Ideal nozzle spacing??

[Peter Betz]

Hello everyone,

Long time since I posted last, still plugging away in the background. 

I was curious about shrinking my head design down a bit, particularly in width. What you do you think is ideal nozzle spacing, off the top of my head I'd want to be able to spin a 100 lead package on one nozzle,

So lets say (16.2mm width actual) bump up to 20mm wide, then rotated 45 degrees it would be just under 30mm wide, so 30mm between tips minimum. Maybe you would want to pick a number that is divisible by standard feeder spacing (like 8mm), so 32mm ??

What would you guys say is a good value to use??

Thanks!

Peter Betz

Betz Technik Industries Ltd.

betztechnik.ca 

[John Plocher]

My chmt48 (2 head) system has a spacing of about 22mm.

With this multi head system (without an automatic tip changer), I typically use one tip for the dandruff parts (R&C) and the other for bigger things, like IC's - it is unusual to have both heads used at the same time for large parts.  This impacts the min head spacing calculations:  like you said, if you wished to handle both heads loaded with a 20mm part at the same time, you would need the full 30mm spacing (~15mm to the right of the left head plus ~15mm to the left of the right head...), but if one head only holds dandruff parts, you should be able to get away with about half that.

You also need to ask if you want to optimize board throughput -vs- large part speeds.  If you have many small parts, 2x small nozzles can almost double your per-board speeds, making it feasible to run a batch of boards through twice, once with 2x small nozzles, and a second time with a single big nozzle to just place big components.  In my workload, this can save 20 to 30 minutes on a run of 20 complex boards.

If placing with 2x same-sized nozzles, you may want to explore having your head spacing match a multiple of your feeder spacing - again raising the per-board speed if your placement of feeders and your part placement planner can agree on an optimal pick order :-)

  -John

--
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/325f8414-6555-46d2-9931-6a50a84f39ef%40googlegroups.com.

[John]

The ideal nozzle spacing would be same distance as on feeder bus bar to allow parallel pick using separate z motors.

But that are dreams of future in diy pnp.

Our Nema8's are 20mm wide, with 2mm needed gap you have 22mm between nozzles.

With 22mm distance most parts up to 16x16mm (like TQFP100) should be no problem (45° placement) or parts up to 22x22mm with 0 / 90° placement angle.

Playing in cad on a simple, light weight, counterweight config belt driven quattro head i discovered that 22mm are not so easy to realize.

John

[Peter Betz]

Looks like 22mm is a consistently pleasing result, which surprises me.

Let me play in CAD, my design might make this quite possible.

Why did you delete the picture John? Looked great!

Thanks for the feedback!

Peter.

[John]

Hello Peter,

I'm so sorry, but i did not want to hijack your thread.

And i was so disappointed, that my animated gif did not show up correctly here in google groups.

But I ordered some parts already and have to check it on a prototype first.

If you do not mind, then i allow myself in case of success to report here again.

John

[Peter Betz]

Hi John,

 

No hijack at all, we are all on here to share! Do you have a mill to make those parts on?

[John]

Yes, i do have diy 2,5D routers ("cheese-mill") with ridiculous 1kw spindles, slow but feel lucky to have it.

Thread whirling (M2) is something on my todo-list if the tools where not that expensive.

John

[Mike Menci]

Passing with belt in middle of 1şt >Nema 8 base:

[John]

Cool Mike. With four or six nozzles, please.

John

[Mike Menci]

Here you will find basic design - some of this files I have in SOLIDWORKS - let me know
http://loadmp4.info/i/6DCaZ7h2QDA.html

Mike

[John]

@Mike


Your link shows a well known design copy (samsung) - it is very complex and not pleasant for diy imho.

If a design consists of parts that are very difficult to manufacture or purchase, then it is not interesting for me.

All parts that you can not either produce yourself, or whose production costs would break the economic burden, you better buy in the far east for some bucks.

So for example motors, rails and carriages, pulleys, belts, nozzle holders and other special parts that are used on the smt head.

All other parts like the main frame etc. should be reproducible with my hobby tools. That is the goal.

John

[Mike Menci]

Hi John

How about this 4x Nema8 + 2x Nema 17;

Mike

[Mike Menci]

Belt Fix - Hex Bolt to spacer inside Stepper support!

[John]

Hi Mike,

Looks very good. Are that 7 or 9mm rails? They are much easier to get as the 5mm rails i planned to use.

Your nema8 brackets solve the carriage mounting problem very elegant (i am using ready made brackets with 8mm mounting distance).

For more accuracy i'm using always two carriages per rail.

I'm sure you know quality differences between original hiwin with cheap copies (i have both).

John

[Mike Menci]

Hi,

No need for double carrage - one good quality is enough - make bracket go higher - there is no load on this 9mm rail in dwgs - just from an old project drawing - enclosed some details - main plate is 10mm and rails are placed in grove to get the line of belt and avoid rail carrige- nema 17 go in as well -  see this details marked..

Mike

[John]

LoL, "one good quality is enough" - ordering cheap rails/carriages is like lotto playing ;-)

I think there is no need for heavy 10mm sheet. 8mm is very stiff and nema17 shaft is long enough.

Don't forget to think about an easy to handle belt tensioner for standard length closed loop belts.

John

[Peter Betz]

22mm nozzle spacing iteration in progress....

- 22mm nozzle center to nozzle center with NEMA8 motors

- NEMA17 Z motor

- Proven and simple to assemble CNC machined rack an pinion Z drive

- Drilled and tapped on all sides, no more nut traps.

- Machined rail reference edges

- Supports 2020, 2040, or 4040 X axis extrusion

- aprox. 42mm total head width

- MGN12C or H X axis rail

- MGN9 Z rails

- inductive homing (although in order to make it as compact as possible it will have to home more traditionally)

Major weight and size reduction over previous design.

Let me know what you guys think!

Peter.

[John]

Hello Peter,

It looks really beautiful, very compact and probably has less weight than i would expect.

If have got no experience with rack/pinion solutions - don't they get a play after heavy use?
Even with hard alloys (e.g. ALZn5.5MgCu1.5) there will be abrasion.

With belts i have after one year (with many z moves) zero play.

Inductive homing is very cheap and has good repeatabilty, but it is very slow (you can not home with 0.5m/s).

A combination optical / inductive for faster homing is something i want to have on my next head.

And closed loop motors for z that prevent loosing exact position when nozzle tip changing.

Currently i have to use event scripting (home z) after tip change because my nozzle tip holder needs a lot of force and sometimes even with strong nema17 (0,59Nm) sometimes steps are lost.

 

I am looking forward to see your head in action.

John

[Marek T.]

I'd check what is typical spacing of Yamaha feeders (finding them the cheapest of professionals) and if it is 15mm do the nozzle spacing 15 or 30mm - to can realize all nozzles picking without x motion necessary. This is how usually industrial machines have this made.

[Trinh Dinh Huu]

Hi Mike Menci!   Hi Mike Menci! I think your design is very good. I want to borrow this idea to design my device. I have a project on making openpnp 4 Nozzle Tip. I am currently doing a 1 head machine for my first project.

Vào Th 4, 30 thg 10, 2019 vào lúc 03:57 Mike Menci <mike....@gmail.com> đã viết:

--

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/136694eb-5d40-4947-8c4b-2df0ed40e498%40googlegroups.com.

--

Đinh Hữu Trình

TDC Bắc Giang

Điện thoại 0978691222; 0912851598

[A J]

Partly on and off topic...

I am setting my nozzle spacing to match my 4 low-res up-camera spacing: approx 45mm with light arrays. I've adopted Nema 11 rotational axes, so 45mm gives me ample room. 

Tangential to topic: Since Opnp does not support OnTheFly multiple cameras, I'm hoping to be able to create a script that might handle this. A possible solution I'm considering is an Rpi for each camera. Opnp would send a component ID and angle info to the Rpi (to the corresponding Rpi for each of 4 possible component picks), and each Rpi would process its camera/component image and control its rotation axis, reporting back the completion of task. Opnp would initiate a Place action when it was at position and had confirmation of rotation completion from the Rpi.

Alf

---

From: ope...@googlegroups.com <ope...@googlegroups.com> on behalf of Marek T. <marek.tw...@gmail.com>
Sent: Wednesday, October 30, 2019 12:51 AM
To: OpenPnP <ope...@googlegroups.com>
Subject: [OpenPnP] Re: Ideal nozzle spacing??

 

I'd check what is typical spacing of Yamaha feeders (finding them the cheapest of professionals) and if it is 15mm do the nozzle spacing 15 or 30mm - to can realize all nozzles picking without x motion necessary. This is how usually industrial machines have this made.

--
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/d42acbdf-613b-4ecf-aeb7-d178360e3d2b%40googlegroups.com.

[John]

@Marek

15mm? Are you sure?

My CL-compatible feeder is about 16mm and spacing on busbar is 16.5mm.

John

[Marek Twarowski]

No John, I don't remember exact spacing, just told from memory only, matter of confirming so. Just head an idea on my mind only. 

As remember different plates are on the market, original and clones, with spacings from 15-18mm. Remember that Bernd told that 16mm is ok for 8mm CL with small gap between. I have started to do my plate but didn't finish it yet, and it seems to me I made 15mm soacing as I wanted to fit there as many feeders as possible at one machine side. Can check it tonight.

--
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/WAHFWVwHtjU/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/544d751b-e7f7-425c-bde4-0c7eedf8d221%40googlegroups.com.

[John deGlavina]

Ah yes, this one: https://grabcad.com/library/smt-pick-and-place-1

I noticed the individual up down motors are staggered, then the turning motors turn a secondary guide shaft that turns a pulley on the primary. You would need to control 12 steppers with that alone! I guess the advantage there is each nozzle is independent (you don't have one going up when one is going down for a pick), and the turning motors can be smaller since they don't need a hollow shaft. But obviously that adds to the complexity significantly. Even if that was a 4 nozzle, you'd still have a lot of parts, and extra weight to throw around.

[Brynn Rogers]

Hi Peter,  

   I like it.    Two things I would like in such a design:

1)  I think the gear on the Z motor should be bigger, made as big as can easily fit this design - Why?  of course to increase the Z speed at the cost of resolution, as Z resolution is not so critical, and speed is always better on PnP.

2) provisions should be there for stacking pairs of this unit side by side to get 4 , 6 or more all on the same spacing as between the first two heads.     By this I mean the carrier can be easily bolted up square to the adjacent unit, with pre drilled and tapped holes and features.   

Brynn

* of course, at the point speed causes problems it is no longer better.

[Peter Betz]

Hi All,

Thanks for the feedback, seems like most think this is a positive change, and I'm excited about it. 

1) About the spacing, it seems like feeders on DIY pnp are far from standard, I think I will keep this at the minimum distance to save space and weight at the expense of requiring a X move to pick from the next feeder. If we ever do settle on something for a feeder, as a group, I can easily change this!

2) Rack and pinion wear. I have not seen any wear on mine after years of light use, and there are many of these out there. The brass and aluminum seem to play well with each other, but perhaps I can send these out for type 3 hard coat anodising as well.....

3) Larger gear. I hadn't considered this, I have a rack cutter so I can have bigger gears made of the same module and not require new tooling on my end. I will give it some thought, might be something I can change later, as you bring up good justification.

4) Linking elements. I like this idea, since all edges of the head will be machined perfectly square, I think I can just add a couple tapped holes on the face near the edges to allow little linking brackets.

Peter.

[ma...@makr.zone]

Hi

This looks very promising. Low weight is very important, IMHO!

What is the weight of the previous design and what is the estimated weight of the new one?

Also: could it be made nozzle tip neutral? E.g. compatible with CP40 nozzle tips?

https://www.robotdigg.com/product/799/OpenPnP-CP40-Holder

https://www.liteplacer.com/shop20/index.php?route=product/product&product_id=63

_Mark

[Peter Betz]

Progress:

You can use 12C or 12H carriages on the main rain, however I slotted the rail attach holes so that if you want to mount two heads (4 nozzles) you can use one 12H carriage (as seen in last image). Small threaded holes near the bottom for linking brackets. 

Regarding other nozzles, they are standard 5mm shafts on the NEMA8 motors so you can mount whatever you want that fits that.

Will have to respond later about weight.

Peter.

[Mike Menci]

Peter might be good adding a gear like in enclosure would give some more torque for changing nozzles, maybe as well triggering a feeder with head ....

Mike

[Peter Betz]

Hi Mike,

I’m sorry I don’t understand what you mean, can you explain in more detail?

Peter Betz
Betz Technik Industries Ltd.

BETZtechnik.ca

[Mike Menci]

Peter did you see the enclosure?

In one of the feeder post a simple feeder is shown on LitePlacer ----- https://www.youtube.com/watch?v=ol07YgZOKgs

- Feeder is triggered by Z axis movement - I wonder in this construction has enough force to trigger such feeder with Nozzle Z going down touching feeder compressing spring and than advancing tape....  

- With little torque on Nema17 and direct gear to rack at changing of nozzles there is as well good possibility to loose steps on Z stepper and than you need to home again and again... Just my concerns .....   
3) Spacing of nozzles - The distance is too small if you want placing 25mm or 20mm qfn/ soic, ...then it is 40mm or 36mm distance might be need.

[Peter Betz]

Sorry I missed the attached picture! I see what you mean. Only issue is greatly slowing the head down. I will test this to see what Kind of power I can get out of it, but there hasn't been any issues yet (current head uses this same configuration, but I haven't done any auto nozzle changes).

Regarding spacing, here is a LQFP100 and 18mm 28 pin SOIC mounted simultaneously on the nozzles:

Seems like a pretty comfortable fit to me? and significantly larger ones can obviously be used as well. 

Peter.

[Peter Betz]

Hi Franz,

1) Sure it’s possible to either isolate the camera from the camera bracket or the bracket from the main plate using plastic screws and plastic washers but I’m not sure this is an issue ? I don’t think the alternate return path through all the rail carriages will be particularly enticing for the electrons? Hasn’t been an issue on the current design but easy to mitigate if it comes up.

2) I will have to double check this, I’m assuming if you mounted the camera on top of the bracket it would be around 80. Will report back.

3) I was seeing great repeatability with the switch, but I was not measuring in microns since there was no use in doing so (because of the spring loaded nozzles). It seems to be well within 0.01mm repeatability.

4) Yes those brackets could easily be machined from solid but would that be worth the extra cost to people (in both material and setup time) ?? To be honest I hadn’t considered it, but this would be an extra 3 setups on the mill.

Peter Betz
Betz Technik Industries Ltd.
BETZtechnik.ca

> On Nov 3, 2019, at 1:15 AM, John <waterpl...@gmail.com> wrote:
>
> Hello Peter,
>
>
> I looked at your new head again and some things do go out of my head:
> - is an isolated mounting of the camera module possible (ground loop) ?
> - how long is the camera distance from lens to board? I think the recommended distance is 80 to 100mm to get better pictures.
> - did you see good repeatability (<10µm) when mounting the proximity switch perpendicular to it's sensing axis?
> - i am sure the siemens in your garage is an accurate toy but would it not be better to make the head bracket as one part to have a perfect right angled motor mount with less weight?
>
>
> Franz

[ma...@makr.zone]

Hi

@Franz, regarding 1)

FYI, the ELP modules' mouting holes are not grounded, well at least mine were not, I checked.

@Peter, the metal parts should all - explicitly! - be grounded IMHO, see also:

https://makr.zone/grounding-the-machine/283/

https://makr.zone/electrical-safety-issues/230/

Or from another source:

https://wiki.apertus.org/index.php/Liteplacer_PnP_Machine_Notes#Grounding_the_Machine

_Mark

[Mike Menci]

OK - I see I forgot to divide by 2! - spacing looks good. 

I made a simple test with my PnP head - kitchen scale pressure with the two heads (no nozzles - there is a bit difference because on N2 Nema11 with screw is attached directly and N1 is a slave...)  This is max what I get when stepper -Nema 11 is already loosing steps.

Pictures enclosed.

Mike

[Peter Betz]

Hi Mike,

 

The old divide by 2 will get you ever time, haha.

 

Great you supplied that data, I just tested the unlocking pressure on the nozzle holders. It releases significantly before the collar is depressed as far as I did here, so it is definitely below 600g- 650g. Looks like a reasonable margin there, which is good to know. I have to test my head as well to know the effect of the rack and pinion vs. What you have. I also sell the heads with pancake NEMA11’s so that will be interesting too.

 

 

This reminds me that I need to design a nozzle holder rack....

Peter Betz
Betz Technik Industries Ltd.

 

[Marek T.]

Smart measuring Mike :-) :-).

I will check what is a power of my pneumatic nozzles advancing old mechanical Yamaha feeders, a lot of them is available on the market for few bucks usually.

Just interresting whether Nema's (on Peter's head) would have enough power to work with them.

[Mike Menci]

Marek borrow kitchen scale from your wife or grand mother - but do not brake it! :-)

Mike

[Marek T.]

That's the plan! ;).