Pick and place machine (the internet told me so!)

[NickPerry]

Hello there.

a random person on a random YouTube video mentioned that if someone can make a pick and place machine for under $200 it'll revitalize the DIY electronic hobbyists.

and since I read it on the internet it must be true!!  

all joking aside, this sounded like a challenge to me, and since I was planning to build a P&P machine anyways, why not give myself a challenge? (sub $500, not sub $200) (on top of the fact that I'm a full time student with a full time job... =D )

My question is this. what would you like to see on a budget P&P machine? what could you live without?

I have some Idea's, but since I'm at work right now I'll wait till I'm at home to draw it up and get you-guy's opinion on it.

Lets hear your desires! (or rude opinions :p)

[BrianHG]

I know another Nick here in Canada who could use a pick and place machine in around 10 days from now...

[NickPerry]

lol, I'm a nick here in Canada, ^.^  but I won't have this project done for a few months :p  (I mean, I could do a shoddy conversion of my cnc machine, but it would be terrible :p)

[mikeselectricstuff]

A $200 P&P will be of absolutely no use to anyone.

[Kjelt]

a random person on a random YouTube video mentioned that if someone can make a pick and place machine for under $200 it'll revitalize the DIY electronic hobbyists.

That is just BS and i tell you why.
1-Any P&p machine to setup reliably for one design/product costs ten times more time than to place it manually. Ergo for hobbyists that do a few pcbs a p&p machine is no improvement.

2- the XY p&p mechanics is common knowledge, tens of designs out there nothing new, openPNP sw available. The real problem is in the parts feeders. The low cost solutions are crap, the only good and reliable solutions are automatic pneumatic or electric feeders and they cost more than $100 each. A decent design you need about 50 to 100 feeders so go figure.

What could be a nice niche product is a machine that is in between a fully automated P&P and a manual semi automatic P&P. What I mean with that is that the machine is guided by joysticks or sliders from the user and the user can see from the camera viewpoint exacly where the nozzle is what the orientation is etc. now if you build something like that for about $500 you might have a nice product for hobbyists.

[Corporate666]

As others have said, there is no reason to try to build even a $500 PnP machine.  It's not like a 3D printer where having it lets you do things you could not otherwise do, even if the quality is shit.

A PnP lets you do something you could do with your hands - put parts on a board.  It only lets you do it with less human intervention - that's it.  And they do that by being very reliable.  A cheap machine isn't going to be very accurate and it's not going to be very reliable, so it's also not going to be very useful.  The only reason to automate PCB parts placement is because you have a lot of it to do.  That means you're almost certainly selling the product that you are placing.  Which means you won't have time/inclination to "save" hours of placing boards which you instead spend babysitting and handholding a cheap machine.

And the main thing....as was written above, everyone who gets into the PnP game totally misunderstands it.  X/Y motion is trivial.  The hard part is FEEDERS.  Everyone thinks about the motion part.  That's like designing a new car and focusing on the size of tires you will use.  Feeders are the ONLY thing you need to solve.  Only after that is in place should you even think about working on anything else.

And the next things should be software (especially ease of use and programming), and flexibility.

There is a reason commercial PnP machines cost so much.  It's not because they are just ripping off everyone buying one.  It's because it's much much more tricky than it seems at the outset.

[mikeselectricstuff]

a random person on a random YouTube video mentioned that if someone can make a pick and place machine for under $200 it'll revitalize the DIY electronic hobbyists.

That is just BS and i tell you why.
1-Any P&p machine to setup reliably for one design/product costs ten times more time than to place it manually. Ergo for hobbyists that do a few pcbs a p&p machine is no improvement.

2- the XY p&p mechanics is common knowledge, tens of designs out there nothing new, openPNP sw available. The real problem is in the parts feeders. The low cost solutions are crap, the only good and reliable solutions are automatic pneumatic or electric feeders and they cost more than $100 each. A decent design you need about 50 to 100 feeders so go figure.

What could be a nice niche product is a machine that is in between a fully automated P&P and a manual semi automatic P&P. What I mean with that is that the machine is guided by joysticks or sliders from the user and the user can see from the camera viewpoint exacly where the nozzle is what the orientation is etc. now if you build something like that for about $500 you might have a nice product for hobbyists.

For small jobs, there may be scope to handle the feeder issue by using vision to pick loose parts from trays, as vision is extremely cheap nowadays.
But you still won't build a machine for $200 that is useful enough to be worth doing.
For $500 you might be able to make something that's useful for a narrow niche of applications, like placing large qtys of the same, not-too-critical part ( e.g. LED arrays), for which traditional machines are overkill.

[KE5FX]

For small jobs, there may be scope to handle the feeder issue by using vision to pick loose parts from trays, as vision is extremely cheap nowadays.

This has always seemed like an obvious thing to ask for, but few people seem to agree.  How hard could it be to design something that does the following, specifically for prototyping purposes?



1) For each part to be placed, navigate to bin or tray containing parts of required type

2) Pick up loose, unoriented part with vacuum pen of appropriate size/power

3) Temporarily drop part on staging surface and capture high-resolution image of dropped part

4) Part orientation recognized (i.e., it didn't land upside-down or on its side)?  If YES, go to 5, else go to 2

5) Pick part up again, rotate X/Y axis as required, translate as required, and place on board

Yes, a machine like this would be slow as hell, unusable for even low-volume production.  (Imagine how long it might take before one of those aluminum electrolytics in the photo above landed on its feet.)  But it would also be cheap as hell to purchase and operate, because it would require no feeders and no supervision.  Put board on table next to part bins, start machine, go to bed, and in the morning a thousand parts have magically been placed as if by elves.

Every time I ask if something like this exists, people laugh at me.  Eventually, the laughter will stop and money will change hands.  Who wants some?

[NickPerry]

wow! didn't expect to get this much feedback before my next break, awesome!

so here are the points I see.

-Figure out a way to have the feeders work automatically (so no taping strips of pats to the work surface and having the machine pick from that limited selection)
-High REPEATABLE accuracy, so basically some form of optical encoder (or similar) that keeps track of any lost steps and corrects for them.
-good software (no worries there, I plan to build my own software, I tend to find that easier than trying to integrate someone Else's soft.)


KE5FX :
I'm one of the many people on here with the knowledge to do that on here, but one of the few nut's enough to actually do it.

HOWEVER, this would (most likely) not make as much money as you seem to think (it would only be useful to hobbyist, and hobbyist don't have much money usually)
luckily I don't do things for the money, I do things for the challenge ^.^

second of all, I wouldn't 'drop' the part and hope it lands right side up, I would use a 5 axis arm.

third of all, programming an AI that can tell one part from another when they are overlapping and jumbled like that is EXTREMELY difficult. but I like a challenge (I already have some Idea's on how to remedy it)

I have to many other projects on the go right now to look at this seriously right now, but when I finish the P&P machine, remind me again and I'll have a look.

Thank you for all the input so far guys! I'm touched ^.^

[MT]

For small jobs, there may be scope to handle the feeder issue by using vision to pick loose parts from trays, as vision is extremely cheap nowadays.

This has always seemed like an obvious thing to ask for, but few people seem to agree.  How hard could it be to design something that does the following, specifically for prototyping purposes?



1) For each part to be placed, navigate to bin or tray containing parts of required type

2) Pick up loose, unoriented part with vacuum pen of appropriate size/power

3) Temporarily drop part on staging surface and capture high-resolution image of dropped part

4) Part orientation recognized (i.e., it didn't land upside-down or on its side)?  If YES, go to 5, else go to 2

5) Pick part up again, rotate X/Y axis as required, translate as required, and place on board

Yes, a machine like this would be slow as hell, unusable for even low-volume production.  (Imagine how long it might take before one of those aluminum electrolytics in the photo above landed on its feet.)  But it would also be cheap as hell to purchase and operate, because it would require no feeders and no supervision.  Put board on table next to part bins, start machine, go to bed, and in the morning a thousand parts have magically been placed as if by elves.

Every time I ask if something like this exists, people laugh at me.  Eventually, the laughter will stop and money will change hands.  Who wants some?

yeah i thought the same, why use feeders at all but a single slow but precise robot arm with a rotating
vacuum pipet/s with inbuilt pressure sensor aided by some clever shape recognition software?
Surely someone have thought about this long time ago? lets laugh together over the idea!

[EEVblog]

A $200 P&P will be of absolutely no use to anyone.

Yep, guaranteed.

[NickPerry]

A $200 P&P will be of absolutely no use to anyone.

Yep, guaranteed.

which is why I mentioned.

(sub $500, not sub $200)

Just watch me well... don't watch me.. I'm a slow worker and I'm sure you have much more productive things to do.. but now I'm very driven to do this, sub $500 machine, 'better*' than the neoden 4.

*Smaller footprint, less reals, but good accuracy, and little to no fiddling out of the box.

lol, I know I'm sounding hard headed now, but how can I say 'yeah you're right, I won't even try' now?

[coppice]

For small jobs, there may be scope to handle the feeder issue by using vision to pick loose parts from trays, as vision is extremely cheap nowadays.

That's how the passives for the biggest jobs are handled now. Large vacuum sealed blocks of passives are simply torn open and poured into hoppers. From there  a vision system pick them out, orients them, and places them.

[Alex Eisenhut]

Make a 200$ this:

[donotdespisethesnake]

(sub $500, not sub $200)

Just watch me well... don't watch me.. I'm a slow worker and I'm sure you have much more productive things to do.. but now I'm very driven to do this, sub $500 machine, 'better*' than the neoden 4.

*Smaller footprint, less reals, but good accuracy, and little to no fiddling out of the box.

lol, I know I'm sounding hard headed now, but how can I say 'yeah you're right, I won't even try' now?

Unfortunately, this is the wrong forum to ask. Anything less than a professional quality machine is treated negatively, and you won't get much constructive advice.

I have had a desire to build a cheap P&P project for a while, so I'm quite interested in a sub $500 build. I would buy one at that price, but not >$1000. If you are building hardware, it is pretty much impossible to beat cheap Chinese manufacturers. I think a practical budget for a "home build" might be $750.

Of the recent DIY P&P machines, I think LitePlacer is the only one that is usable and can be bought, but at €1700 for a kit, it is not that cheap. At that price, buying a Neoden might be better.

There are cheaper ways to handle reels than individual reel feeders, I've been looking at approx $10 per lane. But handling loose strips and bulk would be useful for hobby/prototype builds. I have some ideas how to reduce setup times as well.

Unfortunately, vision systems are not a very good substitute for precision machinery, but at least vision recognition is getting cheaper and easier.

[mikeselectricstuff]

For small jobs, there may be scope to handle the feeder issue by using vision to pick loose parts from trays, as vision is extremely cheap nowadays.

That's how the passives for the biggest jobs are handled now. Large vacuum sealed blocks of passives are simply torn open and poured into hoppers. From there  a vision system pick them out, orients them, and places them.

[Mechatrommer]

Make a 200$ this:

typical problem with people around the globe including the jedi master is they expect to pay a dime for a ferrari... just as asking, make $50 this...

[Skashkash]

What could be a nice niche product is a machine that is in between a fully automated P&P and a manual semi automatic P&P. What I mean with that is that the machine is guided by joysticks or sliders from the user and the user can see from the camera viewpoint exacly where the nozzle is what the orientation is etc. now if you build something like that for about $500 you might have a nice product for hobbyists.

This has been my highly desired device for the last 10 years. Fills that niche between hobby and production.

Been thinking about it a lot.

I've been waiting for someone else to build it so I would not have to do it myself. Make it flexible enough,  and you can also us it for watch repair  . 

[NickPerry]

Before work today I kinda picked through some of my extra parts I had from my cnc machine (I know I have stepper controllers somewhere but I couldn't find them quickly)
these steppers were to small for my mill, but should be just fine for this, and I'll use those bearings to make the slides.

I'm low on angle Iron so I'll have to pick up some more. but I have plenty of aluminum sheeting to make it look pretty :p (it's for my E-bike project, but I should have plenty left over for this)



REPLIES________________

wilfred: just building a simple 2 axis part mover is indeed the first order of business (since I have the parts for that already :p) and I'll build up from there.

coppice: I think I'll stay with reels for now since I have a good Idea what I want to do for that

Alex Eisenhut: lol, just the material cost for that is well over $200 :p pretty cool though

donotdespisethesnake: No worries, their comments may be a bit harsh, but not without information, which is the most important part I'll just have to actually build it, I completely understand that there are a lot of people claiming to do something crazy and then never following through. once I have the frame welded  up and whatnot the conversation will shift, that's always the case ^.^ (Thank you though!)

mikeselectricstuff: that's pretty cool! and uses some of the same techniques I was thinking for KE5FX's Idea (using a shaker plate to separate a small quantity making it MUCH easier for an AI to locate the different parts.)  but no, I'll be using reels for now

Mechatrommer: why? for hobby/small scale just use a normal soldering bath off Ali.

Skashkash: mine should be able to do that (since I plan to have a camera on the nozzle end, and manually being able to control the axis is a given) but it would be much easier (for the end user) if it's automated.


END_________________________________________-

Just a note, I only have time to work on this project Monday's and Tuesdays, and also between the times I'm not working on my E-bike project. so expect a slow build.

Thank you for all the feedback!

[IanJ]

As my production requirements are pretty low (I'll do 25 pcb's as a batch run normally, once every few weeks)........I'll be happy with:-

- Sub $1000 p&p.
- Effective vision system.
- Hopper for passives.
- Tape for everything else (strip of tape).
- Slow is fine, and I don't mind having it process say four 100mm pcb's at one go.
- A nice GUI for setting up the job.

Saying that, I think design & build of a p&p machine is easily under estimated.....there's a lot of tech to be mastered.....so good luck!

Ian.

[Jeroen3]

I've made a simple hand assembly jig. A simple wooden table with a hovering armrest sliding horizontally with two linear bearings.*
When building this, I noticed how incredibly expensive good linear bearings are. For $200, you'll have a hard time getting the bearings.

I also thought of why there isn't a vision based PNP with tray loading for all parts. But then I thought it would be too expensive anyway.
You, as end user, would have to save time or money using the machine. Saving on outsourcing or manual assembly time.
I think you'd be busy configuring the expensive machine, so you're not saving time. It would need very comprehensive software to work. Expensive software.

Assembling high tech small electronics is just an expensive business. It costs ~€5K for 25 board prototype with an 8 week turnover.
Better be right first time!

^{*Similar to this, except you are the Y axis and hold the tweezer.}

[Kjelt]

For small jobs, there may be scope to handle the feeder issue by using vision to pick loose parts from trays, as vision is extremely cheap nowadays.

The video you posted from ASM is usable for 0402 and smaller. For resistors this means there is no value printed on them so either side is ok to place.
For 0603 and larger relatively more of the components can not be picked.
Besides any dirt, hairs or other contamination could bring havoc.

For a starter hobby P&P where reels are to expensive there should be some kind of sorted stock. Like the ics that are placed in a tube and vibrated I was thinking of placing passives (if the same size)in order of placement in a tube so the p&p head can pick them up from one place instead of looking somewhere in a tape that is sticked. However one mispick would ruin all and the shift is to how to efficiently pick and order the components for filling the tube.
I guess the unsolvable feeder issue is what stops efficient low budget P&P machines.

[hlavac]

I would be interested in working on such a hobbyist level p&p - my hands are too shaky to do smd manually

Here are some ideas:

- Tape reels / cut tape are much better for storage and easy pickup. Forget hoppers

- I would split pickup and placing to two independent processes that run in parallel using sort of ammo clips for specific component shapes. One part deals with loading the clips with correct sequence of components from tapes while other moves over the pcb and dispenses components in optimal order

[hlavac]

Also it would be great if the machine was able to put solder paste on.
I imagine something like putting the paste on the component when placing it instead of the pcb
It would do away with expensive stencils and shelf time limitations of pasted pcb.

Or, alternatively, microsolder the components to the pcb

[ejeffrey]

Unfortunately, this is the wrong forum to ask. Anything less than a professional quality machine is treated negatively, and you won't get much constructive advice.

Plenty of people here would be happy for a hobbyist grade PnP.  But unlike a lot of other tools such as a soldering iron where even a $10 firestarter will do more than nothing, to be useful a PnP has to compete against "nothing" which is a perfectly viable option.  Building a placement tool for fun and to learn about motor control and machine vision is great, but if you want to make something that is useful as a product it has to be better than hand placement and cheaper than commercial assembly in some niche.  A lot of people here are skeptical that you can do PnP that is useful on its own merits for that kind of budget.  And the reason is that it is a really attractive idea.  PCB manufacture is now stupidly cheap, but small volume assembly is still relatively expensive, and a lot of the cost is labor / set up.  A hobbyist grade PnP is really attractive and a lot of people have tried and come to the conclusion that it is really hard to do on such a tight budget.