EEVblog #1282 - Design Your Own Membrane Keypad

[EEVblog]

How to design and your own custom membrane keypad and get it manufactured, to make your products look really professional.
It's easier and cheaper than you think!

Subscribe: http://www.eevblog.tv

[VEGETA]

Thanks for the awesome video as always. I've been following uSupply project ever since the first video, I wish we could see schematic soon.

I was wondering about the final estimated price for this product. I guess it could be something like 150 US dollars right? without shipping of course.

The PCB suggests that you ditched the LT3080 design? going for the transistor\mosfet solution? I am really excited to see the final details!

[johnlsenchak]

Dave

Don't  you worry  about  fingerprint wear  on the front  of the keypad   with the numbers and letters  eventually  wearing  off . You didn't mention  in the video
the longevity  of the front  of the  keypad  due to  continuous  use.  If you are  selling a unit  with that keypad   wouldn't  that be   a factor ?

[Clear as mud]

I'm going to summarize the video here, so I don't have to watch the whole thing if I'm looking for something in it later.

Membrane keypads - couple of hundred dollars for a few prototypes, or around $1 or $2 each in bulk.
Optional (more expensive) get a large one to cover entire product, with clear windows over your LCD screens (1:20).  More colors add more cost (5:18).

They consist of a custom flex PCB (2:20); glued to a vinyl decal (3:28) made of either polycarbonate or polyethylene/polyester; with adhesive tape on the back (1:35).  A few types of 3M adhesive (including the 467 type seen earlier) are briefly shown at (21:41).  The keypads can be tactile or non-tactile type (5:37).  The buttons can be embossed, or slightly raised above the surrounding keypad material, for a better feel (9:09).

Tactile keypads use tactile metal domes similar to those sold by Snaptron (5:54), usually embedded in the top decal layer of the keypad.  Several models are shown (the third one is similar to the kind used in the uSupply keypad featured in the video).  A datasheet for another manufacturer shows you can solder them to a PCB, instead of integrating in the decal (7:38).  An example of a product with the tactile domes on the PCB is shown starting at (9:48).  Another method is to buy them in a stick-on array (7:53).  See (10:47) for an in-depth explanation of how tactile domes work.

A non-tactile keypad is shown starting at (13:02).  They will have a longer lifetime because there is no metal fatigue issue.  Reliability is much better, but there's no tactile feedback, so it's better if the product makes some sort of key press feedback noise.  The one shown has two conducting printed circuit layers separated by a small space, but it is also possible to have just one circuit layer, with some conductive material on the back of the keys (like in a pocket-calculator keypad) (14:50).

(15:45) DESIGN PROCESS
(17:41) Get your connector specifications from the datasheet of whatever flex connector you're using.
(19:00) You can get carbon-printed resistors on flex PCB (if you need resistors).
(20:16) Example of a company website selling keypads, with a shopping-cart system for a quote.  Dave goes through all the options.

[HwAoRrDk]

I can't imagine that metal fatigue of the tactile domes is the limiting factor on lifetime of membrane keypads. Based on the number of times I have seen membrane keypads with the decal wearing out (or even through, in some instances!), I would expect the robustness of the vinyl is the primary factor.

By the way, the mentions of activation force brings to mind one thing that has always been an issue of uncertainty for me: how do you choose? This is something with common tactile switches too, that there are a hundred and one choices of activation force, so what do you pick? Do you go by trial-and-error: get a bunch of samples, and pick what you think feels good? Are there some kind of standards?

[Bud]

You only may need 2 or 3 samples with the activation force spec spread apart. You will try and will have a good idea which direction to go and approximately how much in between. This is what i did.

[thm_w]

I can't imagine that metal fatigue of the tactile domes is the limiting factor on lifetime of membrane keypads. Based on the number of times I have seen membrane keypads with the decal wearing out (or even through, in some instances!), I would expect the robustness of the vinyl is the primary factor.

Its not, the limitations I've seen are: plastic wearing out/cracking, and corrosion from moisture ingress. Water gets into the key or between the plastic layers and that conductive silver ink or copper goes bad very quickly.
I've seen many hundreds of membrane keypads fail. Only seen a handful of tactile dome failures, and a lot of those were from extreme force (hammering or kicking the keys). Think about it, say 1 million cycle life dome pressed 100 times per day, that is 27 years of continuous operation.

You can get quality non-tactile membranes with proper waterproofing mid-layers and thick plastic embossed layers. But if you are designing something that is going to be pressed on a daily basis, and not designing cheap crap, go with tactile every time.

By the way, the mentions of activation force brings to mind one thing that has always been an issue of uncertainty for me: how do you choose? This is something with common tactile switches too, that there are a hundred and one choices of activation force, so what do you pick? Do you go by trial-and-error: get a bunch of samples, and pick what you think feels good? Are there some kind of standards?

Yes, the major manufactures will give you a ton of free sample panels. You can use these to gauge the feel, button shape, etc. you want to achieve.

http://www.technigraphics.com/Membrane-Switch-Panels/product-41

[EEVblog]

By the way, the mentions of activation force brings to mind one thing that has always been an issue of uncertainty for me: how do you choose? This is something with common tactile switches too, that there are a hundred and one choices of activation force, so what do you pick? Do you go by trial-and-error: get a bunch of samples, and pick what you think feels good? Are there some kind of standards?

That.

[floobydust]

One thing to watch out for on this design, is the thin perimeter border especially where the cable breakout is. That's where membrane keypads come apart and de-laminate, or peel off the enclosure.
There is hardly any area for the adhesive at the keypad's perimeter. It's also where ESD gets in.

You have to cut a slot in an enclosure for the cable to get in, and so there is nothing to really stick the keypad to the enclosure there. I move the cable breakout in some distance from the keypad edge.

I've been burned twice with outdoor products with membrane keypads- water proofing, UV resistance, temperature, ESD immunity failures. Every time, it was due to not thinking about the edges and seams.

[Xenoamor]

I guess at such a low unit cost Dave could sell replacements for the more heavy users who may wear them out

[VEGETA]

I guess at such a low unit cost Dave could sell replacements for the more heavy users who may wear them out

Yes since this part is so fragile compared to traditional buttons. However, I asked him about approximate price for the complete unit but he gave no answer... so fingers crossed!

Do you think it could be something like 100$ without shipping?

[thm_w]

Yes since this part is so fragile compared to traditional buttons. However, I asked him about approximate price for the complete unit but he gave no answer... so fingers crossed!

Do you think it could be something like 100$ without shipping?

https://www.eevblog.com/forum/supporters-lounge/straw-poll-usupply-pricing/

I wouldn't worry about the buttons, but you can use PC control as well.

[VEGETA]

Yes since this part is so fragile compared to traditional buttons. However, I asked him about approximate price for the complete unit but he gave no answer... so fingers crossed!

Do you think it could be something like 100$ without shipping?

https://www.eevblog.com/forum/supporters-lounge/straw-poll-usupply-pricing/

I wouldn't worry about the buttons, but you can use PC control as well.

I've seen this thread and I am surprised about how low power this thing is. I thought it would be a lot better.

[Xenoamor]

Yes since this part is so fragile compared to traditional buttons. However, I asked him about approximate price for the complete unit but he gave no answer... so fingers crossed!

Do you think it could be something like 100$ without shipping?

https://www.eevblog.com/forum/supporters-lounge/straw-poll-usupply-pricing/

I wouldn't worry about the buttons, but you can use PC control as well.

I've seen this thread and I am surprised about how low power this thing is. I thought it would be a lot better.

I believe this is targetted at the low power MCU design market where there is a huge gap in the market. There's the OTII but it's far from cheap and has no display/control

[VEGETA]

I believe this is targetted at the low power MCU design market where there is a huge gap in the market. There's the OTII but it's far from cheap and has no display/control

I differ with you. We followed the project since the beginning and it was a lab power supply with li-ion batteries as its supply. it uses a switching pre-regulator then linear post-regulator. the ultra low noise thing probably won't be achievable through pre-regulator. Having no batteries is a killer for most people... especially now that you need to connect it to a source...

he has uSupply and a portable one... the portable is the isolated one but even that had a battery.

so fingers crossed

[thm_w]

Having no batteries is a killer for most people... especially now that you need to connect it to a source...

Use a power bank

[VEGETA]

Having no batteries is a killer for most people... especially now that you need to connect it to a source...

Use a power bank

I know I could but not it is 2 separate devices instead of one. we thought it would be only one unit.

[EEVblog]

I differ with you. We followed the project since the beginning and it was a lab power supply with li-ion batteries as its supply.

Actually, nope, the "uSupply" has always been a USB powered supply from day 1, which pre-dates any videos by many years.
Several variants went into a battery powered version.

[VEGETA]

I differ with you. We followed the project since the beginning and it was a lab power supply with li-ion batteries as its supply.

Actually, nope, the "uSupply" has always been a USB powered supply from day 1, which pre-dates any videos by many years.
Several variants went into a battery powered version.

I think you have the first version which had batteries, then another usb powered version according to the videos. Now I know you didn't want it to be battery powered from the beginning.

So will there be 2 versions or just one low powered usb device?

[EEVblog]

I differ with you. We followed the project since the beginning and it was a lab power supply with li-ion batteries as its supply.

Actually, nope, the "uSupply" has always been a USB powered supply from day 1, which pre-dates any videos by many years.
Several variants went into a battery powered version.

I think you have the first version which had batteries, then another usb powered version according to the videos. Now I know you didn't want it to be battery powered from the beginning.
So will there be 2 versions or just one low powered usb device?

There were 3 or 4 USB prototypes before the first video was ever made, I posted a photo of them all once.
Just USB at present.

[VEGETA]

I differ with you. We followed the project since the beginning and it was a lab power supply with li-ion batteries as its supply.

Actually, nope, the "uSupply" has always been a USB powered supply from day 1, which pre-dates any videos by many years.
Several variants went into a battery powered version.

I think you have the first version which had batteries, then another usb powered version according to the videos. Now I know you didn't want it to be battery powered from the beginning.
So will there be 2 versions or just one low powered usb device?

There were 3 or 4 USB prototypes before the first video was ever made, I posted a photo of them all once.
Just USB at present.

Then it should be cheap to be interested, at least to me. However, will it have a limited specs of voltage and current or just power?

I guess you already ditched the LT3080\81 solution... I'd like to watch a video about choices that went through that. Maybe returning to LM317 since 2v difference is what you want. If you want 1v then I guess you need to use a transistor.

[coppice]

Don't  you worry  about  fingerprint wear  on the front  of the keypad   with the numbers and letters  eventually  wearing  off . You didn't mention  in the video
the longevity  of the front  of the  keypad  due to  continuous  use.  If you are  selling a unit  with that keypad   wouldn't  that be   a factor ?

You can't wear off the labelling on this type of keypad. The ink is printed on the back of a clear sheet of plastic with a textured front surface, making it hard wearing and non-reflective. The only place this type of keypad has surface wear issues is when you need a clear window in it, for something like an LCD display. Then you need to use plastic with a smooth surface, and print on a texture everywhere but the display area. Those printed on textures tend to wear off over time, but the keys are still fully readable. The result is just cosmetically poor.

[philcc]

Designing a membrane switch for an application isn't an easy task, but with proper guidelines it become very simple. Here is a membrane switch guide to help engineers design the application easily: https://bit.ly/33b3M3X. I hope the resource is useful.

[thm_w]

Designing a membrane switch for an application isn't an easy task, but with proper guidelines it become very simple. Here is a membrane switch guide to help engineers design the application easily: removed. I hope the resource is useful.

If you (Phil C. Marketing Professional at Melrose) work for Melrose systems, be upfront about it.
No reason to use a bitly link unless you want to track people.