Building a Conveyor for the Pick and Place!
Вставка
- Опубліковано 17 лип 2024
- It's time to build a conveyor! Being able to automatically move PCB panels into the pick and place for assembly, then move them out afterwards is the first step in allowing us to chain multiple machines together, and ultimately automate the whole assembly process!
-------------------------------------------------------------------
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
The little trick at 2:28 is sooooo slick - I love it!
That moment when you see a favorite UA-camrs comment in another favorite UA-camrs video. Love both your channels
The only thing I had to think about was how to get such good accuracy with the digital parts' orientation to match the video. After something thinking this is what I got:
Put the video feed with the part in your hand into a streaming software like OBS, then put your 3D software as a semi-transparent feed ontop of it. Align your 3D software with a still of when you want to grab the part then nab a screenshot
For the aluminum rails, try and get some adhesive (one side) PTFE/Teflon tape. That'll be plenty slick.
For the spring tension, you might consider designing a set screw which changes the spring angle .. assuming this is a torsion spring with the two legs that come up, you'd have the set screw attached to one leg and turning it would push that leg (thus increasing the torsion to the opposite attached to the arm). Otherwise, you might be able to get some inspiration from how tank track suspension is setup.
PTFE tape is exactly what I would suggest, specifically the red Teflon stuff that Mcmaster sells. It's surprisingly robust and is the slickest stuff in their catalog (trust me, I tested basically all of them for work one time. We went with the red stuff.)
@@SpencerPaire do you have a link to this? I can't seem to find this?
I was gonna say the same , you can also buy ptfe sheet and affix it yourself somehow. I would worry a little about using tape, as the adhesive could migrate out or cause issues somehow,. Tape seems to always make things sticky.
in the industry they use double side band ( one on bottom like you use to drive the board and one on top of the board to hold the board beatwin and to eliminate the friction on top side )
that would be my suggestion, add another belt that rides under the angle aluminium, but o top of the pcb.
@@kitingmare yes, the belt with springs I suggest to be placed on top ( to put pressure on the pcb from top to bottom ) in this way when the head of the machine it placing a component on the board this will not shake. The bottom shud be more stiff, to reach the precision you want
You could use servos with disk cams!
I mean, for locking the boards.
I was just going to say this too. Seems like it has the benifit of being able to turn it off and on without drawing massive current
@@benjamincrall8065 yeah, and it is very solid and repeatable.
Or small air cylinders and a rail platform under the belt
Hey Steven!
Your solution tickled my brain cells and I thought of the following:
How about using a C profile instead and embedding a thin spring loaded bar with lots of small diameter rollers on paperclip axles along the entire length? Add some PTFE/IGUS glide tape on the inner-upper face of the rail and you should get fairly even pressure distribution that could eliminate slippage. Also, having the bar embedded inside the rail means there's no need to source fancy torsion springs, you might be able to get away with something as simple as automatic pen springs?
Great addition to the Index, I 💚 it 👍🏼
You can buy ptfe/Teflon tape.. Back in the days we used it for pc mouses when playing FPS games 😅
As others have said, use the aluminium angle to support the underside of the belt, then add a second belt with spring rollers above it to push it down. The top belt can just be geared off the bottom belt to avoid have to add more motors.
Our JUKI PnP Machines in my work run the PCB on top of the belts into the machine. Then there is a Table where you can put Bolts to stabilize the PCB if it's very thin. This table drives up and pushes the Belt with PCB against some Metal that's over the Belt and clamps the PCB firmly in the Machine. This whole assembly works with a big stepper motor and a big Screw. If you have questions, just ask ;)
I think you could get away with simple spring steel rather than rollers with torsion springs, sort of like leaf springs for vehicles. Probably the easiest/cheapest source would be window screen tension springs.
Great stuff Steven. I like how many comments from operators or those who have pnp machines at work. Years ago when I worked with pnp machines the conveyors had infrared sensors to detect boards. The board was supported in the machine by a table that moved up after clamping. On double sided boards we had to position pins (with magnetic bases) on the table, so when it raised it did stab a part on the underside. One first side we also could just use big plastic blocks. I forget now how they were actually clamped - I think the lower half of the channel moved up into place. I also remember some delicate or odd shaped boards rode in a carrier. I solenoid pin stopped the board against a home position so that the board was in the right place. The conveyor was in sections so the machine could que a board to wait for the next machine, be placing a board and have another board awaiting placement. On old Yamaha machines I ran the support table was pneumatic, but I like the idea of a scissor lift table or something like that for clamping. Or servo- driven cams, with a slope on them (like a screw) placed just off center to a rail, so that it pushes board up, and slightly inward. You could even spring load that.
For the slippery top rail, try some PTFE tape for vacuum sealers. I'm looking at a roll 1/2" wide and 33 feet long for $6.99 on Amazon. It's teflon-impregnated fabric with adhesive on one side, so it should be pretty durable and easy to just stick on.
If you end up going with solenoids, look into flipper solenoids for pinball machines. They have two coils, one to get it moving and another much smaller one to hold it in place. They're meant to solve exactly the problem you described.
This Old Tony did a maho CNC conversion and used "Servo Motors" Apparently they are the bees knees for that kind of holding torqe/Positional acuracy you are looking for. might be worth looking into. I would Shy away from the sprung rollers though. Look into making a long flat arm that goes under belt instead without springs have 2 countersunk cap screws on either end so you can adjust the gap between the aluminum rail and belt to match your needed "goldielocks" distance. Looks super good this way though and seems to work fine so could just be a brute force it bearings and just use an offset cam to increase height for pinch/ gap.
In our placement machine it is as follows: the board is moved in and stopped by a rubber stopper. And then a table moves up from below, on which small metal tips can be distributed to prevent the board from bending.
When the table is raised, the holders are then pressed onto the board from above via a lever system. (These holders are on the left and right of the traversing system and protrude about 3-5mm above the board, so they hold the board in place pretty well).
I'm always surprised at how sophisticated our Siplace 80F4 is already, even though it's almost 20 years old.
There are glass measuring rods or measuring devices installed in the most insane places, where you would not expect them.
For example, there is such a marking on the inside of the quills that hold the pipettes. And these are in turn assembled to form a 12-piece turret head...
Great video once again! One idea for how to test how the setup would work with reduced friction is to tension a fishing line under the top aluminium rail. Small surface area = low friction. The suggestions of resting the board on a rigid part and pinching it with rollers from top make a lot of sense so +1 for those suggestions. Anyhow, I'm super happy to see this progress. Keep up the good work!
A couple of observations. A belt above and below each side captures and moves the board. Teflon tape may work as a slipway surface. I also note that the PCB is actually on the unsupported side of the angle rail. This means that the placer pushes against the board on a spring surface. The design will need to go through several iterations but a decent start.
Just want to say I really appreciate your effort Stephen! I'm really looking forward to this machine. I think your design goals are really filling a nice niche! And I appreciate how much you've focused on making this design accessible.
I'm doing a small production run in a few months (maybe ~1000 units). Would love to use this. We'll see which of us finishes our project first =P
Also, an adjustable spring might be a good middle solution. A 3d printed part with a screw to increase and decrease the tension on the spring. ¯\_(ツ)_/¯
I use round polyurethane PU belts. They have a bit more friction and you can splice them by heating both sides and melting them together. They also exist in many different shapes
9:50 igus has some fantastic glide tape in their portfolio.
This shit is mint! I love your enthusiasm and ingenuity. It's contagious.
KISS idea is to add stops for the clamps, so the belt by the rollers doesn't touch the aluminium. If it is adjustable and set to for instance 1.5mm for a 2mm board you get way less hill climbing for the board.
But if the board is far from both rollers the belt may såg too much :/
I love this series!
Hello, I just made a video about the transformation of my 3D printer to a through hole soldering machine. In the video I referenced your channel so people can also see your amazing work. Thanks for inspiring me to move forward with that project.
Hey. Nice setup you've got there! I recommend you other approach:
Instead of spring loaded rollers, use 2 more stepper drivers and motors and add some stoppers for the panel and some clamping brackets to lock the panel in place.
I'm telling you this, because this is how industrial machines are made.
You need a gap between the belt end the aluminum profile when moving the panel.
When you feed a panel into the machine, a sensor shall detect that, and start the conveyor system and put the second sensor into wait mode for let's say 5s (if the second sensor does not detect the panel in 5s, machine shall get in error).
When the panel hits the normal closed stopper, a sensor shall stop the conveyor system and trigger the clamping system to lock the panel against the rails(aluminum profiles), thus keeping the panel as stable as possible during PCBA.
After PCBA is done, the clamping releases, the stopper will get out of the panel's way and the conveyor will start moving.
After the panel will exit the machine, another sensor shall trigger the closing of the stopper.
This is what I've seen working with a selective soldering machine and system is quite reliable. It was compressed air driven(the stopper and clamping system parts) and for the conveyor system there were used 2 servos on each side of the conveyor, because the machine accepted multiple panel widths. Instead of belts, chains were used because of the high temperatures needed for selective soldering (140degC panel temperature and 300deg C wave solder temperature)
I'm waiting to see something close to that from your side also 🙂
Good luck and greetings from Romania! 👍🏻
Teflon/UHMW strip along the underside of the rail. Low friction coefficient, relatively inexpensive, consistent thickness so your Z axis location will be consistent along the entire board length. I believe self-adhesive teflon tape is also available.
That and I'd add the weaker-sprung rollers as tightly along the rail as can fit, so as to minimize how much the board will "bump" as it steps into and out of each roller as this movement could shake parts out of position.
What about a Delrin vacuum table. It would have the low coefficient of friction for when the belt is moving, and would be very stable when vacuum is applied. You would only need to power the solenoid briefly. And as you are not using it for machining you wouldnt need a very expensive pump.
You mentioned probably not pneumatic, but that actually might be a decent way to clamp. Could also solve the sliding problem by adding bearings or something above the board, so it's not actually sliding.
The P'n'P-machines at work uses normal DC-motors for the conveyor, they rely on a pneumatic cylinder and reflective sensors to stop the PCB at the correct position. The conveyor consists of bearings and a flat belt for transportation and pneumatic controlled clamping from the top to secure the PCB (the cylinders is mounted below and outside the conveyor).
And I think that you should separate the conveyor in to three sections (in feed, mounting area & out feed).
I was going to say... Man you really made the camera angles work around the beam!
Ahahaha nope, got insanely fed up and moved my whole shop. The other room is becoming a print farm! :D
Awesome build series... Have really been enjoying it.... Why not try an upside down version of what you made? Have the aluminum rails on the bottom and the belt on top? The board should still slide across the rails just fine and the belts will still pull it, but when you apply any pressure from the top the board will have the aluminum to keep it rigid. As it is now, the belts let it spring down with any pressure.
Good job
This is so cool. It is really looking pro now!
For the rollers that you are going to fit maybe you could use a concentric nut. Or you could change it instead of using timing belt you could use one stepper motor with a lead screw and some sort of clamping mechanism to hold the PCB in place. And use some 8-mm smooth rod with ball bearing like the ones that are used on 3D printers.
Very positive vibe
You'll probably have to use some kind of variable clamp pressure mechanism. larger 55g or 20Kg hobby servo motor and cams or bell cranks should do it. I recommenced through beam IR sensors, light source on the bottom, receiver on the top. A mechanical switch will be problematic.
Just casually using a panel of 2 Formlabs boards ;)
👀
Excellent
You need a lift plate (sometimes called a rising table) with magnetic base pins to lift the pcb up off the belt and to a consistent (top referenced) height so you don't have to constantly change z heights of your placements based on board thickness variation. You also need to plan for pcb warpage, sag, and flexing due to placement force. It would be better to lift the pcb up against 2-4 tabs on each rail that can be loosened and moved (in a t-slot) if there are parts placed close to the edge. Designers, especially noobs, constantly violate IPC design rules and put parts too close to the edge of the board. I would suggest flat belts instead of GT2 timing belts. You will also want three IR proximity sensors, one at the entrance, one at the exit, and one to slow the conveyor speed just before a pneumatic cylinder / pin that serves as a board stop. Research the IPC SMEMA specification so you can properly interface to upstream and downstream machines via four pins / two switch closure signals at entrance and exit.
You also need to add adjustment points on each conveyor rail so you can level the plane that the pcb is in relative to the plane of the pick&place X-Y gantry.
Here's a good view (at the 1:20 mark) of what I am talking about:
ua-cam.com/video/eIvABQLuqWE/v-deo.html
What if you flipped the whole thing upside down? If the angle prevented it from pushing down, gravity would help your spring rollers and would only need to prevent the board from moving laterally.
then you'd need to push the belts down so that they could even have contact to the PCB, instead of the weigt of the PCB doing it for you like now
Did I really just hear freecad instead of fusion360? You man are a shining example
Yeah, he's trying to open-source as much of the process as possible, including the development tools. KiCAD for board layouts, FreeCAD for design, various OSS IDEs, etc. etc. etc.
@@OddlyIncredible 😂
Bicycle tyre puncture repair kits work great for joining GT2 belts.
Really?! Ok details! Thanks for the tip! The glue chemically reacts like pvc "glue"? How do you cut? How strong is it? Does the lack of metal make it snap often? Details please...
@@TheRainHarvester The glue is a non contact adhesive but the puncture patches help add reinforcements. As to strength, I have never put enough tension on it to test.
Good source of a super slippery strip you could put on the underside of the aluminum rails - vinyl cutter teflon protection strip! Cheap and cut into strips for you already lol.
There exist Teflon/PFTE Tape, which you could tape around the aluminum rail, to increase the "slipyness" whiteout adding thingness. Also it might be cheaper to use such tape than bolting a fat piece underneath it.
You may wish to consider some HDPE/Delrin " C" channel and run two belts that the board is sandwiched between, you could also have the lower belt extend beyond the perimeter as an infeed/outfeed tray
As I foresee an armature of sorts, picking up a board from astack, dropping in on the infeed/lower belt, that belt running through a stencil area, then rolling into the PNP area, where it's gripped by the upper belt, parts are placed, then it's outfed into the oven area where you'd offload it to a second set belts running through whatever heating apparatus you come up with
Cool! You could put some PTFE sticky tape on the underside of the Aluminium rails to decrease friction. Those tapes are pretty cheap.
You could use the top vision camera for board positioning on the conveyor
I like the buncha rollers idea
4 belts in two pairs. Two on top side, two on the bottom.
Left side being driven by a stepper with meshed gears to drive top and bottom belts, mirror for right side.
Use teflon tape on the top rail and build a parallelogram with another aluminum angle to support the bottom belt, to raise the parallelogram you can use air cilinder (using the vacum pump as compressor), or some servo with a cam.
Kapton tape is pretty slick
As others have said - belts top and bottom will help hold the boards in position on the conveyer, and can decrease the friction overall. I guess the worst thing that can happen is the board slips relative to the conveyer which would land it out of position relative to the PnP heads expectation and lead to a bigger error to be dealt with for the fiducials. Not an issue - just not ideal. also - with two belts you'll need the bottom one to move up and down to accommodate different thickness boards which adds complications but better to have in than not.
I'll start by saying I love your content! I have 1 question for you on this setup however - what about boards like your automatic feeders or boards that may be round/don't have parallel edges? Thanks again for the content!
Do you really need to put the rollers at the same location on both sliders? I would say put them in zigzag form to always have one roller in contact with the board. It should reduce the required number of rollers. And 4 might work.
Cool idea. May be cut on the length a 4mm PTFE tube (Bowden 3D printer) and put on each side of the Alu rail.
Very cool update! I have a though: what if you left it the way it is with the 4 spring rollers and then have some spring rollers in the middle that directly support the board? Or sprung paddles or rails or something? I think it could be a good idea to have the belt function to move the boards but have a seperate set of sprung parts supporting the boards from sagging. Like you said, it gets a little finicky near the rollers and having an independent structure that handles support may avoid this? What do you think?
on conventional machines the supports is under the belt so you can remove the board at any time except the final stage when there is a moving up an down bad with supports pin that lock the board for assembly.
At 3:52 my brain goes "SPACERS!!!!"
Verry nice video im locking for something similar for my 3d printer
I would use normally off solenoids to hold the board up with spring pressure, you can use larger springs and then put the solenoid 2~4 times pressure point distance away to have leverage to compress the springs when the boards move. Constantly dragging boards against a fixed causes a ton of friction and wears components.
@Stephen Hawes How will the board's be staged or stacked and then moved onto the conveyor? Do you have a drawing or something to help us see how this is supposed to work (end to end)?
Better purchase on the toothed pully: When you get the belt length right, maybe you put the tensioning pully on the other side of the belt and wrap the belt around the pully more instead of less.
As position detecting i would use a reflective photointerrupter.
That doesn't affect the friction
Maybe you can tension the belts in the longitudinal direction in a way that pulls the idlers down, and thus reducing the clamping force on the board.
Would it be possible to have a large, wide, perforated belt, with something like a vacuum table underneath?
i love your videos
What about building it the other way around with the slippery rail on the bottom and the belt on top? Wouldn't that help a lot with flexing while placing parts?
to increase pressure of the idlers it might be a good idea to make the "arms" pushing them against the belt smaller, looks like 5 mm should be possible right away and when making the idlers smaller even more should be possible.
How can you see the fiducials on the PCB rails if they are covered by the aluminum guide rails? Typically the PCB rails are only on two edges of the PCB and the PCB rails are only 5mm wide.
try Teflon for sliding surfaces, as it is the most slickest solid sliding material.
if that doesn't help try another timing belt on top so you dont need any sliding surface at all
I wonder if you can get teflon tape on a reel that you can stick on the underside of the top rail? Can't wait to see what you come up with :D
I personally realized that paint-coated extrusions are hell a lot more slippery than raw aluminium ones ( I used T shaped ones for guiding my rotary cutter and the painted one slides off the material a lot more) so maybe an other model of extrusion could be a solution
So yeah a drive train of a bunch of rollers, also you could add tensioners to the belts, and nylon or ptfe strips to make it slippery... Maybe a belt with teeth in both sides?
Hi, nice work. Respect for this.
Try to use "POM", it is a plastic with good sliding properties.
Good luck and best regards
would it be possible to add a nozzle to the pnp that can pick up a board from a stack an then place it on a belt where it gets clamped into place somehow?
Looks like my neoden's rails. A lot like it 👌
Have you tried using UHMW tape on the top rail? It's and easy fix without any CAD or physical changes, worth a try.
excellent, wondering how much of nightmare diy selective solder would be now
I‘d try something like 3M 5421 on the aluminum surfaces (both top and side)
Or PTFE ultra-low friction tape.
If you flip the Assembly, the Board is pushed in the aluminium. This would be much more stable and the Belt would be on top and push down
Why don't you use rollers by pairs (bottom driven *and* free on top) to remove the friction issue?
It's been my experience with pick and place machines that they is tappered pins into holes, at the final location, to both hold the board up against the rail and register the board.
What if you put the aluminum angle iron on the bottom and the belt sliding on the top with it tensioned? That way the deflection downward is essentially removed because you're pushing against the metal rather than a stretchy belt
Will we also do a reflow oven?
I would invert the channels so the board is resting on the rails, and use the belt just to grab the edge of the board and drag it forward
You could replace the upper rail with rigid rollers, or maybe even put belts on both sides.
Is this still an open source project? I can't find the latest version of the feeder on the github. I really hope to see the pcb file, I enjoy learning them.
This kind of conveyor will prevent to use of fiducial markers on the rails (I'm not sure how OpenPnP supports fiducials), but in this case, it must be placed somewhere within PCB I belive
Stephen! What if you had a belt above and below the board? Like how pitching machines have 2 wheels to maximize grip, and maybe using a stiffer belt; or a wider belt? Or just a series of rollers like in shipping centers running off a single motor and geared in series
Shot in the dark but what about small suction clamps run of a separate pump to hold board. Once its activated it would also act as a known hight with low verification
How will it handle smaller boards or odd shape ones like the glow tie
probably have them panellised with tabs into a large rectangular board, so the Index can do multiple small boards at once, then broken into individual boards afterwards
Panelise or run them on a carrier
Coming along great.
Out of interest, what are those crimpers you're using for the JST-style connectors? ...and would you recommend them?
I have a set of ratcheting crimper, which are fine for larger connectors but tend to mangle the crimp on smaller ones unless you get them aligned just right.
I've been looking for a reliable alternative which aren't insanely priced for a while now but there are so many it all seems a bit of a minefield.
Engineer PA-21, I got some after seeing them on this channel, they're fantastic!
Or maybe some way to atach te board to the belt? I was thinking magnets but ofc that would mean you need to put another one in the board and that means more preparation/work.
Perhaps ptfe tubing split down one side and slid over the rail would reduce friction?
I was about to suggest nylon strips, which yeah, it's about the same thing
I would rest the boards on top of the rails pushed down by spring steel and pushed forward by one belt with a stud in the middle
Instead of a static setting for the springs why not change it up to use a tension screw like you see with 3d printer extruders. simple solution and would allow for a variety of springs to be used.
Look for Murtfeld ! Perfect slippery surface
The belt idea looks amazing, but will it add a lot to the cost? If/when you sell kits, can you do a version without a belt for people who don't need that functionality?
Same feeling here.
PTFE is a good material for slides.
1:55 anyone has that twitch stream? is it on his patreon?
I heard that it is better to use idlers with teeth on them when they are on the teeth side of the belt, does not matter that much imo but just want to reduce possible problems in the future
I also thought about that. he can just print them with teeth I am pretty sure there are already appropriate models for this on thingiverse
I know igus make a foil called Tribo-tape, with has less friction (if you belive Igus) and is made for stuff like this.
Most of my boards are quite small with components close to the edge which would be a problem with this system. Maybe for a large run they would be panelized so would not be an issue but it would be nice if it worked with all types of board.