[daedalus]

Stewi: Have you considered having these made with rapid prototyping? Places like shapeways.com are really pretty cheap now, and it would allow people without machine shops to order one-offs. It would definitely make the sprocket wheel production easy, and I'm guessing could be used for the whole lot with a bit of thought. It also has the advantage that you can integrate features like the spring holding the incoming tape, and the motor mounts, reducing your part count.

[stewi]

Thank you for the input.
The last time I checked on rapid prototype (quickparts.com), it wasn't that inexpensive.
For a small production I would consider making a silicone mold and use aluminum epoxy cast.
I'm currently building another CNC machine, this time with the capability of tilting the router head. Together with the 4th axis, I'll set up a rotary chuck and can hopefully mill a stack of discs and do the sprocket wheels as well.
Else, I would also attempt to cast these wheels in a silicone mold, once I have one golden part.

I'll certainly continue building a few feeders to proof the concept, but I didn't mean to start a production of component feeders to diminishing potential customers in US.
I was hoping somebody would pick up from here, build a few parts and share these parts within the home built P&P community.

[daedalus]

shapeways is about 1.35$/cubic cm of material in part for their economy plastic (SLA process). That means that the wheel you are building would be a couple of dollars, and the surface finish is pretty good. It also means you can do the whole wheel (spokes + hub) as one piece. Their minimum order is a few dollars, so its hard to go wrong.

The reason I suggested doing as much as possible this way is that shapeways have a store feature where you can put your design on their site, then people can just order the whole set made for a fixed fee. Effectively its a kitting service less motors, shafts, and bearing bits.

I have done a fair bit of epoxy casting at work over the years, and these days I use shapeways to make my moulds in RP plastic, then just cast into them under vacuum, the part quality is pretty good that way, and no more waiting around for rubber molds to vulcanise.

[stewi]

Thanks again. I'll send some drawings for a quote.
In order to incorporate the motor/gear mount into the base, I'll remove the material in the rear and have just the 4 fins sticking out. See attached drawing
Making the coil spring out of plastic as well, which supposed to take the shock load from the tape reel, looks like quite fragile to me, although I like the idea of reducing the number of parts.
After all, I like the idea of bringing a third party vendor in and folks order directly there. I think this is the true spirit of open source.

[stewi]

I'll upload the other parts as well.
Here is the first one basewtabs by stewi on Shapeways

[stewi]

O.k. here it is: generic feeder by stewi on Shapeways
I think particular the smaller parts and the sprocket wheel are very attractive priced. There is no markup.

[stewi]

The 2 motors I selected for this feeder:
Pager motor activating the component cover:
Kysan Electronics Product Detail Page
and gear motor pulling the cover tape:
Kysan Electronics Product Detail Page

There is a $100 minimum purchase from Kysan and shipping costs are quite substantial with no value added.
However, even with pins and screws added to the BOM, I met my personal goal to make a feeder for under $50.

[scsi]

Good progress, stewi! Is that gearmotor 12mm wide? Would it still fit?

What are your plans for the driving electronics? Do you want the feeers to be controlled individually with single-ended TTL lines? Maybe an inteligent multiplexer of some kind?

[daedalus]

Hi Stewi,

That looks great, I will have to order myself a set of parts to play with

Do you mind sharing the original cad file for the assembly on here? The shapeways part viewer is a little 'basic'.

Did you ever consider taking the waste cover tape out below the paper carrier tape? (lowering the angle coming off the peelwheel, passing tape in slot in backplate). In my mind this would simplify the drive mechanism somewhat. Maybe this is for ease of loading?

The only other suggestion would be that 9-pin d-sub male connectors are pretty cheap, and have benefits over pin-headers if this is going to be removed / inserted frequently. (mechanical protection on pins, tapered shell, polarisation, etc)

[stewi]

Wow, I'm getting some attention.

Yes, I'll be sharing the drawings in dxf and step. Not sure, if I can post it here or upon request on PM.

The electronics I like to leave to the user. Again, all machines are different in their control logic and keep that part quite secretly.
With the pin header, I thought it would be a neat idea to plug in the feeder directly into a control board. You can plug in as many feeders as you want into one control board. Do you have a better suggestion? Gold leads?

The gear motor is supposed to be 10mmX12mm, I'll add my dxf file (if I can) I made from the spec sheet.

Disposing the cover tape underneath, I've thought about it. When I introduced my concept to Siemens, their concern was, that the cover tape is too flimsy to reach to the cutting knife. If I lay the cover tape on top of the empty carrier tape, I should not have this problem.
Also, I agree with your observation, that it will be easier to load the tape with the knurl wheels exposed.

Thanks for the feedback.

[stewi]

Do you mind sharing the original cad file for the assembly on here?

Apparently, I can post dxf files. Haven't checked the file size yet and may have to zip it up. I'm using ACAD/MDT 2004 and Alibre 2011.
Which CAD format shall I post? dxf, step, stl ...

[daedalus]

I would prefer step or stl, as I'm running solidworks and those two are easy for me to import.

[stewi]

Will do per personal e-mail.

[stewi]

Hi Stewi,
The only other suggestion would be that 9-pin d-sub male connectors are pretty cheap, and have benefits over pin-headers if this is going to be removed / inserted frequently. (mechanical protection on pins, tapered shell, polarisation, etc)

I get your point with the 9-pin d-sub. If I find one 10.5 mm wide, I should go for it.

[daedalus]

I have a link you might find interesting, it details how to make interlocked parts, and might be of interest for the cover part (as it could remove the two rivets). Shapeways | designing mechanical parts for 3D printing part 2: The Whoosh

on the connector front, I think I might have steered you into a dead-end with the de9 connectors, I forgot they have a 1mm rim, so are more like 12.5mm then 10.5mm. There is a miniature version of these sockets which would work, but at $7/plug its probably too expensive.

The next cheapest option for this might be something like spring loaded battery terminals, which are also quite robust, something like Digikey: 478-4677-1-ND @ $1.40/slot on the motherboard side, then just some pcb contact pads on the reel.

[stewi]

I know, 3D printing offers new opportunities in the design, but also can't be manufactured otherwise.
Instead of the rivets there should be flared posts and the cover snaps in.
Same with the rivet for the peel wheel.
Ultimately, the parts are die cast or injection molded for lower costs.
I don't have the expertise to make a decision, which features in the mold will drive up the cost to the point, that it is better to use rivets.

You have the drawing file (I hope you got it), make changes and let us know.

[chome]

Would it be possible to make fully autonomous feeder? I'm not an expert in feeders (EE by trade), but it seems to me that the only thing it does is advance a tape when the pick position gets emptied. This event can easily be detected with a photosensor.

I can design a circuit to do this - driving two small motors and reading a handful of photosensors can be done for about $10 in parts.

[chome]

This is the source of 12x10 gearmotors in single quantity -> Pololu - 30:1 Micro Metal Gearmotor HP

[stewi]

Thank you for the interest in the feeder project. The motors are likely purchased directly from China or through a direct importer like Kysan Electronics I selected A12FT-5V400RPM
Empty feeder detection or fully autonomous feeder should contain a RF ID.
That a feeder is empty is recognized by vacuum detection from the pick up nozzle, but we rather need an early warning system, which alerts the operator to splice in a new tape before the old tape runs on empty. An optical sensor in a fixed position of a feeder wouldn't help too much if the pick head picks at a rate of 100 ms just from this one feeder track. This tape would advance by 40 mm in a second. Within 10 second the tape would disappear inside the machine and can no longer be accessed by the operator to splice in a new tape.
The RF ID should store the information of the tape content (component name) as well as the number of components in a reel. The machine will read the component information for super fast set up time.
1.A LED in the feeder table will blink as soon the operator approaches with this particular feeder, indicating where this feeder needs to be placed on the machine.
2. The number of components in the tape are read by the feeder table controller
3. Upon each feeder activation the number of remaining components are stored back to the feeder RF ID.

Some customers require a time stamp for each component used on a board. This info could be stored on the ID tag as well.

Last, but not least, the entire idea of this feeder concept is to make the feeder small and inexpensive to store the remaining components in the reel together with the feeder dangling at the end of the tape. If you ever need to search for this particular component in the inventory room you only need a mobile reader to find it.

[cittadinilaser]

HELLO Stewi , others

Congratulations for your hard and good work
I´m from Spain and also interested in P&P machines , as others because the high costs of a commercial machine I´m looking to make one by myself ...

I like very much your feeders ( I see some of them ) and I want to ask if you can send me the files (Acad ) to figure how to mount them ,
As parts I´ll take the option of buyng across shapeways, electronics is not very difficult for me ,
Thanks in advance

[email protected]

Fernando Cittadini