PnP Feeder Addressing: 1-Wire to the Rescue!

[@stephen_hawes]

It's good to be back gang! Excited to dive into 2021 with a bunch of new stuff!

[@ChrisFredriksson]

You got me the best possible spot on Bench 1 !! For that you'll have me as an eternal Patreon! Thanks for yet another awesome video 😊

[@jamess1787]

What do you mean convinced to use flux? Did you only rely on rosin core before? :8

[@chetanshinde8504]

Stephen Hawes You are amazing 😎

[@Gentle_Chu]

If im remembering it correctly you switched to a larger motor in order to have enough power to drive both the tape peeling and tape movement with the same motor. Now that these functions are separated wouldnt it also be better to revert to the older smaller right angled motor in order to get a slimmer feeder design and therefore having more space for feeders on the machine. Also 32 feeders seems a bit small. Maybe adding a bit or 2 extra to the address would be beneficial to future proof the system for people who wanna expand on your project. It is after all open source. Anyways great design! I wasnt a fan of the mechanics of the last revision of the feeder but i really like this new design :)

[@justinkurban6708]

Definitely keen to see what's happening with the PNP this year.

[@TheCebulon]

This is so awesome. Also your happiness is uplifting. It’s getting better and better and will be a great machine to produce lots of little happy finished pubs. Greetings from Munich, Tom.

[@simonrichter3950]

Holy moly, it's awesome how far your project has come. And your video quality is outstanding! Music and cuts and everything is on spot

[@ViniciusCZanini]

Finally!! I can't wait to see everything working

[@wolfganglienbacher]

Congrats on the flux achievment-unlock! Glad you like it! Next ditch the pointy tip of your iron for one with a bezel. Having a flat area in the tip helps keep a little reservoir of solder there, which will help heaps with drag soldering. And if you bridge it, clean the bezel and surface tension will have all excess solder stick to to the iron instead of the bridge.

[@ViniciusCZanini]

Thank you very much, Stephen!! This will definitely change the way we make boards as a DIY community! I can't wait to see the awesome projects that this machine will assembly!

[@pioneer1943]

You just made it in time. I almost went crazy from the wait lol. Happy new year and good luck on your projects!

[@henriburau6293]

Me too! I checked his YT like every second day because I thought I might miss the notification :D

[@jimamos8960]

The end of the tunnel approaches and I am seeing the light! Great work Stephen and thanks to all the other contributors that are collaborating and pushing on all the aspects of these efforts. I listened in to the trailing edge of last weeks xoom session after work and suffice it to say we have a very talented bunch of individuals in what I'd characterize as a global consortium. I'm still stuck in limbo until I've moved, otherwise I would be actively building an index right now. "Cheers" to all those actively engaged in their builds, and keep engaging and sharing on the discord going forward.

[@shyleshsrinivasan5092]

Yes ! It always feels awesome when it blinks in the FIRST TRIAL !

[@danedwards_ee]

@6:53 RIP pin 12 of the motor driver

[@draglorde]

it scared me

[@georgmik2548]

maybe I can't count but I count pin 13 😉😋

[@mikebergman1817]

Oh yea.. saw that:(

[@UhloMuhlo]

Yessss! Honestly, I was on a bit of a withdrawal. It's great that you're back! Favorite moment: 2:25

[@qwertzuiopqwertzuiop2107]

this is incredible! also i really appreciate that you a keeping all the designs clean and dont just force sth together which will only last for the video! happy i found this channel:)

[@headbanger1428]

Your enthusiasm is legit because what you’re doing really *is* that exciting. Can’t wait for you to hit 100x your current subscribers because you 100% deserve it. I hope your corporate sponsors are reading this twice.

[@JonathanKayne]

It seems you have finally learned the third rule of soldering: Flux is magic! 1. More Heat 2. Less Solder 3. Flux is magic

[@XiCKwY]

More heat transmission, but less heat in component. Your components be safe.

[@ModelLights]

@@XiCKwY 'More heat transmission, but less heat in component.' Exactly, you only need good heat at the surface of the component leads, heating the components to death like some things suggest is only a crutch against making cold solder joints. Learning to make good joints without putting crazy extra heat into the part is far better..

[@vylbird8014]

Good old one-wire bus* *Minimum number of wires: Two.

[@danielwood4786]

Love your channel, the pnp is coming along. Love to see some content specifically on board design and choosing components, at least your approach to it. Can't wait for your next vid.

[@kyream3507]

Hey Steven its Jordan congratulations on the project I was plesently surprised when your video popped up on my feed. Keep up the awesome work.

[@kyream3507]

Jordan Pesce

[@randomname3894]

The guy at pcb way will definitely ask himself why you still use no solder mask between the pins on your micro footprint.. At least I do;)

[@Soundbeforelight]

Your break was well used--this week's video quality is crazy awesome! Can't wait to see new grounds with the conveyor belt!

[@Dr_Frog]

I am impressed with your skills, the absence of flux was driving me nuts, great to see you have now implemented it.

[@itavero]

Another approach for the addressing could have been an addressing pin going from "feeder base" to feeder base, with a cheap AND gate in between. The addressing input goes to the AND gate and the feeder itself, and an additional output goes to the other input of the AND gate. The output of the AND gate goes to the addressing pin on the next feeder base. Via RS485 you can define a command that states "if you don't yet have an address and your addressing pin is high, take X as address". After a feeder receives an address, it simply switches the output to the AND gate high so that the next feeder can be addressed. In order to deal with "unoccupied" feeder bases, you can add a pull-up to the input of the AND gate on the base feeder base board. You can even feedback the output of the last base to the main board, so it immediately knows when it is done addressing and therefor how many feeders are present. Hope this makes any sense. It sounds cheaper than using an EEPROM to do this stuff and we've done something similar on an commercial HVAC system we've build parts for.

[@georgmik2548]

I just hope there wont be another hand soldering montage next time, because that is the whole reason of this project isn't it? I want to see it in action

[@hansdietrich83]

Bruh

[@rycudas]

I've basically just binged this entire project. When you were talking about using your own protocol I was wincing. So glad to see you shift to something a bit more standardised, not having to worry about maintaining that custom thing will make life way easier. Did you think about using the serial bus for address programming though? Keep one address (I'm a network gal, so I suggest going for the top address) as broadcast. Send out a broadcast 'I want to set an address to F6' message, have all the lights go blink blink, press a button on the feeder to select the one you want to program. Send a "I've been programmed with my address as F6' back to the host from the feeder you selected. Either have the other feeders watch for that message as well as the button press, or have the host send a broadcast releasing the others from their wait state.

[@rycudas]

Heck, you could program a sequence that way....

[@my1987toyota]

great Video. A quick word about solder paste if you use up the container quickly it's consistency should stay the same and it's great for dispensers. However I have personally found that it will eventually separate ( after about a month of being opened ) and even after stiring it, it will never be quite homogeneous again. This makes it's use in solder paste dispensers frustrating . I prefer and recommend solder paste stencils instead . It makes paste consistency much less of an issue. That said best of luck on your project.

[@3v068]

Good vibes, Lots of energy, Entertaining, Subscribed. EDIT: Also lots of valuable information you can apply to other projects, if you are able to follow along

[@canadianavenger]

Nicely done! I'm liking the progress you've been making. One thing you might run into is RS485 really wants to run over a twisted pair, so you may have some signal integrity issues running over the parallel strand IDC cable you are using. You can get twisted pair IDC, cable for this, which has flat sections every so-often for crimping. Looking forward to the next installment!

[@SeanHodgins]

Super cool. Once you get these really dialed in, I might have a fun torture test for your PNP. Next soldering upgrade, get a chisel tip. I can't solder anything with a conical tip!

[@rohanbandaru]

Why do you do SMD soldering with an iron rather than a heat gun+paste or a reflow oven?

[@nikkismith8750]

So very, very impressive! I've really enjoyed following the progress of this project. Are the BOM files up to date so we can order parts and build one?

[@ryanp5607]

The kicad files aren't in master - they are in the feeder-05 branch

[@brawndo8726]

11:30 How does one go abouts creating a "programming flash sequence". Is there an atmega on that thing?

[@semperior]

Stephen Hawes, Have you thought about adding uart going to your motor drivers so you can utilize the features of the tmc2209 motor drivers in place of what looks like the a4988s? It would remove the need run ends tops due to "stall guard 2" plus, you wouldn't have to mess with the pot(can modify voltage in Marlin) and has the benefits of being nearly silent! May reduce your BOM and work! Just a thought Working on my own Pnp and you've prevented me from running into a few issues, Thank you!

[@YagoFernades]

What a cool project, you are a really good presenter! Keep up the good job :D

[@FP3DStudio]

Congratulations for this incredible project, man! For the head crash, if it's now a problem, you could consider to change the camera and head position. For example, invert camera and head position alon x axis, or rotate it by 90°, so they're perpendicular as the the feeder axle. Hope to be clear :) Good job, man!! Waiting for next video!!

[@bart8P]

Crazy how few views you get, your video's are so much better then GreatScott for example

[@ConnerHall]

I love great scott. But his videos are like inspecting individual grains of sand on the beach. As a beginner, his stuff was super dense /hard to conceptualize. This sort of middle ground, practical engineering makes a lot of sense to me. I also love his forward thinking when it comes to open source projects / keeping things accessible. Keep up the great work guys!

[@zchrode]

i have no idea what is going on but goddamn this video made me happy

[@dagr8kerek]

I'm really enjoying this video without even knowing what you're making lol

[@mmcnama4]

I saw you taped the end of the peeled waste film onto the waste bobbin... why not model some kind of cut/slot/friction fit for it to grab that w/o tape?

[@Naturesyouth]

this is a really cool system. but instead of using the feeder floor and 1 wire eeprom, just have the feeders stack with pogo pins, common the 12v, gnd and rs485 lines, and daisy-chain a uart with each feeder just passing the next its address+1 then you have each feeder remember which part it has and its width, then when ever you add or remove a feeder the stack can just resort its-self, complete with free support for any width feeders.

[@joer7151]

When your boards are at their final revision, will you do a crowd supply? I'm in awe of all the work put into this

[@markkovacs4241]

Hi Stephen, I must say this project has turned out really awesome, it was fun to follow and now I have gotten the confidence to start designing my pnp. If I can have a suggestion tho, I think a vibratory feeder design implementation would be useful too for this project. Keep it up!

[@DominicClifton]

agreed, for the parts that come in tubes, like connectors.

[@qwertzuiopqwertzuiop2107]

im just porting this to this vid in the hope ull see this comment. i just watched the vid about the camera module. im not sure if the hotspot problem has been completey resolved yet. if not, i can give u a great tip that ive used with our vision projects. you can buy polarized film really cheaply online. if you put a polarizer over the camera and a different one over the LEDs it will completely block out any reflections!

[@robfegley1121]

Would it help in addressing the feeders to have address pins [log2(N) of them] on the feeder base and each feeder, meaning the slot in the feeder base has a binary-coded address? Then, the 1-wire EEPROM should have a unique 64-bit serial number built in from the manufacturer. So, say you want the feeder to have as many as 16 feeders, you rework the board for the feeder base to have each slot have its four address pins tied to ground or Vcc according to the binary value of that slot number. Then, as you move a feeder from one slot to another, it automatically learns what slot it is in and can store that in its EEPROM, communicate the slot and its own 1-wire SN back to the chassis microcontroller, or both. I was thinking an I2C PISO DAC IC on the feeder to read those 4 pins. This way, it takes the possibility of human error out of the setting of the feeder address and it effectively works like a handshake where the registration of the feeder into the slot of the feeder base is such that the slot has a permanent address wired into the backplane and the feeder EEPROM serves as the virtual address of the feeder itself. The chassis microcontroller queries the 1-wire bus and receives responses from each unit with its slot number, its 1-wire SN, and maps the logical feeder address to the physical slot address and perhaps a temporary store of what components are currently on the reel. That last part would involve human input unless there is a barcode on the reel and you had a reader on each feeder or on the chassis itself. Just spitballing here although I'm out of my depth in terms of making it work myself, meaning its a theory on my end.

[@Blamm83]

You think you'd ever need to go above 32 feeders? Since you soft-limited the machine because of the binary programming mode with leds?

[@johankorten2797]

In theory he could just add a second bus of course.

[@DominicClifton]

in my experience of running a PnP machines with designs that fit on a 35x35 square double sided PCB you'll need more than 32 feeders. an entire side of my machine is taken up with just passives (0,10,75,300,470,1k5,4k7,10k,100k,10nf,22nf,100nf,470nf,1uf,2.2uf,4.7uf,10uf,22uf,47uf,100uf is 20 feeders. then you need all the ICs, leds, didoes, connectors, inductors. i can't use a machine with less than 60 feeders with some additional tray spaces for parts that don't come on tape.

[@asronome]

You could have the feeders take a command from the serial interface that tells them what ID to write to the EEPROM. Then use the key combo to enter "programming mode" and send the ID from the computer. That would make the process much easier, considering you're gonna write custom software.

[@GlebHripkov]

Please use BC2 soldering iron tip. It is thin enough for soldering microcircuits in a QFP package, high enough heat capacity for soldering elements in 1206 packages (it is very convenient to solder 0603 with such a iron tip)

[@adarmaori9853]

@12:10 what is this connector called? I couldn't find it online and it's perfect for something I'm working on

[@moertelruehrer]

I know it's quite late, but "ribbon cable connector" brought me to the answer. They are called IDC connector, they have their own category on LCSC for instance. :)

[@Ben-gd6wl]

Whatever happened to the sweet biotech Flame thrower arm...😕

[@fluiditynz]

I love the excitement at positive results of ideas. I get a lot of those too.😎

[@AnotherByteData]

Kudos for using FreeCAD and KiCAD bro!!!

[@movax20h]

Really cool project. I think you should try to make it a more narrow design. Resistors, capacitors, can easily eat 20 different feeders, if you are not careful with a circuit / PCB designs, and they are usually very small, a so bunch of more narrow feeders would really help waste less space. Not sure what is the limiting factor in your design, but it looks it might be the motor.

[@RebelPhoton]

What's the name of the contacts you solder on at 9:40?

[@afourtrackmind]

I hope the novelty of a PNP being assembled by hand for the world to watch isnt lost on you heathens.

[@ghostman982]

just a small idea, have you tried taping some string or ribbon to the foil that gets peeled off to reduce the amount of wasted parts when putting the foil onto the spool?

[@brunokc100]

love this projekt. i would never need it, but its fun watching you build it. keep up the good videos.

[@dummypg6129]

Dude, new sub here. Are you a full time youtuber? i mean the stuff in creating Pick and Place machine is godly awesome. i mean i do have some engineering knowledge (worked as Tape and Reel Engineer for short time before becoming a developer) about this mechinism, creating one is really awesome. Wish i have the money and time CAD skills to do this.

[@asidesigner8542]

Thanks for sharing, stephen what's the part number for Feeder spring connector?

[@EdHayes3]

I feel like the crystals you are using have capacitors built in, and you're also adding them externally... AMAZING job on all this though BTW... I'm so impressed.

[@LucasHartmann]

You can use a modbus broadcast command to program floor addresses. Should be easier to automate, just have the user move a single feeder across all slots.

[@hansdietrich83]

You have to uodate some of those footprints to get silkscreen between some of the pins

[@tld8102]

Having a buck converter on each feeder circuit mechanism isn't very energy efficient.... perhapse the different voltages can be pre stepped down and the feeders can just pick up from a common shared power bus rail using something like pogo pins

[@charetjc]

I must have missed something in a prior video. What was the reason for the eeprom? It seems like you have a bus controlling the feeders, and need addressing that remains fixed to the feeder slot while the feeders are swapped out with refills. This wasn't all that clear to me from the video. If I guessed correctly, well done.

[@gnumone]

can you share a link where you get your spring finger interfaces ?

[@womacko]

Now I regret, that I've watched your latest episode.. no more episode to go, for now.. 🤐 Great stuff, very inspiring!

[@AnimaMaestro]

Really nice project Stephen! Could use one of these to assemble all the PCBs I have in my animatronic project !! Found you through your MakerCast episode. Just started my KZbin channel showcasing my animatronic and would like to get in touch with this community.

[@codemakeshare]

awesome! just found your channel, also started building a simple PnP machine over christmas. Re. EEPROM, will the feeder itself also remember which parts it has in it? I guess that can be stored in the uC of the feeder?

[@nicholaslabrecque]

your making me get back into some electronics engineering

[@buxx5334]

hey dude. i like youre pcb holder for soldering. where i can get it?

[@justascernevicius3881]

Love your channel. Please try stencil for your prototypes. It will be much much faster to assemble everything. As soldering oven you can use clothes iron that is upside down or pan filled with sand.

[@MAYERMAKES]

awwwww this is so awesome, time to get some eeproms going, I may need to start aloner project just so i cen get into details like this level

[@etch3130]

That PCB vise looks great, where did you get it from?

[@MarcelRobitaille]

Finally! Your videos are always so good.

[@mehrdadfeller]

I was wondering if you have looked at PushPull feeder design? It looks much simpler and costs around a few bucks to make...If so, what are the pros and cons of using it?

[@Motion_0112]

where did you learn and find all the things you know?

[@HobkinBoi]

Damn, maybe I gotta look into making one of these.... maybe sometime in the future. Looks hella fun and could make prototyping a breeze.

[@robrocco5420]

There is some guy's job is to look over your schematic.. to make sure that your tolerance is within spec of production and to place the cut lines and to place the board into a larger board for production.. the real boards are about 2'x4' in size.. you also have the film people that have to check the films for etching and cutters at the end.. along with the boxer's that pack up the products.. and the QA people too

[@randycarter2001]

I would have used the same optical sensor as the sprocket for the tape peeler. Then a little silver paint or silver foil tape on the arm. No need to purchase a pill magnet. I had to buy a box of 100 on Amazon just to get 1.

[@aidan2360]

I think it would be really cool if you made a video or two going in depth with the code. I understand though that a lot of people would find it boring and you might not want to do it. Your videos are great keep it up!

[@hedefalk]

Maybe this has been adressed but I can't seem to find anything about it: 1) What is the feature difference of this feeder design over the somewhat already popular and foss 0816 one? docs.mgrl.de/maschine:pickandplace:feeder:0816feeder 2) Isn't this one a lot wider meaning less of them can fit on a machine?

[@weizeng3643]

Hello Stephen, Can you tell me how to program STM32F103xx? I don't know how to download feeder-firmware to the chip? thank you~

[@_yannis2707]

Wow this guy got some serious motivation.

[@AlexMadinger]

Wow! So proud of all the achievements you and the community are making! Who finally convinced you to use flux??

[@jeanlesperance5215]

Hey how do we get into contact with you? Patreon?

[@wtechboy18]

this man's laugh is infectious

[@tankvdv]

Чувак, никогда не женись! Я не хочу ,чтобы твой блекс в глазах погас. Спасибо тебе!

[@danstone_0001]

Wow, just wow, where can I get one or 4

[@RinksRides]

I imagine it prob would be in the PCB house to invest in pulsed fiber laser cutting machines, at least for the crazy routing we can come up with.

[@martshal]

Truly awesome once again. I hate to admit how lazy I am (or how much I don't like Patreon), but have you considered using KZbin's "join" button and abilities to have subscriptions here. (Come to think of it though, I'm not overly fond of this flatulent giant either... maybe I'm the one that should rethink.)

[@randycarter2001]

Where are you posting the files for this project? The git hub doesn't seem to have anything useful. Not even the BOM.

[@FredMF]

whats the issue with fusion ?

[@ayanvaidya2727]

last scene is so weird who was that at call is that like a collab or something?

[@RinksRides]

where can I buy one/make one?!

[@yasirhassan8557]

Can you please tell me which theme for KiCAD Schematic is this? Thanks

[@vasiliynkudryavtsev]

Wow, the same T-short Dave from EEVBLOG has. Good choice.

[@mikesnapper9001]

finally a new PnP video! i was start to enter PnP progress withdrawal

[@rosskingii]

Should have put one on feeder also so you can store what product is in feeder

[@afourtrackmind]

Your positivity is nauseating. Your videos are fucking great. Thank you.

[@80lab38]

First FreeCAD, now flux and IDC connectors, where's this going to end? I like it!

[@yourmumisacrumpet4882]

this guy needs more credit

[@SudaNIm103]

ᗯOሠ ᗩᗯᗴᔕOᗰᗴ ᑭᖇOᒍᗴᑕT ❢ 𝚀𝚄𝙴𝚂𝚃𝙸𝙾𝙽: Am I correct in assuming the feeder’s UI is programming the EEPROM directly, not via the controller and isn’t globally aware of which addresses are already in use on the bus? What happens if two “feeder-floor” EPROMs, on the bus are set to the same ID, perhaps by accident? Is the bus controller up stream able to detect such address conflicts? I appreciate that this kind of problem might be of little consequence, given the scope of the project, I am just particularly interested in such things. 𝑹𝒆𝒂𝒍𝒍𝒚 𝒆𝒏𝒋𝒐𝒚 𝒚𝒐𝒖𝒓 𝒗𝒊𝒅𝒆𝒐𝒔  👍 ᐛ ⌚︎👍

[@hassanzahin1534]

I was googling about 1-wire and found your new video 🤣