well it's almost the same ! Here's mine, if someone else want to see it.

M566 X1200 Y1200 Z1200 E100 ; (JERK) Set maximum instantaneous speed changes (mm/min) 
M203 X18000 Y18000 Z18000 E3600 ; Set maximum speeds (mm/min)
M201 X3000 Y3000 Z3000 E200 ; Set accelerations (mm/s^2)
M906 X1100 Y1100 Z1100 E800 I70 ; Set motor currents (mA) and motor idle factor in per cent

My "R" is currently a bit more since I screw tighter the piezo mount and increase the sensitivity.

M558 P5 I0 X0 Y0 Z0 H2 F300 R0.4 ; Set Z probe type to unmodulated, the axes for which it is used and the probe + travel speeds
G31 P10 X0 Y0 Z-0.10 ;

before someone tells me about my Z acceleration, I have to precise that I'm under the 1.19beta9 firmware. There's a substantial interesting feature since the beta 8 :
"Z probing moves now use 250mm/sec^2 acceleration unless a lower Z acceleration limit has been configured. This is to avoid triggering nozzle-contact sensors at the start of a probing move."

Edited 3 time(s). Last edit at 07/15/2017 06:15PM by Hergonoway.

Interesting. I'm using beta9 now but obviously didn't read the documentation.

It's very useful that the duet team are working on a nozzle contact sensor themselves, the firmware is getting some nice optimisations for it which does us no harm at all.

As for other firmware marlin seems like we can get some tweaks in there. Smoothie are open to adding something if I can get someone to fork it and make changes.

---

Simon Khoury

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

Quote
DjDemonD
As for other firmware marlin seems like we can get some tweaks in there. Smoothie are open to adding something if I can get someone to fork it and make changes.

Initial attempt at adding a pause (dwell) before probing to smoothie. I don't actually have a pizeo yet, but testing on my delta with a long delay of 2 seconds and it does work with G30, G31 (delta_grid) and G32 (delta_calibration). I tacked it onto smoothies M670 since it is already how they setup the zprobe params, and went with D (dwell) since R in smoothie was already used as ReturnFeedrate.

If setting via Config file

zprobe.dwell_before_probing   0.2    #Dwell time in seconds

Setting via GCode

M670 D0.2

Here is the firmware.bin for testing (along with the usual warnings of possible printer destruction etc etc). If someone can test on other printer types / leveling strategies we can then do a pull request and hopefully get this into smoothie. The changes are very minimal so hopefully it shouldn't be a big issue.

Edited 1 time(s). Last edit at 07/16/2017 01:40AM by sdavi.

Fantastic work thank you.

I haven't got a smoothie compatible board at present so if anyone can feedback on this change, then we can consider making a PR to smoothieware.

Genuinely impressed by the progress here.

---

Simon Khoury

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

Okay so current thinking in terms of firmware setup for use with piezo probe is limited to two settings:

Z acceleration at 200mm/s² maximum.
Set M558 Recovery time (RRF), Dwell (Smoothie using experimental version above), Pause before Probing (Marlin using experimental version above) or Z_probe_delay (repetier). 0.2 seconds has proven to be enough for most but feel free to set a longer pause if you wish.

Notes: In RRF 1.19 probing in undertaken with a default 250mm/s² Z acceleration now.
Please feedback to us if you have tried the Marlin fork or Smoothie Fork we now have to test. We can hopefully get these changes merged with the main branches.

Edited 1 time(s). Last edit at 07/16/2017 06:10AM by DjDemonD.

---

Simon Khoury

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

@DjDemonD Yes the support is quite good from duet, currently happy of my switch to Duet wifi with the piezo. By the way I've just saw your youtube how-to, you should definitely promote these on your web site!


General question : does anyone used the piezo mount with a flex3drive or a nimble? I'm currently thinking about leaving the world of long bowden to more direct drive approach.

Edited 4 time(s). Last edit at 07/16/2017 06:40AM by Hergonoway.

Yes I do on one machine a flex3drive. I have a nimble but I haven't tried it yet.

They work quite differently but the print quality is amazing. Truly high speed (100mm/s+) printing is a challenge as the stepper has to turn at high rpm and reverse to retract. At lower speeds this isn't really a problem with the right settings.

Recently someone proposed up gearing the stepper to address this issue.

Edited 2 time(s). Last edit at 07/16/2017 07:06AM by DjDemonD.

---

Simon Khoury

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

Are you satisfied with it ? do you have any recommendation regarding using your piezo with one of these 2 systems ? I thought that a dedicated mounting option may be interesting but if the stock one (yours) is enough, well.


EDIT : damn you've edited your message when I wrote mine :p

Edited 1 time(s). Last edit at 07/16/2017 07:12AM by Hergonoway.

That's exactly what I thought, if I remember the current ratio is something like 40:1, maybe with a 4:1 (reversed) gear on the nema would improve reliability at high speed, I can't imagine my poor nema working at 100mm/s with the corresponding extraction and retraction with a 40:1 ratio

Edited 1 time(s). Last edit at 07/16/2017 07:17AM by Hergonoway.

Sorry for the editing I do tend to do that so give me a minute or two after posting its probably going to change.

I'm using the Lykle designed effector see thingiverse which has a nimble mount option but mounting a flex3drive is just a few holes in the effector. This works with a printed or laser/water jet cut aluminium effector the dxfs are listed in my Piezo20 listing in thingiverse. You could use a flexible drive extruder above the effector and the piezo 20 module below (or integrated if printed effector).

Edited 1 time(s). Last edit at 07/16/2017 02:17PM by DjDemonD.

---

Simon Khoury

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

Quote
Hergonoway
General question : does anyone used the piezo mount with a flex3drive or a nimble? I'm currently thinking about leaving the world of long bowden to more direct drive approach.

I am running a nimble with the piezo on my delta. Loving it.

@ElmoC are you satisfied with the lock up system ? seems a bit "not durable"

I'm currently looking to adapt the mount to my effector and the piezo

Speak with Lykle at zesty he will do custom mounts for people when he has time.

---

Simon Khoury

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

Quote
Hergonoway
@ElmoC are you satisfied with the lock up system ? seems a bit "not durable"

I'm currently looking to adapt the mount to my effector and the piezo

Are you referring to the part that holds the filament against the drive gear? If so, it feels durable enough. My only concern is how small it is and with no tether, it could be dropped and lost. I haven't check about replacement units yet.

@ElmoC Yes the "red piece", the lock up seems easy but, I don't know, it doesn't seems stiff

Quote
Hergonoway
@ElmoC Yes the "red piece", the lock up seems easy but, I don't know, it doesn't seems stiff

It can be a bit tricky at first (I was trying to put mine in backwards to start with), but I have done several very long print jobs without any problems.

I did write to Zesty about replacement blocks and got a reply back that they are sending out replacements ones. Not sure why, but will ask.

I believe they had a few that didn't quite fit and decided to reissue them with a slight design tweak.

---

Simon Khoury

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