Precision Piezo Z-probe Now available.

Quote
executivul
.......................................
You must see the timings and speed of rise, if the ADC can cope with it. I would rather use all analog/simple circuitry for it. Don't trust isolated smart chips (see BLTouch).

The ADC is plenty fast enough. If the nozzle touches the bed at a speed of 2mm per second and the piezo test to a trigger point after say 20 microns then that will be reached after only 10 milliseconds. The PIC is quite happy doing a conversion in 100 microseconds and is able to do some quite complex signal processing at the same time. Where MCUs seem to do badly it is often because they are being programmed by people who were only taught high level languages - to make an MCU product rock solid normally needs somebody who knows every register like an old friend and this normally (but not always) means somebody writing in assembly.

Mike

Got my sensors and already got one in my current prototype carriage for my CoreXY build. Now I just need to figure out how to mount the PCB and a few other bits and get the rest of this thing built!

Quote
Moriquendi
Inverting the wires shouldn't have fried anything. The potentiometers are pretty delicate, it's just a function of their size. If you look carefully at the right hand pot you should see a flat side on the twiddly bit, turn that flat side to the bottom. This is the normal setting, it should work for almost everyone. The blue led should normally be off and flash on when triggered.

Idris

Ok you are correct. I played with the knobs again until finding a position that worked

I am now trying to deal with the fact that the cables connected to my 27mm piezo make the thing trigger. I tought it was a bad contact somewhere because if I played around with the cables it would trigger, but I installed a different piezo and same thing happen (when I move the board around or try to fit it somehwere, it triggers.

So, it's a bad contact somewhere, or I am doing something stupid. Probably the second one. I will try with a smaller piezo and or find a way to glue the cables so they don' move.

Which type of PCB are you using? v1.22 with three inputs or the PP20 PCB with one input on the back?

It's worth measuring the resistance of VR1 (how you do this depends on which board you have). If VR1 is damaged it will read very high resistance (>1M) which will make the circuit oversensitive.

It sounds like you have it set too sensitive, the cure for this is to move VR2 slightly further away from the trigger point. If you can't find a compromise between triggering when the wires are moved and not triggering reliably when probing then you may have to reengineer your mount to produce a better signal.

Idris

---

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

Quote
Moriquendi
Which type of PCB are you using? v1.22 with three inputs or the PP20 PCB with one input on the back?

It's worth measuring the resistance of VR1 (how you do this depends on which board you have). If VR1 is damaged it will read very high resistance (>1M) which will make the circuit oversensitive.

It sounds like you have it set too sensitive, the cure for this is to move VR2 slightly further away from the trigger point. If you can't find a compromise between triggering when the wires are moved and not triggering reliably when probing then you may have to reengineer your mount to produce a better signal.

Idris

Yeah I glued piezos everywhere on my mount in the last 4 hours trying to figure it out. I can get it to consistently respond to taps on the nozzle, but my bed slowly approaches the nozzle and pushes it all the way up (and the nozzle just digs into the bed) without 'tapping' the thing. I will do the measures, but I will also go back to my original thinking of "I need a carriage that supports this".

Thanks for the help!

EDIT: I am using the PP20

Edited 1 time(s). Last edit at 10/28/2017 05:11AM by coredump.

The circuit responds to how quickly the signal from the piezo changes rather than the magnitude of the signal, this is probably why it's responding to a sharp tap but not the slower bed contact.

Idris

---

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

Yeah I can't see a solution aside from making my bed travel faster (unsure if it's possible) to have it 'tap' the nozzle. I risk shoving the nozzle hard on the bed and if it decides not to stop I am screwed.

I'd avoid that if you can as it's detrimental to accuracy and repeatability, I currently probe at 2mm/sec and I have probed as slow as 0.5mm/sec. When everything's engineered right it'll just work, you wont have to fiddle with tuning it.

Idris

---

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

Well back to the drawing board to design a F***** carriage that works.

THanks Idris.

It's worth saying that there is a big difference between underbed setups (which are more sensitive on account of having 3 or 4 piezos contributing to the signal) than hotends. I'd recommend 4-7mm/s probing speed for hotends.

Its a case of using low acceleration on the bed but higher speed. Use a probing height of about 7mm or more to ensure you generate the speed you need before contact.

To limit the potential for head crashes causing damage lower your bed motor current to the minimum you need, if this can be done via Gcode then set low current, probe then restore to your preferred current level, obviously this is less practical if you are setting motor current via potentiometers. If a tap by hand will trigger it then the bed should be able to as long as you get the accel/speed right.

If you do come up with a novel carriage for your machine then please post it here [forums.reprap.org] and if it works and its publically available you get a free piezo kit. I sent two out this morning. We can then build up a set of designs to suit many popular printers.

Edited 1 time(s). Last edit at 10/28/2017 06:08AM by DjDemonD.

---

Simon Khoury

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

I use 5mm/sec on my corexy with 1605 screws for bed. No problems so far.
Delta also 5mm/sec.

In an earlier post on this thread Alexander Mundy was concerned about piezos giving an out of phase signal.

Quote
Alexander Mundy
................

The one thing I was thinking about with the analog board was multiple comparators instead of the single input. Would be useful for a delta toward the outer edge where one or some of the piezos give180 degree out of phase signals.

Back when I started looking at piezos I was concerned that there may be this "phase reversal" on a sensor but found that no such thing happened. It took a while to get to grips with what was happening but this is it in the diagram below.



A print bed with three piezo sensors A, B and C detects a nozzle coming into contact with the bed. If the contact is directly above piezo A then a charge will be generated in that piezo but not in piezo B or Piezo C. Some of the charge from Piezo A will flow into the other two and the voltage across the whole set will be 1/3 of what it would be with one.

If the nozzle makes contact in the middle of the bed at point "M" then the three piezos will each experience one third of the load and will generate one third of the charge. There will be no flow of charge between the piezos and the resulting voltage will be exactly what it was in the case of contact with only one piezo.

If the nozzle makes contact in the middle of the fulcrum line between piezo A and piezo B then half of the load will be experienced by each of these piezos and none by piezo C. A portion of the charge will flow to piezo C and again the voltage will be the same as in the first and second case.

The last case is where the nozzle makes contact beyond the fulcrum line at point "P". In this case there will be an upward pressure change at piezo C which will give a reverse charge. Piezo A and Piezo B will each experience more than half of the load and a greater flow of charge from these two piezos to piezo C will result, once again giving the same voltage across the set of piezos

Edit: I noticed that the plots for the last case showed the wrong piezo receiving a negative charge. The diagram has been redrawn to show this correctly.

In the real world this is exactly what happens. FSRs are not as good but a somewhat similar mechanism means that they also should not suffer from phase reversal.

Mike

Edited 2 time(s). Last edit at 10/29/2017 02:51PM by leadinglights.

Quote
coredump
Yeah I can't see a solution aside from making my bed travel faster (unsure if it's possible) to have it 'tap' the nozzle. I risk shoving the nozzle hard on the bed and if it decides not to stop I am screwed.

All of my research has used a speed of only 1mm/sec although I use 2mm/sec for my printers. I have never tried using piezos as pure microphones but I suspect that one approach that may stretch out the response time and increase the peak voltage is to fix a weight onto the top of the piezo. The idea is to squeeze the element between the irresistible force of the vibration and the immovable object of the weight. I think a weight of about 5 grams (say 1/4 oz) should do it. Try fixing it with a bit of double sided tape and if it looks promising then fix it permanently with epoxy - the reason for the change to epoxy is that DS tape is compliant and will kill a bit of the signal.

Mike

I suspect the phenomenon of teeter-totter which often dogs FSR use is due to the bed mounts. I just changed the mounts on my microdelta to reduce this effect as the bed was lifting out of the mounts when contact occurred in your area "P" on your diagram above. I now have some fixings to hold the bed firmly in the mounts.

Inevitably a lot of the bed mounting ideas in use with piezo or FSR are essentially bed "floating" on the mounts which can generate issues.

Nicely explained though.

---

Simon Khoury

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

Damn son, got this thing working.

[www.youtube.com]

I ended making it like this: [www.thingiverse.com]

I totally stole the idea from OutcastZeroOne's mount, and DjDemonD under the bed support (initially I was going to make it with 3 parts like his bed support but decided it was unnecessary complex).

Duet doesn't has a M48, but I made a macro with G30s and threw the 30 results on google sheets, and it gave me a std dev of .00488, with maximum -0.100 and minimum of -0.12

There's still things to do and simplify, but atm it works well enough.

I tried measuring the PP20 VR1 resistance and either measured at wrong points or just did something wrong because I didn't get anything near 0.4MΩ, so I just did what Moriquendi said and turned the little chanfer on the trimpot down, and from there I just regulated VR2 until it seemed to be ok. I was actually kinda surprised that it worked right away (was with my finger on the emergency stop button because I was sure it was going to crash the bed).

Thanks for all the help!

PS: ALSO, I watched the 6 videos about installing and configuring and there is very important information there that isn't on the manual, for example I was totally unaware that the piezo polarity made any difference. Also better information on the measure procedure would be great

Edited 1 time(s). Last edit at 10/31/2017 02:28AM by coredump.

Glad you got it working, and working well by the looks of it.

VR1 seems to be causing problems for people and I think we'll be getting rid of it in the next revision. VR1 is a hangover really from the early days when we didn't know how consistent piezo disks would be and wanted to allow for the widest possibly implementations both mechanical and electrical. Current thinking is that boards for the PP20 module will lose the VR1 pot while v1.22 boards and it's descendants will keep it to allow for maximum flexibility.

We'll have a look at the manual, writing documentation that is complete, clear and concise has been one of the surprisingly challenging aspects of this project and something we always appreciate feedback on.

Idris

---

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

Quote
coredump
PS: ALSO, I watched the 6 videos about installing and configuring and there is very important information there that isn't on the manual, for example I was totally unaware that the piezo polarity made any difference. Also better information on the measure procedure would be great

Yes, the polarity does need a special mention in the docs. It's one of those difficult things to track down as everything appears to work fine when tapping by hand, doing a probe and tapping by hand above the bed, but frustratingly just crashes into bed when doing a probe - which results in more fiddling with the pots trying to get it to work and/or assuming its config or probing speed related. It was driving me crazy until I saw a mention buried deep somewhere in this (now very long) thread. Although I suspect in my case, it was from designing a custom mount and having the piezo oriented "up-side-down".

Been using the piezo for a few weeks now and its working great - first layer goes down perfect every time.

Polarity is mentioned in the manuals in several places but I will accept perhaps it is not prominent enough and correct that over the next few days.

Glad you got it working, very few people, if any have been unable to get it going and achieve great results. I think its a system which takes a few days to familiarise yourself with and get a feel for, but I feel worth the time spent to have super accurate first layers, and better bed adhesion.

---

Simon Khoury

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

Quote
DjDemonD
Polarity is mentioned in the manuals in several places but I will accept perhaps it is not prominent enough and correct that over the next few days.

The power connector polarity yes, not the piezo polarity. I think that a mention on the manual that watching the videos will complement the information would be good too.

Quote

Glad you got it working, very few people, if any have been unable to get it going and achieve great results. I think its a system which takes a few days to familiarise yourself with and get a feel for, but I feel worth the time spent to have super accurate first layers, and better bed adhesion.

yeah I haven't done any test print yet, but I used the 3 screws to level my bed and seems to be pretty good.

Also, does this mount qualifies for the bounty? Should I post it there too?

Perhaps what I should have said is adding a line about testing for, or looking out for, the pizeo connection if its behaving oddly. I did have the red piezo wire (own one i drilled out) to the +ve so the idea of the piezo being in "backwards" wasn't even a consideration, especially since it "appears" to work when tapping it. Your comment from July is what I eventually found and is something what should be mentioned, specifically the pressure applied/released test.

So far I have tested about 50 Murata discs and somewhat more no-name piezos. The Murata discs are consistent in polarity and each one has much the same output while the "no-name" discs are not so good although still capable of giving excellent results.
The main problems with "no-name" are:-

• Polarity and output are variable both between batches and in a batch. I think these must come from several factories and made to several different standards.
• The high temperature behavior of "no-name" piezos is very variable with some giving no output at all at 60°C
• The thicker wires can result in the wire fracturing at the soldered joint or even pulling the solder right off the disc.

So buy Murata if available. The only other tip I can think of is to buy piezos with wires on as soldering to the upper surface is an arcane art bringing much frustration to the unwary.

Mike

Quote
coredump

Quote
DjDemonD
Polarity is mentioned in the manuals in several places but I will accept perhaps it is not prominent enough and correct that over the next few days.

The power connector polarity yes, not the piezo polarity. I think that a mention on the manual that watching the videos will complement the information would be good too.

Quote

Glad you got it working, very few people, if any have been unable to get it going and achieve great results. I think its a system which takes a few days to familiarise yourself with and get a feel for, but I feel worth the time spent to have super accurate first layers, and better bed adhesion.

yeah I haven't done any test print yet, but I used the 3 screws to level my bed and seems to be pretty good.

Also, does this mount qualifies for the bounty? Should I post it there too?

Okay so there are two manuals - one for the Piezo20 (which until very recently) was only a ready-to-run product that I made and tested myself. And the manual for the piezo universal kit.

In the kit manual it has a big section on polarity. What I have not done is update the Piezo20 manual from the point of view of someone building it, preferring at this point to rely on the build video.

I will take another look at them with this new perspective.

As for the bounty have a look at the thread [forums.reprap.org] and if you can post a video, the details of where the mount is shared publicly and the accuracy/repeatability results. Then PM me and let me know which kit you want.

---

Simon Khoury

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

Quote
DjDemonD
As for the bounty have a look at the thread [forums.reprap.org] and if you can post a video, the details of where the mount is shared publicly and the accuracy/repeatability results. Then PM me and let me know which kit you want.

Does the google sheets thing I did count as repeatability since I don't have a marlin for M48? Or there's another method that is available on Duets?

Manuals for Piezo20 now updated.

---

Simon Khoury

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

Quote
coredump

Quote
DjDemonD
As for the bounty have a look at the thread [forums.reprap.org] and if you can post a video, the details of where the mount is shared publicly and the accuracy/repeatability results. Then PM me and let me know which kit you want.

Does the google sheets thing I did count as repeatability since I don't have a marlin for M48? Or there's another method that is available on Duets?

Yes thats fine, only marlin has a specific probe test, for other firmwares just do 20 probes and record the values and then give the range and deviation.

There is a method on duets see this thread [www.duet3d.com] where there are some sample gcodes that can run a test sequence. Just check it first before running it though as one of these went outside the travel limits on one machine I tried it on.

Edited 2 time(s). Last edit at 10/31/2017 11:47AM by DjDemonD.

---

Simon Khoury

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

Just to come back to this. I'm presently iterating on this design, which should work as a replacement carriage for any of the 2020 extrusion based printers (Mine is going on a TronXY X5S, but I can't see anyreason it wouldn't work on a CR10 etc.)
[cad.onshape.com]

So far I have what appears to be reasonable repeatability (I'm testing in a jerry-rigged dremel press as I need the printer working) but have some work to do on repeatability. Suspect it might end up somewhat sensitive to print settings/material usage.

Unfortunately I've popped the solder off the only disc I had - so going to be a while before I can do much more on it (yay for lack of electronics outlets in Belgium!).

Edited 1 time(s). Last edit at 11/05/2017 04:32AM by jebk.

I can send you more Piezos if you need them, sounds promising.

---

Simon Khoury

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

Quote
DjDemonD
I can send you more Piezos if you need them, sounds promising.

Thanks for the offer, but I already have a batch on order that should be here before I get chance to work on it properly again anyway.

I've been searching the forums, but I'm failing to find dimensional data for the Pizeo20 board. I've made attempts to measure the STL files for the groove mount, but would rather build a 3D model of the PCB as space is limited and my STL measurements are suspect.

This a variant of a D-Bot with ideas taken from the HyperCube design. The clamped hot end pivots on a buried 4mm shoulder bolt (axis is perpendicular to the front plate) and a 20mm piezo is pinched between brackets.

I'm intending to attach the Pizeo20 in the empty area above the right hand belt clamp.

Links to dimensional data or a 3D model would be appreciated. With sufficient information, I would be more than happy to generate a full 3D model of the Pizeo20 and share it with the community.

Pm sent

---

Simon Khoury

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