Piezoelectric disks for Z contact detect and bed levelling
Posted by leadinglights
|
Re: Piezoelectric disks for Z contact detect and bed levelling February 16, 2017 06:20PM |
Lykle's effector system for piezo hotend, working:
Will be published shortly.
No nozzle wobble and clear triggers.
We will adapt this design for some other mounting options soon.
Simon Khoury
Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions
Will be published shortly.
No nozzle wobble and clear triggers.
We will adapt this design for some other mounting options soon.
Simon Khoury
Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions
|
Re: Piezoelectric disks for Z contact detect and bed levelling February 16, 2017 10:18PM |
Any chance of getting that setup as openscad customizable or thingiverse customizer or anything like that? I'm drooling to try this but my rod spacing is unique.... 
|
Re: Piezoelectric disks for Z contact detect and bed levelling February 18, 2017 06:38AM |
I have been investigating the behaviour of piezoelectric disks used in the configuration that DjDemonD uses and have got some useful traces which show just how sensitive this can be.
To do this I modified a jig that I made for earlier work. Instead of being compressed between two surfaces the disc was suspended around the outer edge and pressure was applied to the centre, the disc having been drilled 4.5mm and a small pad made to fit the hole. This should be a fair approximation to the way DjDemonD uses. The jig was preloaded with a force of 5N, about 500 grams to approximate the weight of a hot-end and small extruder. The drilledd disk with edge ring and centre button is on the left in the picture below.
The center pad was pushed down at 1mm per second for 20µm, 40µm, 60µm, 80µm and 0.1mm at a speed of 1mm/sec. The traces that I got are shown below. Scope set 100ms/cm and 2V/cm
The second picture shows the output change with different speeds, the traces being at 0.5mm/sec, 1.0mm/sec, 1.5mm/sec, 2.0mm/sec, 2.5mm/sec and 3.0mm/sec. Scope set to 20ms/cm and 2V/cm The small wiggles in the line are caused by the steps from the stepper motor - 1.8°, half stepping and a 1mm leadscrew so each wiggle is 2.5µm.
Incidentally, I tested the three piezo disks that I fitted to one machine back in March last year and have used regularly since: One disc has lost about 10% of it's mojo but the others are not measurably different.
Mike
To do this I modified a jig that I made for earlier work. Instead of being compressed between two surfaces the disc was suspended around the outer edge and pressure was applied to the centre, the disc having been drilled 4.5mm and a small pad made to fit the hole. This should be a fair approximation to the way DjDemonD uses. The jig was preloaded with a force of 5N, about 500 grams to approximate the weight of a hot-end and small extruder. The drilledd disk with edge ring and centre button is on the left in the picture below.
The center pad was pushed down at 1mm per second for 20µm, 40µm, 60µm, 80µm and 0.1mm at a speed of 1mm/sec. The traces that I got are shown below. Scope set 100ms/cm and 2V/cm
The second picture shows the output change with different speeds, the traces being at 0.5mm/sec, 1.0mm/sec, 1.5mm/sec, 2.0mm/sec, 2.5mm/sec and 3.0mm/sec. Scope set to 20ms/cm and 2V/cm The small wiggles in the line are caused by the steps from the stepper motor - 1.8°, half stepping and a 1mm leadscrew so each wiggle is 2.5µm.
Incidentally, I tested the three piezo disks that I fitted to one machine back in March last year and have used regularly since: One disc has lost about 10% of it's mojo but the others are not measurably different.
Mike
|
Re: Piezoelectric disks for Z contact detect and bed levelling February 18, 2017 07:50AM |
Yes its the mechanical noise I am referring to. There is one development on Smoothie (which I am a total n00b at) you can set debounce_ms 1 on z-probe, this massively improves probing, reducing false triggers to almost 0 at full speed, and to 0 at 60% speed. RRF can do this but I haven't yet worked out how to do this with the piezo setup as I am running it as an analogue sensor on RRF, with the P parameter used to set threshold, whereas David says you can use P as debounce, but this is only when using the sensor in digital mode, which each time I've tried it on RRF it fails, but time to revisit this I think.
Sorry for the crap-arse diagram but hopefully it makes sense, I use a 2mm lip to hold the rim of the piezo, with the piezo-electric element facing downwards, with a 0.5mm-1mm recess above.
The smallest ring possible on the clamp would be 4.5mm ID 6.5mm OD.
If you can try with and without a 5mm hole drilled in the dead centre (spur point wood drill works a charm for me).
Thanks for doing this it should quickly determine if 20mm piezo is feasible.
Simon Khoury
Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions
Sorry for the crap-arse diagram but hopefully it makes sense, I use a 2mm lip to hold the rim of the piezo, with the piezo-electric element facing downwards, with a 0.5mm-1mm recess above.
The smallest ring possible on the clamp would be 4.5mm ID 6.5mm OD.
If you can try with and without a 5mm hole drilled in the dead centre (spur point wood drill works a charm for me).
Thanks for doing this it should quickly determine if 20mm piezo is feasible.
Simon Khoury
Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions
|
Re: Piezoelectric disks for Z contact detect and bed levelling February 20, 2017 03:38PM |
Here's a link to order the boards from OSHpark
If you use the OSHpark link I'll get to know how many people order which I quite like the idea of.
The ones I'm making are no longer financially supporting the charity, I cant keep buying the parts to make them and from my point of view, giving them away.
If you make a donation to the charity that is always welcome Great Dane Care
If you choose to make a donation for the work I've done that is also welcome (paypal, idris_nowell15 at hotmail.com)
On the other hand I made the board layout and schematic freely available to everyone and there is no requirement on you to do either.
Idris
Edited 3 time(s). Last edit at 02/20/2017 06:20PM by Moriquendi.
If you use the OSHpark link I'll get to know how many people order which I quite like the idea of.
The ones I'm making are no longer financially supporting the charity, I cant keep buying the parts to make them and from my point of view, giving them away.
If you make a donation to the charity that is always welcome Great Dane Care
If you choose to make a donation for the work I've done that is also welcome (paypal, idris_nowell15 at hotmail.com)
On the other hand I made the board layout and schematic freely available to everyone and there is no requirement on you to do either.
Idris
Edited 3 time(s). Last edit at 02/20/2017 06:20PM by Moriquendi.
|
Re: Piezoelectric disks for Z contact detect and bed levelling February 21, 2017 11:17AM |
I received the 20mm piezo discs this morning and have checked one of them - scope plots below.
Although there are some anomalous differences, I think that there is enough of a signal to be optimistic.
The first set of traces gives a narrower pulse which is as I would expect with a lower electrical capacity. The only strangeness here is that the tail end of the pulse goes negative - this was not seen on the 27mm disks.
The second set is a bit stranger as the pulse width is not narrowed as much as the first set, I would again have expected a faster decay. There is again the negative going tail end and also a little notch in the three fastest traces, I think this may indicate that the plastic ring was too close a fit.
I checked the compliance of the disks by mounting the jig in a vice and using a DTI to check the deflection when I put on an additional 500 grams of load with a dynamometer (spring scale)
The deflection I got with the 27mm disks was 0.035mm and 0.04mm with the 20mm disks - I would have expected a little less but possibly the brass is thinner.
One other thing that was not too good is that only 4 of the disks would be capable of having a 4.5mm hole put in the middle, the other 6 had the soldered connection too close to the middle.
Mike
Although there are some anomalous differences, I think that there is enough of a signal to be optimistic.
The first set of traces gives a narrower pulse which is as I would expect with a lower electrical capacity. The only strangeness here is that the tail end of the pulse goes negative - this was not seen on the 27mm disks.
The second set is a bit stranger as the pulse width is not narrowed as much as the first set, I would again have expected a faster decay. There is again the negative going tail end and also a little notch in the three fastest traces, I think this may indicate that the plastic ring was too close a fit.
I checked the compliance of the disks by mounting the jig in a vice and using a DTI to check the deflection when I put on an additional 500 grams of load with a dynamometer (spring scale)
The deflection I got with the 27mm disks was 0.035mm and 0.04mm with the 20mm disks - I would have expected a little less but possibly the brass is thinner.
One other thing that was not too good is that only 4 of the disks would be capable of having a 4.5mm hole put in the middle, the other 6 had the soldered connection too close to the middle.
Mike
|
Re: Piezoelectric disks for Z contact detect and bed levelling February 21, 2017 01:20PM |
Nice going Mike!
That ties in nicely with an integrated Nimble effector I have been working on.
The effector houses an integrated peizo sensor which the back of the Nimble rests on. Underneath the sensor is a little space so the sensor can bend.
Normally the Nimble is already used to clamp the hot end and this allows the sensor to sense the tap on the bed. The extruder is used here as the force transmitting device. The nice thing is that the sensor does not need to be modified in any way. Another benefit is that it takes very little space, so I am sure I can fit it to a lot of the other adapters I have designed so far. The biggest problem I had was making sure I had space to fit all the screws and maintain stiffness of the effector, while still allow the fan of the V6 to be positioned correctly. That took a while but I am confident I have it now.
I still have to play with the tolerances in the disc housing and see how much space needs to be left open for the sensor to be able to bend enough, But with a deflection of 0.04 mm, I am not too worried. I am hampered by the lack of a scope, but I will try and cobble something together. Or perhaps just use my DuetWifi as a scope.
So, in theory it will work, but you know what they say. In theory there is little difference between theory and practice. In practice....
Lykle
________________________________________________
Co-creator of the Zesty Nimble, worlds lightest Direct Drive extruder.
[zesty.tech]
That ties in nicely with an integrated Nimble effector I have been working on.
The effector houses an integrated peizo sensor which the back of the Nimble rests on. Underneath the sensor is a little space so the sensor can bend.
Normally the Nimble is already used to clamp the hot end and this allows the sensor to sense the tap on the bed. The extruder is used here as the force transmitting device. The nice thing is that the sensor does not need to be modified in any way. Another benefit is that it takes very little space, so I am sure I can fit it to a lot of the other adapters I have designed so far. The biggest problem I had was making sure I had space to fit all the screws and maintain stiffness of the effector, while still allow the fan of the V6 to be positioned correctly. That took a while but I am confident I have it now.
I still have to play with the tolerances in the disc housing and see how much space needs to be left open for the sensor to be able to bend enough, But with a deflection of 0.04 mm, I am not too worried. I am hampered by the lack of a scope, but I will try and cobble something together. Or perhaps just use my DuetWifi as a scope.
So, in theory it will work, but you know what they say. In theory there is little difference between theory and practice. In practice....
Lykle
________________________________________________
Co-creator of the Zesty Nimble, worlds lightest Direct Drive extruder.
[zesty.tech]
|
Re: Piezoelectric disks for Z contact detect and bed levelling February 25, 2017 02:14PM |
Sorry, only registered users may post in this forum.