Precision Piezo Z-probe Now available.

So there is an output on the 2.85 universal board that is a conditioned analog representation of the raw piezo voltage? Is 0V to VCC the range ignoring inverted polarity signal, or is it 0V (max piezo flex 1 way), 1/2VCC (piezo at rest), VCC (max piezo flex the opposite way)?

The at rest signal will be ~0.5VCC. The output signal is not just the piezo voltage but buffered, it's also filtered. You probably could get the output as high as VCC if you hit the piezo really hard but normal range is much lower. For example, on my printer with the duet, the threshold is set 5 points above the at rest reading which I think is a trigger threshold of 16mV above the at rest reading.

Idris

---

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

But is it differential is what I was interested in? IE swing below .5VCC flex one direction and above .5VCC flexed the other?

It will swing above and below 0.5vcc dependant on which way the board is flexed.

Idris

---

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

I am having trouble getting my Duet to use my Orion module as a z endstop. I am obviously missing something in the configuration but for the life of me I can't figure out what ....
Here are the relevant config.g parameters:


; Endstops
M574 X1 Y1 S1 Z1 S2 ; Set active high (S1) endstops for x and y, Z endstop is the z probe

; set up for piezo z probe
M558 P8 I1 R0.4 F400 ;digital mode, NO/NC, delay, speed
G31 X0 Y0 Z-0.1 P500 ;X, Y, Z offsets, debounce interval

I tried normal and inverted modes for the sensing. I have a mechanical Z endstop connected and that is what triggers the eventual z stop.
The Duet sees the Z probe as I can see the output value change as I tap the nozzle but I just can't get it to use the probe as a z endstop.

Probably something obvious to those of you who have the setup working but the answer has escaped me so far and I could use a hint.

Thanks

Quote
jens53
I am having trouble getting my Duet to use my Orion module as a z endstop. I am obviously missing something in the configuration but for the life of me I can't figure out what ....
Here are the relevant config.g parameters:


; Endstops
M574 X1 Y1 S1 Z1 S2 ; Set active high (S1) endstops for x and y, Z endstop is the z probe

; set up for piezo z probe
M558 P8 I1 R0.4 F400 ;digital mode, NO/NC, delay, speed
G31 X0 Y0 Z-0.1 P500 ;X, Y, Z offsets, debounce interval

I tried normal and inverted modes for the sensing. I have a mechanical Z endstop connected and that is what triggers the eventual z stop.
The Duet sees the Z probe as I can see the output value change as I tap the nozzle but I just can't get it to use the probe as a z endstop.

Probably something obvious to those of you who have the setup working but the answer has escaped me so far and I could use a hint.

Thanks

A Z probe is not the same as an endstop. Test and calibrate the probe as described at [duet3d.dozuki.com], then set it up to home Z as at [duet3d.dozuki.com].

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Thank you for that input! The overall concept in my mind seems to be a bit flawed but I suspect the second link you provided will clear that all up for me.
At the moment I am working on the first link you suggested. I have performed the static test and all is well. When I do the dynamic test however, the printhead never stops when using the G30 command yet a tap on the printhead has the Z probe output going from 0 to 1000 so it obviously sees the probe but just doesn't interpret the resut. I assume something in the config.g file is set up wrong but what ?

Getting closer to discovering what is going on - after homing the printhead, I can see that the z probe is active (by tapping the printhead) . For some reason though, when doing the single probe G30 command, the z probe is shut off (ie not reacting to tapping).
Again, it would appear that I have something set up wrong in config.g

Edited 1 time(s). Last edit at 06/20/2019 12:51PM by jens53.

This is driving me slowly insane. Looking at the probe I see the following sequence of events:
I home all
printhead goes to 0/0
At this point I have no Z endstop defined so it stays wherever Z was
I can tap the head and the red light goes off, the status panel shows the z stop going from 0 to 1000 and back to 0

Red light is on
I issue a G30
Red light goes off (before any movement)
Head travels down 5 mm
Head travels up 5 mm
Red light turns on

After the G30 command is finished and after the red light goes back on, tapping the printhead again triggers a '1000' change of the z probe status in the web panel.

It's as if the sensor is specifically turned off during a probe move.

The above sequence happens when there is (as far as I understand things) no z endstop set up so it happens wherever the printhead is at the moment.

If a z endstop is set up, the 'home' command first sets the head to 0/0/0.

@jens53

The Z probe can also be the endstop.

In my congig.g I have the following:

M574 X2 Y2 S0 ; set active low endstops
M574 Z1 S2 ; Set Z endstop controlled by Probe
M558 P8 H8 F780 I1 T9000 R0.8 Z1; set Z probe to digital, 8mm dive height, 780mm/min speed, 800ms pause and set as Z endstop
G31 X0 Y0 Z-0.1 P100; set Z probe trigger value and offsets

My endstops are active low, so that's different than yours, but the rest of it should work.

Then my homeall,g and mosez.g both use G30 to set the Z=0, and works perfectly well.

Thank you for posting your config details. I commented mine out, copy/pasted yours in and adjusted x and y for active high. I noted that you have a "Z1" on your M558 line which I did not have. The other changes seemed to be immaterial to the situation. Anyway, with your config details set up, the printer behaves as it did before Seems that I am missing something else in the setup. I will experiment with your setup as a starting point and maybe something will develop.
A quick question though: You refer to your homeall.g and movez.g files having G30 in them. I currently do not have those files customized in any way. I issue a G30 directly from the command console. Can you tell me if I should be able to issue that G30 directly or is there something in the other scripts that needs to be set ? Upon issuing the G30, the nozzle moves down but doesn't stop when I tap it .... should it stop like I am expecting it to do?

I stand corrected ... I was just doing another test and it actually WORKED !!!! WOOOOHOOOO
Now to figure out why it didn't work before!

Thank you SO much !!!!

Dang it .... false alarm. When I do a G30 the head only goes down about 5 mm before reversing. I can't manually trigger it to go less than that by tapping it

Additional data point ... I have now moved the G30 into my homeall.g file and it behaves the same as before. When the M558 I parameter is set at 0, the printer reports that the probe was already triggered at start of the probing cycle and when I is set to 1, the error reported is that the probe did not trigger during the probing move. The probe reports fine on the web panel before the G30 command and after the G30 command but just not during the actual G30 command

OK, I think I have it finally worked out. Although the probe seemed to work fine, it would appear it was set too sensitive. The mere act of the head moving down triggered the probe. I did not recognize the issue because the probe, after triggering, moves down another maybe 5 mm or so (why ??). It actually stops after triggering and then starts moving further down before reversing and moving back up. I am assuming this to be another setup issue someplace still to be found. The main story here is that the probe works fine but one has to be careful with sensitivity because it could be set so sensitive as to trigger by a butterfly landing on it
Thanks for the help!

I have another related issue that has cropped up now that I have the piezo sensor working.

When I home z, the printer does it's thing, moves up a bit stops and moves down to find z zero. I find that if I reduce sensor sensitivity enough so as not to trigger on the up/down movement, it ends up putting considerable pressure on the bed before it triggers. The bed probably moves 3 mm or more. I have even had occasion where there is no trigger and the steppers are skipping.

I could use some suggestions on how to change my setup so that the piezo sensor doesn't trigger when the z axis starts to move but yet is sensitive enough to recognize the bed before it depresses it by 3 or more mm's.

BTW, a light touch with my finger triggers the probe.

The most likely solution is to slightly loosed the four assembly screws, start with half a turn each and test again.

If you could edit your posts to add information rather than making a new post each time it would help to keep the thread easier and clearer to read.

Idris

---

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

I am using an Orion sensor with this [www.thingiverse.com] mount which is a derivative of the officially suggested Chimera/Creality mount. Although I don't recall exactly where, I seem to recall that the screws for this sensor can be tightened fully. I have tried loosening the screws as per your suggestion and it seems to have improved the sensing of the 'up' movement (where it is not wanted) with no improvement on the actual probing.
Could there be too much mass in my hot end ? I am using two heavy solid copper Vulcano hot ends instead of lighter aluminum hot ends. The odd thing is that a light tap with my finger triggers the sensor every time. Maybe the heavy mass combined with the acceleration of the z movement causes the initial trigger which requires reducing sensitivity? Is there a way to reduce acceleration for the actual probing move ?
Other suggestions ?

I apologize for making several separate posts in a row instead of editing the first post in the series.

Update: M201 is your friend. I set the acceleration of the Z axis to 1 (yes, 1) in homeall.g which allowed me to turn the sensitivity of the Orion way up. By the time the nozzle hits the bed, it is going fast enough to give the probe a nice tap and so far this seems to work.

G91 ; relative positioning
G1 Z5 F1000 S2 ; lift Z relative to current position
G1 S1 X-505 Y-505 F1800 ; move quickly to X and Y axis endstops and stop there (first pass)
G1 X5 Y5 F1000 ; go back a few mm
G1 S1 X-505 Y-505 F360 ; move slowly to X and Y axis endstops once more (second pass)
M201 Z1 ; use very slow acceleration
G1 S1 Z-505 F1800 ; move Z down stopping at the endstop
M201 Z10000 ; bring up default acceleration back up to some random high value
G90 ; absolute positioning
G92 Z0 ; set Z to axis minimum (to be fine tuned yet)

Note that there is no G30 in my homing files. I haven't got the G30 to work yet.

Edited 1 time(s). Last edit at 06/21/2019 11:42AM by jens53.

Ok, This is starting to make much more sense now.

I think you've got something of a perfect storm of conditions going on.

Additional mass on the hotend adds to the inertia below the Orion. When the printer starts moving (whether the head moves or not) there will be a vibration, this vibration is transmitted through the Orion and with more mass on the hotend it tends to produce a larger signal in the sensors. For this reason it's likely that you will need to reduce the acceleration in Z drastically and reduce jerk to 0. I'm using a prototype with a Titan aero mounted below the sensor, that's something like 200g IIRC and I'm using a Z acceleration of 5mmS^-2 with 0 jerk. I have a line in my homez.g and homeall.g files that reduces the acceleration before probing and resets it after:

; Reduce Z jerk and acceleration
M201 Z5
M566 Z0

Probe code goes here

; Reset Z jerk and acceleration.
M201 Z50
M566 Z5

Second, that Chimera mount you're using is probably too rigid, all the current hotend mounting parts have cutouts to increase Z-axis compliance, try my version (second model) here and see if there's any improvement.

Conversely, the top part of your mount needs to be as rigid as you can make it.

Probing speed does not need to be very high, I probe at ~3mmS^-1 and accuracy generally improves at lower probe speed.

Idris

Edited 1 time(s). Last edit at 06/21/2019 03:48PM by Moriquendi.

---

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

Thank you Idris.

I have not yet considered jerk settings but I will keep that in mind.
I did try the mount you referred to but it did not work for me (unrelated to actual operation). I do not recall any more why it was not suitable. The mount I use is essentially the same as yours (including the cutouts) with the section attached to the Chimera being split into two. The top part of the mount I am using is quit rigid so I am good there.
I will play a bit with probing speed as I have mine fairly high to give more of a jerk to the probe meeting the base. I run at 1000 mm/min rather than the recommended 300-400 mm/min (or yours at 180 mm/min). I will also see if I can increase the acceleration a bit when I reduce the probing speed. Maybe I can match what you have.
The settings as they are now are ok for establishing z zero but if I do a bed mesh calibration that would put a severe crimp into things. I doubt that I could measure 400 mesh points in a reasonable time period with the current settings. I would like to run that many points because the printer is 500 mm* 500 mm and that is a modest 25 mm between points. I have not yet tried the mesh bed calibration ... got to get the rest working first.

I am starting to see light at the end of the tunnel and am quite excited about the new printer I am building. I am even more excited about the new bed sensing technology and the hopefully as close to perfect as is possible first layers.

Update:


The light at the end of the tunnel turned out to be a flare sent up by someone else that got lost in the tunnel
I suspect the issue is with the mount but I can no longer justify playing with this and I do have a BLTouch in my spare parts so that will have to do for now.

The details:
I have tried 200/500/1000 mm/min for speed. I have had the sensor so sensitive that if I touched the workbench the printer sits on it would blink. It is incredibly sensitive.
The problem is that if it is set really sensitive it stops at random positions during the probe run - I am assuming roughness in the mechanics or more likely the stepper motor stepping action. If I reduce the sensitivity enough so it doesn't trigger during the downward probing run, it pushes the bed out of the way before it finally stops (I have the motor drive reduced to bare minimum to avoid nasty surprises).
For a while there I thought I had it sorted but when I looked closer I found out that the head always stopped two or three mm above the print surface. This didn't make sense to me so I searched for another cause. As it turned out, the mechanical end stop was interfering. The switch is on the side of the printer (Creality CR10 S5) and the gantry activates the switch when it drops down low enough. The action of the gantry as it hit the stop was jarring enough to repeatably cause the sensor to activate as the mass of the hotend was brought to a sudden stop. One would think that if that (indirect) process works repeatably then surly the printhead hitting the glass of the bed should cause a similar action ... alas, it doesn't. The glass is sitting on an aluminum plate and the whole thing sits on springs and the nozzle touches the plate gently enough to prevent the jarring action needed.
I did my initial tests at xyz 0/0/0 which has no direct support underneath the aluminum the glass plate sits on and the bending is quite pronounced (both gantry and the build surface bend). I thought that maybe the lack of support at that spot is causing issues so I started testing a ways into the build platform but had the exact same result.
The one thing I have not tried and that is just occurring to me now is that the glass plate has a thin plastic wrapper around it from shipping. We are talking maybe 0.1 mm, only there to prevent finger marks. The nozzle punches right through it. I doubt it will make a difference but in the interest of being thorough I will remove the plastic and try that.
I still think that this system is showing great promise but the combination of the mass and the flex in the system in combination with the mount make this method of Z sensing not practical. Since neither the printer mechanics nor the mass will change (what about people with direct extruders ?), the only possible variable remains the mount. Even changing the mount seems a long shot since I can tune the system to trigger when a fly buzzes past and poops onto the printhead

I am open to comments but will be working on a new Chimera to CR10 mount shortly and then I will no longer be able to experiment with the current setup.

Edited 3 time(s). Last edit at 06/21/2019 06:27PM by jens53.

I have another update (I couldn't tack it onto the last post as the edit window apparently expired)

A fresh day, a top-up of my patience bucket, lots of screwing around and it might look like I have a working sensor setup after all.
I do not know which change in particular did the trick or if it was a matter of all together. Also, it's early days yet and I am not sure about reliability yet.
The z homing routine is now:

G91 ; relative positioning
G1 Z5 F500 S2 ; lift Z relative to current position
M566 Z0 ; set jerk to zero
M201 Z5 ; set Z acceleration to 1 mm/sec/sec to prevent axis jerk that triggers the z probe
G1 S1 Z-505 F700 ; move Z down stopping at the endstop
M201 Z100 ; restore normal acceleration
M566 Z5 ; restore normal jerk parameter

I found that there were some harmonics developing at different Z speeds and I had to work around that but speed settings still need to be optimized.
I will readjust the V wheels of the gantry in case they are not totally right.
I also backed off the mounting screws about 3/4 of a turn as earlier suggested (thank you!)
One interesting thing that was a totally new discovery - the front left foot on the printer, which is close to 0/0/0 was not fully supported letting the frame potentially twist a bit. The sensor just seemed more sensitive to touching the frame in that corner. I ended up shimming it a bit.
Lower probing speed still results in more pushing down the build plate and pushing up the printhead and gantry but at 700 I seem to get a fairly reliable setpoint of -0.1 mm (not incorporated in the script yet).
What I want to do next is do maybe 50 moves to Z300 followed by a z probe to see if it triggers from system vibration. I would also look if Z zero was reliable. If that works then I can run a bed compensation routine and hope that errors stand out when I look at the overall bed map.

Ordered a plain Orion PCB to support the good efforts of the development team, and with the hope of creating a mount to my modified CR-10s.

Using a MGN12H carriage for X with a aluminium T-profile attached as the mount for the hotend. Attaching a rough 2 minute sketch of how i'm thinking of adapting it.

Still in the design-phase of my mount and had a few questions to guide my efforts.

My questions:
- How much do need the PCB to flex for accurate readings? (And what is the max limit of flex you should allow for in the design to prevent flex-damage of the PCB?)
- Any progress on the stock CR-10 hotend mount that could be shared (in design or in concept). Have not figured out a good way of allowing screw-mounted hotends to flex in z-only direction so far.
Might just order a E3D v6, but hopefully not, if i am unable to come up with a good enough design for the swiss micro hotend and it's problematic screw-mount.


Any tip's or other suggestions are welcome, design will be on thingiverse as soon as it's been proven.

I am still struggling with z stop by probe working but homing not triggering the probe.
The one thing that I have completely overlooked is the electrical to Duet connection. Specifically, the duet has 4 inputs and the Orion board has three connection pins. I currently use the 3.3V line of the Duet to go to the V+ connection on the Orion, the ground pin from the Orion goes to the middle connector on the Orion and lastly I use the z-probe-in connector that goes from the Duet to the Orion signal pin.
I have assumed that since the orion triggers well in this configuration, that all is well. It never occured to me that maybe I should use the z_probe_mod pin on the Duet.
Could I please get confirmation of which pin on the Duet connects to the output of the Orion ?

Again, in homing mode the status screen reports a Z probe value of 0 except when it triggers when it goes to 1000. In probing (G30) mode, Z probe value stays at zero even if I have a woodpecker doing his best to get any bugs out of the probe ....

MAJOR UPDATE:

At the risk of getting a reputation for an 'old fart' that makes 'old fart mistakes', I have fixed my problem. No wonder I was unable to find a solution on the internet because nobody in their right mind would admit to reversing the z-probe-mod line with the gnd line which is what I managed to do.
I have never considered the wiring because the homing worked ... go figure ..... key thing is G30 now TRIGGERS when I poke at the nozzle.

I still would like to know what z_probe_mod line is for.

Edited 1 time(s). Last edit at 06/25/2019 11:32PM by jens53.

@jens53: kudos to you for disclosing your error and how you fixed it. we ALL learn from EACH of our mistakes. there is no shame in it.
sorry i cant help with the extra wire.

I have just installed my 2.85 board and piezo, but having a few issues.

I've set it up to be what i believe is optimal sensitivity as in the manual.

If i knock/tap the hotend, i can get it to trigger (change from red to blue), but the printer only responds to this sometimes.

On further investigation, if i tap the nozzle, i can get it to trigger on the leds, but not the printer, but equally, i can also tap the nozzle, and the printer responds but the leds are not indicating a trigger.

when homing Z, the same occurs, sometimes it will react and think its touched the bed, other times it takes several knocks to prevent it from diving in to the bed.

I really could do with knowing - if the LEDs are indicating a trigger, why is it not dropping the voltage everytime? and why is my printer able to recognise a tap (voltage drop) when the leds did not indicate?

Have i got a faulty board? my order number is 11827 if that helps.

thanks in advance!

Simon

Hi Simon,

Can you give me some more information please?

What controller board are you using? What printer?

Which pins do you have connected on the Universal PCB? There are two sets of pins, those marked D and another set marked A.

Idris

---

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

thanks Idris,

Its a heavily modified Geetech A10 - the controller board is a lerdge x 32bit. Been happily using 3dtouch and bltouch probes but fancied the piezo.

Board details:
[www.lerdge.com]

autoleveling details:
[www.lerdge.com]


i didnt realise there was 2 sets of pins - its connected to A (i followed the pictures to connect)

I need to update the manual which still refers to the v2.75 board. For now, use the D pins, they will behave the same way as the v2.75 manual expects.

Idris

---

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

yay!

thank you - thats solved it perfectly!

Excellent, glad to hear it.

To answer your question, the A pins output an analogue signal which bypasses the onboard trigger circuit, you can plug this into an analogue capable pin on your controller (currently only tested on the Duet) and set the sensitivity using G-code or firmware.

Idris

---

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