Precision Piezo Z-probe Now available.
Posted by DjDemonD
|
Re: Precision Piezo Z-probe Now available. November 12, 2019 03:08PM |
Quote
MKSA
Quote
whosrdaddy
Underbed sensing is really only suitable for stationary beds (think Delta's). You can always put a piezo between the carriage and hotend assembly.
Works flawlessly on my P3 steel :
Suitable ? How ?.
For Delta, DC42 has a neat solution, a Wheastone bridge of strain gauges integrated in the effector.
Hi MKSA,
I don't understand your comment?
I was only saying that underbed piezo sensing is very difficult on a moving bed. (And it seems leadinglights might have a solution for that
)I know the smart effector but that was not my point...
|
Re: Precision Piezo Z-probe Now available. November 13, 2019 02:41AM |
Quote
whosrdaddy
Hi MKSA,
I don't understand your comment?
I was only saying that underbed piezo sensing is very difficult on a moving bed. (And it seems leadinglights might have a solution for that
I know the smart effector but that was not my point...
Explain how and why you find this system (when it works) suitable for a 3D printer ? Simpler, better, proven solutions exist.
Of course, for a Cartesian, the simple way to go is to build it properly. Better eliminate a problem than solve it isn't ?.
A sensor mounted in the carriage will make the Z0 limit switch and adjustment unnecessary. Just a plus, not a necessity.
Edited 1 time(s). Last edit at 11/13/2019 02:43AM by MKSA.
"A comical prototype doesn't mean a dumb idea is possible" (Thunderf00t)
|
Re: Precision Piezo Z-probe Now available. November 13, 2019 11:26AM |
Totally agree MKSA, a flat bed and a good microswitch is all you need.
The value is the fact that bed can be used as a probe, in my case I need this for my toolchanger as not all tools aren't perfectly aligned in Z
The value is the fact that bed can be used as a probe, in my case I need this for my toolchanger as not all tools aren't perfectly aligned in Z
|
Re: Precision Piezo Z-probe Now available. December 05, 2019 08:14AM |
Quote
Moriquendi
The amount of movement of the nozzle should be very very small, less than one microstep. If your nozzle has more movement than than then something is wrong.
Any mechanical nozzle probe technology is going to require dome degree of movement of the nozzle, the Orion PCB and mounting parts are designed to minimise this, I have not seen any system that uses the nozzle as the probe that has less movement than piezos.
Idris
I had MUCH more play there and I tweaked the mounting screws until I was ready to do some serious harm to the world in general
I also tried different mounts. The up/down play may not have been visible but the sideways play was easily 0.5 to 1 mm. Mounting screws were around 3/4 to a full turn loosened after they were lightly tightened up.
I have not given up on the piezo setup but needed a break from it. My setup, a CR10-S5, has a lot of mass compared to the mechanical members used and I was using a Chimera arrangement with dual copper volcano hot ends so there was a fair bit of mass. As mentioned, the biggest issue was the probing force required (hence going to the BLTouch). I could tweak things for a few grams on one spot but then get visually obvious bed and gantry deflection from much higher probing forces at a different spot.
Oh, I could easily get a trigger out of the system by looking at it sideways ... extremely sensitive .... but that didn't translate into less probing force for reasons that were beyond me.
|
Re: Precision Piezo Z-probe Now available. December 05, 2019 10:11AM |
I think I may be able to offer some possible reasons why your piezos may be playing up. These are only speculations and it may be something entirely different: Like the little girl in the nursery rhyme, when piezos are good they are very very good; but when they are bad they are horrid.
Warning, this may be Too Long Did Not Read, it was damn near Too Long Did Not Write.
I have used piezos on all of my printers and they have been remarkably consistent, but I have come across the occasional anomaly. The most recent problem is instructive. This was on an underbed setup using three piezos built into the bed leveling jacks and the problem was found during trials of a new conditioning circuit.
The plots below were taken directly from the buffer memory of the conditioning microcontroller. The first set of five plots show the contact of the nozzle with the bed at 3mm per second in the middle of the bed. A good contact occurs if the line between the Y value of 48 and 128 is clean and has no significant bend - in this case, the conditioner gives a green light. The problem occurred on only one corner of the bed at 2mm per second where I consistently did not get a green light. The plots from the microcontroller buffer were also very strange, as seen in the second plot.
Viewing the output from the piezo discs on an oscilloscope I got the same plot for each of the three sensors. I tried tapping with a fingernail, tapping with a fingertip, pushing down with a finger and dropping an 8mm steel ball from 10mm. In each case, the plot was sensibly as expected for all three sensors. With the nozzle making contact with the middle of the bed at 2mm per second the oscilloscope showed what was expected, a baseline with a sharply upward rise at the moment of contact. The first clue came when looking at an oscilloscope trace at the corner: Initially, the plot went down before rising - at very much the sort of rate shown in the earlier plot. The drawings below are from memory as I hadn't set up the scope to record.
When I took the piezo sensor assembly apart I found that I had assembled the pressure pad quite off-centre of the piezo disk. My excuse is that the pad and disk are not visible when being assembled and something slipped without me noticing it.
As to why the off-centre pressure pad gave such a peculiar output I can only speculate. I do know that it is possible for the output from a piezo disk put under pressure to be the opposite to what it is when bent - obvious when you think about it. Possibly the disk was being driven into the Kapton film that was under it so bent slightly before giving a pressure output - the thickness of a Kapton film with the adhesive is about 60 µm which is a lot in the world of piezo disks.
On the last point, I have attached two pictures, the first showing the nozzle bouncing on the bare metal bed at 2mm per second while the second shows that a Kapton film was enough to almost totally dampen the bounce with 4 out of 5 good contacts recorded - 60 µm is indeed a long distance in piezoland. Also notable is that the bounce gives rise to the possibility of the trigger point being found on different bounces each time.
Mike
Edited 2 time(s). Last edit at 12/05/2019 10:13AM by leadinglights.
Warning, this may be Too Long Did Not Read, it was damn near Too Long Did Not Write.
I have used piezos on all of my printers and they have been remarkably consistent, but I have come across the occasional anomaly. The most recent problem is instructive. This was on an underbed setup using three piezos built into the bed leveling jacks and the problem was found during trials of a new conditioning circuit.
The plots below were taken directly from the buffer memory of the conditioning microcontroller. The first set of five plots show the contact of the nozzle with the bed at 3mm per second in the middle of the bed. A good contact occurs if the line between the Y value of 48 and 128 is clean and has no significant bend - in this case, the conditioner gives a green light. The problem occurred on only one corner of the bed at 2mm per second where I consistently did not get a green light. The plots from the microcontroller buffer were also very strange, as seen in the second plot.
Viewing the output from the piezo discs on an oscilloscope I got the same plot for each of the three sensors. I tried tapping with a fingernail, tapping with a fingertip, pushing down with a finger and dropping an 8mm steel ball from 10mm. In each case, the plot was sensibly as expected for all three sensors. With the nozzle making contact with the middle of the bed at 2mm per second the oscilloscope showed what was expected, a baseline with a sharply upward rise at the moment of contact. The first clue came when looking at an oscilloscope trace at the corner: Initially, the plot went down before rising - at very much the sort of rate shown in the earlier plot. The drawings below are from memory as I hadn't set up the scope to record.
When I took the piezo sensor assembly apart I found that I had assembled the pressure pad quite off-centre of the piezo disk. My excuse is that the pad and disk are not visible when being assembled and something slipped without me noticing it.
As to why the off-centre pressure pad gave such a peculiar output I can only speculate. I do know that it is possible for the output from a piezo disk put under pressure to be the opposite to what it is when bent - obvious when you think about it. Possibly the disk was being driven into the Kapton film that was under it so bent slightly before giving a pressure output - the thickness of a Kapton film with the adhesive is about 60 µm which is a lot in the world of piezo disks.
On the last point, I have attached two pictures, the first showing the nozzle bouncing on the bare metal bed at 2mm per second while the second shows that a Kapton film was enough to almost totally dampen the bounce with 4 out of 5 good contacts recorded - 60 µm is indeed a long distance in piezoland. Also notable is that the bounce gives rise to the possibility of the trigger point being found on different bounces each time.
Mike
Edited 2 time(s). Last edit at 12/05/2019 10:13AM by leadinglights.
|
Re: Precision Piezo Z-probe Now available. December 06, 2019 11:27AM |
never mind
Edited 1 time(s). Last edit at 12/06/2019 12:05PM by dasneo72.
Edited 1 time(s). Last edit at 12/06/2019 12:05PM by dasneo72.
Sorry, only registered users may post in this forum.