Neoden 4 pick and place

[harry4516]

...
It's very likely your paste is too much on the pads. Even with hand stenciling and hand pick and place, I don't have shorts even with 0.5 mm pitch TQFP.
Usually for less crucial parts, I left the paste opening at -5% only with 0.125 mm stencil thickness.

with these sizes I still get shorts.
Now I'm using -10% and 0.08 to max 0.1mm thickness, and no more shorts.

The other improvement was using another solder paste. Now the Loctite GC10 works very well.

[Spikee]

Saleae logic tried to do in house assembly of many 0201 components in house with a brand name 60K pick and place and dit not have good luck with that (see blog posts / their theamphour episode).

Doing 0201 reliably could be possible if the software was more advice. But it is not.
Backslash and things like that can be compensated for is software without much issue. And than you are not even using the vision as additional compensation.

Feeding  / pickup issues can also be detected in hardware or in software (via vision). Did the pickup for 0201 passive not work? , just let the feeder peel another part, let the vision detect the position and pick it up.

One weird idea that came up in my head is:
Get it to scan the whole (empty) pcb and make a high detail stitched image of it. The idea of that is so real time vision positioning could be used instead of purely relying on mechanical accuracy. ... or just get it to quickly re-calibrate via fiducials every time you want to place a very small component.
Having a good camera that that can autofocus / has good resolution is key in that case.

or even fancier add more camera's to the head (for example one in each corner) and do a full (realtime) stitched image. In this way you can calibrate the offset for each part using all the fiducials on your board without doing additional movement or even loose placing speed. (assuming you can see the whole board with those four camera's (or just use ... more fiducials))

In other words if you have good vision / software you do not need super expensive fancy closed loop servo systems with glass encoders and exotic granite pnp bases.

[bootstrap]

Mr Bootstrap
you said

" An added irony is, some of those projects would be excellent components for a precision pick-and-place machine (servo controller/driver with encoder inputs and robotics vision system camera at the very least)."

You are clearly inexperienced with all aspects of this stuff- but making wild claims.... especially your belief that servos and closed loop is required. If this is from your own experience, your implementations have been flawed.
 I'd suggest you go and read the following forums end to end, yes all 20 to 40 pages  on eevblog
:
Neoden 4 forum
smallSMT forum
TVM 802B forum

and then go into say the DIY cnc forums and see what those guys are doing with regards to accuracy and precision and what they're using.

then go and consult all the tech manuals of all the mfrs of belts, belt drives, ballscrews, linear guideways.....

and then go and do the numbers on exactly what is required. You'll find if local datums are used and sensible motion practices are used, it is not a difficult job considering that the placement head requires no dynamic accuracy , only static.

You seem fascinated by needing closed loop and servos when the reality is, depending on the choices you make and the speed this may be entirely unnecessary  (or indeed entirely necessary ) if you understand the technology rather than... guessing.

I'm actually suspect of your motives in this forum. Can you let us know if you work for a high end manufacturer?

You don't know me, so don't pretend you do.  I'm perfectly happy to freely admit my weaknesses, so if you want to know something, just ask (don't assume).

I've created a number of precision mechanical systems in the past, including advanced automated telescopes that have to do some rather amazing things very precisely (I'm sure you realize the earth rotates, so both satellites and astronomical objects in the sky are constantly moving, and need to be acquired and tracked precisely if they're being photographed... sometimes for hour+ exposures).  So I understand precision mechanical systems, and I've created both open-loop stepper and closed-loop servo systems before (often with DC brushed "pancake motors").

So I've been though the issue of backlash many, many, many times in a great many devices.  Which means I'm aware of the problem that backlash is, and also various methods to compensate for backlash.  However, there are several things going on in pick-and-place machines, and they interact with each other, so the consequences of an open-loop system on a pick-and-place system isn't definitive or necessarily immediately obvious.  For one, you need to know the mechanical design of the specific pick-and-place machine, plus what factors it depends upon (far beyond just the reference marks on the PCB).

I already have thousands of PDF spec sheets on all the mechanical components you mentioned, plus many other kinds of components, plus numerous articles, and so forth.  In fact, I've been designing and building devices with the mechanical components you mention for decades.

If you think it is possible to make a final and definitive statement about whether what I need can or cannot be done with an open-loop system, then I can answer that from past experience.  With enough tricks, cheats, schemes, methods, algorithms and creativity, I am certain it is possible.  That doesn't mean it is wise, or reliable, or efficient for this kind of device at the precision levels I require.  In my experience, having designed and implemented both kinds of systems (and dealt with the working devices at length), I can say that my experiences with closed-loop servo systems have been MUCH more satisfying.

I never said it is impossible to create a pick-and-place machine with steppers or open-loop design.  However, from past experience, open-loops systems can be problematic and inaccurate for a great many reasons (including even obscure issues like [differential] thermal expansion/contraction for example).  Of course even closed-loop servo systems aren't 100% immune to problems either, but they definitely tend to be more precise and reliable in my experience (and theory).

What I AM NOT experienced with is modern SMT equipment.  I've been doing mostly software for several years, and my return to hardware puts me in a situation where I need to switch over to SMT.  All the advanced and compact electronics components I need for new designs are SMT.  Though I looked at this equipment now and then over the past year or two, I have no operational experience with the equipment (but tons with older though-hole tech).

I have looked at most of those forums (and others elsewhere), and spent two or three days reading other sites recently too.  However, I pretty much move on when I find a thread is discussing systems incapable of 0201 and 0.50mm pitch.  Surely that means I miss worthwhile messages now and then, but I have to judge what is worth reading and what is not.  More than likely I will spend $60K to $80K on SMT equipment in 2 or 3 weeks, and so I need to learn what I need to know in order to make good choices.  That's why I'm here.  I know many people here haven't physically worked with this equipment, but some have.

No, I do not work for a "high end manufacturer".  Since [before, actually] I finished school I've been a self-employed scientist, engineer, inventor and product developer.  Mostly I've invented, designed, develop and prototyped products and technologies that I sold or licensed to companies that already manufacture and market the kinds of devices and technologies I created.  But also from time to time I've taken contracts from outfits like NASA, AirForceResearchLabs and various astronomical observatories to solve problems they have, or design systems they need.  So while I always design in ways appropriate for quality and reliability (and sometimes volume production), I have never been in the situation of needing to worry about volume production (if that's what you mean by "high end manufacturer").  I have always spent a lot on the equipment I've purchased to do my work.  I have enough problems and hassles already, and don't want my equipment adding more.  BUT, "it is just me here", so spending $60K ~ $80K of my savings isn't something I do without a second thought.  So I want to make smart decisions.  So my one and only "motive" here is to figure out whether I should buy a neoden4, a LE40V, or some other pick-and-place machine... and also figure out what are the best stencil-printer and reflow-oven products for my purposes and budget.  If that's not a specific enough answer to your question, inquire further.

I hope you can understand that it is entirely possible to know A LOT about mechanical and electronic systems, but NOT MUCH about these specific devices and processes.  Because that is the case with me here.

The bottom line is this.  I'll appreciate anyone who knows about this area and gives me the benefit of their experience in these equipment and processes.

[glenenglish]

Hi Spikee

Advanced Assembly asked me to put some local fiducials nearby the 0201s I had on JESD204 thin traces.... on a big area board

IE I had global fiducials and local fiducials- which makes alot of sense.

regards

[Spikee]

Hi Spikee

Advanced Assembly asked me to put some local fiducials nearby the 0201s I had on JESD204 thin traces.... on a big area board

IE I had global fiducials and local fiducials- which makes alot of sense.

in that case they are probably doing offset calibration for each placement because they don't want to mess up boards.
Since vision is so fast these days it is possible to do continuous calibration with very little time loss.

Mechanical accuracy is really expensive. Adding some more fiducials is more or less free.

[bootstrap]

Saleae logic tried to do in house assembly of many 0201 components in house with a brand name 60K pick and place and dit not have good luck with that (see blog posts / their theamphour episode).

Doing 0201 reliably could be possible if the software was more advice. But it is not.
Backslash and things like that can be compensated for is software without much issue. And than you are not even using the vision as additional compensation.

Feeding  / pickup issues can also be detected in hardware or in software (via vision). Did the pickup for 0201 passive not work? , just let the feeder peel another part, let the vision detect the position and pick it up.

One weird idea that came up in my head is:
Get it to scan the whole (empty) pcb and make a high detail stitched image of it. The idea of that is so real time vision positioning could be used instead of purely relying on mechanical accuracy. ... or just get it to quickly re-calibrate via fiducials every time you want to place a very small component.
Having a good camera that that can autofocus / has good resolution is key in that case.

or even fancier add more camera's to the head (for example one in each corner) and do a full (realtime) stitched image. In this way you can calibrate the offset for each part using all the fiducials on your board without doing additional movement or even loose placing speed. (assuming you can see the whole board with those four camera's (or just use ... more fiducials))

In other words if you have good vision / software you do not need super expensive fancy closed loop servo systems with glass encoders and exotic granite pnp bases.

Can you provide a link to "theamphour episode" for me to read?  Thanks.

Also, I was also thinking along the same lines as you... doing with cameras what has in the past been done with precision mechanics and linear encoders.  I posted a few comments about that about seven messages back.

I would guess this approach has not been adopted in modern pick-and-place machines for two reason:

#1:  Existing companies resist switching over to new approaches, so they stick with whatever approach they have.

#2:  Doing more with cameras slows down the process, because it adds more steps (some for every placed component).

For a new manufacturer or an advanced DIY project, this approach may have much promise.  And for DIY and low volume projects, the moderate slowdown is of little importance, while the ability to place smaller components will be a great advantage for some to many people... especially as the years pass and the most desirable components continue to shrink in size and pitch.

[Spikee]

http://www.theamphour.com/237-an-interview-with-joe-and-mark-garrison-subtly-spelling-sayleeay/

http://blog.saleae.com/need-raise-logics-price/

A lot of pick and place machines are based on 20-30+ year old products. Having small updates every few x years ( this has been mentioned in that or other electronics podcasts). It is not in their interest to design products that change to much from this base design.

Its all about money. The low cost pnp market is not really a market they want to be in. You want those 0201 specs (for example) well than you just have to pay 100k (for example). They could design a low cost 10k pnp who does that with vision and all using off the shelf feeders but that more or less completely destroys their whole product line.

Vision can more or less be as fast as you want it to be. But if you want ultra fast (than you are looking at a big pricepoint anyway) it is easier to just trow a bunch of money at glass encoders and that kind of stuff than actually doing some work on your software.

For 3k a good base could be made. Extruded alu profiles, hiwin rails, closed loop steppers and china ballscrews , pick and place head.
Add off the shelf feeders to that and you have a good place to start. For 0402 and those kind of things the current openpnp sw is probably good enough already.
For good 0201 placement the said vision steps / capability need to be added.

You'll want to use these somewhat precision components so you get close enough for most parts , and can do vision offset cal for the special ones .
Spending a little money here greatly reduces the time commitment of doing major location adjustments in software.

For cameras it would be good to actually use a machine vision camera as these have features that are quite nice to have in such application. Some even have the capability to do the lens correction itself. (like matlab vision can do) otherwise you might have to do that processing yourself in realtime which can probably take quite some processing time. This feature gives a corrected image where the lens curvature is gone. This can be quite useful.

Tbh at 10-12k ex tax ex feeders one could commercially design and produce such machine but it would cost about a year of dev time for two engineers or so.
The harder part of this is making it foolproof so you are not spending time on customer fault finding / fine tuning. Unless you get paid for it via some on demand remote troubleshoot service.

[TheSteve]

For 0201's and 0.5mm BGA's I think you should look for something other then the Neoden4. Consider spending real money, not buying one of the cheap machines. The Neoden4 is actually pretty capable when dialed in and you figure out the software oddities etc but really doubt you'd get the results you want with it.

Thanks.  Do you have a neoden4 or work with pick and place machines?  I'm just curious what your reasons are for giving this advice.

To answer your first two questions - Yes and Yes. As for the last question you should read through this thread.

[mrpackethead]

I've created a number of precision mechanical systems in the past, including advanced automated telescopes that have to do some rather amazing things very precisely (I'm sure you realize the earth rotates, so both satellites and astronomical objects in the sky are constantly moving, and need to be acquired and tracked precisely if they're being photographed... sometimes for hour+ exposures).  So I understand precision mechanical systems, and I've created both open-loop stepper and closed-loop servo systems before (often with DC brushed "pancake motors").

With all that experience, you should build one.

[bootstrap]

I've created a number of precision mechanical systems in the past, including advanced automated telescopes that have to do some rather amazing things very precisely (I'm sure you realize the earth rotates, so both satellites and astronomical objects in the sky are constantly moving, and need to be acquired and tracked precisely if they're being photographed... sometimes for hour+ exposures).  So I understand precision mechanical systems, and I've created both open-loop stepper and closed-loop servo systems before (often with DC brushed "pancake motors").

With all that experience, you should build one.

I am so, so tempted to do so, partly because I need one (lame excuse), but mostly because I am finding the various aspects of pick-and-place machine operation fascinating.  And as a few of my recent posts indicate, there appears to be enormous opportunities to create different configurations of pick-and-place machines to make cheaper precise [but probably not fast] units.

Frankly, if I had one or two talented partners on such a project, I'd probably buckle and go for it.  I wouldn't be surprised if we could build a new pick-and-place machine from scratch for roughly the same I will probably spend on the complete production line (stencil-printer, pick-and-place, and reflow-oven).  As with most projects, the problem is TIME.

[mrpackethead]

Come and join the fun with the openPNP project,  theres lots goign on, and some amazing results.

[bootstrap]

Come and join the fun with the openPNP project,  theres lots going on, and some amazing results.

Is that your project?  If so, where are you located?

[mrpackethead]

Come and join the fun with the openPNP project,  theres lots going on, and some amazing results.

Is that your project?  If so, where are you located?

No, its not "my" project,  but im using it.. I'm in New zealand, but there are people in Europe, USA, Australia, Canada, UK...

[bootstrap]

Come and join the fun with the openPNP project,  theres lots going on, and some amazing results.

Is that your project?  If so, where are you located?

No, its not "my" project,  but im using it.. I'm in New zealand, but there are people in Europe, USA, Australia, Canada, UK...

I assume OpenPnP is only software, correct?

If I do decide to make a pick-and-place machine, I would certainly be inclined to adopt as much open software, hardware and mechanics as possible.

[mrpackethead]

Come and join the fun with the openPNP project,  theres lots going on, and some amazing results.

Is that your project?  If so, where are you located?

No, its not "my" project,  but im using it.. I'm in New zealand, but there are people in Europe, USA, Australia, Canada, UK...

I assume OpenPnP is only software, correct?

If I do decide to make a pick-and-place machine, I would certainly be inclined to adopt as much open software, hardware and mechanics as possible.

Theres software, several hardware builds and all sorts of stuff.

[rx8pilot]

With all that experience, you should build one.

I am so, so tempted to do so, partly because I need one (lame excuse), but mostly because I am finding the various aspects of pick-and-place machine operation fascinating.  And as a few of my recent posts indicate, there appears to be enormous opportunities to create different configurations of pick-and-place machines to make cheaper precise [but probably not fast] units.

Frankly, if I had one or two talented partners on such a project, I'd probably buckle and go for it.  I wouldn't be surprised if we could build a new pick-and-place machine from scratch for roughly the same I will probably spend on the complete production line (stencil-printer, pick-and-place, and reflow-oven).  As with most projects, the problem is TIME.

Do you need a P&P machine to assemble PCB's in-house or an all new business? I would have a ton of fun with the challenge of designing and building a P&P machine. I have a lot of experience designing and building very high precision mechanics for imaging and motion control. Now I design electronics and write embedded software most of the time. It would still take a mountain of time and money to do it with no guarantees. It would literally be an all new career to go down that path with the remarkably lofty goal of making MyData performance for low cost.

Seriously. Skip the N4 based on what you seem to require. You will be looking for a tall bridge to jump off of if you really need very fine pitch assembly as a reliable process. I don't have an N4, but I do have a Quad 4000C. The design of my machine goes back to the 80's and went out of production in the late 90's. PPM purchased the remnants of the company and started an update program for the software. Without getting too distracted with details - they definitely improved the machine and it is capable of placing 1005's. I am not suggesting you buy the same machine, but something like it. It is ridiculous (in my humble opinion) to consider one of the lowest cost/quality P&P machines on the market and expect it to nail PCB's with 0201 and .3mm without constant fiddling and re-work. You really can't be surprised that the moco subsystem is open loop steppers either - there is no money to put them in there. The PPM machine is steppers with linear encoders and a side scanner for alignment on the fly. Up vision for big parts and BGA. It is not fast, but is precise. With all that said, the small passive parts require the super precision feeders  - there is no way around that. The placement accuracy of the machine is meaningless if it struggles to pickup the parts in the first place. If the nozzle picks up on the edge of the component, it will flip or be at an angle causing a pick failure. You can only solve that by have very precise feeders that put the part in the exact same location every time and it has to be gentle about it. If the feeder is bumpy, the parts will jump out of the pocket before the head even arrives. From what I have read so far in this thread, the biggest weakness of the N4 is the feeders. For the price it seems pretty good. The Quad precision 8mm feeders are about $800ea which is still rather cheap when compared to the high-end machines - but they can index 2mm very precisely and smoothly.

I would not try to satisfy your requirements for under 80k with feeders, printer, oven, training, vacuum/air, accessories, etc. . If you are industrious like I was, you can get a broken machine and fix it up. That allowed me to get a 1005 capable machine with about 60 feeders for under $10k in cash - but it took nearly 10 months of nights and weekends to get it up and running. To buy it from PPM with with all the feeders and other parts would have been about $50-60k and would be ready to go on the first day. I also looked at DDM Novastar and liked them in general. They are targeted towards low volume and only the bigger machines can deal with 0201 (so it seems). What I did not like is that they cannot hold very many parts at a time. The Quad I have now can handle a TON of parts so I can have 6 designs ready to roll (in the machine) at any time and still be able to do prototypes without tearing the machine apart. That is super critical in my case where I do small batches of a bunch of different designs and don't want to have to setup each one every time I need a batch.

How much is your time worth and what business are you in?

[mrpackethead]

I bought two 2nd hand Yamaha machines for exactly the same reason.. Foruantly they did'nt take 10 months, but it did take a month of engineering effort to get them back to running condition..  About $25k each.

Im building a machine, slowly, just becuase it intrigues me. 

[mikeselectricstuff]

One weird idea that came up in my head is:
Get it to scan the whole (empty) pcb and make a high detail stitched image of it. The idea of that is so real time vision positioning could be used instead of purely relying on mechanical accuracy. ... or just get it to quickly re-calibrate via fiducials every time you want to place a very small component.

That won't work as the PCB has paste on it, so you can't see the pad edges accurately. If you had a super-precise paste print that would allow it, chances are you could get away with poorer placement accuracy

Having a good camera that that can autofocus / has good resolution is key in that case.

You don't need AF as the board distance is fixed.

In other words if you have good vision / software you do not need super expensive fancy closed loop servo systems with glass encoders and exotic granite pnp bases.

Unfortunately yes.
I do wonder if it might be feasable to use something like laser interferometry to accurately measure the head position in 2D space, to allow cheaper mechanics, but even then speed might be an issue

[mikeselectricstuff]

For cameras it would be good to actually use a machine vision camera as these have features that are quite nice to have in such application. Some even have the capability to do the lens correction itself. (like matlab vision can do) otherwise you might have to do that processing yourself in realtime which can probably take quite some processing time. This feature gives a corrected image where the lens curvature is gone. This can be quite useful.

Processing time for vision isn't a huge deal as you can do it while travelling to the placement position. CPU time is cheap, and for a PC controlled machine, it isn't really doing anything else during travel. 

[Koen]

I agree with glenenglish, this smells like bullshit. A seasoned linear motion engineer would start small and humble to understand his needs. A "2-3 weeks read of forums" planning is ridiculous and even more ridiculous is asking newbie questions then three sentences later reveal how the industry should do it according to your standards.

[mrpackethead]

I agree with glenenglish, this smells like bullshit. A seasoned linear motion engineer would start small and humble to understand his needs. A "2-3 weeks read of forums" planning is ridiculous and even more ridiculous is asking newbie questions then three sentences later reveal how the industry should do it according to your standards.

My intial thoughts where that this guy might be working for one of the other PNP manufacturers.. Its realy hard to tell, thats for sure.. but anyone returning to hardware thats going straight into 8 layer PCB' with high speed traces, is probalby way too buys doing those designs to have any time to be building bits of hardware...   I'll sit on the fence. this is the internet and well its just hard to tell.


I'd farm this work out out to someone like macrofab.

[EEVblog]

ADMIN NOTE:
As requested I have moved bootstrap's recent posts to a new thread.

[bootstrap]

ADMIN NOTE:
As requested I have moved bootstrap's recent posts to a new thread.

I don't see it.  Can you provide me a link or something to help me find it?  Thanks.

PS:  I assume it would still be within "Manufacturing and Assembly".

PS:  Found it, but this is a thread about ideas, not products.  Can it remain under "Manufacturing & Assembly"?

[uncle_bob]

ADMIN NOTE:
As requested I have moved bootstrap's recent posts to a new thread.

I don't see it.  Can you provide me a link or something to help me find it?  Thanks.

PS:  I assume it would still be within "Manufacturing and Assembly".

PS:  Found it, but this is a thread about ideas, not products.  Can it remain under "Manufacturing & Assembly"?

Hi

Just for the sake of others who are wondering where things went:

https://www.eevblog.com/forum/manufacture/pick-and-place-discussion/msg936104/#msg936104

Bob

[uncle_bob]

Thanks Bob, do let us know if it works out.

Hi

Ok, so here's the summary on the printer:

I ordered it on eBay and it came in via DHL in a few days. Absolutely nothing to complain about in terms of shipping speed. I did not have "CNC machine precision" expectations for the printer. I'd say it's fit and finish are a bit better than I expected. I have not run up any stencils on it, that's a bit in the future. It appears to have working controls and reasonable adjustments. I see no reason why it will not do a very adequate job for the levels (0.5 mm pitch) I'm trying to hit.

The only negative is the rubber feet. They were *supposed* to be in the package. They are nowhere to be found. Given that I have lots of those running around here ... no big deal at all. It came with a steel scraper and a solder spreading knife that I didn't expect. Overall that's a win.

Bob