LitePlacer.com forums

I looked at your design and one thing I noticed:


This would work fine with the thick paper reels, but with simple plastic reals would the switch resistance not be to much (tape will flex) ?
I have not tested this of-course but it is just something that took my attention.

Edit:
Could you save a.STL file of the design on Github?
I can't seem to export it as one. I want to look at it in further detail by importing it in SolidWorks.

Edit:
Video of this concept working:
https://www.youtube.com/watch?v=CPw1u8_RcOc

bobc wrote: Btw, optical sensing of holes doesn't really work for transparent tapes.

Actually, it does, the tape is not transparent enough. Here is what the camera sees:


Zoom in, invert and set threshold:


And it is not even very sensitive to tape bends or reflections.

Another diy feeder at work:
https://www.youtube.com/watch?v=uNBV_dabrDA
This one looks a lot like the professional ones.

If the camera would be used for aligning the parts then an automatic feeder is quite easy to make. Even with a 3D printer.

Edit:

A easy way to solve the precision part of the mechanics is to take tape of multiple parts at once (i.e. Expose 5 resistors).
Align automatically let the camera align with the hole (it now knows precisely where the 5 parts are) -> place 5 parts -> send signal that it needs 5 more parts -> automatically align the hole again.

By doing it this way just a simple stepper motor / hobby servo can be used for advancing the tape and no real mechanical precision is needed on the feeder part.

Edit:
Open source pick and place feeder with design files (3d printable)
reprap.org/wiki/Pick-n-Place_Feeders

the professional looking feeders:
http://blog.protovoltaics.com/diy-pick- ... ce-part-2/

I have seen most of those before, except the protovoltaics one. The thing that strikes me is that they are all quite complicated. That means expensive. In that case, why not just buy second hand feeders off ebay? You can get single feeders for £20, less if you buy a job lot. They are a bit worn, but are still ok for light use. I had a look inside one, and you can see why they retail new for £300! A lot of fine precision parts. But they are designed to work 24/7 so that is expected.

To make DIY ones, they need to be a LOT simpler and cheaper. That might mean they break, but if they cost £5 per lane, then you just make a new one. Reprap shows that you can make a decent FDM printer without spending $20,000. The overriding principle is KISS. Minimise moving parts. Minimise precision parts.

So I think reinventing an industrial quality feeder is a bit futile.

This is why I like Volker Besmens design. There is only one moving part, the motor. The servos have a builtin gearbox and controller, really easy to control using an Arduino. The precision parts are the tape channel, and tape channel cover. I am hoping that they don't need milling, and I can use 3D printer to make them.

I think I can simplify Volker's design by using standard Arduino hardware rather than custom built electronics.

Individual feeders will also be more expensive, by creating a module of 5,10,20 then they can share the same structural frame.

If this works, it will be really cheap and hopefully a reliable design. it might be a bit too simple, in which case it may need some refining.

I really like the idea of using vision system to advance the feeder, that would eliminate the need for the switch.

The continuous servos are pretty much stop/go, so it needs rapid feedback to stop in the right position. I have found that trying to control the advance just by time duration doesn't give reliable results, possibly because the update rate of the servo is something like 50Hz. So if you set it to "forward" for 100ms then that is only 5 pulses. You might get 4, or 6 or 7.

I'm evaluating some other DC motors, also some small stepper motors. The stepper motors require extra hardware driver though.

JuKu wrote:

bobc wrote: Btw, optical sensing of holes doesn't really work for transparent tapes.

Actually, it does, the tape is not transparent enough. Here is what the camera sees:

Ah, I was thinking of simple photodiode/transistor type detector there.

But I am impressed with the capabilities of the vision system though, that is really good. If the camera is already there, why not use it

Spikee wrote:I looked at your design and one thing I noticed:

I can't give enough credit to Volker Besmens, which my design is derived from!

This would work fine with the thick paper reels, but with simple plastic reals would the switch resistance not be to much (tape will flex) ?
I have not tested this of-course but it is just something that took my attention.

The green part is a pressure plate designed to keep the tape against the switch, which hopefully solves that problem.

Could you save a.STL file of the design on Github?
I can't seem to export it as one. I want to look at it in further detail by importing it in SolidWorks.

Which design? I have put SCAD models into github, they can be exported to STL. Volker has 3D PDF models of his design which I found very useful.

Unfortunately, my CAD skills are very minimal.

I can't seem to convert the scad model to a stl as it is not water tight or whatever they call it.

Do you have a link to the 20$ feeders ?
Tbh for that money it is very hard to beat a ready made product ....

Your design but than with stepper motors for rough positioning. Then use the camera to align the next 5 "exposed" parts.
If the stepper motor systems is accurate than the camera vision might not even be needed.

Spikee wrote:Another diy feeder at work:
https://www.youtube.com/watch?v=uNBV_dabrDA
This one looks a lot like the professional ones.

I like their feeders as well!

haha

hello bobc,

Can you post a picture of your idea? I know the feeder of Volker, but I am also curious what you've created.
What about the control of the feeders? how can the software be controlled whit the liteplacer? are you planning to write a plug in for the original program?
how does that work (as you might notice I'm no software expert). Also the feeder's location need to be found by the liteplacer software....
I was thinking of creating something which only uses parts produced with my laser cutter in stead of 3D printer, it's fast and accurate, and plastic parts will do perfect for this application. adjustment of different tape thickness could be done by a simple plastic 'spring' mounted under the tape which can push the tape upwards to somewhat prevent it being pushed down by the pick up, something like this: http://www.skiffy.com/docs/images/index/HP_319.gif?v=87