Upgrading the Pick and Place Frame!
Вставка
- Опубліковано 1 жов 2024
- I've been using the same mechanical design for my pick and place design for a while now, and it was in desperate need of some changes. In this one I go through and redesign all the printed parts, and build up a much better machine than I had before!
Thank you to PCBWay for sponsoring this video! www.pcbway.com/
Check out my Patreon!: / stephenhawes
Keep these passionate videos coming. Every time you post one of these, you get me out of my chair and down to my workshop to try and make something, or at least clean it up to get ready for the next project!
Nice work. I am working on building one too.
When will you update github files?
Camera suggestion: All the close up face stuff is very jarring.
I believe this project will be soon in china as DIY kit :)
Your work is so awesome!
STL for the “clippers thing” cable management clip. Please
I have two nitpicks.
First: I don't like the name Index is way to common and not distinctive enough. It lets it fall under the radar.
Second: I would do a another redesign to use a H-Bot design meaning you could get rid of the Y-stepper
But other than that I really like it 👍
I hate to pain you with a new suggestion but the plastic parts you printed for the corners should be at the same height as the top of the Z axis so you can flip it over and work on the bottom more easily.
Stephen, using the LulzBot TAZ6 - I'm now 1.5 days into printing the legs for the Index frame. Going low and slow for best results. 200 micron, 4 perimeters, 30% gyroid infill, 3mm brim, no supports. Should be done in another 10-15 hours.
Going well so far -- and I'll post the build on The Maker's Workshop. If there are any major changes coming up ( RS485 for the feeders I see ) it would be good if you could give me a heads up before I commit to milling/printing parts or sending the gerbers off to be manufacturered.
Looks great! I bet you can still get rid of the remaining supports though. I'll be great to see this thing going!
STL ?
You should ask for 6 layer flex PCB with rigid FR4 overlay teeth to use as a belt. Probably gonna cost>1000 USD to manufacture. PCBway explodes.
Thanks for the update! Can’t wait for the next board revision and one day I’ll build this beauty! 👍🏻😊
You can buy rack + pinion whatever kind you want : us.misumi-ec.com/
This is not AD, just it was a silly sentence: "I couldn't find a good reliable source of ..."
At Misumi you can buy anything you want for this machine, but they are Japanese firm therefore very expensive.
I know I've said this already, but you're like the Bobby Duke Arts of electronics. Seriously
You don't need more holes in the feed sprocket wheel. Using quadrature encoder technology and placing another sensor on the feeder you get a quadrature encoder. This essentially quadrupedals the resolution. The exact sensor location doesn't have to be right next to the existing sensor. The only requirement is that when the hole is dead center over sensor 1 sensor 2 is right over the edge of a hole.
I had just finished building the frame, got the XY and Z motors, then dropped it on the floor. Snapped one of the Y motors and snapped the down camera bracket :-( time for the glue I think. Will reprint with the Free Cad versions once I can prove it works.
Hi Stephen) can you share STL models of your PnP machine?
Did you change the lengths of the aluminium frame?
Can you provide a link for the GT2 belt for the dual head?
Quick tip regarding drag chains: Make sure to use cables rated for them. They often rub along the side of the chain and can wear through there. Also not all cables handle repeated bending well.
Is there a "rating" to look for? (Like mm²/AWG for square area and thus current handling but this time for use in a cable chain?)
@@wi_zeus6798 I have no personal experience with them. But in my opinion ua-cam.com/video/ZPWoC3N0epM/v-deo.html explains the issue with drag chains (and other cable management methods) very well. (In the context of 3d printers but they're not that different from a pick&place machine.)
@@trilader thanks, a good video on that topic
@@trilader I know that there is a spec available for cables (maybe individual insulated wires) that guarantees that they are fit to be used on industrial robots, where they get bent a lot but don't necessarily rub against other materials (don't know if it's part of the spec). But these cables are pretty expensive
will it be possible to pick and place 0402 components
How about making it an Core XY? For more Speed
I don't know if this would lend itself to a coreXY versus Cartesian, just because of the need for more rigidity in coreXY to prevent the machine racking during operation, plus coreXY is a slightly tougher drive system to both design and code for. It might be necessary to double up the front and back rails, and/or bolt the whole thing down to a giant plate of material (e.g., plywood sheet), in order to reduce flex if using a coreXY drive system.
Plus, if Stephen designs his controller to allow that functionality, setting up accel/decel curves should allow for pretty fast traversals even on a Cartesian machine. My desktop CNC mill is Cartesian and thanks to accel/decel curve setup it'll move over 20m/min during rapids.
@@OddlyIncredible You may be confusing corexy with h-bot; corexy was used almost as a successor of h-bot for 3d printers specifically because it avoids the issue of racking. I still think it would be non ideal for this, as its belt runs are exceptionally long compared to cartesian machines and this is a Quite large machine, and there is more of a susceptibility to small positioning problems as it uses both forward and backward bends in the serpentine belt runs. (this is also a problem with h-bot)
I hate you for not using heat shrink on the limit switches
Only noobs put heatshrink on solder joints.
Real geeks leave them bare so you can clip probes on it, or solder on extra bits later. 😋
Why not use the ends with m5 thread in so you do not have to use all the t-nuts?
How about sensorless homing?
ARM's the way to go!
I hope he is looking at the Samd21 or Samd51 chips, I have been using those for all my projects now, and they even run FreeRtos on them. They are great!
I suggested STM32 because the ecosystem is really inexpensive ($11 dev boards, $2 chips, free IDE from ST and Arduino IDE support is out there, cheap toolchain) and pretty powerful, but SAMDs would work just as well. There are plenty of beastly little ARM Cortex MCUs out there that could run this thing without issue.
Hey very cool upgrade !
I was wondering, have you thought of using brushless motors with a ODrive board for the X et Y axis ? That would make them much faster than with steppers !
While that's certainly useful, what would be the additional cost over traditional steppers and cheap drivers?
@@wi_zeus6798 Yes: Odrive: $119USD, motors 2x @ $30USD each, minimum. Encoder for each motor: Depending on what you get can go from $10-30+USD. You'll get very good performance, however, vs a stepper motor 2x $15 + 2x $5 for a driver.. certainly is cheaper.
@@wi_zeus6798 That will greatly increase the cost of the overall machine yes. The steppers he uses are priced at $8.50 a piece, while a good hobby brushless motor costs roughly $60. Then you have to buy encoders for each motors, so between $10 to $20 each, and finally the ODrive board. The board costs 119$ on the website, and can control 2 motors so it would be enough for the X and Y axis. But as it's a fully open source project, all the schematics and PCBs design are published. Maybe he could produce a cheaper board by doing it himself.
There could be two tier of the pick and place machine. An entry level one with steppers, and an advanced one with brushless motors for faster board assembly.
@@codanael I you / someone really wanted to, the necessary parts could be added to the control board of the PNP, as an alternative version.
Personally I would not want to spend the additional money since I know that I'd not be able to run the machine at these speeds. It would be too noisy. But that's something not everyone has to care about. Also, the additional cost is quite high
Cool updates! You should really look at switching to NEMA8 hollow stepper motors. It's going to be lighter so you can drive your PNP faster. NEMA8 is more than strong enough to rotate a part.
Holy _shit._ Just, well, holy _shit._ Every update is better and better.
I'm already looking at buying a 3D printer just so I can build a clone of this PnP rig.
Use dual end stops on the gantry and dual drivers so it can automatically square up when homing.. also feeders rs485 or maybe CAN BUS. Lastly maybe an optical pickup on feed tape holes as a sanity check have it count the feed holes. If done right it can be fairly accurate
Where are you from?! I noticed the CT and MA plates on your wall!!! I'm from NH :D
Great work!. Just in case you had no though about it you could use I2C if you put the address info on the connector side of your feeders. Something like a voltage divider on the small connector board and an analog read on the feeder board... I know you don't like analog stuff :)
Looks awesome man. I will say. Between you/Great Scott/ EEVBLOG/ and bigclive all got me to get into electronics. I wish I would have done it years ago. I can finally work on ideas that I have had.
How to stop the linear rail waggon from falling of the rail? I'll drill a hole in the ends of both rails, cut a thread, and screw some hard stopping metal piece in there. Has the advantage that the waggon still cant run away from the rail, even when you assemble or take apart what ever machine the rail used on.
For the ring light you might be able to do something like what Carl Bugeja did (ua-cam.com/video/R6fPeYFXOoI/v-deo.html). Just make the slots in the main PCB a bit longer so the sides can be angled before soldering. maybe print at jig to hold the assembly while soldering :)
This was the video that brought me to your channel. You do a very good job with everything! Thanks for what you're doing!
Hello from Russia! Your need used HIWIN linear (or other manufacturer) for speed. Your project now very slow speed for auto pcb assembly. Good luck!
The part introduction/reveal in this video was excellent! 😆 I love seeing your personality show through in these videos.
One way to get the cable chain out of the way (at least a bit) is that the stationary end to the middle doesn’t move. This means you can have the cables come out from the bottom in the center of the axis and have a cable chain half the length. Less length means less floppyness, and I find this much more manageable on my machines.
very nice. you could also have used tmc2209 drivers, then you dont need an endstop ;)
what is that amaizing screw vise you have?
it's called the Omnivice! it's the best thing ever: kit.co/stephentherobot/my-maker-essentials
Oh... this should be good.
I like your design a lot, will this new design will be available on the github page soon?
keep the good work..!
sorry was that a multi wire stripper that removed the insulation from all of the wires at once?! where did you find that?
You are awesome bro can you send me the link from where you got that universal all in one PNP PCB
Really great job Stephen. Thought trough very well. This was the reason I bought a 3D printer.
I thought you would have gone with stepper drivers that have sensorless homing full stop that would have been the better option as it would have been a bit more up-to-date with things. As end stop are slowly going out of use.
3rd frame so far. You pay for it we learn from the case for free. Thanks.
I you are having trouble lining up fasteners put a small magnet on the end of your allen key, helps the fastener say on the key! Have you seen the clear disks that normal printers use for optical location?
Aw yes!! I'm the latest addition to your table! 😍
any one else going to meme that smug "limit switch"???
You should look into core x/y or h type carriages for lighter x axis and super fast movement. Also I would replace the steppers for servos or bldc motors with odrive controller. Your design is more like a lightweight router than a pick an place.
TBH I suspect that there's a reason commercial PnP machines are Cartesian and not coreXY, although with his having open-sourced the whole thing, someone may already be working on a different drive system.
I Just Love Your Videos
Keep Going Man 👌❤
I was waiting for an opto limit switch :(
this is awesome. keep it up
2:42 has me cracking up
GREAT
You dont need limit switches when you have the skr 1.4 turbo 😎
Dude, you are smart
Ha ha
👍
Like it’s over for dude like he just sold his self to be famous
What?
Are you talking about his touting his sponsor? PCBway does make good boards - my main go-to is JCLPCB but PCBway is also good, especially for weirs shit like his belt-drive/optical-pickup board.
I feel bad for this guy he going to hell at the end
marcus. Pratcher VEVO we all do, don‘t we? And it’s going to be fucking awesome!