Hacker News new | past | comments | ask | show | jobs | submit login
Rapid Prototyping with a $100 Inkjet Printer (ygoliya.medium.com)
146 points by SG16 on May 23, 2021 | hide | past | favorite | 35 comments



I wonder if circuits printed on paper would yield themselves to the same specialized applications as flexible PCBs.

For example, there are many variations of custom-made Dactyl keyboards, most of which require hand-wiring to accommodate for the curved shape of the design: https://github.com/adereth/dactyl-keyboard

Innovative designers have come up with alternatives that rely of flexible PCBs. The most recent example I recall is the Bastyl keyboard: https://www.reddit.com/r/mechmarket/comments/jvacs6/gb_basty...

I don't know how to even begin prototyping with flex PCBs, but that barrier to entry may be lessened with paper circuits like these.


>treating the silver ink on a hot plate at 100–120C produces acceptably conductive patterns

Therein lies the rub:

>the ideal resistivity of the silver ink traces would be less than 5 times that of bulk silver.

>The ideal resistivities were not met, and the resistivity of the sample ink traces ranged from five to ten times the bulk resistivity.

https://digitalcommons.calpoly.edu/matesp/162/


Could you explain this a bit more - I think I don't quite know enough about electronics to follow the criticism.

Are you saying the traces are too resistive, or not resistive enough? How would that manifest - i.e. I see working circuits on the video, but are there classes of circuits that this would not work for?


Too resistive, or at least, higher than what they were aiming for.

You can compensate by making traces wider to get a lower resistance from the same resistivity, or by running circuits slower (digital circuit speed is often limited by the RC time constant where C is parasitic capacitance of an input somewhere — a higher-resistance trace takes longer to charge up the input gate capacitance). How hard it is to adjust for the higher resistivity will depend on the circuit, of course.


Too resistive. Copper, silver, and gold are used because they have very low resistance. Having high resistance in the traces makes anything more complex than a few low speed components not practical.


There's a ton of stuff I make (especially for hobby projects) where I'm pretty sure it wouldn't really matter. I'd love to fidget with stuff like this, especially since I just acquired a wide-format inkjet printer.


yes that's true, liquid inks usually have more resistance and this is a drawback. This is because the ink contains non-conducting materials which are required for the right viscosity and to prevent the silver to agglomerate. These can usually be removed by heating.

One can also use high intensity light to "weld" the silver particles together without damaging the plastic substrate. This requires an expensive setup but improves conductivity drastically... https://en.wikipedia.org/wiki/Photonic_curing


While the idea is amazing, I can't help but think about the environment. Until I'm not good enough at electronics and prototyping I think it's cheaper and safer for the environment if I experiment with a breadboard


Breadboards are great for prototyping until you get signals close to 10 MHz. Stray capacitance becomes an issue north of there.


Semi-related: One way to test if you have stray capacitance is to bring your finger close to the circuit and see if readings start to change.


The article shows examples of this technology being used to print components, such as temperature and humidity sensors, and antennas. That's more than a breadboard can do, which is designed for building circuits.


That's cool. Punch holes in these printouts, glue different layers together, add some extra ink into the holes for conductive connections between layers ("vias" [0]) and you can make multi-layer boards easily.

[0] https://en.wikipedia.org/wiki/Via_(electronics)


Wonder how impedance control is with paper in lieu of FR-4.


Probably quite poor.

FR4 boards are built with woven fibers, and you can get measurable differences depending on how tight the weave pattern is, or the angle of your transmission line relative to the weave pattern. Given that paper is not woven, I’d expect a lot more dispersion than on cheap PCBs.

That said, with a good enough process able to press the paper + resin stack reliably, you could possibly get something usable.


Probably non-existent since the thickness of insulation layer isn't controlled.

Surprisingly, silicone rubber has a dielectric constant between 2 and 4 [1], which is pretty close to FR-4's 4.4. So maybe it isn't impossible?

[1]https://res.mdpi.com/d_attachment/polymers/polymers-09-00533...


Any more details on cost effectiveness? In this case the printer isn't the main cost but the ink.

https://store.novacentrix.com/category-s/1836.htm


The cost is probably not that bad on a per-unit basis, a 50mL seems like it should last for a pretty large number of PCBs. The problem is that the resulting PCBs are almost completely useless because you can't solder anything to them without melting the substrate.

The ink that works with standard inkjet printers just doesn't stick to materials like FR4 or kapton. There are specialized machines for printing conductive ink onto usable PCB substrates either with an extruder or screenprinting but they are at a minimum several thousand dollars and have extremely expensive consumables.


That explains why all of their examples are things like temperature sensors, humidity sensors, or antennas. I guess you stick wires on the paper using some sort of conductive glue?


There's silver epoxy but that's outrageously expensive.

Googling shows that you can add fine graphite powder to some kind of epoxy, to get something like that maybe.

Still glue is not super easy to control, and can't be used for fine traces. If the glue leaks across two pins, you can't just remelt it with a soldering iron.


Screen printing is a better choice for printing onto PCB surfaces. Copprint (an Israeli startup) has copper inks specificallý for FR4. Soldering is easier there...

https://www.youtube.com/watch?v=1up5a7z6tjE

Another option is this:

https://www.youtube.com/watch?v=OpObgG522sY


I would be also very interested in the costs. It certainly saves some time, but how does it compare to etching your PCBs at home?


Frankly I'd be more interested where you could print resistors, diodes, and transistors on plastic.

OLED manufacturers are apparently working on roll-to-roll printing on a plastic substrate.


Any idea how I could use a similar process to print glue? I would like to print glue in very specific places on a page so I can run sand over it and make a sand portrait.


Print a negative and use it like a stencil. You could also mix acrylic with sand so you can exclude the glue part altogether


Thanks, but I'm trying to automate the process for mass production.


Not OP, but curious why stencil is not a viable option for mass production? You don't have to roll compound over the stencil manually. Stencils are still used in applications like solder paste.


Yes stencil or screen printing is much faster for mass production. It is faster and you could directly print on a variety of substrates like FR4, plastic, glass and even paper. I'll wirte about that next


Well I'm not trying to reproduce the same picture 100 times but instead 100 different pictures. So it would be 100 different stencils.


You could modify a 3D printer to output glue instead of plastic, a slightly related video I can provide shows how you can print chocolate on to a piece of break[0].

0: https://www.youtube.com/watch?v=oRBxuDvS2gs


Check this out, might be interesting for you:

https://www.youtube.com/watch?v=OpObgG522sY


laser cut a frisket and spray glue through the holes.


Perhaps, there could be a specialized substrate, that absorbs the ink, makingnit suitable for pasting on a blank PCB board. Say an organic porous film, on which we could print the circuit, then paste it onto a base, heat it and get a working circuit.


Yes you can do that.. a simpler way is to use screen printing directly on to a PCB


$100 inkjet and $1000 worth of ink? Or?

We would need to know a price indication (which I guess is impossible at this stage) to know which of the printer and ink make up the largest part of the cost.

Silver sounds expensive...


The "Show HN" tag is usually reserved for shipped products created by the original poster.




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: