Don't dip the boards in the flux. Put the flux in a spray bottle and just spritz the bottom of the board. It won't evaporate in the bottle, the application will be more consistent and no flux will get on the top of the board.
When I was a kid, my father had a company that manufactured boards. They used a wave soldering conveyer system. It was about 8 feet long, had a track that the populated board would run along, and it would enter a solder pool which had a small 'wave' along its width that would just skim the bottom of the board as it passed through. It was pretty fascinating. You really need one of these! It can completely solder a board in about 10 seconds. By the way, you also need to seriously consider your ventilation solution. What you are using currently is not sufficient. If you're going through that process hundreds of times, your entire room will be coated in burned flux and solder particles, as well as your lungs. You don't want to develop problems 20 years down the road. Future you will be thankful that you dealt with it properly now. Love what you're doing! Don't cut safety corners! Cheers!
It's not about the shape of the edges. All the areas you want soldered are in air pockets, preventing the solder from going up to the board. If you hold a cup upside down and push it into water, the water is not going into the cup. That's what's happening here. You got it working, because you're rolling it over the solder, so air can escape on one side. What you could do instead, is make holes or channels to the areas you want soldered, where the air can escape. Then you won't have to grind the corners and you won't have to roll the board over the solder.
AHHH cool thought! I didn't think there was a good enough seal for air to get trapped, but I think you're right. I'll retry with the stock pallet and use the roll technique and see if I get better results, thanks!
I'm reminded of injection molding and various kinds of casting. In these processes you have to include sprue channels in the designs to allow for material to flow in and out and for air to escape as material fills the main volume.
@@rockosgaminglogic Either that, or just put a couple of unpopulated through holes on the pcb which act as the vents. I’m not sure a 0.1mm increase would help much given manufacturing tolerances
I have the same dip soldering machine at the company I work at and also started out with the same issues... We discovered the temperature at the surface is lower than what the display says, probably losing heat from touching air and being irradiated as infrared, or just the temperature sensor is inaccurate. Eventually we set the solder bath to 420C (on the display) to get the process right to perfectly solder the though-hole components. Another way to solve your mounting hole issue is just use non-plated through holes and keep solder mask on the bottom side of the mounting holes which prevents wicking... the top side of the mounting hole can still be connected to your PCB Gnd which will still make contact with the screw head.
Hey mate, in manufacturing you gotta always make sure you preheat things or you’ll not only get the issues you saw here but also premature failures. You should also cool down slowly using the suggested profiles of the solder you use. This typically happens on a conveyor belt in wave soldering but you can do mostly the same in a controlled oven (for prototyping we used a heavily modified fan assisted toaster oven with an arduino to control the heat/pid and allowed us to customise curves
Back in the day we used to call this flow soldering. Strongly suggest you get yourself an extraction hood above this thing. The flux and vaporized lead is super super bad for you. Even if you wear a mask while you’re using the machine the stuff hangs around in the air.
I don't believe that any states allow the use of lead solder in manufacturing environments anymore, thus they typically use silver solder, which isn't nearly as good. However, you are correct about the fumes, the vaporized flux and other vapors are extremely bad for your lungs.
@@kennorman3586 in certain industries which have exceptions, yeah. Like aerospace. I doubt hobby grade pick n place qualifies. Even automotive doesn’t.
@@JasperJanssen almost all hobby soldering is with lead, it's a lot easier and it's not getting ROHS qualified for sale so everyone uses leaded solder for it's better characteristics
Adafruit uses a selective solder machine. It pumps the molten solder up a fountain then moves the nozzle around to cover only the areas needed. An alternative is to move the part around over a fixed nozzle.
As someone that works in the industry, we use selective solder or hand soldering many many years ago we had a wave solder machine. I've not seen people dip boards before. Not to mention dipping looks way too dangerous for a production environment.
As others have said - wave soldering. Or something i have seen from a friend: He used a conveyor-toaster for reflow-soldering (worked great, he later took it apart and made the conveyor dip the boards into a pool of molten solder)
Oooooh! Did they document it anywhere yet? That sounds really neat, and i saw a commercial conveyor oven like that the other day at an auction and *thought* that but didn’t have the need yet! Either way getting them in contact with Stephen may be good. (Edit: Typos)
the way you tilt the board in the dip, reminds me of wave soldering, your ensuring that there is a clear edge for the surface tension to peel-off along, so not to much sticks
Your interview with Thomas Sanladerer brought me here via the magic of the TH-cam algorithm. Your charisma is absolutely off the charts. You speak well and convey your information extremely effectively and with scintillating style. Very happy to have viewed your stuff so far. Very impressive team as well. You guys just ooze professionalism. Curious Marc style of intelligence.
We use an Ender 3D printer to move the board over the solder pot and dip it, move back away from the pot and pause so we can swap the board. Our g code script does runs of 12 boards then we re run it.
At work we use a selective soldering machine with kinda like a fountain of liquid tin. A panel with the PCBs put in moves over it so that it only touches the areas where there are pins to be soldered. Advantage is that there is a smaller amount of tin to heat up (so less energy consumption on that end). We also add nitrogen to reduce oxidation (extracting the nitrogen out of the air unfortunately does require some additional energy). In our previous facility we had a conveyor belt going through a puddle of liquid metal, but the results weren't as good as the selective soldering machine. Also, occasionally parts or boards would fall off of the conveyor belt into the puddle... The smell was horrendous 😂
Related to the screw holes, why not place the screws up top and have the bottoms line up with the bottom of the board and do this soldering job and then use a heat gun to remove the access solder so you can remove the screws afterwords after the circuit board is completely cooled? Experimentation is so much fun! Nice video you guys!
@@Dampfhammer96 yes, you can also find some existing open source designs on youtube, they are very similar to this one, but less accurate, a successful joint is made 60-80% of the time I think, I didn't realise ersa made such a small machine, I expect they will be at the german trade fair that stephen mentions in his video
As the other commenter said its trapping air. Treat it like mold casting give a route out for the air with no contained shapes. Also add some stoppers to the bottom of your tongs or some arms to stop at a depth you are plunging it into the flux without it being hand steadiness dependent.
The process is called wave soldering where the molten solder is pumped out of a fountain, making the solder appear as a wave. There's also a conveyor system to move the board over the puddle at the correct speed. This video is a good demo of the machine and the process: th-cam.com/video/q4pUKfRHd8Q/w-d-xo.html
Better for when you have a proper assembly line (since you can just have the boards all be on a conveyor that runs through component placement, soldering, etc.), but I don't think Opulo is making in _that_ kind of volume yet
So happy to see where it started and where this project has come to! I would love some videos about the process of making this project into a company and all the challanges that come with it, i think you mentioned a little at the end of this video. Maybe something super rough, just talking head style, on a second channel as im sure a lot of people here come for the pcb work aspect of this channel
Small suggestion here to improve the fume issue as the danger puddle is still just chugging out fumes like a smoker who just got off an intercontinental flight. Set up an enclosure in kind of a \_/ shape (wider at the front) around the solder pool with the filter fan on the back (may need a bigger fan) with the walls as tight in as possible. Either vent the air outside (shouldn't be hard to make a vent tube adapter for a window) or through a suitable filter, but having it enclosed on all sides but the front will allow the fan to pull all the fumes to the back. Probably wise to put the port for the vent tube on the box at the top, near the back. This also will help contain any splashing if anything goes wrong, as well as massively reduce the amount of flux and solder fumes in the building.
that method reminds me of wave soldering. with that method, you don't dip the pcb into the solder. you have small motor that creates waves in the solder. and only the maxima of the waves reach the board. works well on industrial scale.
Here's an idea. Instead of a template that only exposes the parts that need solder, how about a template that only protects the parts that need protecting. If that means a template composed of many parts isolated from each other, then attach them to a base that allows solder to flow through it, but also keeps at the isolated template pieces locked in place relative to each other. I think this would allow air to escape easily, and allow solder to reach everywhere it needs to.
I was think a low tech was to improve efficiency of the connector assembly process could be to make some sort of 3d printed sheath that had stacks of the connectors in the correct alignment to slip onto the pcb and then within the sheath thingy you'd need some mechanism to only let one layer of the connectors out. down side would be that if you changed something small on the pcb that affected alignment you have to change the design of the sheath thingy to account for that. think making a good release mechanism could also be finicky... also maybe a future lumenpnp could have a connector toolhead/ little gripper that could place larger through-hole parts/connectors
We use wondermask p to mask our boards. Also paper backed masking tape has worked as well. We use a full wave soldering machine though. I don't think dip soldering should really change much to make them incompatible but I don't have any experience with dip soldering boards.
Nice work! In commercial wave soldering, A pump is used to make a little molten ridge which the boards are passed over. (This gives air pockets a chance to escape, which I think was one of the problems with the chamfered stencil. Cutting vent channels in the stencil and/or drilling vent holes in the board might solve that.) They do wave-soldering with surface-mount parts too, but the pick-and-place machine actually puts little glue dots under each part, to hold them to the PCB during soldering, since they pass through the solder wave upside-down.
I used to work at a place that built controller PC boards for cable boxes, and since this was before SMD came into the fore everything was through-hole. The actual soldering was handled by a belt-fed wave-soldering machine roughly the size of a refrigerator on its back. It had an adjustable width feeder that accepted an assembled PCB (or in this case a palletized PCB with three modules on it), passed it over a foam roller in a trough that applied flux, then over a heater to preheat the board and parts, and then through a "wave" - basically a roughly quarter-inch hump - of molten solder (thus the name) being pumped over an angled barrier to do the actual soldering. The end result was a near-perfect solder job every time. Populated but unsoldered boards went in one end and came out the other end ready to be bench-tested/QCed/pre-tuned (there was a trimpot that needed to be set to "dial out" parts tolerances in one of the critical timing circuits), and then packed for shipping.
If I remember right from when I worked on the board line at NCR many years ago we used Kapton dots to cover the holes and they were applied by the PnP before SMT reflow. Then the boards were stuffed by hand (with only a very few TH parts it wasn't economical to have a TH PnP machine) and then through the wave solder machine. We also had a desktop "mini-wave" machine that was used to replace large items like PCI edge card connectors. You could use the wave to heat up and remove the entire connector in one shot. Then you'd clean the holes out and replace, flux and mini-wave again to re-solder it.
in the industry they used a solder machine and they still have rework to do at the end of that process depending on the boards oh and they use a wave soldering technique so the solder cascades over a crest to touch the board also the boards are on a traveling track
The spatulas look like a great $4 solution - but if you want to go for consistency in dipping the plane of your boards a specified amount in -Z, I'm thinking about modifying something like a "Drill press". The poor-man's drill press is this $50 dremel stand (do not recommend, crappy plastic). But perhaps something like this could be modified to install a board and then use the lever to dip it down a certain amount in Z to be consistent for dipping in both flux and solder pools.
if you were to be able to make air channels in your pallet it might help with getting solder up into the cavities when upside down. (think of placing a glass upside down into water).
When i used to work a solder flow machine and we used a product called seal peel .you put it on the bits you don't want soldered then it just peels off easy .
Wondering about the method used in removing the board is also a key to the solder adhering correctly. A swooping/rolling method rather than straight dipping and then straight removing.
Very cool! Two suggestions… not that I’ve ever done this, but watching your process they came to mind. 1. Use a hot plate or a jig with blown hot air to bring the boards partway up to temperature. You want it lower heat than to melt the solder so no chance to burn stuff. 2. A jig where the board sits over a small contained sprayer chamber (like from a windex bottle or maybe electric pump or piezo) with your alcohol flux might work easier than the dip.
I'm no specialist XD. my guess is that the thermal capacity of the metal is low (in general), so it will dissipate the heat it has really fast into anything that comes in contact with it. In other words, your assumption that the temperature has to be higher is not wrong! However, I believe making it a bit hotter (not a lot) and leave it less time in there will make the "pads" and "component' dissipate the heat of the liquid metal real fast (making it solid) and the other parts on the boards won't have time to solidify the liquid metal and the liquid will just leave back into the pod. Nice video and demonstration :)
The grinding of the PCB has me worried. From what I've heard, powder generated from PCBs is really not good. For tasks like that, I hope you have a bubble set up with a HEPA filter. This keeps the dust out of the main environment. It's super easy to build, I made one for a project I was working on in my bedroom which involved cutting insulation a lot. All it was was a bubble made from heavy plastic sheeting, all sealed with super stick tape. I then attached a box with a fan built into it and attached a car cabin HEPA filter (cheaper than many special purpose ones) to the box so all air flow had to go through it. This created a negative pressure space in the bubble, so any leaks would mean outside air would blow in and particulates couldn't escape. Also, don't forget about your clothes. After you are done and take off your respirator, the particulates can be on your clothes. It would be a real benefit to have a mid chamber to change in. Obviously, in your setup it would be easier to use something like tyvek suits or something that goes over clothes. Still, don't forget about your hair. I had mine in my bedroom, so I just had pair of clothes dedicated for that type of work that I would change in an out of for the tasks at hand. This sort of stuff won't have a noticable effect on health usually for a long time, but it will. My bubble cost me around $100. It's cheap and easy, just takes time to set up. I highly recommend one if you are dealing with particulates. Also, for your project, you could probably easily get away with a bubble in a box. That sort of thing where you stay outside and just put your arms in through attached gloves.
You will get less dross if you cover the solder pot surface with some peanut oil. But conveyorized wave solder with preheat is the professional way to go.
I'm pretty sure manufacturers are not allowed to use leaded solder anymore. Also... lead doesn't vapourize at those temperatures so you wouldn't be able to breath it in but the flux vapours aren't good either.
I thought you had to heat up the board with components before dipping. You need to look into a wave solder machine. Board edges ride on a conveyor belt over a wave of solder. Board is heated before the wave and then cooled after as it rides the conveyor.
A bit unrelated, but it was neat seeing you get interviews at ERRF and whatnot. Hopefully I’ll get there in person one day myself! I hope it wasn’t too exhausting to do that all day, but i think that will really draw in a bunch of people!
Congratulations on getting to this step, net is research for " Wave Soldering " which you almost were doing by hand. You just need a bar to ' tap ' enough lead to create a wave to wick up above the mask at the right temperature.
Get yourself a small space heater with those infrared/ceramic elements and mount it flat (you may have to disable a tilt cutout switch) to stage the board over to preheat it before dipping, this can increase your workflow as you can have multiple boards preheating and then dip them in succession. If you're interested in "hacking" one together, checkout what people are doing for DIY vacuum forming, its a similar "heatbox" idea, just different implementation.
I've never done anything like this before at all, but a suggestion might be to put a few unplated drill holes in the pcb that you're soldering for venting purposes.
I waaaaant to automate this. I want to build basically a hanging kinetic platform, the basic structure is a pillar mounted near the edge of the basin which has a rocker arm horizontally across the top, and further outboard this rocker arm is actuated by a servo, then inboard a linkage hangs down. 3 of these hold up a platform which receives this pallet or one like it. Push a button and the servos go through the routine of lowering the board near the solder to preheat, then doing that rolling dip maneuver (or whatever maneuver proves best). Afterwards it lifts the pallet back up so the operator can grab it. Advantage: probably almost none, would allow for more consistency, would be cool, definitely more complexity, somewhat of a distraction, will probably need some design tuning to deal with being next to this hot thing all the time. A real advantage would be seen if it could be built into an automated workflow where panels are palletized, dipped, and depalletized so the operator can focus on populating boards without switching tasks, but that's definitely downtown distractiontown
@@fischX This is simply not true. In a wave solder machine you have a reservoir of solder and a submerged pump that propels solder out of a linear nozzle across the reservoir to create a region where the solder is higher as it spills back down into the reservoir. This raised section is "the wave". You then conveyor PCBs across the nozzle and only the wave is high enough to reach the board but also the constant flow of solder is a constant flow of heat that prevents shock cooling like Stephen found was happening. Wave soldering is great, I assume Stephen has done or has had done the research and decided not to go that route, presumably because of cost.
One thing that I wonder, when looking at the design of the board, is why are there so many thru-hole components in the first place? You could have used SMD connectors in many places, and I'm still not sure why all the motor drivers are on little breakout PCBs instead of being integrated directly.
yeah those headers could all be SMD, a bit more expensive but the process would be cheaper. The sockets for motor drivers is presumably because this looks like a glorified 3d printer mainboard, which have used those "Polulu" style socketed stepper drivers for at least a decade. Easier to let somebody else design a good stepper driver board when starting up. Thermals are better when on a single PCB but plenty of time to do that better.
Group B racing was incredible. The main reason it was so dangerous was the difficulty of convincing people in South American countries that waiting in the middle of the road for the cars go cars to go by wasn't a good idea. My favorite car was the Lancia Stratos. A friend of mine had one of the limited production models...
I was thinking silicon studs for fans, to push into the holes on the pcb. Or some other silicone that fits the holes. then you could just push it down onto the 4 silicon things and dipp in solder.
One of my friend tried this and experienced 2 fatal mistakes: 1. Dunk the board but can't pull it up because he had no idea where's the board, oh wait there's $400 ARM SoC in it 🗿 2. Dunk and pull success! then the ARM SoC become loose and fall into the puddle 💀 Total loss is around $1100. Lesson to learn: PCB oven is a lot safer than tin puddle!
Hello! I have no insights on the process, other than maybe ask adafruit some pointers? they do a lot of in house manufacturing and are pretty friendly, perhaps one of their engineers can tell you to look at reference PDF X that will help... I mostly came to say that you shouldn't worry about the videos schedule, as those kind of videos is insanely interesting, and comes only from doing what you do. I think many of us are really interested in this aspect of the manufacturing chain (i mean problem solving, engineering, etc), and if it takes less frequent/less regular videos, so be it! When i equip my fablab, i'll definitely search for funds for an Opulo!
Just curious why you didn't get a wave soldering system instead of the puddle? Way more easy to control, not much larger, and will give you the perfect solder joint every time.
Good question, we thought about it! The whole point of the company and LumenPnP project is to find ways of making electronics accessibly at a mid-scale. While we could buy the more fancy, expensive machine, we want to start with the thing a lot of our customers might be using first. $300 for the solder pot and some flux and solder bars is a much easier pill to swallow than a few grand. We want to find solutions to these problems for everyone! That being said, I wouldn't be surprised if a wave soldering machine is in our future...
Obvious one, but don't use plated through holes for mounting; just a hole through the fibreglass. Not only will that remove your solder problem, but in general you can't rely on a good electrical connection from the PTH to the chassis there over time unless you are very carefully controlling the surface chemistry to prevent oxidation. Also the act of tightening down the screw/bolt can delaminate the copper or introduce non-conducting hairline fractures.
Just a thought here, what about taking one of your pick and place machines, changing the head to a soldering iron and adding a solder reel, and converting the machine to a solder machine that autofeeds the tin and visually checks each connection. Slower then the molten bath method but less Rework needed and if done right can use the same mounting system /trays for the pick and place
I haven't seen oil used since Electrovert acquired Hollis - over 20 years ago. Hollis had a wave solder machine with oil injection to reduce bridging - a pretty messy machine as I recall.
Hi Stephen, your assumption about the pre-heating is right. If you look at a wave soldering process the PCB is much longer in the oven (for pre-heating) compared to the time it‘s going though the wave itself. It also guarantees that the solder goes deep enough into the through-holes. If you have to solder massiv power components that need big ground planes then it get‘s even more difficult.
Surface mount connectors have a big mechanical disadvantage over thru hole connectors. If you have a static board and cable strain relief in place to minimize mechanical forces on the connectors, SMD would be fine. The current Lumen does not have any of that, and the issue that the motherboard is mounted directly on the machine that constantly vibrates and shakes because of the mechanical movement. It would require a complete machine redesign having the controller somewhere away from the machine frame or at least mechanically decoupled.
Perhaps air pockets are preventing the solder from reaching the solder points. Have you tried carving channels in your templates for air to escape when you dip it into the solder bath?
Did you think about doing the mechanical holes not plated/npth? When you could avoid your just dip the whole board without the pallet. So instead of working with the pallet you avoid the problem completely. Also, if you get some pins not soldered I would be kinda skeptical if the solder properly flows through the plated holes or just creates a fillet at the bottom. As these mechanical connectors will be used by the end user there will be some stress at the pins and you obviously don't want them to fail.
What if, for preheating, you had some sort of hotplate to place the boards on? I have a hotplate, The Charlatan, that just gets hot enough to seal cellophane and activate dry mounting tissue. I feel like that would be absolutely perfect for this, it would streamline production quite a bit more.
Shortly after I started at my job, everyone (this was a very large company) received a "safety" mug for drinks. Per rumors, a tech spilled a drink into a solder pot. "Danger puddle" indeed!
You guys would have seen it by now but have a look at core electronics selective soldering machine. There’s also lots of cheap used selective soldering and wave soldering machines in the USA even on eBay as they’re not super commonly used here anymore.
9:10 you can use a 3d printer or simular device it will be not fast but acceptable automatization, TIP create rotary hed for lifting different components.
IMHO, fluxing and good (pre-heat, dip, cooling) profile is key for reliable and good (wave) solder bumps. See also (examples): a) "Wave Solder Profile Supplement" notes from aimsolder b) "Wave Soldering and Thermal Profiles" chapter from Microelectronics Cloud Alliance (MECA) ...
Modify a cheap 3d printer kit to make a solder dipping machine made one for my last company and was successful but eventually they outsourcing the job. For the bigger board they had a wavesolder machine which was cool.
I bought one of the first "print your own parts kit". I'm still waiting for further information on the promised upgrade kit to convert the machine into a V3. Any news? Thanks
If you think this is that amazing, you should look into wave soldering. Also, use capton circles for your areas you don't want soldered instead of tape. The little circles are easier.
![[TH]FFWS SEA Spring 2024 Point Rush - Day 2](http://i.ytimg.com/vi/uTzfNlXO9bg/mqdefault.jpg)
Every LumenPnP we make is danger puddle approved: opulo.io/products/lumenpnp
What is needed to turn a V1 kit into a V3/V4 kit?
Don't dip the boards in the flux. Put the flux in a spray bottle and just spritz the bottom of the board. It won't evaporate in the bottle, the application will be more consistent and no flux will get on the top of the board.
@@MartinBogomolni Good question, I was asking that question multiple times. It was promised in the V3 presentation video!
Yep, first comment I saw said air pockets. Glad we're on the same page
When I was a kid, my father had a company that manufactured boards. They used a wave soldering conveyer system. It was about 8 feet long, had a track that the populated board would run along, and it would enter a solder pool which had a small 'wave' along its width that would just skim the bottom of the board as it passed through. It was pretty fascinating. You really need one of these! It can completely solder a board in about 10 seconds. By the way, you also need to seriously consider your ventilation solution. What you are using currently is not sufficient. If you're going through that process hundreds of times, your entire room will be coated in burned flux and solder particles, as well as your lungs. You don't want to develop problems 20 years down the road. Future you will be thankful that you dealt with it properly now. Love what you're doing! Don't cut safety corners! Cheers!
It's not about the shape of the edges. All the areas you want soldered are in air pockets, preventing the solder from going up to the board. If you hold a cup upside down and push it into water, the water is not going into the cup. That's what's happening here. You got it working, because you're rolling it over the solder, so air can escape on one side. What you could do instead, is make holes or channels to the areas you want soldered, where the air can escape. Then you won't have to grind the corners and you won't have to roll the board over the solder.
AHHH cool thought! I didn't think there was a good enough seal for air to get trapped, but I think you're right. I'll retry with the stock pallet and use the roll technique and see if I get better results, thanks!
@@stephen_hawes The flux IS the seal.
redrok
@@stephen_hawes just enlarge the throughholes by 0.1mm and they should let the air escape.
I'm reminded of injection molding and various kinds of casting. In these processes you have to include sprue channels in the designs to allow for material to flow in and out and for air to escape as material fills the main volume.
@@rockosgaminglogic Either that, or just put a couple of unpopulated through holes on the pcb which act as the vents. I’m not sure a 0.1mm increase would help much given manufacturing tolerances
I have the same dip soldering machine at the company I work at and also started out with the same issues... We discovered the temperature at the surface is lower than what the display says, probably losing heat from touching air and being irradiated as infrared, or just the temperature sensor is inaccurate. Eventually we set the solder bath to 420C (on the display) to get the process right to perfectly solder the though-hole components. Another way to solve your mounting hole issue is just use non-plated through holes and keep solder mask on the bottom side of the mounting holes which prevents wicking... the top side of the mounting hole can still be connected to your PCB Gnd which will still make contact with the screw head.
Hey mate, in manufacturing you gotta always make sure you preheat things or you’ll not only get the issues you saw here but also premature failures. You should also cool down slowly using the suggested profiles of the solder you use. This typically happens on a conveyor belt in wave soldering but you can do mostly the same in a controlled oven (for prototyping we used a heavily modified fan assisted toaster oven with an arduino to control the heat/pid and allowed us to customise curves
Exactly. Preheat all boards first before even attempting to dip them. 100C should do, while still being low enough to not melt anything.
Back in the day we used to call this flow soldering. Strongly suggest you get yourself an extraction hood above this thing. The flux and vaporized lead is super super bad for you. Even if you wear a mask while you’re using the machine the stuff hangs around in the air.
I don't believe that any states allow the use of lead solder in manufacturing environments anymore, thus they typically use silver solder, which isn't nearly as good. However, you are correct about the fumes, the vaporized flux and other vapors are extremely bad for your lungs.
To add on to this, if you’re not venting outside, I’d get a combination activated carbon + HEPA filter.
@@cnc-maker Lead solder is still used by some manufactures.
@@kennorman3586 in certain industries which have exceptions, yeah. Like aerospace. I doubt hobby grade pick n place qualifies. Even automotive doesn’t.
@@JasperJanssen almost all hobby soldering is with lead, it's a lot easier and it's not getting ROHS qualified for sale so everyone uses leaded solder for it's better characteristics
Adafruit uses a selective solder machine. It pumps the molten solder up a fountain then moves the nozzle around to cover only the areas needed. An alternative is to move the part around over a fixed nozzle.
If only he had access to some robotics that could move a board around...
As someone that works in the industry, we use selective solder or hand soldering many many years ago we had a wave solder machine. I've not seen people dip boards before. Not to mention dipping looks way too dangerous for a production environment.
@@UTRG-UnderTheRain agreed ... wave solder is The Way(tm)
I'm pretty sure it even has a nitrogen atmosphere to reduce dross, but the dross cleaning always looks Sisyphean...
Rule number 1 of soldering: more heat
Rule number 2: less solder
Rule number 3: flux is magic
Looks like you realized how important rule number 1 is!
That thing. It scares me. Especially seeing your energy around it and lack of glasses.
As others have said - wave soldering.
Or something i have seen from a friend: He used a conveyor-toaster for reflow-soldering (worked great, he later took it apart and made the conveyor dip the boards into a pool of molten solder)
Oooooh! Did they document it anywhere yet?
That sounds really neat, and i saw a commercial conveyor oven like that the other day at an auction and *thought* that but didn’t have the need yet!
Either way getting them in contact with Stephen may be good.
(Edit: Typos)
the way you tilt the board in the dip, reminds me of wave soldering, your ensuring that there is a clear edge for the surface tension to peel-off along, so not to much sticks
Your interview with Thomas Sanladerer brought me here via the magic of the TH-cam algorithm. Your charisma is absolutely off the charts. You speak well and convey your information extremely effectively and with scintillating style. Very happy to have viewed your stuff so far. Very impressive team as well. You guys just ooze professionalism. Curious Marc style of intelligence.
We use an Ender 3D printer to move the board over the solder pot and dip it, move back away from the pot and pause so we can swap the board. Our g code script does runs of 12 boards then we re run it.
At work we use a selective soldering machine with kinda like a fountain of liquid tin.
A panel with the PCBs put in moves over it so that it only touches the areas where there are pins to be soldered.
Advantage is that there is a smaller amount of tin to heat up (so less energy consumption on that end).
We also add nitrogen to reduce oxidation (extracting the nitrogen out of the air unfortunately does require some additional energy).
In our previous facility we had a conveyor belt going through a puddle of liquid metal, but the results weren't as good as the selective soldering machine.
Also, occasionally parts or boards would fall off of the conveyor belt into the puddle... The smell was horrendous 😂
Related to the screw holes, why not place the screws up top and have the bottoms line up with the bottom of the board and do this soldering job and then use a heat gun to remove the access solder so you can remove the screws afterwords after the circuit board is completely cooled? Experimentation is so much fun! Nice video you guys!
the advantage of adapting a 3d printer design to make a selective soldering robot is that you'd have another complementary product to sell
Something like th-cam.com/video/8BRFVnV-hus/w-d-xo.html
@@Dampfhammer96 yes, you can also find some existing open source designs on youtube, they are very similar to this one, but less accurate, a successful joint is made 60-80% of the time I think, I didn't realise ersa made such a small machine, I expect they will be at the german trade fair that stephen mentions in his video
As the other commenter said its trapping air. Treat it like mold casting give a route out for the air with no contained shapes. Also add some stoppers to the bottom of your tongs or some arms to stop at a depth you are plunging it into the flux without it being hand steadiness dependent.
Preheat makes so much sense! It's pretty much make or break for every soldering joint Ive ever made!
Wave soldering is much less hassle, maybe there are home grow way to do it with sound or a spinning wheel to get a standing wave in a danger puddle
The process is called wave soldering where the molten solder is pumped out of a fountain, making the solder appear as a wave. There's also a conveyor system to move the board over the puddle at the correct speed. This video is a good demo of the machine and the process: th-cam.com/video/q4pUKfRHd8Q/w-d-xo.html
That is a really informative video. Lumen should definitely use this technique rather than dangerous looking manual dipping.
They don't have a wave pool, its just a pool.
@@rockosgaminglogic Would need tiny pcb lifeguards then for something as dangerous as a wave pool!
@@mvadu Wave soldering is way more expensive and involved than dipping. The uploader obviously knows about wave soldering.
Better for when you have a proper assembly line (since you can just have the boards all be on a conveyor that runs through component placement, soldering, etc.), but I don't think Opulo is making in _that_ kind of volume yet
So happy to see where it started and where this project has come to!
I would love some videos about the process of making this project into a company and all the challanges that come with it, i think you mentioned a little at the end of this video.
Maybe something super rough, just talking head style, on a second channel as im sure a lot of people here come for the pcb work aspect of this channel
Small suggestion here to improve the fume issue as the danger puddle is still just chugging out fumes like a smoker who just got off an intercontinental flight. Set up an enclosure in kind of a \_/ shape (wider at the front) around the solder pool with the filter fan on the back (may need a bigger fan) with the walls as tight in as possible. Either vent the air outside (shouldn't be hard to make a vent tube adapter for a window) or through a suitable filter, but having it enclosed on all sides but the front will allow the fan to pull all the fumes to the back. Probably wise to put the port for the vent tube on the box at the top, near the back. This also will help contain any splashing if anything goes wrong, as well as massively reduce the amount of flux and solder fumes in the building.
that method reminds me of wave soldering. with that method, you don't dip the pcb into the solder. you have small motor that creates waves in the solder. and only the maxima of the waves reach the board.
works well on industrial scale.
Here's an idea. Instead of a template that only exposes the parts that need solder, how about a template that only protects the parts that need protecting. If that means a template composed of many parts isolated from each other, then attach them to a base that allows solder to flow through it, but also keeps at the isolated template pieces locked in place relative to each other. I think this would allow air to escape easily, and allow solder to reach everywhere it needs to.
Next up is automating the solder dip! Better consistency! But alot more work.. Nice to see how you are moving forward!!
At that point, just use wave soldier. It's been an industry standard for years
I was think a low tech was to improve efficiency of the connector assembly process could be to make some sort of 3d printed sheath that had stacks of the connectors in the correct alignment to slip onto the pcb and then within the sheath thingy you'd need some mechanism to only let one layer of the connectors out. down side would be that if you changed something small on the pcb that affected alignment you have to change the design of the sheath thingy to account for that. think making a good release mechanism could also be finicky... also maybe a future lumenpnp could have a connector toolhead/ little gripper that could place larger through-hole parts/connectors
We use wondermask p to mask our boards. Also paper backed masking tape has worked as well. We use a full wave soldering machine though. I don't think dip soldering should really change much to make them incompatible but I don't have any experience with dip soldering boards.
Nice work! In commercial wave soldering, A pump is used to make a little molten ridge which the boards are passed over. (This gives air pockets a chance to escape, which I think was one of the problems with the chamfered stencil. Cutting vent channels in the stencil and/or drilling vent holes in the board might solve that.)
They do wave-soldering with surface-mount parts too, but the pick-and-place machine actually puts little glue dots under each part, to hold them to the PCB during soldering, since they pass through the solder wave upside-down.
I used to work at a place that built controller PC boards for cable boxes, and since this was before SMD came into the fore everything was through-hole. The actual soldering was handled by a belt-fed wave-soldering machine roughly the size of a refrigerator on its back. It had an adjustable width feeder that accepted an assembled PCB (or in this case a palletized PCB with three modules on it), passed it over a foam roller in a trough that applied flux, then over a heater to preheat the board and parts, and then through a "wave" - basically a roughly quarter-inch hump - of molten solder (thus the name) being pumped over an angled barrier to do the actual soldering. The end result was a near-perfect solder job every time. Populated but unsoldered boards went in one end and came out the other end ready to be bench-tested/QCed/pre-tuned (there was a trimpot that needed to be set to "dial out" parts tolerances in one of the critical timing circuits), and then packed for shipping.
If I remember right from when I worked on the board line at NCR many years ago we used Kapton dots to cover the holes and they were applied by the PnP before SMT reflow. Then the boards were stuffed by hand (with only a very few TH parts it wasn't economical to have a TH PnP machine) and then through the wave solder machine. We also had a desktop "mini-wave" machine that was used to replace large items like PCI edge card connectors. You could use the wave to heat up and remove the entire connector in one shot. Then you'd clean the holes out and replace, flux and mini-wave again to re-solder it.
in the industry they used a solder machine and they still have rework to do at the end of that process depending on the boards oh and they use a wave soldering technique so the solder cascades over a crest to touch the board also the boards are on a traveling track
The spatulas look like a great $4 solution - but if you want to go for consistency in dipping the plane of your boards a specified amount in -Z, I'm thinking about modifying something like a "Drill press". The poor-man's drill press is this $50 dremel stand (do not recommend, crappy plastic). But perhaps something like this could be modified to install a board and then use the lever to dip it down a certain amount in Z to be consistent for dipping in both flux and solder pools.
Super fun! You should add locking pins to the next round of boards so the through hole parts stay in place
Really love you work, enthusiasm and energy! Keep up that great and inspiring work Stephen. And perhaps till Electronica in Munich 👌🏼
Stephen: *shows a crazy contraption thing* "this is for another video"
The 10th Doctor: "it's a machine that goes... Ding..." *Ding*
if you were to be able to make air channels in your pallet it might help with getting solder up into the cavities when upside down. (think of placing a glass upside down into water).
When i used to work a solder flow machine and we used a product called seal peel .you put it on the bits you don't want soldered then it just peels off easy .
Really fun to watch!
Wondering about the method used in removing the board is also a key to the solder adhering correctly. A swooping/rolling method rather than straight dipping and then straight removing.
Happy that you figured it out, but come ON!
SMT parts everywhere!
Very cool! Two suggestions… not that I’ve ever done this, but watching your process they came to mind. 1. Use a hot plate or a jig with blown hot air to bring the boards partway up to temperature. You want it lower heat than to melt the solder so no chance to burn stuff. 2. A jig where the board sits over a small contained sprayer chamber (like from a windex bottle or maybe electric pump or piezo) with your alcohol flux might work easier than the dip.
I'm no specialist XD.
my guess is that the thermal capacity of the metal is low (in general), so it will dissipate the heat it has really fast into anything that comes in contact with it. In other words, your assumption that the temperature has to be higher is not wrong! However, I believe making it a bit hotter (not a lot) and leave it less time in there will make the "pads" and "component' dissipate the heat of the liquid metal real fast (making it solid) and the other parts on the boards won't have time to solidify the liquid metal and the liquid will just leave back into the pod.
Nice video and demonstration :)
The grinding of the PCB has me worried. From what I've heard, powder generated from PCBs is really not good. For tasks like that, I hope you have a bubble set up with a HEPA filter. This keeps the dust out of the main environment. It's super easy to build, I made one for a project I was working on in my bedroom which involved cutting insulation a lot. All it was was a bubble made from heavy plastic sheeting, all sealed with super stick tape. I then attached a box with a fan built into it and attached a car cabin HEPA filter (cheaper than many special purpose ones) to the box so all air flow had to go through it. This created a negative pressure space in the bubble, so any leaks would mean outside air would blow in and particulates couldn't escape.
Also, don't forget about your clothes. After you are done and take off your respirator, the particulates can be on your clothes. It would be a real benefit to have a mid chamber to change in. Obviously, in your setup it would be easier to use something like tyvek suits or something that goes over clothes. Still, don't forget about your hair.
I had mine in my bedroom, so I just had pair of clothes dedicated for that type of work that I would change in an out of for the tasks at hand.
This sort of stuff won't have a noticable effect on health usually for a long time, but it will. My bubble cost me around $100. It's cheap and easy, just takes time to set up. I highly recommend one if you are dealing with particulates. Also, for your project, you could probably easily get away with a bubble in a box. That sort of thing where you stay outside and just put your arms in through attached gloves.
Hope your new PCB kits will be available soon)
You will get less dross if you cover the solder pot surface with some peanut oil. But conveyorized wave solder with preheat is the professional way to go.
A solder fountain machine would be perfect if a bit expensive and you could automate it with one of your pic and place motion systems
Silicone plugs is also a solution to protect your mounting holes, given you have the volume to warrant fabricating a ton.
Amazing work, I hope you get a good fume extractor for those fumes , its a leaded solder so enhaling those fumes is no bueno.
Good luck man!
How do you know that he is using leaded solder?
I'm pretty sure manufacturers are not allowed to use leaded solder anymore.
Also... lead doesn't vapourize at those temperatures so you wouldn't be able to breath it in but the flux vapours aren't good either.
@@Jehty21 what i know is all solder wicks and pastes contain lead.
@@SAILxKamal well, your knowledge is wrong.
I thought you had to heat up the board with components before dipping.
You need to look into a wave solder machine. Board edges ride on a conveyor belt over a wave of solder. Board is heated before the wave and then cooled after as it rides the conveyor.
Came here to suggest he look at wave soldering. Very similar process, and there are a lot of techniques he could learn just by watching it work
Pretty sure those machines are an extra zero or two more expensive though, as this is just a literal heated vat.
@@zerumsum1640 yeah for sure, but I think he could make his own with only a couple more components
@@JB-fh1bb it's plywood, some ducting, and a little hardware. shouldn't be hard.
It seems like you need vents so that the air doesn't get trapped under the board and prevent solder from filling in.
A bit unrelated, but it was neat seeing you get interviews at ERRF and whatnot. Hopefully I’ll get there in person one day myself! I hope it wasn’t too exhausting to do that all day, but i think that will really draw in a bunch of people!
What a legend. Taking away power from the big manufacturers and giving it to the people!
Congratulations on getting to this step, net is research for " Wave Soldering " which you almost were doing by hand. You just need a bar to ' tap ' enough lead to create a wave to wick up above the mask at the right temperature.
Get yourself a small space heater with those infrared/ceramic elements and mount it flat (you may have to disable a tilt cutout switch) to stage the board over to preheat it before dipping, this can increase your workflow as you can have multiple boards preheating and then dip them in succession.
If you're interested in "hacking" one together, checkout what people are doing for DIY vacuum forming, its a similar "heatbox" idea, just different implementation.
I've never done anything like this before at all, but a suggestion might be to put a few unplated drill holes in the pcb that you're soldering for venting purposes.
I waaaaant to automate this. I want to build basically a hanging kinetic platform, the basic structure is a pillar mounted near the edge of the basin which has a rocker arm horizontally across the top, and further outboard this rocker arm is actuated by a servo, then inboard a linkage hangs down. 3 of these hold up a platform which receives this pallet or one like it. Push a button and the servos go through the routine of lowering the board near the solder to preheat, then doing that rolling dip maneuver (or whatever maneuver proves best). Afterwards it lifts the pallet back up so the operator can grab it. Advantage: probably almost none, would allow for more consistency, would be cool, definitely more complexity, somewhat of a distraction, will probably need some design tuning to deal with being next to this hot thing all the time. A real advantage would be seen if it could be built into an automated workflow where panels are palletized, dipped, and depalletized so the operator can focus on populating boards without switching tasks, but that's definitely downtown distractiontown
That's basically what a wave soldering machine does. Problem is those are big expensive devices
@@fischX This is simply not true. In a wave solder machine you have a reservoir of solder and a submerged pump that propels solder out of a linear nozzle across the reservoir to create a region where the solder is higher as it spills back down into the reservoir. This raised section is "the wave". You then conveyor PCBs across the nozzle and only the wave is high enough to reach the board but also the constant flow of solder is a constant flow of heat that prevents shock cooling like Stephen found was happening. Wave soldering is great, I assume Stephen has done or has had done the research and decided not to go that route, presumably because of cost.
One thing that I wonder, when looking at the design of the board, is why are there so many thru-hole components in the first place? You could have used SMD connectors in many places, and I'm still not sure why all the motor drivers are on little breakout PCBs instead of being integrated directly.
yeah those headers could all be SMD, a bit more expensive but the process would be cheaper. The sockets for motor drivers is presumably because this looks like a glorified 3d printer mainboard, which have used those "Polulu" style socketed stepper drivers for at least a decade. Easier to let somebody else design a good stepper driver board when starting up. Thermals are better when on a single PCB but plenty of time to do that better.
oh, and get an actual ir pre-heater for pcbs should help with consistent temp pre-dippping
Group B racing was incredible. The main reason it was so dangerous was the difficulty of convincing people in South American countries that waiting in the middle of the road for the cars go cars to go by wasn't a good idea. My favorite car was the Lancia Stratos. A friend of mine had one of the limited production models...
I was thinking silicon studs for fans, to push into the holes on the pcb. Or some other silicone that fits the holes. then you could just push it down onto the 4 silicon things and dipp in solder.
i just love your energy when it comes to solving very specific problems most people won't understand.
this is just refereshing to watch :)
keep it up
10:15 aka the " French fry machine"
One of my friend tried this and experienced 2 fatal mistakes:
1. Dunk the board but can't pull it up because he had no idea where's the board, oh wait there's $400 ARM SoC in it 🗿
2. Dunk and pull success! then the ARM SoC become loose and fall into the puddle 💀
Total loss is around $1100. Lesson to learn: PCB oven is a lot safer than tin puddle!
Typo fixed, sorry. I'm typing this at 3am
Hello!
I have no insights on the process, other than maybe ask adafruit some pointers? they do a lot of in house manufacturing and are pretty friendly, perhaps one of their engineers can tell you to look at reference PDF X that will help...
I mostly came to say that you shouldn't worry about the videos schedule, as those kind of videos is insanely interesting, and comes only from doing what you do.
I think many of us are really interested in this aspect of the manufacturing chain (i mean problem solving, engineering, etc), and if it takes less frequent/less regular videos,
so be it!
When i equip my fablab, i'll definitely search for funds for an Opulo!
Just curious why you didn't get a wave soldering system instead of the puddle? Way more easy to control, not much larger, and will give you the perfect solder joint every time.
Good question, we thought about it! The whole point of the company and LumenPnP project is to find ways of making electronics accessibly at a mid-scale. While we could buy the more fancy, expensive machine, we want to start with the thing a lot of our customers might be using first. $300 for the solder pot and some flux and solder bars is a much easier pill to swallow than a few grand. We want to find solutions to these problems for everyone! That being said, I wouldn't be surprised if a wave soldering machine is in our future...
Obvious one, but don't use plated through holes for mounting; just a hole through the fibreglass. Not only will that remove your solder problem, but in general you can't rely on a good electrical connection from the PTH to the chassis there over time unless you are very carefully controlling the surface chemistry to prevent oxidation. Also the act of tightening down the screw/bolt can delaminate the copper or introduce non-conducting hairline fractures.
Just a thought here, what about taking one of your pick and place machines, changing the head to a soldering iron and adding a solder reel, and converting the machine to a solder machine that autofeeds the tin and visually checks each connection. Slower then the molten bath method but less Rework needed and if done right can use the same mounting system /trays for the pick and place
Stephen I LOVE the kitchen tongs!😂
Next step, get a small wave-soldering machine. Makes it soooooo much easier.
Super sick. i real like to see you build an automated dipper
While you are at electronica you can find a cheap wave soldering machine. Also in industry we use oil in molten solder.
I haven't seen oil used since Electrovert acquired Hollis - over 20 years ago. Hollis had a wave solder machine with oil injection to reduce bridging - a pretty messy machine as I recall.
I saw the change over from oil in the early 1990's.
Hi Stephen, your assumption about the pre-heating is right. If you look at a wave soldering process the PCB is much longer in the oven (for pre-heating) compared to the time it‘s going though the wave itself. It also guarantees that the solder goes deep enough into the through-holes. If you have to solder massiv power components that need big ground planes then it get‘s even more difficult.
If you like to remove the flux from the top, just a toothbrush and alcohol, then blow away the excess with heatgun.
If you don't wanna worry about flux an ultrasonic bath in branson ec cleaner will make it look brand new
Nice job!
I would imagine that surface mount editions of all those connectors exist, and would therefore be less work than old tech danger puddles
Surface mount connectors have a big mechanical disadvantage over thru hole connectors. If you have a static board and cable strain relief in place to minimize mechanical forces on the connectors, SMD would be fine. The current Lumen does not have any of that, and the issue that the motherboard is mounted directly on the machine that constantly vibrates and shakes because of the mechanical movement. It would require a complete machine redesign having the controller somewhere away from the machine frame or at least mechanically decoupled.
Perhaps air pockets are preventing the solder from reaching the solder points. Have you tried carving channels in your templates for air to escape when you dip it into the solder bath?
Did you think about doing the mechanical holes not plated/npth? When you could avoid your just dip the whole board without the pallet.
So instead of working with the pallet you avoid the problem completely.
Also, if you get some pins not soldered I would be kinda skeptical if the solder properly flows through the plated holes or just creates a fillet at the bottom.
As these mechanical connectors will be used by the end user there will be some stress at the pins and you obviously don't want them to fail.
I'm hoping that "this thing" is a machine for automating this. Because that would be so frickin awesome
What if, for preheating, you had some sort of hotplate to place the boards on?
I have a hotplate, The Charlatan, that just gets hot enough to seal cellophane and activate dry mounting tissue.
I feel like that would be absolutely perfect for this, it would streamline production quite a bit more.
hope the ideas i just ordered PCBs for are successful so i get a reason to build a lumen
Shortly after I started at my job, everyone (this was a very large company) received a "safety" mug for drinks. Per rumors, a tech spilled a drink into a solder pot. "Danger puddle" indeed!
Back then, there was a wave-solder conveyer in the bay next to mine. I guess that was a "danger lake." 😉
You doing a great job!
6:24 :D TH-cam recomended this video , Note :that is why SMT Owens has heterophile before flow soldering.
You guys would have seen it by now but have a look at core electronics selective soldering machine. There’s also lots of cheap used selective soldering and wave soldering machines in the USA even on eBay as they’re not super commonly used here anymore.
Ultrasonic cleaner for the boards with excess flux you can fill the cleaner with very high percentage alcohol (99%)
I want to put my hand in the puddle of _forbidden gallium_
Temp is always critical. Just wondering why you are so reluctant to crank the temp. I’d try at least 340, maybe even 350
9:10 you can use a 3d printer or simular device it will be not fast but acceptable automatization, TIP create rotary hed for lifting different components.
It's *almost* like you're manually imitating what a solder fountain does, by rocking it into the molten solder and out again.
Man you are an inspiration
commercial systems use a wave in the solder bath
IMHO, fluxing and good (pre-heat, dip, cooling) profile is key for reliable and good (wave) solder bumps.
See also (examples):
a) "Wave Solder Profile Supplement" notes from aimsolder
b) "Wave Soldering and Thermal Profiles" chapter from Microelectronics Cloud Alliance (MECA)
...
Would a sponge in your flux tub help with applying the flux?
I think so to, push down on a soaked sponge.
Not a bad idea, what if you masked a sponge? That way you don't have to clean up as much flux
@@kelna2 Yep.
also wave soldering machines mist the pcb with flux, maybe you could mist it inside a tupperware so you don't inhale too much
Modify a cheap 3d printer kit to make a solder dipping machine made one for my last company and was successful but eventually they outsourcing the job. For the bigger board they had a wavesolder machine which was cool.
I bought one of the first "print your own parts kit".
I'm still waiting for further information on the promised upgrade kit to convert the machine into a V3.
Any news? Thanks
If you think this is that amazing, you should look into wave soldering. Also, use capton circles for your areas you don't want soldered instead of tape. The little circles are easier.
You should get a fan for all the fumes.
you need a solder pump and a slotted insert to form a solder wave.