Squink Lets You Print a Circuit Board for the Price of a Cup of Coffee(techcrunch.com)
techcrunch.com
Squink Lets You Print a Circuit Board for the Price of a Cup of Coffee
http://techcrunch.com/2014/07/14/squink-lets-you-print-a-circuit-board-for-the-price-of-a-cup-of-coffee/
40 comments
I can't speak for this specific solution, but veroboard drives me (personally) bonkers and it's not enough of a hobby for me to justify investing in photochemical etchers. Being able to design a project on a computer, and get a PCB for a few dollars, would be great. I'd definitely play around with electronics projects a lot more if I could do more prototyping on a computer.
The other thing that would be great is something like a Shapeways for PCBs. People could upload their own projects, then they print the PCB for you (and even ship you a bag of components), for you to build yourself.
The other thing that would be great is something like a Shapeways for PCBs. People could upload their own projects, then they print the PCB for you (and even ship you a bag of components), for you to build yourself.
>The other thing that would be great is something like a Shapeways for PCBs.
The service already exists. Numerous companies will produce small quantities (>5) of a PCB design from a Gerber file at a reasonable price and many will also do assembly. Prices vary depending on quantity and lead time; Amongst the cheapest are Itead, who offer ten 5x5cm PCBs (two-layer, with soldermask and silkscreen) for $14 shipped.
The main reason to etch your own boards is lead time, as it can take a week or two (plus shipping from China). European and American PCB houses can turn around boards in as little as 24 hours, but at much greater cost.
http://imall.iteadstudio.com/open-pcb/pcb-prototyping.html
The service already exists. Numerous companies will produce small quantities (>5) of a PCB design from a Gerber file at a reasonable price and many will also do assembly. Prices vary depending on quantity and lead time; Amongst the cheapest are Itead, who offer ten 5x5cm PCBs (two-layer, with soldermask and silkscreen) for $14 shipped.
The main reason to etch your own boards is lead time, as it can take a week or two (plus shipping from China). European and American PCB houses can turn around boards in as little as 24 hours, but at much greater cost.
http://imall.iteadstudio.com/open-pcb/pcb-prototyping.html
Where do I buy the photoetch machine and is there an easy way to use it? It is automated based on plan files you give it?
I can see how someone who is familiar with placing electronic components would be sceptical of waiting a long time. But if there is a convenient setup even a slow automated placer would help me since I am not familiar with electronics or have steady hands.
Are there automated pick and place machines that are general purpose? I think if you take this concept and improve the speed and capabilities then you can replace a lot of human labor.
I can see how someone who is familiar with placing electronic components would be sceptical of waiting a long time. But if there is a convenient setup even a slow automated placer would help me since I am not familiar with electronics or have steady hands.
Are there automated pick and place machines that are general purpose? I think if you take this concept and improve the speed and capabilities then you can replace a lot of human labor.
>Where do I buy the photoetch machine and is there an easy way to use it?
There's no special machinery necessary, and it's so easy that a child can do it. The technique involves applying a resist material to the copper-clad board where traces are needed, then using a corrosive liquid to etch away the unwanted copper. Simple PCB etching requires only a plastic tray, a couple of bottles of chemicals, an ordinary clothes iron, a blank copper-clad board and access to a laser printer or photocopier. A slightly more advanced method requires a UV exposure box, which can be bought for as little as $100. There are dozens of excellent tutorials available, just search online for "PCB etching".
Pick-and-place machines work fantastically well in production quantities, where components are supplied on tape reels that can be automatically fed into the machine. In small quantities, correctly aligning the components is a major problem, requiring either a great deal of human labour or a very sophisticated machine vision system and a complex multi-axis CNC machine.
Placing SMD components by hand is a fiddly but simple job, especially with the aid of a magnifying lens. For experimental and prototyping purposes it often makes more sense to use through-hole components, which are somewhat bulkier but much easier to handle. Through-hole components can be used with solderless breadboard or veroboard, which allow you to build and test circuits without producing a PCB.
There's no special machinery necessary, and it's so easy that a child can do it. The technique involves applying a resist material to the copper-clad board where traces are needed, then using a corrosive liquid to etch away the unwanted copper. Simple PCB etching requires only a plastic tray, a couple of bottles of chemicals, an ordinary clothes iron, a blank copper-clad board and access to a laser printer or photocopier. A slightly more advanced method requires a UV exposure box, which can be bought for as little as $100. There are dozens of excellent tutorials available, just search online for "PCB etching".
Pick-and-place machines work fantastically well in production quantities, where components are supplied on tape reels that can be automatically fed into the machine. In small quantities, correctly aligning the components is a major problem, requiring either a great deal of human labour or a very sophisticated machine vision system and a complex multi-axis CNC machine.
Placing SMD components by hand is a fiddly but simple job, especially with the aid of a magnifying lens. For experimental and prototyping purposes it often makes more sense to use through-hole components, which are somewhat bulkier but much easier to handle. Through-hole components can be used with solderless breadboard or veroboard, which allow you to build and test circuits without producing a PCB.
> and it's so easy that a child can do it.
...
> then using a corrosive liquid to etch away the unwanted copper
Do I then dump the wastewater into the stream in the backyard?
...
> then using a corrosive liquid to etch away the unwanted copper
Do I then dump the wastewater into the stream in the backyard?
Ferric chloride etchant can be safely poured down the drain if it is neutralised with washing soda and filtered to remove the precipitated copper. Alternatively, it can be mixed with a proprietary neutralising compound and disposed of as solid waste. The etchant itself is quite harmless; It is the copper ions which pose an environmental hazard.
Many alternative etchants (e.g. cupric chloride) have an indefinite working life and so pose no disposal issues.
Many alternative etchants (e.g. cupric chloride) have an indefinite working life and so pose no disposal issues.
"The circuit board is designed through a web-based portal usable only with Squink plugged in."
Whoa there, cowboy. Its a kickstarter that not only may or may not produce a working machine, but even if it does, I have to depend on their perpetual existence just to use said machine? I... have trust issues.
Whoa there, cowboy. Its a kickstarter that not only may or may not produce a working machine, but even if it does, I have to depend on their perpetual existence just to use said machine? I... have trust issues.
Their actual kickstarter page just says their machine takes gerbers, so it should be completely platform agnostic.
as someone else pointed out, the portal seems to be hosted on the squink it self. so the portal is jut a way to access the squinkk like u access ur linksys router
Can anyone confirm whether this is a portal 'in the cloud' or if it's simply a web server hosted on the Squink itself?
Very cool project, I wonder if future iterations by competitors will drive the cost down.
Very cool project, I wonder if future iterations by competitors will drive the cost down.
Home-printed (not etched) PCBs? This is awesome. Sure, the traces look thick as hell and you're obviously limited to 1-2 layers, but it's what progress looks like.
Just as the barrier for entry has come down for software -- in the form of better tools, frameworks, and languages -- so too is it coming down for hardware in the form of easier, more accessible and better-quality prototyping. The specialist in me worries (occasionally), but what a wonderful moment of innovation we live in..
Just as the barrier for entry has come down for software -- in the form of better tools, frameworks, and languages -- so too is it coming down for hardware in the form of easier, more accessible and better-quality prototyping. The specialist in me worries (occasionally), but what a wonderful moment of innovation we live in..
This is way more interesting than the title makes it sound. From the title, I expected something that just made a PCB, and then you would have to place the components on it and attach them.
What they are proposing will make the PCB, then put conductive glue at the places where the component leads are supposed to attach, and then pick the components from a tray, rotate and aline them, and place them on the glue.
What they are proposing will make the PCB, then put conductive glue at the places where the component leads are supposed to attach, and then pick the components from a tray, rotate and aline them, and place them on the glue.
There was a Kickstarter that did exactly this a while back:
https://www.kickstarter.com/projects/1597902824/agic-print-p...
I guess they're supposed to start shipping in a couple weeks (Aug 1).
https://www.kickstarter.com/projects/1597902824/agic-print-p...
I guess they're supposed to start shipping in a couple weeks (Aug 1).
Nice, I like that it uses commodity printer hardware - inkjets have a really high resolution nowadays, and they claim that 0.4 mm BGAs still work, which is usually out of reach for the typical DYer and prototype services like OSH Park.
Also, they specify the resistance that can be achieved, something I sorely missed from this Kickstarter.
Only downside is that the ink price seems quite high - 100$ for a bottle that can print 50-100 sheets.
Also, manual pick & place is probably messy.
Still, at $2,499 early birg price, Squink is only affordable for hacker spaces and the like, while the AgIC Print DIY kit sells for $299. Exciting times!
Only downside is that the ink price seems quite high - 100$ for a bottle that can print 50-100 sheets.
Also, manual pick & place is probably messy.
Still, at $2,499 early birg price, Squink is only affordable for hacker spaces and the like, while the AgIC Print DIY kit sells for $299. Exciting times!
> Exciting times!
Agreed!
Yeah, the Squink Kickstarter is sorely lacking in information. They don't give trace width and spacing specs, etc.
I also like that AgIC uses off the shelf printer hardware and can manage high resolution. However, I'm skeptical you can do much more than OSHPark with 0.4mm BGAs. The challenge with OSHPark design rules is the breakout, and here you have one (maybe two?) layers, so again, you can't manage much of a ball grid.
Although these projects are exciting, I can't seem them replacing cheap PCB fabs + reflow oven for my own prototyping projects. So then question becomes, how do you take the next step? Maybe you can print on FR-4 or pre-preg, use optical alignment to handle registration and glue multiple layers together with conductive glue to achieve multi-layer boards. For ground planes, maybe you choose some layers pre-clad with copper. I'm not sure what to do about vias. You could think about using layers with a selection of various pre-made vias, but again, it is hard to see how that would work with large BGAs. So you have to include a drill and find some way to plug/fill the vias.
Agreed!
Yeah, the Squink Kickstarter is sorely lacking in information. They don't give trace width and spacing specs, etc.
I also like that AgIC uses off the shelf printer hardware and can manage high resolution. However, I'm skeptical you can do much more than OSHPark with 0.4mm BGAs. The challenge with OSHPark design rules is the breakout, and here you have one (maybe two?) layers, so again, you can't manage much of a ball grid.
Although these projects are exciting, I can't seem them replacing cheap PCB fabs + reflow oven for my own prototyping projects. So then question becomes, how do you take the next step? Maybe you can print on FR-4 or pre-preg, use optical alignment to handle registration and glue multiple layers together with conductive glue to achieve multi-layer boards. For ground planes, maybe you choose some layers pre-clad with copper. I'm not sure what to do about vias. You could think about using layers with a selection of various pre-made vias, but again, it is hard to see how that would work with large BGAs. So you have to include a drill and find some way to plug/fill the vias.
I agree, multi-layer gluing and pre-made ground planes would probably work. The big issue is how to connect the layers, I don't see how to do vias either.
Maybe there will be a symbiosis of this and 3d printer, where the single pcb layers will be printed, using conductive material and non-conductive filler.
But however this is going to turn out, we are witnessing the birth of home electronics manufacturing. This will be significant, at least as much as the availability of high-quality consumer paper printers in the 80s/90s.
Maybe there will be a symbiosis of this and 3d printer, where the single pcb layers will be printed, using conductive material and non-conductive filler.
But however this is going to turn out, we are witnessing the birth of home electronics manufacturing. This will be significant, at least as much as the availability of high-quality consumer paper printers in the 80s/90s.
I'm guessing the real innovation here is coming in the form of new conductive inks and glues. Once there is freely available conductive ink that can create a trace with better than terrible resistance, there will likely be a cambrian explosion of projects like this.
This isn't just a PCB printer it also does pick and place and solders component with conductive glue! It's essentially an entire SMT assembly line on your desktop. Wow, this is cool.
Lots of issues will need to be worked out for this to be viable. 4 layers minimum (ground, power, 2 routing planes). High pitch TSOP and BGA need tight pitches and traces. The width of traces and spacing and dielectric material need tight tolerances for impedance control for signal integrity.
But they are not asking for much of an investment so it may be worthwhile to try it out.
But they are not asking for much of an investment so it may be worthwhile to try it out.
That's assuming a level of complexity this doesn't appear to be shooting for. Sure, you probably couldn't build an RPi or BBB with this guy, but it seems that they're targeting a lower end shop trying to spin a few iterations of a low-cost IoT device before sending it out to a fab house. The level of complexity needed for that hardware isn't huge.
I could also see e.g. the guys handbuilding guitar effects going nuts over something like this.
Yes! That would be great. So much more consistent than handsolder.
Squink is on Kickstarter:
https://www.kickstarter.com/projects/botfactory/squink-the-p...
https://www.kickstarter.com/projects/botfactory/squink-the-p...
Something to think about is a miniboard (a possibly trademarked cheap small board supplier of great reliability, fast speed, and excellent reputation) sells boards for $17 each shipped the next business day, so printing 250 boards yourself costs 3500 + 250*2= $4000 vs $4250 for professionally made boards. Hmm I donno.
250 boards is a lot of work. I don't think "most" hobbyists are into the hobby that long. For the sake of argument, we'll assume I've ordered that many PCBs as a hobbyist since 1980. How many computers from 1980 are still supported with software updates etc? I could eventually make 250 PCBs in a couple decades, but whats the realistic "clock time" lifespan of this thing?
Note that if you're going into serious production rather than the prototype market, you can get a hundred small boards much cheaper than $17 a piece.
Another interesting thing to think about is this is a printer... you need a substrate. First thing I thought of it using polished flat marble tiles from home depot. So there are other costs.
Also its an ink emitter. Much like ink jet printers which could theoretically print at 5 cents per page IF you use the ink up within 3 months of opening, reality is you try to print a report, realize your cartridge is dead / dried out, install a new $50 cartridge, print your report, and repeat six months later. So ink jets actually cost like $10 per page as a real world on the ground cost.
Given the above, its quite possible you could pay $50 for a cartridge and a board every six months in a couple hours, plus $3500 capital cost, or pay maybe $40 for a pro shop to overnight airmail a prototype PCB to you every six months.
The TLDR is the financials only work over an extremely narrow range of production rates under very peculiar requirements.
Its interesting in the marketplace there are no serious efforts at a full kit for traditional hands on mfgr. There are numerous partial kits that don't quite contain everything.
You need a B/W laser printer that is stably calibrated to precise dimensions thats really good at transparencies, a box of transparencies, whatever gear you use to verify calibration (A machinists $5 class 1-2-3 block and some software?). A UV box (basically a giant eprom eraser). Some photo trays and etchants and other chemicals.
Some "maker" cultural groups have an intense aversion to hands on, however weird that sounds, and will pay anything or put up with anything, to avoid using their hands, and that's probably the ideal customer for this product given that you can get better results cheaper and faster if you use a traditional PCB process.
There is probably a market for someone to sell a complete, vetted, verified photosensitive PCB etching kit.
250 boards is a lot of work. I don't think "most" hobbyists are into the hobby that long. For the sake of argument, we'll assume I've ordered that many PCBs as a hobbyist since 1980. How many computers from 1980 are still supported with software updates etc? I could eventually make 250 PCBs in a couple decades, but whats the realistic "clock time" lifespan of this thing?
Note that if you're going into serious production rather than the prototype market, you can get a hundred small boards much cheaper than $17 a piece.
Another interesting thing to think about is this is a printer... you need a substrate. First thing I thought of it using polished flat marble tiles from home depot. So there are other costs.
Also its an ink emitter. Much like ink jet printers which could theoretically print at 5 cents per page IF you use the ink up within 3 months of opening, reality is you try to print a report, realize your cartridge is dead / dried out, install a new $50 cartridge, print your report, and repeat six months later. So ink jets actually cost like $10 per page as a real world on the ground cost.
Given the above, its quite possible you could pay $50 for a cartridge and a board every six months in a couple hours, plus $3500 capital cost, or pay maybe $40 for a pro shop to overnight airmail a prototype PCB to you every six months.
The TLDR is the financials only work over an extremely narrow range of production rates under very peculiar requirements.
Its interesting in the marketplace there are no serious efforts at a full kit for traditional hands on mfgr. There are numerous partial kits that don't quite contain everything.
You need a B/W laser printer that is stably calibrated to precise dimensions thats really good at transparencies, a box of transparencies, whatever gear you use to verify calibration (A machinists $5 class 1-2-3 block and some software?). A UV box (basically a giant eprom eraser). Some photo trays and etchants and other chemicals.
Some "maker" cultural groups have an intense aversion to hands on, however weird that sounds, and will pay anything or put up with anything, to avoid using their hands, and that's probably the ideal customer for this product given that you can get better results cheaper and faster if you use a traditional PCB process.
There is probably a market for someone to sell a complete, vetted, verified photosensitive PCB etching kit.
Why don't I just grab a conductive pen and draw my circuit??
You can't just draw a printed circuit board, it's like in the name and stuff.
I'm not sure the price of a cup of coffee is a good indication of something being inexpensive nowadays:
Sexagintuple Vanilla Bean Mocha Frappuccino, $55 [1].
[1] http://www.dailymail.co.uk/news/article-2641608/Is-worlds-ex...
[1] http://www.dailymail.co.uk/news/article-2641608/Is-worlds-ex...
I'm not sure picking the world's most expensive cup of coffee is a representative sample of what a cup of coffee costs.
How much does a representative sample of a cup of coffee cost?
The coffee I drink costs about $0.10 a cup.
The coffee I drink costs about $0.10 a cup.
In Seattle, a ~350ml cafe-latte (~310ml milk, 40ml coffee) is around $3.50 (can be $4.50 downtown, $5.50 at airport) - this commonly called the 'Double Tall Latte' - our local news media uses it as a cost comparison for dang near everything.
I never understood why people spend so much when you can make at home for so much less. (Apparently 1/10 to 1/50th of the cost?!?)
Do people really consider a cup of coffee cheap? It seems expensive enough to be considered an occasional treat, not a daily purchase.
Do people really consider a cup of coffee cheap? It seems expensive enough to be considered an occasional treat, not a daily purchase.
It's definitively a small luxury item, but can't really argue with anyone's spending habits.
More relevant, "cups of coffee" is not a good unit of measurement and can be misleading. Just say the number and trust that people can cope.
Other offenders: Measuring distance in "cars lined up from here to the moon", expressing data in "Libraries of Congress", or expressing cost in $/day instead of $/month or $/year.
More relevant, "cups of coffee" is not a good unit of measurement and can be misleading. Just say the number and trust that people can cope.
Other offenders: Measuring distance in "cars lined up from here to the moon", expressing data in "Libraries of Congress", or expressing cost in $/day instead of $/month or $/year.
I currently spend about a dollar a cup to two dollars a cup for black coffee. Mostly because seems like it would take a few hundred dollars worth of experimental purchases and a reasonable number of hours to match the quality I get form local cafe.
If I move to another city or If I decide I have time/money for the hobby I will do.
If I move to another city or If I decide I have time/money for the hobby I will do.
A representative cup of coffee costs $3-$5 and is not the same as a cup of coffee you would enjoy at home, which is not measured in cost per cup.
> This is said to be the world's most expensive Starbucks coffee featuring 60 espresso shots and costing $55 (£33).
That's a lot of shots! How many espresso shots is enough to kill someone?
That's a lot of shots! How many espresso shots is enough to kill someone?
Death by Caffeine[1] says 172.84 espresso shots to kill a 195lb human.
[1] http://www.caffeineinformer.com/death-by-caffeine
[1] http://www.caffeineinformer.com/death-by-caffeine
I'm upvoting this because coffee should cost fifty cents a cup.
Cheaply and quickly producing basic single-sided PCBs really isn't hard. Photochemical etching can be done with $100 worth of equipment, a couple of feet of bench space and a little practice. Using a $3500 machine to produce poorer quality boards doesn't make much sense to me. Likewise, the pick & place functionality just isn't very useful in producing prototypes - the setup time will be equal to or greater than placing the components by hand with tweezers.
There are a swathe of recent projects based on conductive ink, but I don't think that they are solving a real problem. I use PCB fabrication services because I need soldermask and through-hole plating, not because photoetch is particularly inconvenient. Someone downthread discusses the problem of printing resolution and fine-pitch parts, but that's really a secondary issue. Without soldermask, it is very difficult to get good joints without bridging on fine-pitch ICs; Without plated vias, it is extremely difficult to route traces for even a relatively simple SMD IC, let alone a BGA.
The boards shown in their publicity materials are trivial and would be easily prototyped using veroboard. I believe that represents the limit of their technology. I can just about see a niche for their product in education, but even that is quite tenuous; The throughput of this machine is likely to be a major bottleneck in a classroom setting versus etching 30 boards on the same panel. I learned to etch PCBs in secondary school, and the process is now much cleaner and safer thanks to modern etchants. Three bubble tanks and a UV box cost considerably less than $3500 and are far more reliable than any CNC machine.
If photochemical etching isn't an option for some reason, you can buy a basic CNC router for far less than $3500, which can produce good basic PCBs while also taking on a vast range of other tasks.
Frankly, the whole effort smacks of "3D printing" hype. There seems to be a generation of designers who have a nozzle-shaped hammer and see everything as an additively-manufactured nail. There are some compelling applications for additive manufacturing, but it is by no means a silver bullet.