O-M-G cheap pick and place machines coming!

[Precipice]

I can visualise that it would be possible to make a very clever linkage that moves a pin up, pulls along, then drops down again from a single servo.

I always end up wondering about a few shark teeth, with ramps. Slide it forward, and it spring-loadedly drops out from the sprocket holes. Drag it back, and it lifts back into the holes, and drags the tape along.
It's be self adjusting for distance, too, no need to be super accurate with the throw length trying to line a pin up with a hole. It'd also spread the load across a bunch of holes. It could also be right at the front of the feeder, for more pull, less push, and less (zero?) need for leader.

[sleemanj]

If you did use 2 servos, you could potentially share the pull-along one across multiple lanes.
 

Small solenoid, horizontal under each lane.
When activated pushes on a linkage which extends a pin vertically through the tape.
Linkages and solenoids run on a horizontal sled under the lanes, driven by a single stepper (can be very thin for desktop, think like one of those flat geared steppers in modern scanner heads on printers, it only has to move a couple cm and speed records need not be broken)

Advance tape = actuate solenoid for lane, drive stepper tape pitch of that lane forward, deactivate solenoid, rewind stepper

If you could live with the added height, ditch the linkage and just mount the solenoids vertically under the tapes and they can go through the tapes themselves (with an appropriately pointy tip).  You can get really small cheap solenoids, very low cost.

Take up of the cover tape in my mind is the biggest problem, some sort of "clutch" mechanism is necessary there.

[EEVblog]

Just thinking about the minimum requirement for a standalone feeder, and taking a bottom-up approach.
What's the cheapest suitable actuator?

I've never really looked into it, but how tough are the mylar tapes that have to be pulled back anyway? Tough enough to stick between a pinch roller and the tape body itself and then drag back to advance to the next part?
No need for anything to drive the index hole.
Pretty crude, just throwing the idea out there.

[EEVblog]

Open Pick-N-Place Tape Feeder
http://www.thingiverse.com/thing:7291

[Precipice]

I've never really looked into it, but how tough are the mylar tapes that have to be pulled back anyway? Tough enough to stick between a pinch roller and the tape body itself and then drag back to advance to the next part?

Not sure they're mylar, and not quite strong enough. I'd say I get away with it about 90% of the time when pulling tape through my feeders if I can' be arsed to do it properly. I should really learn...
The tape also stretches and springs back a bit, which may also be a problem. (Spotting the sprocket hole(s) with a photodiode would be easy enough, though if anyone came up with a non-indexed drive solution. Vision could then compensate for any residual offset by looking at the sprocket holes before it picks.

Another thing to think about is that, if you're exposing more than one component at a time, you need to be somewhat gentle with both the acceleration of the tape, and the picking of the previous component, or you get components everywhere.

[Precipice]

I thought the U shaped path was a bit extreme and a straight through sprocket driven tape feed would be simpler.

You need the U shape, otherwise you end up with empty tape all over your PCB. this is a bad thing.

[mikeselectricstuff]

Just thinking about the minimum requirement for a standalone feeder, and taking a bottom-up approach.
What's the cheapest suitable actuator?

I've never really looked into it, but how tough are the mylar tapes that have to be pulled back anyway? Tough enough to stick between a pinch roller and the tape body itself and then drag back to advance to the next part?
No need for anything to drive the index hole.
Pretty crude, just throwing the idea out there.

Nope - on my machine the feed pin is about 4" from the feed position, so to get the last  few parts off tapes I manually pull the cover tape to advance parts, and there are a few issues that make this a nonviable approach.
Some cover tapes are quite stretchy.  If you get a kinked paper tape, you need significant force to pull it through, and on plastic tapes, it tends to pull off with 'stutter' in small sections instead of smoothly, and this would make the tape bounce, throwing parts out.
However a friction drive to the main tape, with optical feedback of hole position might work, though clear tapes could be problematic to detect reliably. Maybe you could do it with a small air jet (from a fan) , detecting pressure change.
You still have the issue of thickness and drive position - you couldn't friction drive plastic tapes as you'd crush them, and driving from the edge means you'd need to deal with the different thickness and texture of plastic and paper tapes.
 
 

[mikeselectricstuff]

I thought the U shaped path was a bit extreme and a straight through sprocket driven tape feed would be simpler.

You need the U shape, otherwise you end up with empty tape all over your PCB. this is a bad thing.

Not necessarily - look at the cheap Chinese machine - that shoves the waste tape under the PCB
If folding back, You need a moderately big U radius to deal with tapes with large pockets, which will be rigid.

[EEVblog]

Not necessarily - look at the cheap Chinese machine - that shoves the waste tape under the PCB

To me that's the really obvious and most elegant way to do it.
I'd put a polycarb sheet or whatever under the board area so tapes don't curl back up and ruin your day. The sheet catch all the dropped parts from the head, and you could pull it out from the end to get the dropped parts (the crumb tray)

[mikeselectricstuff]

If you did use 2 servos, you could potentially share the pull-along one across multiple lanes.
 

Small solenoid, horizontal under each lane.
When activated pushes on a linkage which extends a pin vertically through the tape.
Linkages and solenoids run on a horizontal sled under the lanes, driven by a single stepper (can be very thin for desktop, think like one of those flat geared steppers in modern scanner heads on printers, it only has to move a couple cm and speed records need not be broken)

Advance tape = actuate solenoid for lane, drive stepper tape pitch of that lane forward, deactivate solenoid, rewind stepper

If you could live with the added height, ditch the linkage and just mount the solenoids vertically under the tapes and they can go through the tapes themselves (with an appropriately pointy tip).  You can get really small cheap solenoids, very low cost.

That's exactly the mechanism used on the Versatronics feeders - they use electromagnets with hinged plates instead of solenoids though.
A gearmotor with a cam may be a cheaper alternative to a stepper, though gear wear may be more of an issue.

 

[mikeselectricstuff]

Not necessarily - look at the cheap Chinese machine - that shoves the waste tape under the PCB

To me that's the really obvious and most elegant way to do it.

It also significantly reduces the maximum pull force you need as you're not fighting the tape - it's going where it naturally wants to go.

[mikeselectricstuff]

I bookmarked this guys blog site some time ago since I thought his tape pull idea was interesting. Essentially it mimics the sprocket driven film feed of a 35mm camera except he has a "U" shaped feed path where the tape ejects  back in the direction of the incoming tape.
This is one image on his site. ( I thought it best to not rip his photo off, so I link to his site)

The URL for the site: http://tim.cexx.org/?tag=pickplace

I thought the U shaped path was a bit extreme and a straight through sprocket driven tape feed would be simpler. In his design prototype the two plates with the tape guide path are held together by a rubber band, which suggests a means where it can be adjusted to different width tapes.

I don't think a simple sprocket wheel is viable as it will inevitably pull the tape downward at some point, which can make plastic tape bounce. Thick paper tapes may also present a lot of friction, which will end up pulling a lot of paper dust out & into the mechanism.
It really needs to be a pretty clean up-along-down motion, ideally with a little backward motion before the downstroke so it doesn't rub along the edge of the hole
 

[Precipice]

It also significantly reduces the maximum pull force you need as you're not fighting the tape - it's going where it naturally wants to go.

Yeah, into a massive tangled snarl under the PCB. (I'm envisioning feeders on 2, maybe 3 sides).

[Richard Crowley]

massive tangled snarl under the PCB. (I'm envisioning feeders on 2, maybe 3 sides).

Can't they simply re-spool the tape back onto a reel (or simply a hub with removable flange)?
Even back in the days of magnetic audio and then video tape, we spooled the tape onto a hub, flanges optional

[mikeselectricstuff]

I don't think a simple sprocket wheel is viable as it will inevitably pull the tape downward at some point, which can make plastic tape bounce. Thick paper tapes may also present a lot of friction, which will end up pulling a lot of paper dust out & into the mechanism.
It really needs to be a pretty clean up-along-down motion, ideally with a little backward motion before the downstroke so it doesn't rub along the edge of the hole

I'm not sure how much of the site you looked at, and it isn't obvious at  first glance in the single image I linked to, but the tape runs in a narrow channel and the sprocket holes are exposed only at the drive wheel position. If the tape continued on in the channel beyond the drive wheel then it could not bend. Not as in the picture where the channel curls around at the diameter of the drive wheel and ejects the tape at the bottom.

With a rotary sprocket you'll always get some downward force on the tape as it's being driven forward

[mikeselectricstuff]

massive tangled snarl under the PCB. (I'm envisioning feeders on 2, maybe 3 sides).

Can't they simply re-spool the tape back onto a reel (or simply a hub with removable flange)?
Even back in the days of magnetic audio and then video tape, we spooled the tape onto a hub, flanges optional

Yes, but you then need enough leader to fix it to the takeup reel

[Neil Jansen]

Yep, unfortunately it's not going to be easy or quick.

Nothing worthwhile ever is.

Just look at Makerbot, arguably one of the major drivers of the out-of-the-box success of 3D printers. Took them 3+ years, a lot of money, and a lot of people before they started to get it right, and many say they still haven't. My Makerbot Replicator(1) for example is a dog when it comes to extruder reliability.
And 3D printers are relatively simple devices compared to PnP machines, and that's why, realistically many are always going to be quite skeptical about a PnP.
Not that it can't be done, it certainly can, it's just whether is can be done for low enough cost?
Although remember, the 3D printer price points are above $1K or even $2K for decent machine that works well.
No one expect a $300 PnP machine. Heck, you can't even get a half decent 3D printer for that. You can argue it's pointless to even try and meet a price point, better to focus on getting a solution that is robust and works. And then it costs what it costs.

Makerbot isn't a great example of a great turnkey printer, IMO.  It's got real flaws, like it's inferior extruder and lack of a heated bed, when compared to other designs.  I speak from experience here, as I've got a ton of 3D printing background, that's what I did before starting the PnP stuff.  I've used the Makerbot plenty, as well as many other commercial and open-source versions.  I'm intimately familiar with the ways that each of them attempted to solve the various problems of 3d printing, and what worked, and what didn't.  If you want an example of a real turnkey 3d printer that is useful, the Lulzbot TAZ 4 is a great machine (and currently retails for less than the Makerbot).  Also, for DIY, the Prusa i2/i3 with J-Head and MK2B heated bed is a champ, if you're careful with sourcing your parts.  I built a Prusa i2 for $250 in one weekend, back in January, and have put about 25 kg of PLA and ABS through it, without a single clog (I print so much that my filament vendor knows me by name and sends me free tailings of neat colors).  I don't know where some people have this idea that you have to spend more time tweaking a 3D printer than using it.  I certainly don't.  Mine prints almost 24/7 sometimes, for weeks on end, and I've never really had to do anything to it since I built it. 

As far as the $300 price goes, it's not as absurd as it sounds.  That price would be for a barebones kit of parts, mostly sourced from China.  You would be responsible for building it, wiring it, and printing your own feeders using the 3d print nozzle.  The delta design uses considerable less parts and less motors than the Prusa i2 that I build for $250, and that wasn't sourced from China.  I don't plan on making any profit from the cheap kits, I plan on making my profit from taking the same design, and selling a full-featured, fully-assembled commercial machine with a full set of feeders/nozzles, with support and warranty, for $5000 - $10000.  That's obviously assuming that we can meet all the technical challenges that we face.  As Dave said, it won't be easy or quick, but that's never stopped me before.  This is what I want to do with my life, I'll figure out how to do it one way or another, I don't just give up so easily.  The advice we've had here is great and we'll definitely try to work some of it in as we move forward.

[EEVblog]

If you want an example of a real turnkey 3d printer that is useful, the Lulzbot TAZ 4 is a great machine (and currently retails for less than the Makerbot)

Yeah, but that's still a $2K printer, and that's my point.
People will it seems happily pay $2K for a 3D printer that really works. So it's easy to think that will pay at least the same for a PnP machine that really works.

As far as the $300 price goes, it's not as absurd as it sounds.  That price would be for a barebones kit of parts, mostly sourced from China.  You would be responsible for building it, wiring it, and printing your own feeders using the 3d print nozzle.  The delta design uses considerable less parts and less motors than the Prusa i2 that I build for $250, and that wasn't sourced from China.  I don't plan on making any profit from the cheap kits, I plan on making my profit from taking the same design, and selling a full-featured, fully-assembled commercial machine with a full set of feeders/nozzles, with support and warranty, for $5000 - $10000.

And you'll most likely make a killing, if it works reliably.
A lot of people are already paying close to $5K for that Chinese one.

One potential problem I can see with the $300 DIY version, is that it's just that DIY, and people will, with absolute certainty, screw it up and it won't work right.
And well, ok, so be it, but the thing is that that could reflect badly upon your higher end fully built model if "word gets around" (even if it's not your fault) that the product has issues. Just something to be cautious of by trying to shoot for a low end low cost DIY model first.

The 2nd issues about shooting for a low target price for the base model is as I have mentioned before, having that in the back of your mind may ultimately cause a compromise in the design that turns out to be the Achilles heel for reliability or accuracy etc that unwittingly flows into your high end model.

e.g. you've chosen and presumably locked into the Delta design, and that looks great now, but what if it turns out to just not be viable right on the edges of performance where you need it and people expect a $5K machine to work at?
And even though I don't know much about this stuff, that's what my engineering spidey sense is telling me might possibly happen here. So personally I would likely have gone for pretty much what the Chinese design has, and then figure out ways to either tweak the cost, or add a killer feature etc.

But hey, I could be wrong, take with a grain of salt. I think what you are doing is great.

[Neil Jansen]

If you want an example of a real turnkey 3d printer that is useful, the Lulzbot TAZ 4 is a great machine (and currently retails for less than the Makerbot)

Yeah, but that's still a $2K printer, and that's my point.
People will it seems happily pay $2K for a 3D printer that really works. So it's easy to think that will pay at least the same for a PnP machine that really works.

As far as the $300 price goes, it's not as absurd as it sounds.  That price would be for a barebones kit of parts, mostly sourced from China.  You would be responsible for building it, wiring it, and printing your own feeders using the 3d print nozzle.  The delta design uses considerable less parts and less motors than the Prusa i2 that I build for $250, and that wasn't sourced from China.  I don't plan on making any profit from the cheap kits, I plan on making my profit from taking the same design, and selling a full-featured, fully-assembled commercial machine with a full set of feeders/nozzles, with support and warranty, for $5000 - $10000.

And you'll most likely make a killing, if it works reliably.
A lot of people are already paying close to $5K for that Chinese one.

One potential problem I can see with the $300 DIY version, is that it's just that DIY, and people will, with absolute certainty, screw it up and it won't work right.
And well, ok, so be it, but the thing is that that could reflect badly upon your higher end fully built model if "word gets around" (even if it's not your fault) that the product has issues. Just something to be cautious of by trying to shoot for a low end low cost DIY model first.

The 2nd issues about shooting for a low target price for the base model is as I have mentioned before, having that in the back of your mind may ultimately cause a compromise in the design that turns out to be the Achilles heel for reliability or accuracy etc that unwittingly flows into your high end model.

e.g. you've chosen and presumably locked into the Delta design, and that looks great now, but what if it turns out to just not be viable right on the edges of performance where you need it and people expect a $5K machine to work at?
And even though I don't know much about this stuff, that's what my engineering spidey sense is telling me might possibly happen here. So personally I would likely have gone for pretty much what the Chinese design has, and then figure out ways to either tweak the cost, or add a killer feature etc.

But hey, I could be wrong, take with a grain of salt. I think what you are doing is great.

Cool!  Sounds like we've hit some common ground, more-or-less    I believe that the uber-EE's that frequent this forum will like our $5000-$10000 machine more than our $300 kit.  That's always been the plan, however, the Hackaday version is focused solely on the cheap kit, which has made for some confusion.   The two-tiered approach may seem a bit strange, but software companies do that all the time as a business model (e.g. Github, Trello, and other companies having a 'freemium' model and an paid enterprise version; they take a loss on the former and profit on the latter).  As for going low-end and having design compromises / achilles heel / etc, that's always a concern for any company that wants to make a consumer product.  There was that first engineer that wanted to start using plastic parts under the hood of an automobile, and was probably laughed at, but now other companies have followed suit and even well-built Toyotas and BMWs have all sorts of cheap clever solutions made out of low-grade parts.  That's innovation.  It's not as glorious as other types of innovation, but it's still progress.  I'm actually on the third revision of the delta mechanism, and the fourth revision of the component feeders (still need one more rev after the first full system tests).  If we find something that works, then others can copy us (abiding by the CC-BY-SA, hopefully).  If we fail, we'll be an example of what not to do.  We'll definitely be cautious and will be sure to document the crap out of whatever we come up with... Including troubleshooting, etc.  The delta design is REALLY hard to screw up, we're doing lots of camera calibration before, and even during the job, to account for any misalignment that could happen.  This design errs on software complexity more than hardware complexity, which is a unique approach to PnP, but is an approach that's being used more these days in other fields like transportation.

Edit: editing after seeing edits to your last post.  If our delta mechanism is a dead end, we'll try CoreXY or something like that, we won't either (A) just give up or (B) try to make it work even though it's flawed.  From my calculations, which I've documented on my HaD blog, I believe that I can make it work.  So far it looks promising, but it does require a bit of work envelope calibration using a multi-point Z-probe and XY grid points for the camera to inspect and calculate a vector error map.  But again, that's all software, and once we get the calibration stuff in place, it will work pretty well.  Definitely not as easy as Cartesian machines with ballscrews and glass scales.

[fcb]

Pulling the cover tape.

I thought Vbesemens had a neat way of doing this, that would kill two birds with one stone - assuming that the cover tape is strong enough.



If this was done with multiple steppers (offset, so you could use the bigger cheaper units), I think this would be quite viable - even using those very cheap motor/gearbox assemblies.  I'd consider using optical hole detection (clear plastic tapes???) though or perhaps a jockey wheel/encoder thing, the wear on a microswitch could be high (paper tape is quite abrasive), leading to occasional replacement.

The other thing, I think it will only be a matter of time before a Chinese firm replicates the TM220A/240A 'success' story with a vision based machine. These guys seem to have already done it.

http://www.hothotsmtmachine.com/

[mamalala]

Just thinking about the minimum requirement for a standalone feeder, and taking a bottom-up approach.
What's the cheapest suitable actuator?
RC micro servo has to be it by very long way. Has the right amount of movement and torque, and a doddle to control very cheaply.

Hmm, i would go for a simple, tiny DC motor with a worm gear, driving a falt wheel with sprockets on one end. Some motor like these: http://www.ebay.de/itm/1-Stuck-Motor-Johnson-Kleinmotor-13-V-mit-Schnecke-/310803801758. And then add a LED plus phototransistor to detect the holes on the side of the strip, that way it can advance one, two, three, or whatever many holes to have the next part. I guess that a servo would have a too limited range for larger parts.

Just send it a pulse, and it will advance as many holes as are set (configurable by dip-switches, for example). Only requires a very small micro and a FET to switch the motor. Should be really simple electronics.

Greetings,

Chris

[mikeselectricstuff]

Pulling the cover tape.

I thought Vbesemens had a neat way of doing this, that would kill two birds with one stone - assuming that the cover tape is strong enough.

This will fail as soon as you get a kinked tape (like you often get when you order from farnell), and also bear in mind that torque will decrease as the takeup reel fills up

Thinking about it, a reflective optical sensor could probably be made to work on all tape types - diffuse reflection from paper, specular reflection from plastic.  This would also have the advantage of not having to have something each side of the tape

[Precipice]

Just send it a pulse, and it will advance as many holes as are set (configurable by dip-switches, for example). Only requires a very small micro and a FET to switch the motor. Should be really simple electronics.

And a feed pushbutton, on the actual feeder unit, pleeeeease!

When threading, often while balancing on one foot and contorted into a weird shape to reach the feeder in the far corner, I really can't guide the mouse (or trackpad, thanks, Mike!) onto the feed button on a screen a few feet away.

[mikeselectricstuff]

Just send it a pulse, and it will advance as many holes as are set (configurable by dip-switches, for example). Only requires a very small micro and a FET to switch the motor. Should be really simple electronics.

And a feed pushbutton, on the actual feeder unit, pleeeeease!

Absolutely. And of course it always feeds a single index regardless of setting, so you can get it in phase.

And no need to bother with mechanically half-indexing unless the feeder design make it easy to do - you can do that by offsetting the pick position.

[Precipice]

This will fail as soon as you get a kinked tape (like you often get when you order from farnell), and also bear in mind that torque will decrease as the takeup reel fills up

Thinking further about inconvenient tape supplies, short or absent leaders and all the other things that plague short runs...
Could one build a feeder that (optionally?) ran with a plate over the top, and peeled the cover tape away some distance back, to avoid exposing multiple pockets, or requiring extension pieces to be attached to the cover tape?
Just jamming your 100-piece tape in the back of the feeder, pulling the tape off to feed the carrier into the feeder, and accepting that, at the end of the job you toss the remainder of the tape and components... Pull the tape from the front, to make sure you can feed all the parts.
The passive strip holder trays that are used with Versatronics (and others) are almost OK, but I always fear the components pinging everywhere. Since proper feeders are expensive, and many customers won't have full reels for a lot of their parts, supporting cut strip as well as possible would be a win! (Hmm, have I just talked myself into a few proper feeders, and a load of drag feeds with fishing weights for the one-part-per-board poxy passives?