Running Multi-Feeder Jobs!
Вставка
- Опубліковано 17 лип 2024
- We are finally running multi feeder jobs! OpenPnP is controlling the whole shebang, and Marlin is handling everything, including feeder control! We got some great results with the repeatability of the feeders. With some bottom vision for alignment, we're in good shape!
-------------------------------------------------------------------
Check out my Patreon!
/ stephenhawes
Join the STR Discord Server!:
discordapp.com/invite/TCwy6De
Index Pick and Place Github Repo:
github.com/sphawes/index
-- FOLLOW --
Twitch bit.ly/32q4pDG
Instagram bit.ly/2FrreOJ
Twitter bit.ly/2FArY3G
-- PO BOX --
Stephen Hawes
PO Box 44248
Pittsburgh, PA
15205
-- GEAR --
These are affiliate links so if you buy anything through them I get a small percentage which helps out the channel. Thank you!
PCB Holder: amzn.to/344qhVD
Microscope: amzn.to/348aSnw
Soldering Iron: amzn.to/2KlHDGs
Wire Strippers: amzn.to/2IESgDV
Cutting Mat: amzn.to/3oNCwOz
Flush Cutters: amzn.to/3ncFwnc
Digital Calipers: amzn.to/3gIzZSL
Leatherman Sidekick: amzn.to/3a4sQuY
Part Organizer: amzn.to/3a9uDyV
Soldering Iron Tip Cleaner: amzn.to/3a3wZiD
Oscilloscope: amzn.to/2W4zrNt
Multimeter: amzn.to/2JVp6Ba
Crimping Tool: amzn.to/37WymwP
SpaceMouse: amzn.to/3gCWR64
Tweezers: amzn.to/3a1auL5
Solder Wick: amzn.to/347kUVI
Files: amzn.to/37bMXp0
Dremel: amzn.to/3gIADj9
Headlamp: amzn.to/3gE5VYw
Pelican Case: amzn.to/3ncFpbg - Наука та технологія
I'm soooo waiting for a kit for this to be released.......ready to buy, and sorely need it!
hey Ian, caught you here
Tip to not waste so many parts during threading of SMT tape. Take a piece of string or thread and tape it to the end of the cover strip. Run that up and wrap around the spool. Don't throw away used cover tape as it can be used for the same purpose.
it is bonkers how these days at home you can just pump out something like that feeder, damn excited for this to be ready to go!
Pnp: "what's my purpose?"
"You make bow ties"
OOOH BOY, HERE I GO ENJOYING THE NEXT 12 MINUTES AGAIN!!
Where's the other 1.5 minutes?
Love the little feeder shot with the parts going in a circle!
You and the guy, I like your videos, everything you do here is open source, I come to thank you for the initiative because there are many here who see your project and leave wanting to sell something that is open source
that black trim and baseboard looks good af. i have been looking at doing that to my house and you randomly had it this video. bet you convinced me. thanks
The Index gets a devlog! This is the greatest day!
I am definitively looking forward to the future of your pnp. It's the main thing missing for me. Good job sofar!!!!
Very cool project! You could permanently attach a piece of tape to the feeder spool that can be connected to the other tape, so you don't lose as many smd parts in the beginning ;-)
You really improved your editing skills! Good job!
It's amazing how contagions and understandable your excitement is when these projects work. 😜😊👍
Love your excitement!
watching the PNP work is most satisfying
Ill buy this as soon as it is available! looks awesome
I had to subscribe... you are such a nutter in such a good way!!!!
One day, all of us will say "I knew Stephen when he had 20.000 subscribers" when he unboxes his golden play button...
A conveyor belt, cant wait.
Awesome work! Love the shirts, may I ask where their from / can buy?
Awesome video Stephen! Maybe you should invent a PnP machine to build up shop when you move to the next place? hehe.. Jokes aside, for the 485, don't overcomplicate it, as simple as possible is more than perfect and will not make troubleshooting a nightmare. I'm working with my own 485 network at work, its not fun when it sometimes isn't working as expected.
Wow! I'm waiting)
Deffo do videos on the openpnp vision pipeline tuning.
The problem with having wide feeders is that as you stack up more units the pick head has to travel much further to get to the part and back to the board. That takes longer. Also, you'll need at least 20 feeders for a medium size board and you won't be able to fit them all in unless your machine is huge.
@@kestvvv 8.6mm, very nice. Lets see a picture of it. I've been 3D printing my own but its 20mm for an 8mm feeder.
Great! Can you reverse the "placing process", so it picks the parts and fills the feeders, after you are done testing? I can imagine it's getting quite messy with loose components all over the place :D
Genuinely surprised you didn't opt for CAN to talk between the motherboard and the feeders.
cool feeder
Do you think that at some point you need ad another chip to the board to split the process load?
Rig-A-Ma-ROLL!! Love this channel :)
Hi Stephan can you make a pnp machine like your mini pcb miller sized?
Did you discount CAN bus for any particular reason when looking for a bus for the feeders?
Physically its very similar to RS485 (twisted pair with 120 Ohm termination at each end), but, you also get a nice message based protocol with arbitration and node IDs...so your feeders and controller can freely communicate bi-directionally without needing a custom protocol that deals with arbitration and collisions and addressing. The STM32F103 has a built in CAN controller and just needs a transceiver like the Texas Instruments SN65HVD230. The only real drawback I can see is that RS485 allows a slightly higher node count. CAN is still good for at least 120 nodes on a single bus though.
FINALLY 🎉
Stephen is becoming a soldering god
Can I please see a demo where it stands now
Try Modbus. There's an arduino library and it's pretty much got everything you need
Inspiring work. I'm late to the party so forgive me if this has been discussed. You only need one pnp machine but many feeders. Your feeder design is very cool, but to keep cost down, it seems to me feeders should be dead simple with no electronics. I've seen passive feeders where the pickup head pushes a lever for the next part. I imagine that slows things down. If so, what about a device running on a rail under the feeders that can move to any feeder and dispense its next part?
"Ticka ticka vrrrt"
I'd recommend some smoother stepper drivers to reduce vibration, something like TMC, in stealthchop they'd be perfect, and even allow you to go super fast if you wanted, they're very powerful
I don't get why he uses stepper for that - odrive servos are cheap too but with much higher speed and probably even better precision. If you search for odrive pick and place you probably get what I mean.
Hello, I'm Goblin Goul from Korea.
It was very helpful because I referred to the video a lot.
Can I get a list of the parts used in the pick and place machine you are currently making?
I would like to know which parts you used, such as camera, head, motor, controller, etc.
And if possible, I would like to receive the Marlin firmware or Smoothie firmware library, can I get it?
🔥
WhenI stumbled across your video series I was thinking you did any "invention" on this. Meanwhile I found some >1 year old videos from the guys with the idea to built one.
I feel like this would be easier with Klipper
nice
I am confused... what is RS45?
Rs485 , a differential signal so the signal can travel far without getting corrupted. I used it on my 30' robot hand video. It worked 50' + away.
@@TheRainHarvester Ah, thanks! That was quite confusing to me.. Perhaps say the "eight" for clarity? I don't think it is common to omit it. :-)
I preferred JLCPCB, it is much cheaper and the Quality is the same...
IS IT GOING TO BE SICK
The feeder waste too many parts , some room for improvement
yeah, my first thought too. Affix some cover tape so one doesn't need to feed until cover tape reaches spool.
It's a shame I wasn't watching for this stuff about a year ago, probably $200 worth of this could be designed out. Some good ideas but seems to be 'designed toward the $10,000 and $15,000 pnp machine' ideas, where they've done a lot of extra things I'd consider unnecessary especially for a low cost design.
There's a better way to arrange the rails, and cut of 15 or 20 percent off the floating rail's weight with the same reach and better accuracy.
Kinematic tool changer set up so you can take off the head and put on a laser cutting head, and probably light CNC capable.
Double rail both sides, for up to 128 feeders. (Realistically I'd do 64 on one side, and not have to access the other side of the machine unless I really needed more than 64.)
Dallas is great but expensive, instead of their $1.30 part I'd use the Microchip 22 cents part. I'd design the pad to take all 3, Dallas form factor, TSOP, and TO-92 for whichever is cheapest at the time.
Non-intelligent floor, not sure what the purpose of having an essentially '3rd party' knowing and communicating really is, I'd bypass it entirely.
Feeder motors separate from the cassette. Only need as many drives as you're using, while having lots more parts with the cheap plastic cassette part ready, just snap off one and snap on another. No motors or electronics sitting around wasted if you have a lot of different designs, but still quick to change and use parts.
Rest of the plastic cassette printed, with the ID chip in that part. Print them, add chip, and they stay married to your part reel forever. Snap them on and off the feeder motor board.
Positions picked up by the head, no need to keep track of what's where manually. Head goes down the line touching the ID chip pads, and knows exactly where parts are. Swap 2 feeders it can pick up the changes. Only needs a simple switch for 'something changed', the head can pick up what went where.
I'd look at a part tray attached to the head. Go to the parts, pick one up put on the tray, times 8 or 10, then go place 8 or 10 parts. Much less wasted travel. You could go slow and still beat 'faster' machines.
No up camera. 4x down cameras, 90 degrees around the head. Align the parallax so your positioning is dead accurate. 4 pics your aligned correctly, place, then 4 pics and you know the part is aligned correctly.
I'd realign the feeder design so the pickup area is directly over the rail, that's what needs maximum accurate position. Then push everything else down, so it hangs like an upside down U over the rail, with the side 'in' as short as possible so there can be a second inner rail with nose to nose feeders.
Those are just the main things I'd rearrange off hand, probably forgetting a few bits here and there..
Well there's the obvious other thing, easily squeezed back in to a $30 BTT SKR 1.4 turbo board while still retaining basically all the functionality.
If you haven't guessed, I'm looking at doing exactly this. Can't escape $200, $250, or $300 worth of hard parts, rails bearings etc. But I'm pretty sure you can make a $400 to $600 unit that will actually be better than some of the $10,000+ machines. 4 cross aligned cameras and get a parts tray working and it could be very nice.
Ok. This is making me install Discord App and Twitch so I can get more of your work. Video is just a tease! Again, your videos make me want to get up and make something even more than #ivanmiranda.
I don't think speed is that important. Accuracy is 100 times more important than speed.
Oh no, looks like you lost access to the FormLabs printer when you left! Ugly brown FDM. :(
no weird hacks? wtf is up with that tape..
Github needs a full update. Feeder design is 10 months old as of March 6th for example.
Check the FAQ! Everything is up to date in feature branches. github.com/sphawes/index/wiki/FAQ
just to let you know YT is again unsubscribing people randomly..for example me from your channel.