Z-Probe with touch plate / pin probe

188 views
Skip to first unread message

Fabian Kropp

unread,
Feb 14, 2023, 12:29:01 AM2/14/23
to OpenPnP
Hi,

I know there has been some development for a contacting probe nozzle.
As implementing the probe in the nozzle seems impossible for the LumenPNP, I wonder if there is support for an external touch probe.

Two possible things I got in mind:
1. touch plate like used in CNC machinery to measure / compensate tool length
2. BlTouch style "poker" that is offset from the nozzle(s).

Regarding accuracy I assume both could do fine. (less than 0.1 mm should be plenty)

Is there any support build in for one of these scenarios?

BR

Fabian

mark maker

unread,
Feb 14, 2023, 2:30:59 AM2/14/23
to ope...@googlegroups.com

> I wonder if there is support for an external touch probe.

No.

However, note that a contact probe nozzle is much more useful, as it can probe when placing parts on the board (measure parts heights and/or the precise board+solder height on first placement), or when picking (for the first time) from feeders (measure the exact feeder height). And it's fast, no extra movements needed.

But it would not be impossible to add touch probe support.

_Mark

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/ac20eeb0-4664-4280-a261-9dc99229f408n%40googlegroups.com.

fabian kropp

unread,
Feb 14, 2023, 2:34:18 AM2/14/23
to ope...@googlegroups.com
Is there a suggestion / sample to equip a contact nozzle on some machine like the Lumen?
Any ideas how that could work?
The Lumen has a sturdy spring - in contrast to what I assume the lightplacer has.

I really would like to upgrade to a contact probe, esp. with your explanation.

BR

Fabian 

mark maker

unread,
Feb 14, 2023, 4:43:17 AM2/14/23
to ope...@googlegroups.com

I always thought that I would try a piezo strip on a part that is somehow under force when the nozzle tip touches down. IMHO, that should work even without (much) modifying the design, especially when using 3D printed parts, as these will surely flex. I'm sure you don't need the huge mechanical contraption the Liteplacer uses.

But I have no actual experience or proof it works.

_Mark

Ian Arkver

unread,
Feb 14, 2023, 6:12:13 AM2/14/23
to OpenPnP
My fevered imagination pictures something like a modified Juki nozzle with an opto-interrupter fixed to the side of the top green disc and a sliver of metal strip attached to the side of the nozzle tip such that the opto is **just** clear when the tip is fully extended.

No use for measuring during actual placement though (unlikely to enjoy being rotated), and possibly tricky to tune precisely.

Ian

Aaron Eiche

unread,
Feb 14, 2023, 9:13:57 AM2/14/23
to ope...@googlegroups.com
My 3D printer uses a device very much like this to probe the bed height. The specific device is the PrecisionPiezo Orion, but there are some other implementations as well. It essentially uses the vibration of the nozzle hitting the bed to induce the piezo effect. I don’t know if you would get the same effect with the compliance of a PnP nozzle. 

fabian kropp

unread,
Feb 15, 2023, 2:02:59 AM2/15/23
to ope...@googlegroups.com
Hi,

I wrapped my mind around all of this bit now.

To me proper z alignment of the z axis would be by touching of on the eternal reference with a bl touch. Then touching off a precision touch plate - to have both ends measured in.

Tips and parts can then be measured with the touch plate. 

Height of feeder strips etc. Can still be touched of with the bl touch.

Looks like I need to get into Java development again after 15 years of absence O.o

Seeing this as a long term idea I would like to gather your thoughts on that. Or maybe reasons why this is not worth the trouble …..

Cheers 

Fabian

mark maker

unread,
Feb 15, 2023, 3:21:14 AM2/15/23
to ope...@googlegroups.com

Hi Fabian,

Not sure I understand exactly what you mean, some inputs anyways:

  1. Nozzle tip Z need not be super exact, because the nozzles or tips are spring loaded. Pro machine bang them down surprisingly hard. Some older ones even just used binary pneumatics!
  2. Some stiff nozzle tip changers (like the Liteplacer's) need rather precise Z, but that's just a question of the changer design and the lack of designed-in play and/or springiness in the material used (the Liteplacer's is made out of stainless steel, extremely unforgiving).
  3. Some DIY feeders suspend naked tape strips between two prongs, without any support under the part. They need super exact Z, otherwise picks will just "trampoline" parts out of the strip. But that's just a flawed feeder design, sorry.
  4. So if you design the rest of the machine right, there is no need for super exact Z.
  5. OpenPnP already has the Contact Probing Nozzle and the Z Probing function, two facilities you should be aware of, and not confuse.
  6. What you explained below as "bl touch" is likely already covered by Z Probing, although you'd lose the nice contact-less quality when compared to a laser displacement sensor.
  7. One thing that I would also explore: maybe Trinamic stepper drivers Stall Guard could be used for Contact Probing. Not sure how fast it could go when probing. That should work on direct drive nozzles, like belt, rack and pinion, or cam, but likely not on a Liteplacer with its screw drive (too much force there, its ancestry from a mill shows).
  8. Nozzles could also be Z calibrated using Auto-Focus: actually this is already done in calibration, but for the Z of the camera. Watch the video moment here. It is slow, but I guess that would be OK for a once-after-homing operation.
  9. I frankly, really don't see justification for an active touch plate. Surely not one of these expensive and ridiculous-precision CNC ones. 😉

_Mark

Ian Arkver

unread,
Feb 15, 2023, 4:04:52 AM2/15/23
to OpenPnP
One comment below:

On Wednesday, February 15, 2023 at 8:21:14 AM UTC ma...@makr.zone wrote:

Hi Fabian,

Not sure I understand exactly what you mean, some inputs anyways:

  1. Nozzle tip Z need not be super exact, because the nozzles or tips are spring loaded. Pro machine bang them down surprisingly hard. Some older ones even just used binary pneumatics!
  2. Some stiff nozzle tip changers (like the Liteplacer's) need rather precise Z, but that's just a question of the changer design and the lack of designed-in play and/or springiness in the material used (the Liteplacer's is made out of stainless steel, extremely unforgiving).
  3. Some DIY feeders suspend naked tape strips between two prongs, without any support under the part. They need super exact Z, otherwise picks will just "trampoline" parts out of the strip. But that's just a flawed feeder design, sorry.
  4. So if you design the rest of the machine right, there is no need for super exact Z.
  5. OpenPnP already has the Contact Probing Nozzle and the Z Probing function, two facilities you should be aware of, and not confuse.
  6. What you explained below as "bl touch" is likely already covered by Z Probing, although you'd lose the nice contact-less quality when compared to a laser displacement sensor.
  7. One thing that I would also explore: maybe Trinamic stepper drivers Stall Guard could be used for Contact Probing. Not sure how fast it could go when probing. That should work on direct drive nozzles, like belt, rack and pinion, or cam, but likely not on a Liteplacer with its screw drive (too much force there, its ancestry from a mill shows).
In my experience with StallGuard it's very hard to tune it to make it reliable and sensitive. Either you end up with it tripping out on normal acceleration, or it needs *quite a lot* of force to stall it.

Admittedly this experience wasn't with Pick&Place - it was a camera pan/tilt unit with worm gear drive, but still - I think it would be impossible to detect the *tiny* force at the start of a Juki nozzle spring compression using StallGuard.

Ian

mark maker

unread,
Feb 15, 2023, 4:27:42 AM2/15/23
to ope...@googlegroups.com

> In my experience with StallGuard it's very hard to tune it to make it reliable and sensitive. Either you end up with it tripping out on normal acceleration, or it needs *quite a lot* of force to stall it.

What driver IC? I think there a big difference between models.

_Mark

fabian kropp

unread,
Feb 15, 2023, 5:00:54 AM2/15/23
to ope...@googlegroups.com
Looks like I need to read and understand some more before going down that path.

Spring on the lumen is too strong and with the pcb as mounting board the board is a spring in itself (therefore I have higher pcb height as it is mounted flush on a HDF board.

Still I would like to improve my understanding and my equipment. ( esp as I have to modify my machine in any case due to narrow top camera….)

I will try to come up with a sketch of what I have in mind after reading and understanding the z probe and contact probe in a better way.

Ian Arkver

unread,
Feb 15, 2023, 5:11:59 AM2/15/23
to OpenPnP
TMC5160A

mark maker

unread,
Feb 15, 2023, 5:38:19 AM2/15/23
to ope...@googlegroups.com

Hmmm... external FETs. Not sure they're as sensitive as the integrated ones.

Like I said, I have no practical experience, except my Prusa 3D printer homing. It does that fast and reliable, but of course banging in frame is not the same as "feeling" minor spring forces.

_Mark

Shai

unread,
Feb 18, 2023, 1:58:43 AM2/18/23
to OpenPnP
My thought would be to just use a resistive sensor placed on the bed, put a component on it and push down. As soon as a force is felt, it will trigger. Very easy to implement. https://www.digikey.com/short/rtwh729q
You just need to set the initial 0 location precisely.

mark maker

unread,
Feb 18, 2023, 4:20:59 AM2/18/23
to ope...@googlegroups.com

If you insist on extra moves, i.e. don't care about it being slow.😉

_Mark

Reply all
Reply to author
Forward
0 new messages