PCB Technology

[Sani300]

Hi,

Yesterday I disassembled my wireless keyboard. And I found this PCB (see photo). I would like to know if anyone knows this kind of PCB! This technology has a name? Where can we produce PCB like that ?

Sani300

[mikeselectricstuff]

It's silver-based ink printed on plastic film.

[Sani300]

Thanks ! Who can do that ?! Is there someone who already do that?

[In Vacuo Veritas]

That's an awful layout. Without length matching all the traces, how can you be sure your keypresses get to the SDRAM at the same time? Lazy designer!

[SiliconWizard]

That's an awful layout. Without length matching all the traces, how can you be sure your keypresses get to the SDRAM at the same time? Lazy designer!

 

[SiliconWizard]

Thanks ! Who can do that ?! Is there someone who already do that?

Unless you can buy a ready-made one, getting such a keyboard *membrane* custom made for small quantities will probably cost you a fortune.

The "membrane" keyword should help you find some manufacturers though.

[mikeselectricstuff]

I isn't an inherently expensive process - just printing ink onto plastic.
However as it's generally only used for high volume products, the manufacturing process will be geared to that, probably a roll-to-roll print process and custom die-cutting,so high entry cost but very low unit cost.

There are places that do custom membrane keypads in low volumes - I'd imagine they use screen printing, maybe with lasercut or drag-knife cutting of outlines.

[Buriedcode]

That's an awful layout. Without length matching all the traces, how can you be sure your keypresses get to the SDRAM at the same time? Lazy designer!

A fine example of Poe's Law

[Sani300]

Do you know exemple of place to do prototype and high qty ?

[ajb]

I've never ordered a kepyad like that, but when we've ordered printed plastic overlays (think tact switch/LCD overlay type stuff), the price curves were insane.  We could pay something like $500 and get 50pcs, or we could pay $800 and get 500pcs.  I'll let you imagine what it would have cost for a single "prototype"!  You have to be working in the thousands to tens of thousands for the curve to flatten out, because the setup costs are fixed and the marginal cost of making 100 instead of 1 is so low.  Also, our parts were made on a 100% digital process, so if the silver ink requires a physical screen then the setup costs will be even higher.  It's not unlike PCBs in that respect, except even more so because the materials and running costs are so low.

As far as where to get them made, generally the same companies that do "lexan" overlays and labels also do overlays with embedded LEDs and buttons, so I imagine they would also do the kinds of printed membranes you're after.  Where in the world are you located, and do you care where the vendor is located?

[Sani300]

Ok thanks for those informations and advices.
At this time, the best would be to find a company in France (that why my english is not really good ) or in Europe. But I can't find company on google with "silver base ink printed on plastic".

[2N3055]

Try conductive ink printing..

[Mario87]

That's an awful layout. Without length matching all the traces, how can you be sure your keypresses get to the SDRAM at the same time? Lazy designer!

Not so much a lazy designer, but probably more likely not wasting time "engineering" just for the sake of it. No-one will notice if the key presses reach the SDRAM fractions of a second later (into the nanoseconds or possibly even less I suspect??), it's not high frequency data or differential signaling which will benefit from trace length matching, so why do it just for the sake of it? Saves the company engineering costs and it works as intended. Everyone wins, so no need to over complicated it.

[rrinker]

That's an awful layout. Without length matching all the traces, how can you be sure your keypresses get to the SDRAM at the same time? Lazy designer!

 That's it! You solved my mystery of how I always get reversed letters for certain key combinations. Clearly, the path from the second key is shorter than the path from the first key, so despite hitting them in the correct order, the keyboard registers them the opposite way. Very happy to find out I'm not such a poor typist after all. 

[Mario87]

https://www.merriam-webster.com/dictionary/humor

https://www.merriam-webster.com/dictionary/context

Impossible to determine how someone is phrasing something via text without it....

[dzseki]

Genesink is a French company who produces tools for that kind of work, not sure if they do printing actualy, maybe you can request some demo at least?
http://www.genesink.com/

[Buriedcode]

That's an awful layout. Without length matching all the traces, how can you be sure your keypresses get to the SDRAM at the same time? Lazy designer!

A fine example of Poe's Law

[Bud]

That's an awful layout. Without length matching all the traces, how can you be sure your keypresses get to the SDRAM at the same time? Lazy designer!

 That's it! You solved my mystery of how I always get reversed letters for certain key combinations. Clearly, the path from the second key is shorter than the path from the first key, so despite hitting them in the correct order, the keyboard registers them the opposite way. Very happy to find out I'm not such a poor typist after all. 

This has been a problem for me for years. When i fast typing, i mostly get "teh" instead of "the". Yes i press the keys in correct order.

[In Vacuo Veritas]

That's an awful layout. Without length matching all the traces, how can you be sure your keypresses get to the SDRAM at the same time? Lazy designer!

 That's it! You solved my mystery of how I always get reversed letters for certain key combinations. Clearly, the path from the second key is shorter than the path from the first key, so despite hitting them in the correct order, the keyboard registers them the opposite way. Very happy to find out I'm not such a poor typist after all. 

Glad I could help. The alternative of getting all your nerves stretched to the same length could also work.

Not so much a lazy designer, but probably more likely not wasting time "engineering" just for the sake of it. No-one will notice if the key presses reach the SDRAM fractions of a second later (into the nanoseconds or possibly even less I suspect??), it's not high frequency data or differential signaling which will benefit from trace length matching, so why do it just for the sake of it? Saves the company engineering costs and it works as intended. Everyone wins, so no need to over complicated it.

I was making fun of someone who was insisting that length-matching the traces on a two-layer PCB with a 12MHz microcontroller is somehow important, and indeed, NOT doing so is laziness. My head is still spinning.

[SiliconWizard]

I was making fun of someone who was insisting that length-matching the traces on a two-layer PCB with a 12MHz microcontroller is somehow important, and indeed, NOT doing so is laziness. My head is still spinning.

You can't say no to a good laugh!