[suske]

The method of tape feeding is certainly a neat solution. The length of the tape feeding block is probably to get enough friction on the tape that is won't move (pull out), and it looks like the tape is aligned in channels.

How are the tape covers striped with this approach?

I think tape cover need to be removed manual with this approach.

In video from http://www.apsgold.com/l-series.html I get many ideas for some solutions. I like their components centering way. That centering tool mounted on Z axis is good idea and not so complicated. It improve speed by fact that centering is don "on the fly" and no need for additional movements.

If you wont simple and low cost solution for tape feeder then you need to escape individual actuator for separate tape. That is one reason why is P&P from "hobby cnc" site good solution. Simple tape cover remover can be made to improve tape feeder and so on.

I hope all this can help.
Sorry, it is hard to me to express myself in English.

Best regards,

Suske.

[Eclipze]

Nice video... great to look at for ideas!

The squaring fingers were nice.

Your english is coming through very well suske

[suske]

Nice video... great to look at for ideas!

The squaring fingers were nice.

Your english is coming through very well suske

Yes it is video that give me many ideas. :idea:

About my English. We obviously do not share same opinion, but thank you anyway.

Truly, I am hopeful that you guys while be tolerant to my misspell, bad grammar and so on.
It is not a problem until you catch what I wish to say.

I see lot of work for us ahead...

Best regards,

suske.

[ciccio]

Have you seen the new Manncorp ECM-EXP on ebay?
http://cgi.ebay.com/ws/eBayISAPI.dll...ayphotohosting
Some interesting things:
Professional machine, but not so heavy ( only 330 lbs)
T-slotted aluminum structure, similar to lot of homebuilt CNC machines
Toothed belts
Do you have more infos?

[suske]

Have you seen the new Manncorp ECM-EXP on ebay?
http://cgi.ebay.com/ws/eBayISAPI.dll...ayphotohosting
Some interesting things:
Professional machine, but not so heavy ( only 330 lbs)
T-slotted aluminum structure, similar to lot of homebuilt CNC machines
Toothed belts
Do you have more infos?

More interesting pics here http://www.manncorp.co.uk/.
This thread is slightly quiet in last few days.

Best regards,

suske

[ciccio]

Hi Suske,
tell us more about your P&P project
Manncorp EXP pick and place machine , for the first time, uses the very same technology used in lot of hobby machines: toothed belts, T-slotted aluminium profiles, etc.
I have never seen a professional P&P so similar to home-made CNC machine.
Now the problem seems to be:---software---
any idea about video cameras?
what software do you think to use?
saluti
ciccio

[ppBuilder]

More interesting pics here http://www.manncorp.co.uk/.
This thread is slightly quiet in last few days.

Best regards,

suske

I tried to have a look at the pics but they wouldn't load...tried 3 different browsers...probably one of those MSIE ultra super optimized sites...
Could you skim the pics and send'em to me, or post them in the files area maybe?

[suske]

I tried to have a look at the pics but they wouldn't load...tried 3 different browsers...probably one of those MSIE ultra super optimized sites...
Could you skim the pics and send'em to me, or post them in the files area maybe?

I can't see pics of "MC-383" type but pics of other two types
are loaded in my Opera browser. It is problem with their web page I think.
If you can't load pics for other two types of P&P machine, I can send it to you.

Best regards,

suske

[suske]

Hi Suske,
tell us more about your P&P project
Manncorp EXP pick and place machine , for the first time, uses the very same technology used in lot of hobby machines: toothed belts, T-slotted aluminium profiles, etc.
I have never seen a professional P&P so similar to home-made CNC machine.
Now the problem seems to be:---software---
any idea about video cameras?
what software do you think to use?
saluti
ciccio

I am in planing phase. Not so much to show now.
I plain to build multi function CNC table, not only P&P machine. Not so happy with wood solutions so I while build an metal CNC machine. Yesterday I start with some experiments. For now I only weld part of X table to see how it looks like. Can't wait to go ahead.
General, I plain to use servo instead stepper motors. Working area is something about 800 x 600 mm.

Software, hm, I don't plain to use camera centering. It is too high tech for hobby CNC. I more like mechanical component centering way.
Have no ideas about software for now. If there is additional electric devices as feeder actuators, squaring fingers etc. than usual CNC software while not work.
I will think about that later. For now hardware is in first place.

Best regards

suske

[ppBuilder]

I can't see pics of "MC-383" type but pics of other two types
are loaded in my Opera browser. It is problem with their web page I think.
If you can't load pics for other two types of P&P machine, I can send it to you.

Best regards,

suske

Of course I only tried the MC-383! Had a look at the big one and the pics load ok.

[ppBuilder]

I am in planing phase. Not so much to show now.
I plain to build multi function CNC table, not only P&P machine. Not so happy with wood solutions so I while build an metal CNC machine. Yesterday I start with some experiments. For now I only weld part of X table to see how it looks like. Can't wait to go ahead.
General, I plain to use servo instead stepper motors. Working area is something about 800 x 600 mm.

Software, hm, I don't plain to use camera centering. It is too high tech for hobby CNC. I more like mechanical component centering way.
Have no ideas about software for now. If there is additional electric devices as feeder actuators, squaring fingers etc. than usual CNC software while not work.
I will think about that later. For now hardware is in first place.

Best regards

suske

Same here. The software for a thing like this is almost trivial, except maybe for the bottom camera image recognition & processing. But that too should not be too hard if one could eliminate the background clutter of the head.

Mechanics first , then software!
Tape feeders have until now been a major headache. Have worked out a simple consept that should work well, and I might actually build one such feeder, just for the hell of it... But making 20 or more as I think would be required for a usefull machine just isn't worth the effort. And then comes stick and tray feeders. Only way I can see around this problem is to use simple strip feeders and pealing manually.

[suske]

Same here. The software for a thing like this is almost trivial, except maybe for the bottom camera image recognition & processing. But that too should not be too hard if one could eliminate the background clutter of the head.

Mechanics first , then software!
Tape feeders have until now been a major headache. Have worked out a simple consept that should work well, and I might actually build one such feeder, just for the hell of it... But making 20 or more as I think would be required for a usefull machine just isn't worth the effort. And then comes stick and tray feeders. Only way I can see around this problem is to use simple strip feeders and pealing manually.

Yes it software is trivial but think there is no hobby version of such software. It going to be lot of work for some programmer guy.
Elimination of background isn't so big problem. On the professional P&P there is one disc over the head which cover camera scanning area to make easy component recognition.

Tape feeder can bee made as an smart device which sense when head pick the component and then feed tape forward, but this is complicated solution and expensive too.
Simple way is to pool tape with Y or X axis. Pealing can be done with help of some rubber roller coupled with a tape by friction and on the other side over belt with other pulley roll tiny tape cover. Easy to say huh.
Hope you understand me.

btw. I can't see way folks on this forum more like wood than metal? I like to weld rather than cut and glue wood. My old arc welder is just fine.

Best regards,

suske

[Eclipze]

I don't think the precision required by a PnP could be acheived with wood, particularly if you using two rails on an axis. Wood swells with humidy and is not precision cut material.

[ppBuilder]

Yes it software is trivial but think there is no hobby version of such software. It going to be lot of work for some programmer guy.
Elimination of background isn't so big problem. On the professional P&P there is one disc over the head which cover camera scanning area to make easy component recognition.

Tape feeder can bee made as an smart device which sense when head pick the component and then feed tape forward, but this is complicated solution and expensive too.
Simple way is to pool tape with Y or X axis. Pealing can be done with help of some rubber roller coupled with a tape by friction and on the other side over belt with other pulley roll tiny tape cover. Easy to say huh.
Hope you understand me.

btw. I can't see way folks on this forum more like wood than metal? I like to weld rather than cut and glue wood. My old arc welder is just fine.

Best regards,

suske

Yes wood is for carpenters! A 'lumber' p&p machine is doomed to fail...better not even try it!

Tape feeders: Can't be made simple enough! And the small size is a problem, as they have to be as narrow as possible to make room for more of them. Most components comes on 8mm tape and such a feeder I think should not be wider than 16mm...not much room for motors and mechanics inside.
The problem of different feed indexes have to be addressed. Most common are 1, 2, 3, 4, 5 (X4mm). Holes are always 4mm apart.
I have considered:
Pneumatics: although the simplest it will cost a fortune in pistons, solenoid valves, fittings and other stuff. Even the smallest pistons might not fit inside. A 'large' shop air compressor is needed.
So pneumatics are ruled out!

Electric motors: may be cheap, but complicated. DC motor, stepper, DC servo...none are simple, and all will require electronics,optostuff, maybe even a computer chip for control. Again space is a big problem.
So motors are ruled out!

Solenoids: fairly simple, medium cost. Big problem with speed...the chips may jump out of their pockets. For larger components with larger indexes the solenoid will have to be 'clicked' several times to fully feed the next component...noisy and time consuming.
So solenoids are ruled out!

Mechanical: This is by far the most attractive proposition for an automatic feeder...fully passive, no electrics, not even for signalling empty tape. Energy is taken from the X-axis (head). NO extra head movement is required for feeding so no time is wasted. A head mounted plunger can plunge into the feeder and drive the tape forward and at the same time turn the peel reel. An even better solution is to have a spike or protusion on the head, and as it moves over the feeder it engages a small lever and cocks a spring in the feeder. When the head moves away the lever is slowly released and the spring moves a rachet forward pushing the tape. An manually adjustable stop infront of the rachet takes care of the indexing. End of tape can be detected mechanically inside the feeder and locking the lever...this is sensed by the head mounted plunger (spring loaded with force sensing u-switch).
Mechanical is not ruled out just yet...but even though it is the simplest of all designs it may still be too complicated considering one need 20 or more of these feeders.

[Eclipze]

I'm certainly in favour of use the head to feed the tape along. It simplifies a great deal and provides the most flexibility and most likely reliability.

The stripping of the tape however is the hard one. I'm not sure how to accommplish this in the design. I don't want to spool it, as that implies variable rotation and possiblity difficultly in removing later. I'd prefer a solution that ejects it out the back of the table and falls in a heap (better yet, shreds it into a bin!

[ppBuilder]

...........
The stripping of the tape however is the hard one. I'm not sure how to accommplish this in the design. I don't want to spool it, as that implies variable rotation and possiblity difficultly in removing later. I'd prefer a solution that ejects it out the back of the table and falls in a heap (better yet, shreds it into a bin!

The peeling process itself is simple; First peel off a length manually, bend it backwards over a roller or pin, and keep it under tension. When the tape moves forward the peel lifts up...until the pin is reached.
It does not HAVE to be spooled, you could just hang a weight on the end of it and that would work just as well....until the weight reaches the floor...

Instead of spooling the peel could be passed trough a set of rubber rollers with constant torque. But the peel is thin and pesky and difficult to control...no telling where it might end up (entangled). I'd prefer to spool it. Variable rotation as you call it is basically keeping the peel under more or less constant tension. I envision the spool driven at overspeed via a friction clutch (like two plates springloaded towards eachother with a piece of felt between them...you'll find this type of clutch in every tape deck and VCR).

But as I've said before...making ONE is doable, but making 30 would probably require as much effort as developing the rest of the machine and software...

[Eclipze]

For a hobby approach, I'm keen to avoid the multiple tape feeding mechanisms like yourself

Perhaps the head could have one pin for feeding the tape, and a servo to pull the strip via friction. A really neat way would be if the strip is pulled aside perpendicular to the tape and re-joined after the pickup location.

[nullbert]

Hello All,

I'm happy to come across this thread, despite the lack of activity. I'm working on a DIY pick and place as well.

My starting point is an old Celmacs table top pick and place machine which will receive an upgraded head. So I do have a big head start on the X/Y table.

My current plans include three heads: a solder dispensing head, and two different vacuum nozzle sizes for the major components I'll be working with.

One thing I'm not sure is how little paste I'll be able to dispsense- I'm hoping to be able to handle .020 pitch LQFP pads. we'll see if I can achieve that fine of control.

The Z axis will be handled by an airpax linear stepper motor with with a miniature linear slide. A separate linear stepper motor will select one of the 3 heads in a linear fashion- kind of like a Y axis within a Y axis. The Z axis linear stepper will engage wichever head is selected.

For theta rotation, I will use one stepper motor to engage gears on each of the heads, allowing one motor to rotate both heads simultaneously.

My original inspiration came from the youtube video showing the milling head adaptation using the centering pit. However, reading this thread provided me with several other other great ideas!

I've attached a couple of pictures of what I'll be starting with.

[blighty]

spooling can be done by pulling the tape back over the feeder onto a spool that is sat on top of 2 rollers. once she has picked a motor connected to 1 of the rollers is turned on, after 5-10 sec's after picking motor turns off. unloading a spool takes seconds.

components shouldn't jump out. the p&p i ran would do 40'000 lays an hour, all picking components are uncovered.

[aaron800]

Hello all,
This is my first post in a long time (if ever). Anyways, I have been wanting to build a pick and place machine for years after tedious hand assembly.

I see there are a lot of good ideas here and elsewhere. I have a few questions in my mind about the actual x-y-z-etc mechanics. I have been mulling over different options for the actual linear bearings and transmission, looking for mostly speed and low cost, as well as low noise. I would think that we wouldn't want reciprocating ball bearings as these are noisy in high speed. I have been sort of leaning towards either igus plastic bearings on a 20mm steel rail (from vxb) or some kind of other slide. For the transmission, I think belt drive would be optimal for this.

I would want to have a two or three independent z-axis for a solder dispenser, small nozzle for 603/805's/etc and a larger nozzle for tssops, soic, qfns, tqfp's, etc. I would also have a right angle mounted (so more in the y-direction) solenoid-operated piston for indexing whatever tape drive mechanism there would be, or at least some kind of pin sticking out and the y-axis or x-axis movement would engage some kind of mechanism.

I was looking at this link, and what if instead of just indexing one part at a time, you did five or more. I think this would save a little time.
http://www.hobbycnc.hu/CNC/Mech29/Mechanika29.htm