SMT 0816 Feeder w. onboard logic - Prototype

[Philip Kristoffersen]

Hey,

Took the SMT 0816 Feeder and did some re-design due to it having a great baseline to work off, but we would like some more logic to continuelly measure operation.



About 120 of them need to be produced for the machine. The idea is that as many parts as possible need to be easily replaceable, aka standard parts / 3D / lasercutting. The backside will be PCB with different components and connections on it and (somewhere) on the front it will click into a bus in which we can talk to all feeders.
Increasing to larger tapes is easy given all that is needed is to increase the size of spacers.

For error control the feeder has photodiodes to ensure movement of tape and that the reel is operational. Magnetsensor is used to ensure correct tightness of the peel. Seeing whats wrong with a feeder would only take a few seconds, but this allows easy gathering of data to over time to ensure continues operation. Due to using a bus with 5V supply there is no lose wires.

Materials are easily acquired given the tape reel can be laser cut, internal parts are 3D printed, structural parts are cheap / local - only "downside" is the PCB + SMD components, but it will be a bulk order for a 160ish. 

Design approach has been to make easy access and swapping of parts if any repair / re-design is needed. It took 8 hours to print a single SMT 0816 Feeder, which had to be reduced significantly. 

The company consists of only 4 consultants whos work consists of innovation and R&D, hence why they cant afford a bulky industrial machine, but also cant hold the machines hand all the time. Hence, why the feeder is wierdly somewhere inbetween the DIY / Industrial.

Any feedback or things I may have missed would be greatly appreciated. I wil update with videoes and stuff ones I get to testing it (3D printed prototype), hopefully next week :)

- Philip

[Philip Kristoffersen]

Darn it,

I forgot to add that the actual mounting mechanism and bus contact is being added later. It will attach to a 40x40 rail.

There will also be status LEDs, reset button, counting the amount parts picked (to warn that a roll is empty soon), etc.

- Philip

[bing luo]

You can trigger the switch of the N20 motor without using the method of compressing the spring with Tape Film. Print a blue push rod. When retracting to the maximum value, trigger the normally closed switch of Omron（RED）, and the N20 motor will not rotate. When pushing in, the push rod leaves the switch, and the switch closes. The N20 motor will rotate and start peeling off the tape film. Without the spring, reliability will be higher.

[Philip Kristoffersen]

Hi Tian,

I see where you are going, however, we have removed the switch in this design. Instead, behind the peel pressure spring, there is a SMD magnetsensor. Inside the peel pressure spring a magnet will be attached.

This way we have feedback systems on each moving part of the operation. We can see with photodiodes down by the tape feed (2) wether or not the arms succesfully moved the tape. We have a pressure spring telling (1) us wether or not the tape is tight enough and last another photodiode (3) telling us wether or not the tape reel is working by looking at the holes movements. This way we can monitor the feeder constantly and also adapt to / try to fix an error purely in software if one occurs while a job is running.

Also there will be sockets for the motor wires on the back plate PCB as well. The 5V, GND and TX will be received up by the tip of the feeder - we just haven't exactly designed the bus system yet :)

Cheers for the input!

- Philip

[Philip Kristoffersen]

I just realized I didn't write anywhere that we are adding a microprocessor on the board for local control of all parts on the board as well as being able to hook up with TX and communicate with the bus. This way we can program the feeder to what parts it has, weather or not it is operational, having errors, etc. - probably the most important detail and I missed it, my bad :(

- Philip

[Mike Menci]

Hi Philip

Please note that in MHO hight above tape picking up area should be minimal, most today feeders are flat very near top of tape... 

I have 4 nozzles  (2 x 2 in X direction) and when Head is approaching feeders first nozzle in loaded - problem with hight is when 2nd nozzle is about to load - bouth nozzles are horizontal and 1st nozzle would hit a feeder before 2nd starts loading from feeder. This is if feeders are East and West - parallel with X axis but can be a problem as well if feeders are South and North and 4 nozzles on 1 Head ...... See sketch enclosed

Mike 

[Niclas Hedhman]

I was fiddling with my own design, very similar (PCB walls, bus
connector, 4040 rail, ++) and a few thoughts I had then to carry over
here;

* Photodiode at (2); You know that there are clear/transparent plastic
T&R right? I scrapped photodiodes and aimed for a ratchet system, where
a steel pin pushed the tape forward and and "ratcheted" over the holes
on the way back. Tape should always end up in the same pos, no
measurements needed, and the wear is on the disposal tape not a 3D
printed part.

* Bus connector; I concluded that PCIexpress edge connectors was my best
choice. Relatively cheap, easy to fit, fairly sturdy and many pins.

* Bus Tx/Rx; Use a RS-485 line, and then encode each slot via
pull-up/pull-down on the machine bus motherboard. On-feeder MCU will
then always have a small unique address and machine can easily query
what is mounted.


Niclas

On 2023-03-24 20:25, Philip Kristoffersen wrote:
> Hi Tian,
>
> I see where you are going, however, we have removed the switch in this
> design. Instead, behind the peel pressure spring, there is a SMD
> magnetsensor. Inside the peel pressure spring a magnet will be
> attached.
>
> This way we have feedback systems on each moving part of the
> operation. We can see with photodiodes down by the tape feed (2)
> wether or not the arms succesfully moved the tape. We have a pressure
> spring telling (1) us wether or not the tape is tight enough and last
> another photodiode (3) telling us wether or not the tape reel is
> working by looking at the holes movements. This way we can monitor the
> feeder constantly and also adapt to / try to fix an error purely in
> software if one occurs while a job is running.
>

> Also there will be sockets for the motor wires on the back plate PCB
> as well. The 5V, GND and TX will be received up by the tip of the
> feeder - we just haven't exactly designed the bus system yet :)
>
> Cheers for the input!
>
> - Philip
>
> fredag den 24. marts 2023 kl. 17.27.20 UTC+1 skrev tianl...@gmail.com:
>
>> You can trigger the switch of the N20 motor without using the method
>> of compressing the spring with Tape Film. Print a blue push rod.
>> When retracting to the maximum value, trigger the normally closed
>> switch of Omron（RED）, and the N20 motor will not rotate. When
>> pushing in, the push rod leaves the switch, and the switch closes.
>> The N20 motor will rotate and start peeling off the tape film.
>> Without the spring, reliability will be higher.
>>

>> 在2023年3月24日星期五 UTC+8
>> 22:28:12<philip.k.kr...@gmail.com> 写道：
>>
>>> Hey,
>>>
>>> Took the SMT 0816 Feeder and did some re-design due to it having a
>>> great baseline to work off, but we would like some more logic to
>>> continuelly measure operation.
>>>

> --
> 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/87b21e65-8716-46c5-a7b3-df7dd24366d5n%40googlegroups.com
> [1].
>
>
> Links:
> ------
> [1]
> https://groups.google.com/d/msgid/openpnp/87b21e65-8716-46c5-a7b3-df7dd24366d5n%40googlegroups.com?utm_medium=email&utm_source=footer

[bing luo]

You have added so many sensors and even processors to the feeder. You go to the factory to see a second-hand JUKI placement machine that has been in use for 20 years. Hundreds of feeders are driven by cylinders, and there are no electronic components on them. This does not affect their efficient production, nor does it require someone to monitor them nearby. Are they really necessary？

[Philip Kristoffersen]

@Niclas - You make a solid point! I've been tasting purely with white paper tape, I completely forgot about the clear ones. Did you test a photodiode with those? Did it work - or I guess it would have to be Analog to see any difference (if any substantial one). I will test and redesign that if needed, thank you for sharing.
Also, for communication the owner is bent on it being a one-line communication. However, they have supplied multiple different systems for industrial systems where it they use one-line communication and it seems to work flawlessy, so I trust him on that one - I'm just the intern :) (I have only worked with RX/TX and protocols like that at Uni, so surprised he doesnt want both lines when they are literally on the chip we are putting in?)

@Tian - I would say yes, they are. Most of these solid feeders had a high degree of R&D which we dont have the time for. The base of the design is that it can be easily tweaked and changed since the R&D will have to happen over time (if needed) as the machine runs and collects data through the sensors. It only needs to pump out 5 - 100 boards pr. job and those jobs may be once a week, so it doesnt need to be *super* effecient. Costs are also kept super low, as well as maintence, due to all parts being local / local produced. I totally get what you are saying though - but just an example an PNP company not far from here just had their machine go offline for 1 week and it cost them over 25.000$ to have maintence stop by and spend a day fixing it. With this design you can literally just throw out the feeder, pick a new one off the shelve and not spend your paid hours trying to fix it. They will cost about the same (or less) as the paygrade of 1 hour of an Engineer here (Denmark). So a heavy focus on keeping the maintence and cost of the machine as cheap and easy as possible.

@Mike - Damn it ... you just gave me the idea to completely re-design it to a flat top.. I totally see what you mean, but we arent using 4 nozzles. I will keep the idea in mind because I actually think I could make that change in 1-2 days have a prototype running. Do you know of any other benifits to keep it as flat as possible that I might have missed? I have noticed that almost all industrial ones are completely flat.

- Philip

[Mike Menci]

Hi Philip,

Thanks for making notes on my comment - I hope you do the re-design and keep Top of tape free for longer distance that you have now... 

Your sensor for transparent tape - Could some other movable part on feeder be used to Close/open the sensor view! ??  

Next comment would be the a "shutter"  is needed on top of tape window - for keeping parts on tape and not jumping out (or turning in tape pocket) before they are picked up from tape with nozzle. 

And why is your Collecting spool for cover tape so large ? I look at some old mechanical feeders and they keep the Spool ID apx 35mm OD = 65mm and in addition to cover tape motion advancing they use frictional coupling to compensate for any over-pull of the cover tape... 

Mike

(I will PM you with details)

[Mike Menci]

On this video you can see shutter in operation - you can see as well how parts "jump" in tape pockets - but not the one under the "shutter" ! 

https://www.youtube.com/watch?v=ZrSAd9tvUTQ      2:23min  when speed is higher than earlier on video... 

Mike

[bing luo]

The reason why components jump out in this video is that the spring preventing size reversal is too strong, and the strong rebound force produces serious vibration and noise. But because there is a shutter working, it can at least access the first element. After I replaced the smaller spring, this problem has disappeared. Look at my new 20 minute test video： https://youtu.be/RA_qQIqzpII.

This shutter is necessary for any feeder, and it can reduce the probability of component jump out  from 0.1% to 0%. Which means that at least one component will not jump out at all.

So when you design an electric feeder, the cost becomes higher because the rotation of the worm motor is unidirectional, not reciprocating, so you have to use an electromagnetic relay to drive the shutter.

[Philip Kristoffersen]

@Mike - I mathed it out:

Worst case length is 20M roll
Tape peel is about 0.12mm thick
Center beam in the tape reel 6mm in diameter.

Roughly, once a entire roll has been peeled, the diameter of the tape reel center goes from 6mm -> 55mm.
It is calculated by increasing the radius of the center by the tape thickness each time the tape has traveled one time around the center perimeter.
Then just looping that until all tape has been peeled.

Code snippet:
   centerBeamPeri =  (centerBeamRa * 2) * M_PI;
   tapeLength = tapeLength - centerBeamPeri;
   centerBeamRa = centerBeamRa + tapeThickness;

Again, very rough estimation. The tape reel is therefor 60mm in diameter, roughly matching the 65mm OD that you said - guess it just looks big in the image?

In regards to the photodiode, my worry is basically that if you use a PWM one and you use clear tape, well then I guess the sensor is constantly going to be HIGH. I was going to use the sprocket holes to go from a HIGH -> LOW -> HIGH during movements, but maybe you can see a difference if its analog instead? Guess I will just have to test, maybe the clear tape is reflective enough, but I dont know currently. We specifically want to know weather or not the tape is moving as anticipated.

I have added a pressure layer to this feeder, just like the original 0816 had. However, this one is wider and will be spring loaded. I am hoping this will remove any bounce, but a shutter isnt a bad idea .. and I think it would be easy to add if the feeder was flatened.

- Philip

[Mike Menci]

Philipe, 

Noted above - just one thing to add - parts in tape move or rotate in worst case when the tape advances for 2 or 4mm fast and when tape STOPS!. 

So,  when tape is under the spring loaded part - parts just move as much they can in closed down chamber - but when cover tape is removed parts "jump" when movement of tape stops!  

Mike

[Mike Menci]

@ tianl 

Have a look at your setup - 

parts do not stop at the same spot from time to time: 

sketch with part for pick up: 

- see if this stopping works as it should - it should Close down exactly in the middle of sprocket gear to centre part perfectly.

I think you need to lower this stopper a bit more to get a perfect stop on tape

Sketch 2

Mike

[Niclas Hedhman]

On 2023-03-25 10:15, Philip Kristoffersen wrote:
> @Niclas - You make a solid point! I've been tasting purely with white
> paper tape, I completely forgot about the clear ones. Did you test a
> photodiode with those? Did it work - or I guess it would have to be
> Analog to see any difference (if any substantial one). I will test and
> redesign that if needed, thank you for sharing.

I had not gotten far enough in design, when I received a transparent
reel with LEDs and realized that it would probably not work, so I
abandoned that prior to testing it.
If you try it, try to find the longest wavelength IR possible, as the
plastic may change transparency in non-visual range and the further away
you are, the higher the contrast might be. But there is still the risk
that another material shows with different physical properties.
Other non-mechanical approach that requires research and fiddling;
measuring the change of permittivity (?) between two plates, "plastic"
or "air", with some high frequency current. Maybe not well-tried and
readily available tech.

> Also, for communication the owner is bent on it being a one-line
> communication. However, they have supplied multiple different systems
> for industrial systems where it they use one-line communication and it
> seems to work flawlessy, so I trust him on that one - I'm just the
> intern :) (I have only worked with RX/TX and protocols like that at
> Uni, so surprised he doesnt want both lines when they are literally on
> the chip we are putting in?)

"Industrial systems" typically refers to PLCs and similar equipment
(which I work with daily) and RS-485 is used heavily, which is a single
pair of bidirectional (half-duplex) communications, where typically one
(a computer) acts as the master to which all others respond according to
their address. I suspect that he has that in mind, and not one-way
comms, which would be outright idiotic; no error/status reporting and no
inventory tracking possible.

Niclas

[bing luo]

Do not compare the position of the shutter. The position reached by the shutter forward and backward is not fixed.   You should compare whether the final resting position of each component is the same

When push the beak  forward, even if the servo's stroke is a bit larger, the gear wheel advances a bit more. Rely on  the spring force the hold beak  to compress the gear to the correct position. (Whether it's a push beam or a hold beam, there are springs.).

When the servo pulls the push beak back to the starting point, the hold beak will prevent the gear from reversing under the pressure of the spring, and the gear will not move. This position is also the position at which the element stays each time, so this position is fixed. At this point, the push beak can stop anywhere between the two teeth (This range is 1mm or 2mm), and the shutter you see is not in the same position. But this does not affect the next time it pushes the gear forward to the same position.

[bing luo]

Do not compare the position of the shutter. The position reached by the shutter forward and backward is not fixed.   You should compare whether the final resting position of each component is the same

When push the beak  forward, even if the servo's stroke is a bit larger, the gear wheel advances a bit more. Rely on  the spring force the hold beak  to compress the gear to the correct position. (Whether it's a push beak or a hold beak, there are a spring.)

When the servo pulls the push beak back to the starting point, the hold beak will prevent the gear from reversing under the pressure of the spring, and the gear will not move. This position is also the position at which the element stays each time, so this position is fixed. At this point, the push beak can stop anywhere between the two teeth (This range is 1mm or 2mm), and the shutter you see is not in the same position. But this does not affect the next time it pushes the gear forward to the same position.

The stroke of the SERVO is fixed, and the stroke of the cylinder is also fixed. It is the spring that solves some accuracy problems.If the mechanical structure can be made tighter, you can see that the shutter is  also  in the same position every time.

在2023年3月25日星期六 UTC+8 22:26:44<mike....@gmail.com> 写道：

[bing luo]

When using the air cylinder, you may see that the shutter stops at the same position. This problem is most likely caused by the universal joint transmission rod of your poor man feeder. It is not a hard link, so every time it moves back and forth, it will have a slight tolerance. Fortunately, the spring can correct this tolerance.

在2023年3月25日星期六 UTC+8 22:26:44<mike....@gmail.com> 写道：

[Mike Menci]

A simple DIMA mechanical feeder in operation with OpenPnP (slow motion): https://www.youtube.com/watch?v=FOffg7-mMO0
and drawings - https://grabcad.com/library/dima-feeder-assembly-1

Mike

[Mike Menci]

And the DIMA 2mm advancing principle here: https://www.youtube.com/watch?v=VyLRNqCEI5U

Mike

[bing luo]

0816  feeder ’  s  cover tape structure  can work well, it won't get bigger and bigger, and the pulling torque is also fixed. There is no need to use a reel.    

The feeder in the factory also does the same. I don't understand why you want to change to the current structure.

[Philip Kristoffersen]

@Mike - Thoughts on this approach?

First note that the attachment for this on the front isnt added yet - this is simply just an idea.

You had a good point with my "tape compression" thing basically doing nothing since the part it was compressing was the tape with peel still on it. So I wanted to try to get around it and found some inspiration from other industrial manufacturers.
The idea is simply that this is an attachment that fits all feeders, but can be adjusted in its depth so that it can match the tape you are working with specifically. So if you have 0.4mm, 0.8mm 1mm, etc., you can simply just swap the attachment to one that is sized for you needs. This way, there is good contact with all types of tape in your feeder and will prevent parts from jumping before being picked.

So basically, it just slides in, a latch on the front clicks into place and latches it. Then you have a perfect fit for your tape and peel, making it so parts cant jump while moving to the pick-up location. This could then be extended if you wanted a longer feeder for your Quad Head :)

Thoughts?

- Philip

[Philip Kristoffersen]

Just an update:

- Philip

[Mike Menci]

Hight looks good - acceptable to me - how much is the hight over the top of the tape ? 

Just a comment 

the Tape bed could be spring loaded 3D printed part pushing tape up. This could be easy way of doing it ? 

Like above old Dima has it ; 

Have in mind that there are tape tolerances for production - Min. bending radios, - 8mm tape is actually les or more than 8mm - see attached for reference.... 

[Philip Kristoffersen]

Hi Mike,

Heigth above the tape would be adjustable. So it would require a bit of testing, but it should be able to get to really low tolerance point? So you could basically have 0.1mm between tape and the attachment - I dont know if lower can be done due to printers accuracy.

You gave me a good idea with the spring loaded stuff ... I shall return :)

- Philip

[Philip Kristoffersen]

Thoughts on this?

- Philip

[Philip Kristoffersen]

Forgot to add these,


- Philip

[Mike Menci]

Philip - to complicated  & should touch the tape on L +R side of the tape pocket. 

Very nice example is this: https://www.youtube.com/watch?v=IhV4RMNl3Z8 

What you think ? Can be done ? and spring end point adjustable to increase or lower the friction/tension on tape

Mike

[Philip Kristoffersen]

Hey Mike,

I saw you sent me an email, would you rather have the conversation through there than here?

Also, I was thinking that the spring loading would be on both sides (L + R), this way it scales with the tape size without having to change any measurements, printing, spring, etc. The one you linked would require to grow with the feeders, creating an unnesecary amount of printing, imo. However, I do like the idea, quite simple for what it does.

So this is what it looks like now. The whole "Feeder Lane" and spring action would simply just be mirrored to the other side with a few minor tweaks for tape allignment. This way it touches the edges of the tape and allows passage of compononts of all types and sizes. Do you see any potential issues with this?

- Philip

[Mike Menci]

Hi Philip, 

It seams like there are only me and you following this - might be better to do it over personal mails. 

It is your prototype - you do as you like. 

Mike

[Niclas Hedhman]

On 2023-04-13 13:56, Mike Menci wrote:
> It seams like there are only me and you following this - might be
> better to do it over personal mails.
> It is your prototype - you do as you like.

I bet that it is not only I who are watching, learning and see this
evolve in public.

Keep it here.
Niclas

[Philip Kristoffersen]

@Niclas - Thanks for the pointer, discussion will continue here :)

@Mike - So this is how it looks now:


I am getting very close to a printable state now in which I can start testing it out. However, your point is valid - now that I see both of them next to eachother it seems .. overcomplicated. If they had a shared structure and the pressure came from underneath you, you could simply adjust the tension of the spring with a screw + screwdriver. I think I would redesign it to be one solid component with adjustable pressure instead.

- Philip

[mark maker]

plus one.

[Mike Menci]

OK Philip- carry on let us see how will it do... 

enclosed - mini feeder drive with Attiny 10 for perusal...and comments. 

Pcb 40 x 10mm - 1.2mm thick. 

Cheers

Mike

[Philip Kristoffersen]

Printing now and testing will start on Monday,


The spring mechanism is temporarily removed - I want to test weather or not the new top component solves the issues I had with the baseline SMT 0816 Feeder first. If it then seems like it would add benefit / solve issues I will then go ahead and add the spring mechanism underneath the tape.

- Philip

[Mike Menci]

Great Philip - patiently waiting for results! 

[Philip Kristoffersen]

Hey again,



So I got it printed and started playing around a little. It works as expected so far :)
It is super easy to install new band now with component on top - just slide it in, put the tape through the slit, click into place and you are good to go.
This is the first top part and there is still room for the parts to jump, so some tweaking is in order.

Next up I will look into the best way to pull the tape back. Does anyone have any inputs for this? Prior design was to use 2 gears, but I am in doubt if this is the best way. I had into issues with the original SMT 0186 a lot where the tape would slip. 

- Philip

[Mike Menci]

Google Search: 

Best would be ; FEEDER FINGER type of peeling off sideway tape - no wheel and drive needed for cover tape storage....

https://www.youtube.com/watch?v=avoaRpr62RY

https://www.youtube.com/results?search_query=FEEDER+FINGER+

Working seams to be ; 

https://github.com/SebG3D/TapePuller

Mike

[Philip Kristoffersen]

Hey Mike,

I have looked at the FEEDER FINGER and my god would that be awesome to have - you would avoid all the mechanics of removing the tape. However, don't they have that on patent? Since I'm producing a product for a company, I'm not sure it would be legal to replicate it?
I have seen others attempt to 3D a version of it, but haven't found anyone doing it successfully.

I will look into the tape puller you linked tomorrow :)

- Philip

[Philip Kristoffersen]

Update on WIP,



Printed a new top peice, seems to follow flush with the 0.6 band and therefor parts can not jump - and if they do, they cant continue underneath and gets caught in the extraction hole. Changed it to magnetize to a metal rod at the end, making it super easy to instal.

Gears seem to grip the tape really nicely and the spring latch for the tape works great, just needs a minor change. Currently waiting on the worm gear to print to continue tests, but overall it seems to work as intended so far. Once the tape gears are confirmed to be working I will add a "container" at the back in which all the tape is collected. A latch will be added so you can open it up and empty it when desired.

A full print is somewhere in the 3.5 hours range vs. 8.5 for the original SMT 0186 Feeder. Taking advantage of a Resin printer as well, printing time pr. feeder can be reduced to about 1 hour in total (When scaling it and making many parts at the same time), so things are on track :)

Currently, I am just finding some minor flaws here and there that I am tweaking and re-printing as I go. Next up is to add components such as photodiode, LEDs, Button controls, click switch, etc. for testing and functionality purposes. As discussed earlier, photodiodes most likely wont work, but I will do testing with different tape types and see if analog can work across the board - for now, the photodiode is simply for testing data.

- Philip

[Mike Menci]

Good progress - I like what I see .. :-) hope you get it to work good. 

Just a note : there are servos with feedback like this one 4 wire! 
https://www.pololu.com/search/compare/267/0,s,8,1,40,21,22,12,11,4/5,6,7,3,2/x
https://www.pololu.com/product/3436 cheepest ....

Mike

[Philip Kristoffersen]

Hey Mike,

That isnt a bad idea - but unfortunately that will only inform us about the functionality of the motor, not if the tape moved or not. The motor can fully retract the arm and advance without catching the tape and we wouldnt know. A photodiode is probably the way to go since it will catch all types of errors from the motor. I for sure will keep them in mind though, thank you for sharing :)

- Philip

[Philip Kristoffersen]

Quick update,



Lots of refinement being done to ensure everything works as it should. The N20 motor has been tested to run at a constant current and it works really well. It ensures that the motor pulls just enough at all times so that the tape gets pulled every single time and does not continue to travel with (and on) the paper. This means that (knock on wood) I have eliminated the common problems with the SMT 0816 Feeder - which I also experienced myself when testing it.

Adapting the SMT 0186 Feeder to a more scaleable solutions has been more tedious than I anticipated though .. this is also my first time working with 3D modeling and 3D printing.

More stuff to come, but hopefully I can very soon just move on to long-term testing rather than tweaking millimeters every day :)

- Philip

[Philip Kristoffersen]

Update:



Getting real close now to something usefull :)

- Philip

[Philip Kristoffersen]

Finally - a video!!

https://www.youtube.com/watch?v=WqCj0JBzLas

So the thing actually works now. There has been a lot of little kinks that I had to fix in order to make sure it didnt keep falling apart / making mistakes. Still some work to be done, but now it just runs without any issues.

The core issues of the original SMT 0186 Feeder has been fixed:

1) Tape no longer accidently fails to unhook and follows the band.
By not controlling the DC motor based on sensor input this does not occur anymore.  Instead its a constant pull for a set time.

2) Parts jump due to the band having too much wiggle room at extraction.

With the now top piece (which can be printed in multiple depth sizes) it is always insured that parts can not jump after tape removal.

3) Gears and tape jumps when hitting the "perfect spot".

I was able to recreate a issue several times in which the tape would jump the gears and then release the switch and re-activate the motor, causing a constant loop of the tape being pulled. This does not occur anymore due to the removal of spring.

The new feeder is built for expansion and sizes in mind. It can be used for all sizes of bands and compononts. Max size depth of band is 2 cm, size of band thickness doesnt not matter since you just adjust the top peice, band width should go as high as the motor can drive - simply just expand the feeder.

Next up I will make several in many different band sizes and start prolonged testing (picking 10.000 components) and test for reliability issues, etc.

Fun times :)

- Philip

[Mike Menci]

Great job Philipe! Pls Keep us posted!

[Philip Kristoffersen]

Another one!

First fully functional and done prototype! Next step is the mounting to the bus :)

- Philip

[Philip Kristoffersen]

Mounting is pretty much done now.

Hopefully ordering PCBs this week.

- Philip

[Philip Kristoffersen]

Another update - PCBs being ordered.




Getting there :)

- Philip

[serge a]

great ;-)

--
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/d581cb8f-f96a-4fe8-9261-0a083866f6fdn%40googlegroups.com.