Instead of thse switches you could maybe use foil on the actuator (or make the actuator from PCB with solid copper) and capacitive sensing electrodes on the PCB.
You don't really need tape detection- the vision system will tell you when you run out of parts. That film take-up spool looks a bit small - I doubt it would hold a full reel's worth of tape before it fills up too much. Also, have you checked the operating life of those switches - some small switches like this are only rated for 10-50K operations. There can be quite a difference in peel strength between different tapes, so worth testing on as many different types as you can get. Also, paper tapes can behave significantly differently, and also create a bit of paper dust that can clog things up and the paper itself can abrade surfaces.
I think you could get away with just a tape pull feeding it through. Keep the encoder wheel for feedback but remove the drive motor and let the tape pully do the work of driving the tape. No need for switches or a tensioning lever thing. (However, as a learning/educational/side-project this is pretty impressive.)
Also when it comes to various tapes there is paper and plastic. Paper tapes used for resistors is thicker and stiffer than the 12mm plastic you've been demonstrating. One addition I'm doing to my design is a 0.91 inch OLED display. I got my last batch from Amazon for a whopping $2.85 each. More useful than blinking lights. Adafruit has libraries that work.
Optical sensors would last way longer than relying on mechanical switches. An LED light on the rear that changes colour when you are approaching running out of stock (based on pnp program using initial qty vs what it is now). This helps operators change to a new reel on another feeder to minimise down time. Always test your feeders with the smallest parts you intend to be able to pick and place. We do expect to handle LEDs, what we really want it to be solid using small parts. The kind we don't want to place by hand. They'd be way cheaper to test with too! Agree with @mikeselectricstuff you don't need tape detection but need to know if tape is jammed/not feeding.
The film peeler only needs to be active for a brief period after tape advancement. Once the advance commend is issued check film tension. If low then run peeler, once tension is restored or a time out period has elapsed shut down peeler. This will keep the motors etc. from burning up if there is slippage.
Personally I would be testing tapes of different materials, especially the difference in clear plastic vs paper. Looking forward to seeing your pnp in complete action.
It's probably a good idea to add an extra spring for the tension arm rather than depend on the one built into the switch as it will invariably fail over time. EDIT: good to know you added one in the end (commented this during the video before the end where it all just worked)
First of all - congratulations getting this far. PnP feeders have an endless list of unexpected challenges and critical to the overall success of any PnP system. Curious if you will be able to control tape weave and other mechanical challenges needed to place small parts. Big LED's are one things, but all circuits need a lot of passives that are notoriously hard to deal with as they get smaller. My commercial feeders have a MUCH larger tape pickup spool - Like @mikeselecricstuff mentioned you will be tending to that tiny bobbin rather often. If you decide the out of tape sensor is actually needed - an optical gate will last a lot longer and be more reliable.
While tempting to use the limit switch's spring to actuate your arm, it isn't smart as the only thing the spring inside the switch is designed for is moving the very tiny plunger. I would have been very surprised if the limit switch could swing that huge tension arm. It works on the tape detection as gravity does the work in one direction and the tape in the other direction. I would suggest a small stud on the arm and "base" to hold the spring in place.
Take a look at the design for manual feeder magazine by Robin Reiter as he elegantly solved pealing and piping the foil. His design made me think - instead of driving stepper wheel and by that applying the motor force onto the tape before peeling the foil, it could theoretically work with only one motor to apply force when winding up peeled off foil while leaving stepper wheel in the same place but changing it into free rotating.
I have a question.As for this kind of tape guide when feed plastic tape and paper tape the thickness of them are different.If there is not something like a spring to press the tape.Will the position of the cpmponent become not reliable enough to pick?
I think the angle of the plastic when it's unsticked from the tape is a but to big, you will have problem to unstick, maybe to have an angle something like 30° or less is better, no ?
Stephen, I might be willing to put together a website for this project when you’re ready to go online with it. You’ve really done amazing job with it and it deserves a well done website. Is there one for it already?
I really enjoy the videos on the pick and place. Awesome. 😝. Great work. I was thinking of building my own pick and place machine and I just wanted to know if you believe the current board is sufficiently working. Or do you think a redesign might be required in the near future and it is worth waiting for that.
There is a better way to peel the tape and that is to copy how the FUJI AIMEX does it. Essentially what it is is a gear attached to the little motor and another free spinning gear on a tensioner that mesh together with the tape peel tape sandwiched in between and that eliminates the need for a spool. and as for the peel detection mechanism your lever thing should suffice. I would replace the sharp angles with rollers though because when I worked at an electronics factory we had a lot of issues with the peel tape breaking if the feeder wasn't set up properly.
Hey Stephen, I'm curious where you learned all of this pcb design and 3d modeling. Do you have an engineering degree or just self taught? Reason I ask is I currently do pcb design for a startup but I was fully self taught, and am always looking for resources to learn new skills. I have almost zero 3d modeling experience and would like to develop that too
Robust against electrical noise over modest distances (better than I2C or SPI), and multidrop (unlike RS232), and very simple protocol (unlike ethernet or USB).
@@SimplyStuart94 You can have many devices connected on the same wires / bus. So no matter how many devices you got (~256 usually), you only need 4 wires (if we count ground and power).
@@SimplyStuart94 For RS232, you'd need to daisy chain the devices (have 2 UART per device, one in, one out, and connect each device's in to the previous out. That's doable, but you need 2 uart, and the cabling & software is a bit more complex. I2C wouldn't work very well, because you quite limited on the number of devices per bus. Not because of the address space (255), but because of the capacitance of the bus. Usually, that's around 10-20 devices, IIRC, that's quite limited for a PNP. SPI needs an additional Chip Select pin for each device. That's really annoying. There are quite a few other protocols that would work (eg: CAN could be good too). RS485 is just very simple, robust, and the transceivers are quite cheap too.
That doesn't work, because when the film builds up on the spool, the diameter will increase and therefor change the gear ratio and pull the film until it breaks.
@@sm3vlc no need to use a spool. Just let it hang and cut the film every now and then. Alternatively, if you insist on a spool, having a flexible split gear driving the spool from the inside will allow it to slip when the tension increases, thus preventing it from breaking.
True. The machines I have worked with either use a spool (Fuji CP-series, completely mechanical model) or just leads the tape to a bin below (Fuji NXT-series, with built in motors). I must say that I do love and admire the engineering of the old CP-series. The drawbacks are power consumtion and high noise when running at high speed.
Hey Stephen are you going to try to place 0402 parts. If so I will consider you a gift from the heavens above, because you are going to save me a lot of money that I can instead spend on a lathe. I'm working on designing a real injection molding machine that will cost $2k-$3k. Its been a heck of a ride. I have been working on acquiring sponsors. I'm very impressed with your PNP design, its going to change a lot of lives.
Please disregard my comment if it doesn't make sense, but wouldn't you be better off with a brushless motor? Aren't those better controllable? Love your project by the way! Keep those projects coming.
The most appropriate day for the "hello my goblins and ghouls" intro
Your enthusiasm is so contagious!
PROJECTS LIKE THIS
projects like this
prooojects like this
Imagining a growing pile of LEDs accumulating on the table as you tested the film spooler! ;D
Instead of thse switches you could maybe use foil on the actuator (or make the actuator from PCB with solid copper) and capacitive sensing electrodes on the PCB.
or go optical
You don't really need tape detection- the vision system will tell you when you run out of parts. That film take-up spool looks a bit small - I doubt it would hold a full reel's worth of tape before it fills up too much. Also, have you checked the operating life of those switches - some small switches like this are only rated for 10-50K operations.
There can be quite a difference in peel strength between different tapes, so worth testing on as many different types as you can get. Also, paper tapes can behave significantly differently, and also create a bit of paper dust that can clog things up and the paper itself can abrade surfaces.
I think you could get away with just a tape pull feeding it through. Keep the encoder wheel for feedback but remove the drive motor and let the tape pully do the work of driving the tape. No need for switches or a tensioning lever thing. (However, as a learning/educational/side-project this is pretty impressive.)
Monitoring power and back EMF of motor would give much better results than a switch and a spring.. Mechanically it is free..
02:05 "button is depressed", I know you feel little button
They make Micro Stepper motors. Could be another option for the motor issue. That way you have precise position control.
Also when it comes to various tapes there is paper and plastic. Paper tapes used for resistors is thicker and stiffer than the 12mm plastic you've been demonstrating. One addition I'm doing to my design is a 0.91 inch OLED display. I got my last batch from Amazon for a whopping $2.85 each. More useful than blinking lights. Adafruit has libraries that work.
7:42 Describing a PWM as impact driver is genius. Gotta keep that in mind 👍
Today's equivalent of Eureka! "It f*cking works!"
I have to say this is an impressively methodical and thoughtful R&D.
Give me some of your excitement energy 😂
"Bobs your uncle"
Sounds like someone's been watching EEVBlog
It's good seeing someone taking contingency measures in early prototypes. Case in point accommodating a placeholder for the spring
Optical sensors would last way longer than relying on mechanical switches. An LED light on the rear that changes colour when you are approaching running out of stock (based on pnp program using initial qty vs what it is now). This helps operators change to a new reel on another feeder to minimise down time. Always test your feeders with the smallest parts you intend to be able to pick and place. We do expect to handle LEDs, what we really want it to be solid using small parts. The kind we don't want to place by hand. They'd be way cheaper to test with too!
Agree with @mikeselectricstuff you don't need tape detection but need to know if tape is jammed/not feeding.
That is really slick Stephen. Well done.
The film peeler only needs to be active for a brief period after tape advancement. Once the advance commend is issued check film tension. If low then run peeler, once tension is restored or a time out period has elapsed shut down peeler. This will keep the motors etc. from burning up if there is slippage.
Isn't that what Stephen's design does already?
@@Graham_Wideman If there is any slippage in the tape the tension arm may turn on the motor, and keep it on.
Personally I would be testing tapes of different materials, especially the difference in clear plastic vs paper. Looking forward to seeing your pnp in complete action.
It's probably a good idea to add an extra spring for the tension arm rather than depend on the one built into the switch as it will invariably fail over time.
EDIT: good to know you added one in the end (commented this during the video before the end where it all just worked)
First of all - congratulations getting this far. PnP feeders have an endless list of unexpected challenges and critical to the overall success of any PnP system. Curious if you will be able to control tape weave and other mechanical challenges needed to place small parts. Big LED's are one things, but all circuits need a lot of passives that are notoriously hard to deal with as they get smaller.
My commercial feeders have a MUCH larger tape pickup spool - Like @mikeselecricstuff mentioned you will be tending to that tiny bobbin rather often. If you decide the out of tape sensor is actually needed - an optical gate will last a lot longer and be more reliable.
While tempting to use the limit switch's spring to actuate your arm, it isn't smart as the only thing the spring inside the switch is designed for is moving the very tiny plunger. I would have been very surprised if the limit switch could swing that huge tension arm. It works on the tape detection as gravity does the work in one direction and the tape in the other direction. I would suggest a small stud on the arm and "base" to hold the spring in place.
'pulse width modulation' for the motor speed
Take a look at the design for manual feeder magazine by Robin Reiter as he elegantly solved pealing and piping the foil. His design made me think - instead of driving stepper wheel and by that applying the motor force onto the tape before peeling the foil, it could theoretically work with only one motor to apply force when winding up peeled off foil while leaving stepper wheel in the same place but changing it into free rotating.
Low frequency PWM (
"it's just a weird shaped microcontroller right now" -- God that is exactly how I feel about most of my projects.
@Stephen, Add a compliant, built in, "Spring" arm to your film tension arm and get rid of the metal spring.
Awesome work! Love your persistence!
The cad modeling, programming and circuit design are cool and all but that spool holder...
Lovely
Still with the RS FOURTY-five, eh? Just kidding. This is a great project and you have every reason to be proud of your achievements!
Thanks for sharing. Brilliant work.
Super cool!
I love your channel. This is a really cool serie.
Have a nice day.
This is FANTASTIC! You're such an inspiration!. Is the feeder going onto PCBWay anytime soon please? I am sitting in a puddle of drool here! LOL
Dude this is awesome!
you should probably not use the switch as mechanical end stop
I have a question.As for this kind of tape guide when feed plastic tape and paper tape the thickness of them are different.If there is not something like a spring to press the tape.Will the position of the cpmponent become not reliable enough to pick?
I think the angle of the plastic when it's unsticked from the tape is a but to big, you will have problem to unstick, maybe to have an angle something like 30° or less is better, no ?
Fantastic! I love this project :D With all the components that the feeder design now has, what is the rough BOM looking like? (per feeder)
Stephen, I might be willing to put together a website for this project when you’re ready to go online with it. You’ve really done amazing job with it and it deserves a well done website. Is there one for it already?
Would it be possible to integrate the moving parts into the main body using a flexible section as a sprung pivot?
So, how much does it cost in BOM parts for a single feeder? That PCB (size and shape) would not have been cheap ... or was it?
I really enjoy the videos on the pick and place. Awesome. 😝. Great work.
I was thinking of building my own pick and place machine and I just wanted to know if you believe the current board is sufficiently working. Or do you think a redesign might be required in the near future and it is worth waiting for that.
There is a better way to peel the tape and that is to copy how the FUJI AIMEX does it. Essentially what it is is a gear attached to the little motor and another free spinning gear on a tensioner that mesh together with the tape peel tape sandwiched in between and that eliminates the need for a spool. and as for the peel detection mechanism your lever thing should suffice. I would replace the sharp angles with rollers though because when I worked at an electronics factory we had a lot of issues with the peel tape breaking if the feeder wasn't set up properly.
Nice work !
Every time I hear someone say "do the thing" I think of Varrick from Legends of Korra. " do the thing Zhu Ki!"
like your video that full of passion and your hard works ! very good and pretty attractive to me to see more videos like this ! thank you ^^
Happy Halloween, Goblins and Ghouls.
🔥🔥🔥🔥👌👌👌👌 glad you got this edited in time 😂
Thanks for your videos! A little question : Where do you have buy your PCB Holder ? Thanks !
More reliable feeder than charm high 😂
Hahaha I was looking for comments to this effect.
Very nice
>Projects like this
Oh I felt this pain a little too much.
These videos are awesome!!
smashing it boiiiiii!
Hey Stephen,
I'm curious where you learned all of this pcb design and 3d modeling. Do you have an engineering degree or just self taught? Reason I ask is I currently do pcb design for a startup but I was fully self taught, and am always looking for resources to learn new skills. I have almost zero 3d modeling experience and would like to develop that too
Feed me with Feeder
Why use rs45 over other comm protocols?
Robust against electrical noise over modest distances (better than I2C or SPI), and multidrop (unlike RS232), and very simple protocol (unlike ethernet or USB).
@@Graham_Wideman I'm not familiar with multidrop, what is it?
@@SimplyStuart94 You can have many devices connected on the same wires / bus. So no matter how many devices you got (~256 usually), you only need 4 wires (if we count ground and power).
@@cynnfu yeah, but multiple protocols enable multiple devices, I was wondering why use rs45 over rs32, or i2c, or others
@@SimplyStuart94 For RS232, you'd need to daisy chain the devices (have 2 UART per device, one in, one out, and connect each device's in to the previous out. That's doable, but you need 2 uart, and the cabling & software is a bit more complex. I2C wouldn't work very well, because you quite limited on the number of devices per bus. Not because of the address space (255), but because of the capacitance of the bus. Usually, that's around 10-20 devices, IIRC, that's quite limited for a PNP.
SPI needs an additional Chip Select pin for each device. That's really annoying. There are quite a few other protocols that would work (eg: CAN could be good too). RS485 is just very simple, robust, and the transceivers are quite cheap too.
Your work I just awesome!!! And you are too 😃
WHOOP! NEW VIDEO :)
Has tu emoción contagia , me identifico
projects like this
PWM - Pulse Width Modulation
Look at dem views !!
couldn't you just put a gear on the first motor to peel the plastic?
That doesn't work, because when the film builds up on the spool, the diameter will increase and therefor change the gear ratio and pull the film until it breaks.
@@sm3vlc no need to use a spool. Just let it hang and cut the film every now and then. Alternatively, if you insist on a spool, having a flexible split gear driving the spool from the inside will allow it to slip when the tension increases, thus preventing it from breaking.
True. The machines I have worked with either use a spool (Fuji CP-series, completely mechanical model) or just leads the tape to a bin below (Fuji NXT-series, with built in motors). I must say that I do love and admire the engineering of the old CP-series. The drawbacks are power consumtion and high noise when running at high speed.
Hey Stephen are you going to try to place 0402 parts. If so I will consider you a gift from the heavens above, because you are going to save me a lot of money that I can instead spend on a lathe. I'm working on designing a real injection molding machine that will cost $2k-$3k. Its been a heck of a ride. I have been working on acquiring sponsors. I'm very impressed with your PNP design, its going to change a lot of lives.
Hey ! Any plan on open-sourcing the injection molding machine ?
🇮🇹 ottimo amico sei veramente Bravo👍 io ti seguo
Please disregard my comment if it doesn't make sense, but wouldn't you be better off with a brushless motor? Aren't those better controllable? Love your project by the way! Keep those projects coming.
SHUT UP AND TAKE MY MONEY!!!!!
Jason street
0:37 just chill out 😌