Simple Feeder for DIY

[mike....@gmail.com]

I enclose herewith a sketch for a simple feeder for Home use. It should be simple to produce it as mainly all parts can be 3D printed except Sprocket with Push Gear and N20 motor.

I have no time to make a prototype but there might be guys out here that are interested in doing it.

Any questions - ask me and let me see your progress!

Mike

[Wolfgang Lienbacher]

I like that! But where to get the motor and sprocket with push gear?

[Mike Menci]

Google Samsung sprocket sm 8mm in AliExpress feeder spare parts .... use N20 motor and 3D print the clutch on end ? 

Mike 

On 13 Nov 2020, at 18.29, Wolfgang Lienbacher <wlien...@gmail.com> wrote:

I like that! But where to get the motor and sprocket with push gear?

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/5cae0cff-b509-449c-ab17-0f65ea83535dn%40googlegroups.com.

[mike....@gmail.com]

Sprocket : https://www.aliexpress.com/item/32618518515.html?spm=a2g0o.productlist.0.0.46084231gIAImQ&algo_pvid=87c4c413-a069-44d6-a22f-bd818ecb242b&algo_expid=87c4c413-a069-44d6-a22f-bd818ecb242b-2&btsid=0bb0623416052916474771336ed6b2&ws_ab_test=searchweb0_0,searchweb201602_,searchweb201603_

N20 DC motor:

https://www.aliexpress.com/wholesale?catId=0&initiative_id=SB_20201113102154&SearchText=N20+DC+motor+

Mike

[Niclas Hedhman]

Also check out Stephen Hawes on YouTube, as he is getting the sprockets from the PCB house and puts slits in that to measure the advancement.

https://www.youtube.com/watch?v=2foSlnGWeu0

https://www.youtube.com/watch?v=EnK9bekxUig

https://www.youtube.com/watch?v=pVGoxflJ6o4

Personally I am striving for a 16mm (or less) width, but will go with https://www.aliexpress.com/item/4001045335860.html + Stephen's approach if my current idea doesn't work.

Niclas

--

[mike....@gmail.com]

90 deg shaft out from gearbox will take to much space between feeders....

Developing a feeder (in 2018) takes a lot of time (printing 3D parts and getting all that done might costs more than getting used feeders...) - my solution was purchase of used Schultz feeders...

Mike

[Niclas Hedhman]

Developing a feeder; Yes, true... There are a few aspects of what way to go; a)  how many does one plan to buy/make, as 100 of them quickly adds up cost,  b) reliability in short and long term  c) the fun of creation itself (my main motivation) and possibly others.

Space; Well, the motor I pointed to use up 18mm (+/- tol) and it should be quite doable to get 20mm spacing between feeders if one fiddles a bit.

Niclas

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/337133b8-388c-4e67-9cc1-49d2e5d733b4n%40googlegroups.com.

[johann....@9mal6.de]

In another thread I showed my very first try on doing advances using a simple solenoid. Ever thought of that?

I mean, pneumatic feeders work very similar and 10mm of stroke should be fine for a 2 or 4mm advance.

From the drawings they have a width of 13mm.

However, if you measure them including the (ugly wound) electrical tape around the coil, they are more like 15mm.

But maybe my solenoids were just out-of-tolerance parts like a lot of the stuff you get on aliexpress... 

Maybe if I get proper ones, they have 13mm total width.

Meanwhile, I got some cheap one-way bearings as were suggested in that other thread.

I hope to give them a try this afternoon...

If they work on a 3d printed axle, they would make the design even more simple and, due to their smaller diameter, I could use much higher components.

Stephen Hawes is a little over enthusiastic in my opinion (at least he acts like that in his videos... personally, I find them very hard to watch).

Sure, its quite some accomplishment what he built so far, but I would not do several of his design decisions just because of personal experience.

However, in engineering, very often there are several working solutions...

Regards

[Marek T.]

Hi Mike,

So N20 in your "design" need to be controlled with some encoder?

Why not to use microservo SG90 instead and have always the same returned position? Do you think it might be too weak?

[johann....@9mal6.de]

Ok, I did a quick test and those one-way-bearings (HF1012 - around 30-50cents on aliexpress) seem to work fine on 3d printed axles.

I used a random seam to get a relatively smooth perimeter and very lightly sanded it with 80 grid to remove the few remaining blobs.

Sadly I screwed up some dimensions of the solenoid mount, so I was only able to do the movement by hand to prove that it works.

So many thanks for that brilliant idea!

I hope I'll find some time during the week to update my overall feeder design...

Regards

 

 

[Marek T.]

Hi Johann,

Sounds great :-). Can you remind what the solenoid do you use?

[mike....@gmail.com]

Solenoid might not be strong enough to turn the Tape Wheel full with SMD parts, and its very fast - which might couse parts to jump out of tape and shuter might be needed...

[Niclas Hedhman]

On Sun, Nov 15, 2020, 15:30 johann....@9mal6.de <johann....@9mal6.de> wrote:

Stephen Hawes is a little over enthusiastic in my opinion (at least he acts like that in his videos... personally, I find them very hard to watch).

Hahaha... So it is not only I who thinks that. I thought maybe it is just me who is mellow.

Niclas

[Johann Wilhelm]

You can find them using the search term "0530 solenoid".

Googling around, I found a current-force diagram here:https://www.tinytronics.nl/shop/en/robotics/accessories/solenoid-push-pull-12v-300ma-jf-0530b

So I'd say, including the mechanical advantage, they should be able to pull with around 1kg. Considering the low dutycycle in this application higher currents should be possible as well.

The speed can be regulated using PWM. So I see no bigger problem as with steppers there.

Regards

You received this message because you are subscribed to a topic in the Google Groups "OpenPnP" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/openpnp/R2svcqGzrs0/unsubscribe.
To unsubscribe from this group and all its topics, send an email to openpnp+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/ae636850-71d3-4c22-ac86-c8d1d19a83adn%40googlegroups.com.

[Marek Twarowski]

Thx. I'll Google it. But agree with what Mike said about shutter. Have you made the rest with tape with parts and no problem like that? Full speed=dynamic=parts jump out?

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CAKkwgeahRnLphDc_2vuHrR35zrRWhjykXGX_cH9XjFxgWYxPJA%40mail.gmail.com.

[Johann Wilhelm]

Well, these tests come next...after I modified my feeder design to accommodate the one-way-bearing.

The first problem was how to get cheap, reliable, and simple mechanics.

Next will be the mounting block (I think I start with a design similar to the ploopy feeders) and some reliability tests.

I doubt that flipping parts will be a show-stopper.

The force is proportional to the current and position.

So the worst-case scenario is to measure the current and have a control loop limit the acceleration, speed, and such.

As I intend to have USB-C for power and control, I think I'll have an STM32F072 on there (that's my go-to for USB as it works without crystal).

So a fast ADC and a fast PWM unit are available anyway.

If it's really needed, I'll have a proper gate-driver to push the PWM clock.

Sure, this adds some complexity in software, but having photo-sensors which can be affected by stray-light, strange gear-ratios which call for some calibration, ...,  also introduces some challenges.

Any idea which components would be the best test candidates?

I expect something light and small in plastic tape... however, what's the cheapest part?

I couldn't find small resistors in plastic tape.

74VHC1G are still around 3cents from TME.

BAS21 in SOT23 is 0.6cents on a 10k reel from RS-components... BC848 in SOT23 is 0.5cents @10k.

Regards

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CAJDP_hf0MpHGqUc24HY7yUg7NtMuTZc_yM1SHZfOS-a7Vy-%3Dag%40mail.gmail.com.

[johann....@9mal6.de]

Well, that went quicker than expected...

I found some shitty tape, that was rewound and hold in a coil with packing tape.

So it's not exactly perfectly sliding.

As you can see, there's really nothing special to the mechanics.

Solenoid, one-way bearing and some 3d printing fun.

I tweaked the PWM settings a bit and all-in-all it works.

However, there are some bends in the tape and some parts on the edge are slightly sticky.

At these parts it gets stuck from time to time.

Ramping up the current does help but the solenoid is getting hot.

I searched the usual sources and could not find any stronger solenoid which would fit (they are all at least 20 mm wide).

So I'd say the concept is definitely worth trying if you find some stronger magnet in a compact form-factor cheaply.

I would expect, that for plastic tape this should work much better since it should slide much easier.

So for me that means back to the beginning...

Regards

[Jonathan Oxer]

Hi Johann,

I really like the concept behind this design. If the solenoid strength is a problem, perhaps a tiny linear actuator could do the job. In some ways this approach is similar to the servo-pusher on the MGRL0816 feeder, but backwards. So either a mechanism like that with a servo, or an N20 motor with a threaded shaft like this:

https://www.aliexpress.com/item/1005001405970915.html

A nut on the shaft could be used so that it acts like a ballscrew, connected to a piece that catches on the tooth of the wheel and pulls it on each stroke. The motor would need to run forwards (push the nut) to reset the position, then run backwards (pull the nut) to advance the tape. It would have plenty of torque and also move the tape more smoothly than a solenoid, without a sudden jerk.

The problem is that the control system would be more complicated, and it would need to sense the position of the nut on the shaft.

Please keep going with your design. I want to see how it works out!

Jon

[Marek T.]

The problem with these linear actuators is their speed, they're very slow if you need have some relatively high torque.

I have the one and the way takes near 2sec. For me it's not acceptable. If pick fails and you want to pick the part again - 2sec await time sounds shity. But maybe no other option...

I've also taken a look on these solenoids. 0420 looks nice (dimensions) but is still weaker than 0530, if we need a stronger one - the size goes up :-(

[Johann Wilhelm]

Well, there's also the possibility to use the mechanical advantage...

Utilizing the 10mm stroke of the actuator, I could get twice the torque... maybe 3 times of the currently availble.

There are cheap linear actuators like this: https://www.ebay.de/itm/DC-5v-2-phase-4-wire-Mini-Stepper-Motor-linear-lead-screw-shaft-nut-slider-Stage/142971445578

The question is how many cycles they would last....

Those very flat steppers look also interesting:

https://h5.aliexpress.com/item/4001204915653.html?spm=a2g0n.seo-amp-detail.moreToLove.4001204915653

However, they would require 32 or 64mm wheel radius for 2mm/4mm advance per step

Hmm I'll need some beer to think about this,I guess.

Regards

--

You received this message because you are subscribed to a topic in the Google Groups "OpenPnP" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/openpnp/R2svcqGzrs0/unsubscribe.
To unsubscribe from this group and all its topics, send an email to openpnp+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/f4144e34-de28-423a-8b77-85d42ac4845dn%40googlegroups.com.

[Marek T.]

Hi Johann,

What are "after the beer" results? ?

Maybe you need two? :-)

[Johann Wilhelm]

Well, I had some beers after all :)

My next best idea was to move the actuator in a "base-plate".

So the size of the solenoid or motor does not really matter.

Obviously, there's a part missing in the image. There will be some sort of tape-guide and cover in this empty space...

I'm not fully settled on the final design...

I'm not sure if the hooks and mechanism will work, but that's just my first try on this concept.

Basically, the feeder consists of some 3d printed parts and a single on-way-bearing.

In a separate base-plate, some servo/stepper/solenoid will do the movement.

This concept would allow me to have not a single electrical component in the feeder and the with will be 15 mm, only!

Well, it could be even tighter but the 8 mm reels measure 14.4 mm, so there's no point to go below 15 mm, I think.

Another interesting point would be, that I have multiple options for the actuators.

If I find the stronger solenoids as too aggressive, I could have a different base-board, which allows for better control of the acceleration.

To get a higher torque, I could use the actuators of 2 adjacent positions.

At least that is my theory :)

Currently, I'm printing the first feeder (and I already spotted a little mistake that will need some cleanup in the print)...

BTW, print time is around 4h hours altogether on a Flashforge Creator 3... I'd say that'll be 8h on an Ender3

Regards

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/973268e1-2fda-4ae8-b67c-4a2d28bf8c35n%40googlegroups.com.

[mike....@gmail.com]

Johann

I would do inner gear wheel larger - your 4mm advancing is now movement of coil les than 2mm ...

Note that one-way bearing will not allow you to pull the tape out - beck... Old mechanical feeders have a spring loaded wedge on mechanical gear (90 deg off - from pull/push bar) which stops sprocket at wanted location...

Just my comments

[Niclas Hedhman]

Please keep us in the loop... Personally, I ruled out solenoids before trying, because I suspected that control is "not trivial" and risk for too much vibrations. But I would love to be wrong.

*I* decided to try with cheap servos that move a small push-pin (blue) back and forth (pink block), directly in the tape hole. Don't know yet if this will work (too busy with unrelated stuff), but I thought I should share.

You received this message because you are subscribed to the Google Groups "OpenPnP" group.

To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CAKkwgebFzSnyPPW%2BkVGGyOJEAgrkq8TWaTjX9FjsJCi1UqkcrQ%40mail.gmail.com.

[Johann Wilhelm]

Well, I'm eager to have 8 mm clearance from the inner teeth to the tape to support the maximum depth of 16 mm embossed tapes.

If I find a problem, that's a simple change. Only that push/pull thing and the wheel needs to be replaced.

To pull out the tape, I have some ideas. That's why I have that empty space.

The easiest thing to do would be a single screw or latch which allows you to remove the cover to frees the tape from the wheel.

Regards

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/e1148bdd-99be-46fa-a9a7-5ce857b5f88bn%40googlegroups.com.

[Johann Wilhelm]

I have some of those already on my table :)

The design idea of having the actuators in a baseboard actually came to me when I was trying to build kind of a test-setup to compare solenoids, hobby-servos, and these geared steppers...

Maybe even these N20s could be interesting candidates...

I think these single pins are fine with new tape, but with the makr.zone-feeder I saw problems with some old and ugly tapes I had lying around.

Having that said, I'm not fully convinced, that my ugliest tapes are really the best thing for testing...

Anyways, I saw holes getting larger if your tape already had some handling.

With a sprocket wheel, you'd have at least 3 pins engaged => less stress on the holes.

But maybe this does not matter in reality... and using mangled cut-tape in a feeder is maybe not the greatest idea in the world :)

You may be right about the control of a solenoid... Hobby servos are for sure the easiest thing to use and are my 2nd choice :)

At the moment, I'm just trying a lot of different things...

I even had some thoughts of having something like these old-school lineshafts... :)

Regards

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CADmm%2BKeZUhvZmoRr2NyTD4KBPwC40B-4BWaY2aNNNJBxKecFZw%40mail.gmail.com.

[Niclas Hedhman]

> You may be right about the control of a solenoid... Hobby servos are for sure the easiest thing to use and are my 2nd choice :)

> At the moment, I'm just trying a lot of different things...

That's what is so cool... Different people trying different things and sharing the outcome.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CAKkwgeYqHzh-ibyNjcRc4TROCB8s-%2BnQVDq1zH%3DB6OVDYLXNTQ%40mail.gmail.com.

[Marek Twarowski]

Hi Johann

Even if the reel has 15mm it's worth to have a feeder more narrow. The reels don't have to hang directly on the feeders but lower and each second mounted on other level than others. Do the feeder with width like the tape is - still makes sense.

If the actuator is "aggressive" you can think about the shutter being closed for the time of advancement - this is "style" as Yamaha does in their pneumatically actuated feeders. And Siemens as well. 

But, maybe using stronger returning string on the solenoid can let to get a bit more smooth motions?

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CADmm%2BKcR%3Di6jyqrAg8%2B3yq7z_rGfPzs0%3Dgq7p6x-Nk57poP25Q%40mail.gmail.com.

[Mike Menci]

Interesting motor for feeders :

[Mike Menci]

Stepper motor versions:

[Duncan Ellison]

Oooooh.... now that really is interesting.  Well done for spotting that, Mike.

I reckon these are probably designed for driving door locks by the look of them.

I would need the N20 version, shame that the pcb connector comes out at the wrong angle, but maybe that can be Maguiver'd.

Duncan

[johann....@9mal6.de]

Hi there!

Well, for a very compact, low-cost feeder these motors are IMHO still not really suited... But by all means, give it a try :)

I moved towards a "central" actuator for the feeder bank.

Like some builds you see where there's an additional axis with a pusher.

You see them e.g. here: https://www.briandorey.com/post/diy-pick-and-place-v2-project-complete

But I want to have the actuation done from below to have a tidy look and easy access to the feeders.

So that's my current state:

 

The only non-3d printed parts are the screws and a cheap one-way bearing.

All needed variations can be handled in the second half...at least in theory :)

The part with the moving bits should be fine for any tape style.

BTW: to get the tape out again, you just remove the springy-thing, get e.g. a utility knife under the tape (or over the wheel), and pull the tape :)

Sounds crude but works well in my tests.

For the cover-tape, I intend to have something like the cheap desktop machines... or even gravity and some weights for the first tests :)

I settled on a base with 10 feeder slots (which is designed to be chained).

Reason is I have "spare" Ender3 which I use for non-mission critical things in my business and around 150x150mm fits nicely :) 

So both, the feeder and the base can be printed on this small and cheap printer.

The section of the first version of the base-plate shown above just finished printing and the fit is not perfect, but also not too bad...

Actually it has a very nice and snappy sound :)

I think, the pushing thing will be a variation of my pnp-head design... so that should be a relatively easy design.

This pushing/pulling thing should fit in this "tunnel" in the center of the base where all these fingers from the feeders get inserted.

So that's to come next...

Regards,

    Johann

[Mike Menci]

Duncan - how about stepper option ?

basic dwg I made for DC motor:

[Duncan Ellison]

Hi Mike,

I'm ahead of you.  Here's a picture of my feeders:

You'll probably see where I'm headed here.

I already have the laser cut sprockets, but the weak point is the 3D printed spur gear.  It works well enough, but it could be improved with a factory made bevel gear like this and the N20 motor is just a straight drop in to my design.

I've already ordered a couple from China.

Duncan

[Mike Menci]

Good news! 

Any gap of 3D printed spur gear will result in inconsistent setting point of/for pick up - but you know that...

Carry on & good luck  :-))

Cheers

Mike

[Niclas Hedhman]

Mike,

I think that's why there is a sensor on the sprocket for position; 2 holes per tooth; and the play in the gear is less of a concern.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/91db66d2-8e99-467b-80d1-fe1a3462da78n%40googlegroups.com.

[Duncan Ellison]

Niclas,  yes that's correct.  

It works reasonably well and it's nicely self indexing, but Mike is right as well, slop in the gear train doesn't help.

I took apart one of my commercial feeders yesterday and found that it has a one way bearing on the sprocket ensuring that motion can only be in the forward direction.  

Duncan

[Mike Menci]

Cheap feeder - poor man feeder - https://youtu.be/8OarEi7-H-E

[Johann Wilhelm]

In-deed, an interesting idea!

I'll look into that...

Regards

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/ca1a65a0-96e0-4ba8-a228-2cdd473240a2n%40googlegroups.com.

[Shai]

This looks super cool! I posted on your YouTube video but I'll post it here too :) Have you tried testing these magnetic encoders side by side to see if they cause magnetic interference with each other? Also, what's your pulse sampling rate per 2mm?

I think you could also just use a sensor on the sprocket directly then if you're using a ratchet design? Otherwise your error on the N20 encoder could accumulate over 1,000 parts and prematurely trigger a move.

[Mike Menci]

More smooth - no vibration see my Option II with servo - see here: https://youtu.be/Bwa_LqPQQdY

Shai I have DIMA feeders on East side and Siplace on West side of my PNP  - this "poor man feeder(s)" are just un inspiration for someone to carry on from this basic ....

I hope to see your feeders in operation soon !

Mike

[Mike Menci]

Johann if you need anything do not hesitate to contact me here on forum or by PM.

Mike

[Edward Casati]

Maybe you can move the servo to the other end, so the front of the feeder is as short as possible?

[Mike Menci]

Sure !

[Mike Menci]

A bit more done - https://youtu.be/H0V0MdhomQk

Comments are welcome.

[Mike Menci]

and this one : https://youtu.be/gu02oFeB6Ik

[Shai]

Looks cool! If you want to test it out using our injection molded sprocket, let me know and I'll send you a few :)

[Mike Menci]

Spool motor drive test

https://youtu.be/FrhA3Gl97JY

[ma...@makr.zone]

Hi Mike

Make sure two or more sprocket holes are visible from the top, so you can use the ReferencePushPullFeeder in OpenPnP with smart vision location/feed position calibration, OCR/QR-code slot change detection and elaborate auto-setup:

  →  + 

NOTE: Contrary to what its name implies, it can be used without any "push-pull" movement, as it supports Boolean-actuating two actuators for feed and peel-off.

https://makr.zone/openpnp-new-referencepushpullfeeder/393/

https://github.com/openpnp/openpnp/wiki/ReferencePushPullFeeder

_Mark

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/93064cb6-068a-4afe-9b68-9f06d5baa076n%40googlegroups.com.

[Mike Menci]

Updated feeder all 3D printed parts inclusive sprocket - https://youtu.be/8tr2TBuLHiw

Control software can be used from: https://docs.mgrl.de/maschine:pickandplace:feeder:0816feeder:controller

let me know your thots!

Mike

[Shai]

Mike,

What's the reason you didn't pursue your N20 worm geared feeder anymore? Or is it still being worked on? Have you tried checking precision under camera?

[Mike Menci]

Working on this version first - see files - https://grabcad.com/library/smt-feeder-servo-drive-1

[Mike Menci]

One more option :-) : https://youtu.be/0t5jrSymPpU

Mike

[ma...@makr.zone]

Hi Mike

I liked this one much, much better:

https://youtu.be/FrhA3Gl97JY

with the 3D-printed wheel, here:

https://youtu.be/8tr2TBuLHiw

It was the first automatic feeder that I saw so far (ever), that seems simple to make, compact and cheap enough to be a real alternative to buying second-hand professional feeders.

Its ratchet mechanism makes it perform fully repeatable, without complicated photo-electronic contraptions or (worse) "hope-for-the-best" drive time calibration. Any remaining imprecision can be fully calibrated and OCR/QR-discovered by the camera too.

It has the potential to be further simplified by driving the spool friction wheel with the back-going ratchet motion, a bit like my Push-Pull feeder but it would need some mechanical link going up to the spool. This way you have one motor only, and a slim design. The end-switch is IMHO not needed with a servo, right?

Whole banks of this feeder could be driven by simply multiplexing the PWM signal for the servo. No feeder PCB needed. One wire per feeder, plus power rails. One could design a veeery long plug-in PCB

Great work, Mike!

(That I prefer this one over the new one is just my 2 cents 😎)

_Mark

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/0e6d0d9d-dcad-464b-accf-6a5338295985n%40googlegroups.com.