Re: Juki versus Samsung Nozzles

[Jason von Nieda]

Okay, this conversation re-piqued my interest, and it's been a few years since I dug around in the nozzle box, so I decided to present sort of a visual history of the various nozzle systems I've played with. I'm also cross posting this to the OpenPnP mailing list for reference purposes:

From left to right:

1. Madell Nozzle System which uses Fuji CP6 nozzles. Purchased as a comparison to #2.

2. 5mm to 5mm shaft coupler for use with Fuji CP6 nozzles.

3. Press fit Juki nozzle adapter. This one used set screws to hold the nozzle in. I think these were made by Thomas Langhans.

4. Custom quick change Juki nozzle adapter. I think this is the one I was referring to as having custom made in a shorter length.

5. Brass quick change Juki nozzle adapter. I believe this is an early protocol from Peter Betz, as evidenced by the 3D printed spring holder.

6. Threaded Samsung CP40 nozzle holder. This is available on Robotdigg, and I believe is the same one Juha sells with the LitePlacer now.

7. Threaded Samsung CP45FV nozzle holder. From Robotdigg [1] [2]. I've never used this but it seems quite nice.

8. Threaded Samsung CP45NEO nozzle holder. From Robotdigg [3] [4]. Also never used. Note that the main difference seems to be on the 45NEO the spring is in the nozzle, and on the 45FV the spring is in the holder.

9. (In separate image) This is the Juki holder on the Charmhigh CHM-T36VA. Seems to be quite similar to #3, except uses ball bearings and a rubber band to retain the nozzle.

If anyone has any questions about any of these, let me know. Happy to take measurements, provide sources, etc.

[1]: https://www.robotdigg.com/product/796/CP45NEO-or-CP45FV-Holder

[2]: https://www.robotdigg.com/product/801/SMT-CP45FV-pick-up-nozzle

[3]: https://www.robotdigg.com/product/796/CP45NEO-or-CP45FV-Holder

[4]: https://www.robotdigg.com/product/797/SMT-CP45NEO-pick-up-nozzle

Thanks,

Jason

[Lance Endres]

I suggest this be the first entry in the future Museum of OpenPnP.   :-)

Lance

[betzt...@gmail.com]

#5 would have come from Ray as evidenced by the 3D printed part. Mark Harris worked on the design, Ray did the first run, and then I took over manufacturing as the runout was not within design spec (this is not easy to accomplish, even now I have troubles getting this done within spec). They are now made on stupid expensive Swiss style CNC lathes. 

Peter.

[ma...@makr.zone]

Is there any reason all the Jukis are on small steppers and the Samsungs on large ones? Thinking about what they have to do, aren't the latter ones overmotorized?

_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/2642eeea-708c-4df0-8b8c-b5095e0ce2c3n%40googlegroups.com.

[Jason von Nieda]

Thanks Peter! And, of course, sorry to anyone I've forgotten or misattributed. It's been quite a few years :)

Jason

[Jason von Nieda]

Hi Mark,

All of the Samsung ones have external threads, and they thread into the motor shaft. The Juki ones are either press fit or slip fit with set screws. Since the Samsung ones have external threads they need larger shafts to thread into, thus larger motors.

That's the technical answer: the pragmatic answer is that that's how Robotdigg sold them to me :)

Jason

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/effdb21d-0e53-81ad-ee15-522ebbab7cef%40makr.zone.

[ma...@makr.zone]

Interesting, thanks! So that's a point for Juki, I think.

_Mark

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CA%2BQw0jzVPNbCPG57hJPNJYoK_db-vwOKAEeq5AQz9bP4%3D3u%3Dxg%40mail.gmail.com.

[Lance Endres]

>> Interesting, thanks! So that's a point for Juki, I think.

I'd go so far as to say the ability to connect your nozzle adaptor to the motor is a 3-pointer.  :-)

On that topic, would it be correct to say that Peter and the others developed the Juki nozzle before OpenPnP had run-out correction?  Not to take away from the impressive work they did or demission the importance of manufacturing accuracy, but I am wondering if this is, strictly speaking, necessary now?  As Peter states, it is difficult and as we know difficult equals cost.  I'm thinking here of an adaptor for the Samsung holder (and how accurate/costly it would have to be).  If we had an adaptor that worked on the smaller stepper motors it might "level the field" so to speak and maybe we would see more use of the Samsung nozzles?

-Lance

[Jason von Nieda]

Hi Lance,

IMHO yes - if we had a high quality Samsung nozzle adapter and a good source of nozzles, I think they would take off. They make nozzle changes much easier and the nozzles are *much* simpler mechanically. I would expect that the Samsung nozzles could easily be made on a lathe to good tolerance much easier than a Juki nozzle could, for instance. The Samsung nozzles are a single part while the Juki nozzles are 5 or 6 parts that must all be concentric.

And yes, one of the driving factors in Peter's very high quality adapters was fixing runout, before we had runout compensation. I made an early attempt at writing runout compensation and it worked for me, but it seemed it did not work for everyone. Tony and Mark, IIRC, then made massive improvements to it. I'm not sure how many people use it. I personally do not use it, and I prefer not to have to.

Jason

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/5dcad666-132b-4e64-94f7-7a7c06809b12n%40googlegroups.com.

[Betz Technik Industries Ltd.]

Well in my mind, the spring loaded Z axis is the major advantage for the Juki. If I were to make a Samsung holder (CP40?), I would think I would need to make it have a similar sliding Z axis. Otherwise Z height while running needs to be a lot more accurate.

 

Am I missing something? Or am I overestimating the importance of the sliding axis??

 

Peter.

[Jason von Nieda]

The Samsung adapters in my pictures have a spring. For the CP40 the spring is in the nozzle adapter, and for the NEO it's in the nozzle. I think we'd want to target the CP40 since the nozzles are extremely simple. So, yea, we'd need the nozzle adapter to be spring loaded.

Here's a better picture of the commonly available one so you can see the spring area:

I haven't measured the runout on this adapter. It could actually be fine. It feels quite nicely built and I can't feel any play in the spring mechanism. I'll measure with a dial indicator this evening.

Jason

--

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/D4879A91-E6F3-4672-873F-6FED6F8D7A8B%40hxcore.ol.

[Anthony Clay]

I'm curious, what was the runout on the CP holder? - I'm considering using four of them attached to machined hollow shafts for a custom head.  

[Betz Technik Industries Ltd.]

I’m curious too, but getting next to no TIR between two sets of threads seems unlikely.

 

Peter.

[Jason von Nieda]

So, I went to test the runout and it turned out I didn't need to, because the thing flops about so wildly it's clear the runout makes it useless.

I think centering with threads is a dead end.

Jason

--

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/45B3AFE3-1CED-4881-B0AC-B2BF529FDA57%40hxcore.ol.

[tony...@att.net]

I believe the CP holder has a tapered seat that is designed to fit into a matching taper in the threaded shaft.  So it is the concentricity of the tapers that control the runout of the nozzle holder rather than the threads themselves.  I can confirm that the CP holder is indeed spring loaded (in the Z direction).

Tony  

[Jason von Nieda]

The one I have has a tiny bevel but it doesn’t seem to do much. See attached.

--

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/ab4d0a05-b60f-48c8-848f-fc6a27802257n%40googlegroups.com.

--

Sent from my BeOS enabled toaster

[Peter Betz]

Yes that is what I’d expect for sure.

So what do you think... should I design a CP40 spring loaded holder that fits on a 5mm shaft??

Peter Betz
Betz Technik Industries Ltd.
BETZtechnik.ca

> On Mar 9, 2021, at 8:21 AM, Jason von Nieda <ja...@vonnieda.org> wrote:
>
> 

[Thorsten Wilmer]

Hi Jason

I'm sorry to see this with the bare eye. I didn't install my cp40 holder yet, so that I can turn it above the camera. But just from manual turning I don't see such a bend as in your video. I'll  post some photos once it's installed.

But I guess one would need more data to know what is the chance of garbage if you order one from robodig/aliexpress.

Kind Regards

 Thorsten

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CA%2BQw0jyPQ%2Bsi-shOUHYep5-63zKYpi72JaKunTk3zqOVU4nEdA%40mail.gmail.com.

[Michael Damschen]

Lance and I have developed and adapter for the NP40 to a NEMA 8 motor.  He posted some drawings earlier.  See link to discussion below.

I don't believe it would need to be spring loaded as Peter mentioned as the nozzle itself is already spring loaded.

https://groups.google.com/g/openpnp/c/7J8TxfwpKX4/m/dCFGyPXgAQAJ

[Jason von Nieda]

Hi Michael,

It looks like this would have the same issue as what I showed in the movie and images, in that there is no reference surface to make the nozzle holder concentric with the motor shaft. Have you found a way to resolve this?

Thanks,

Jason

[Michael Damschen]

We are building prototypes now, but gut feeling is it shouldn't be too difficult to solve.  Commercial products used the NP40 nozzles and address run out issues, if any.  But I could be overlooking something, but running these tests are the fun part. :)

[Jason von Nieda]

Hi Michael,

The nozzles are fine, I think. It's just this nozzle holder that is sort of junk. It's possible I just got a bad one.

Just so I'm clear, you are designing essentially a 5mm slip fit to M5 thread coupler, right? And the intention is to use the commercial nozzle holder and nozzles from e.g. Robotdigg?

Jason

--

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/9b8c3a4a-0a5c-4d3b-9309-c10dd9374c8fn%40googlegroups.com.

[Michael Damschen]

Yes, that is the adapter.  The plan is to use it with any NEMA 8 motor with any NP40 (Samsung) nozzle.

[Lance Endres]

I think some of the nomenclature got confused.  I have been using the term "coupling" for what we are working on to avoid confusing it with the holder/adaptor (see image below).  But, yes, it is essential a slip-fit to M5 connection for the commercial nozzle adaptor.  As Mike said, since these adaptors are used on commercial machines (and the LitePlacer), we are inclined to believe they work fine but the issue is getting them connected to a NEMA motor with enough accuracy.  Perhaps we are wrong, but we intend to find out.

[zart...@gmail.com]

Why not have an area between the coupling and adapter that contacts the bevel just past the threads and/or the surface of the tightening/wrench area? If those surfaces are more consistent, would it not straighten itself out, once tightened?

[John Plocher]

On Tue, Mar 9, 2021 at 10:02 AM zart...@gmail.com <zart...@gmail.com> wrote:

Why not have an area between the coupling and adapter that contacts the bevel just past the threads and/or the surface of the tightening/wrench area? If those surfaces are more consistent, would it not straighten itself out, once tightened?

What about adding a wrap of teflon tape to eliminate the wiggle and slop?  It should also help center the adapter...

 -John

 

[tony...@att.net]

Jason,

From your movie it doesn't appear that the bevel on the holder is in contact with the mating bevel on the motor.  Could the threads be bottoming out before the bevels make contact?  If so could you just shorten the threaded portion of the holder to get a better fit?

Tony

[Jason von Nieda]

Hi Tony,

The shaft is threaded through - it's not bottoming out. The bevel does not appear to be designed to mate. I think it's just a regular chamfer left there to avoid a sharp edge. If I screw the adapter into the shaft it only starts to grab as those surfaces start to touch.

I took a couple runout test videos: https://www.dropbox.com/sh/oa47epxoxmefjm4/AADpoyjDglAi_RYTyzdpUOYSa?dl=0

These show that it's not the mating that is the issue, but the adapter itself being a piece of junk :)

Jason

--

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/b92a015d-d893-4aa5-8fde-3a255511fd4dn%40googlegroups.com.

[Shai]

Are Juki and CP40 nozzles the main ones that OpenPNP users use? Or is there other type of Samsung nozzles? 

[ma...@makr.zone]

> So what do you think... should I design a CP40 spring loaded holder that fits on a 5mm shaft??

You mean the whole thing, not just a coupling to the Robotdigg one, right? 

Yes please ;-)

_Mark

[Anthony Clay]

Is Jason's the Robotdigg model, or are there variants of the same part?

Anthony Clay

ZarthCode LLC
4652 E 144th Pl S

BIXBY, OK 74008-8008

http://www.zarthcode.com
anthon...@zarthcode.com
PH: 918.860.CODE

--
You received this message because you are subscribed to a topic in the Google Groups "OpenPnP" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/openpnp/NssZzLKkEk8/unsubscribe.
To unsubscribe from this group and all its topics, send an email to openpnp+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/25a3e3eb-e103-f5ef-bdb0-b5a21f34b412%40makr.zone.

[ma...@makr.zone]

Could it be centered on the low side?

Btw, IMHO runout compensation nowadays works without any relevant down-side, ever since the new Directional Backlash Compensation was introduced:

https://github.com/openpnp/openpnp/wiki/Backlash-Compensation

_Mark

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CA%2BQw0jz-iDXT14QMJnJWbxFWEqt%2BCHHkDYcDxoGObr%3DutWFhSw%40mail.gmail.com.

[Clemens Koller]

Hi!

I think the runout of the Samsung CP40 could be low-ish (<100um_pkpk) by default as it's a turned part. But I would expect clones to exist in the market and a wide variation in quality. The spring in my CP40 (as delivered with Liteplacer) is quite hard - on the one side. It does not squash components, but it could be much more soft for the PnP operation, IMO. The o-rings are quite hard (or: the spring is quite soft) when it comes to automatic nozzle changing. The o-rings are supposed to wear out slowly, quite slowly, but I don't see that yet. I've added some lube to the o-rings to help in the nozzle changing.

The steel bolt is blocking the bottom part (piston) to jump out of the cylinder, the play in between the cylinder and the piston will be consistently forced to one side because of the spring tension and the bolt is blocking only on one side. The play is visible and I guess it is in the range of 100um..200um.
The bolt is press fit in the piston, so it cannot be disassembled easily.

The force caused by the spring is (in my case) ranging from 5N to 10N over its spring travel range of 3.0mm.

If you crash the nozzle it will likely bent below the thread (M5) and the runout will be bad.

I polished the nozzle tips to be matte reflective / shiny to have high contrast to work on in the image processing side and to achieve best results with nozzle calibration.


Clemens

On 02/03/2021 18.14, Jason von Nieda wrote:
> The Samsung adapters in my pictures have a spring. For the CP40 the spring is in the nozzle adapter, and for the NEO it's in the nozzle. I think we'd want to target the CP40 since the nozzles are extremely simple. So, yea, we'd need the nozzle adapter to be spring loaded.
>
> Here's a better picture of the commonly available one so you can see the spring area:
>

> IMG_9094.JPG
>
> I haven't measured the runout on this adapter. It could actually be fine. It feels quite nicely built and I can't feel any play in the spring mechanism. I'll measure with a dial indicator this evening.
>
> Jason
>
>
> On Tue, Mar 2, 2021 at 11:10 AM Betz Technik Industries Ltd. <betzt...@gmail.com <mailto:betzt...@gmail.com>> wrote:
>
> Well in my mind, the spring loaded Z axis is the major advantage for the Juki. If I were to make a Samsung holder (CP40?), I would think I would need to make it have a similar sliding Z axis. Otherwise Z height while running needs to be a lot more accurate.
>

> __ __

>
> Am I missing something? Or am I overestimating the importance of the sliding axis??
>

> __ __
>
> Peter.____

>
> --
> 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 <mailto:openpnp+u...@googlegroups.com>.
> To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/D4879A91-E6F3-4672-873F-6FED6F8D7A8B%40hxcore.ol <https://groups.google.com/d/msgid/openpnp/D4879A91-E6F3-4672-873F-6FED6F8D7A8B%40hxcore.ol?utm_medium=email&utm_source=footer>.

>
> --
> 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 <mailto:openpnp+u...@googlegroups.com>.
> To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CA%2BQw0jzDAAq%3DwKV_uCkvxEfeoVNq%2BBgAqAUwqgA49SYyDp1GUg%40mail.gmail.com <https://groups.google.com/d/msgid/openpnp/CA%2BQw0jzDAAq%3DwKV_uCkvxEfeoVNq%2BBgAqAUwqgA49SYyDp1GUg%40mail.gmail.com?utm_medium=email&utm_source=footer>.

[Hrishikesh]

I recently bought a few CP40 nozzle holders and nozzles. They are a mix of copies and original items pulled from the Samsung machines. I did a runout test on all the holders and this is what it looks like. For the copies, the runout is about 0.35mm. For the original, it's a little less than 0.1mm. 

[Mike Menci]

The problem of run-out starts from threaded part screwed into stepper shaft. With M5 normal metric thread there is a gap that will couse the runout on that long lenght to Nozzle. I would shorten the threaded part on nozzle holder so much that the end of stepper shaft would toutch the cilindrical part of nozzle holder

- or option 2 would be to insert washers inbetween there and measure runout again if it is less this way...

Mike

[Mike Menci]

Or option 3 - Increase the lenght of threaded section in stepper shaft ...

Mike

[Hrishikesh]

That is good advice. I added a few precision shims between the motor shaft and the nozzle and, what a difference! Near the threaded part of the holder, the runout dropped to 0.02mm. Near the nozzle end, it was 0.08mm. It is possible to extend the thread inside the shaft but I may not be skilled enough to do that. 

[Mike Menci]

Great Hrishikesh,   to extend the thread you get M5 thread bit - like this one:

https://www.ebay.de/itm/Durable-Spiral-Thread-Tap-Tapping-Drill-Bit-for-Metric-Aluminum-Tapping-Machine/383994020043?hash=item5967d3c0cb:g:lgEAAOSwlq1gPidf

and just screw it in a bit more - as far as needed only - 1 -2 turns only..

Mike