The MHz keyboard, next gen thoughts

[matman]

I recently built a programmable footswitch macro keyboard for myself and it got me thinking about possible advances in keyboard design.

Here are some ideas on how I think next gen keyboards could lift the game.

+ Internal key press response in the order of 1us (microsecond)
This is achievable using double throw switches and an SR latch debounce. This would require some logic directly attached to every switch, something already in place for keyboards that offer fully independant RGB key lighting. As far as I can tell all current PC keyboards use a single throw switch and delay based debouncing with response times roughly 100us to 1000us(1ms) or more.

+ Keypress to software input delay in the order of 10us using USB3.
Current USB HID keyboards have a best possible average input latency of about 2ms. A next gen design could support HID reporting for compatability but switch to a custom USB3 mode with a suitable driver. I believe some premium keyboards do this today for the purpose of NKRO support.

+ Programmable piezo keypress feedback
Haptic and/or audible feedback could be precisely linked to key press and/or release detection. Intensity of the feedback could be user adjustable per key. Current mechanical feedback designs only approximate coincidence with actual contact.

All of the above is achievable without significant R&D and minimal cost increase over existing premium keyboard designs. I suspect there is a viable market for such a product for around $200-300.

A more far fetched feature would be a programmable resistance force curve. For the limited extra value on offer I doubt it's worth the develop/manufacture cost increase.

I realise that for many, the performance of current keyboards is just fine but I'd bet there are plenty of folks around who feel they could take advantage of a more responsive keyboard.

Just to be clear, I have no irons in the fire. I just wanted to get this thought out and see what people think.

 

[Erasmus354]

The keyboard being a physical interface device, the most important thing is how it feels. This is why mechanical keyboards have become so popular. They simply have a much better feel to the user. I don't see any of your improvements as something that is noticeable to anyone. They are minor mostly unnoticeable tweaks at best. When you are talking about response time, that logic is not the limiting factor with a keyboard, it is usually the physical actuation method that is the limiting factor. The logic attached to it is just fine for the hardware switch it is interfacing to.

 

[Supercharged_Z06]

As to next gen keyboards, why are almost all good mechanical keyboards wired?

Why not add a USB port just to power backlighting for the fancier models which would be completely separate from the actual keyboard interface - this could be dedicated to also charge a battery just for automatic backlighting when needed. Provide another battery that is dictated to the actual keyboard interface link along the lines of Logitech's universal "nano" USB receiver for the keyboards actual interface.

Hell, just give me a top tier mechanical sans lighting that can hook into something similar to Logitech's nano receiver - use 4AA's so that it lasts a good year or two on a set of batteries. Use E-ink for the caps/num/scroll lock indicators as well as a comms link indicator to minimize battery draw.

 

[sharknice]

As to next gen keyboards, why are almost all good mechanical keyboards wired?

Why not add a USB port just to power backlighting for the fancier models which would be completely separate from the actual keyboard interface - this could be dedicated to also charge a battery just for automatic backlighting when needed. Provide another battery that is dictated to the actual keyboard interface link along the lines of Logitech's universal "nano" USB receiver for the keyboards actual interface.

Hell, just give me a top tier mechanical sans lighting that can hook into something similar to Logitech's nano receiver - use 4AA's so that it lasts a good year or two on a set of batteries. Use E-ink for the caps/num/scroll lock indicators as well as a comms link indicator to minimize battery draw.

There are wireless gaming mice because cord resistance hinders accuracy. That obviously isn't an issue with keyboards.

Why would you want a wireless mechanical keyboard? I can only think of a two reasons. You game from your couch with a keyboard and mouse, or you're an extreme neat freak and don't want any cords showing on your desktop. There are very few people like that, and even fewer of them willing to pay the extra 50% cost for gaming quality wireless and deal with any battery issues.

 

[jojo69]

you have a buildlog for the physical implementation you did on that footswitch setup?

 

[matman]

I don't see any of your improvements as something that is noticeable to anyone. ... The logic attached to it is just fine for the hardware switch it is interfacing to.

The target audience I was thinking of is the professional / serious gamer. If a 3 orders leap in key reporting precision boosts anyone's game by even less than 1% that's a nice advantage. I would be surprised if it didn't result in a touch typing wpm increase too. Don't underestimate the ability of our senses to appreciate sub millisecond improvements.

This concept revolves around a change to the hardware switch, to one that is 1000 times quicker with the right supporting logic. I have no argument with your assertion that current logic is fine for current keyboard switches.


@jojo69 No online worklog as such. I have most of the code revisions for it and a few wiring diagrams that somewhat reflect the final design but most of my material is more like a log of experiments and mistakes than a build log.

 

[Archmage]

I think that to have a chance of marketing this, some comparison tests would need to be done to demonstrate practical, real-world superiority.

I'm not sure who, other than highly competitive gamers, would care about reducing input lag to such a degree.

The good thing is that there is a decent amount of money in competitive gaming now, with many top teams being sponsored by the likes of Logitech, Corsair, Razer - and many of their fans taking notice.

You may have already heard of Cherry's RealKey - I hope to see some tests.

http://hexus.net/tech/news/peripher...d-60-mechanical-keyboard-debuts-realkey-tech/ -
http://semiaccurate.com/2015/01/12/cherry-makes-radical-advance-new-keyboard/ - more detailed
http://hardforum.com/showthread.php?t=1849340 - discussion of above, some skepticism

I agree, however, that the feel of the peripherals is paramount, even for gamers who may be obsessive about responsiveness.

 

[trick0502]

+ Keypress to software input delay in the order of 10us using USB3.
Current USB HID keyboards have a best possible average input latency of about 2ms. A next gen design could support HID reporting for compatability but switch to a custom USB3 mode with a suitable driver. I believe some premium keyboards do this today for the purpose of NKRO support.

I know my kb (corsair k70) has a 1000 MHz polling rate. That would put the input delay exactly the same as a 1000 MHz gaming mouse.

 

[matman]

Interesting and relevant. Thanks Archmage.

Firstly a couple of people have mentioned keyboard "feel" as most important. I assumed this was so obvious that I didn't think it was worth addressing. Far am I from suggesting that you could take a $5 OEM keyboard put some new switches in and suddenly have the "best" keyboard. Take the keyboard that you think feels the best and imagine that it was 1000 times faster at turning your finger movements into changes on screen. That's all about the feels .

I wasn't aware of Cherry's new push for precision but it feels comforting to know that I'm thinking in the same direction as the designers of some of the highest quality keyboards and switches around. It's difficult to interpret their market speak but from what I can tell they're using the the same old MX switches, employing an analog keypad matrix with DSP for anti ghosting and using the DSP to improve debounce performance by a factor of maybe 5 over existing cherry MX units. This possibly warrants a doubling of the HID max polling rate of 500Hz. It's seems like a decent attempt to keep the MX switch relevant.

@trick0502 it's 1000Hz not MHz and that's not its total input delay. That's just its polling rate. For comparison 10us is 100000Hz.

 

[Zolishoru]

...
+ Internal key press response in the order of 1us (microsecond)
This is achievable using double throw switches and an SR latch debounce. This would require some logic directly attached to every switch, something already in place for keyboards that offer fully independant RGB key lighting. As far as I can tell all current PC keyboards use a single throw switch and delay based debouncing with response times roughly 100us to 1000us(1ms) or more.
...

AFAIK, HAL sensors doesn't need debouncing, you can service interrupt to the micro immediately. If you can, give them a try.

 

[Supercharged_Z06]

There are wireless gaming mice because cord resistance hinders accuracy. That obviously isn't an issue with keyboards.

Why would you want a wireless mechanical keyboard? I can only think of a two reasons. You game from your couch with a keyboard and mouse, or you're an extreme neat freak and don't want any cords showing on your desktop. There are very few people like that, and even fewer of them willing to pay the extra 50% cost for gaming quality wireless and deal with any battery issues.

Not really a neat freak, but I have a PC located in the kitchen on a high shelf (above the cupboard) that isn't anywhere near where the keyboard sits. Running a cable down to the keyboard would look pretty horrible... see pic:

Would love to sport a mechanical keyboard with this setup - but they are all wired. With my main gaming rig in my man cave, this isn't an issue, but for the above application a wireless mechanical would be ideal.

 

[Butcher9_9]

I think that to have a chance of marketing this, some comparison tests would need to be done to demonstrate practical, real-world superiority.

I'm not so sure about that. Put in the hands of the right marketing company and even if there is no real world advantage it will sell.

Look at all of those 9999999+ dpi mice out there. Yes they are technically better but if you ever use them on that setting they will make you PC unusable.

1000 times faster at turning your finger movements into changes on screen. That's all about the feels .

The bigger issue is how fast is everything else. Sure you keyboard might be super fast but then your monitor/mouse/PC whatever adds a delay as well. If you went from say 1ms to 10us, would anyone notice if the total delay was say 50ms, maybe, what about say 100us to 10us, most likely not.

 

[dandragonrage]

Not really a neat freak, but I have a PC located in the kitchen on a high shelf (above the cupboard) that isn't anywhere near where the keyboard sits. Running a cable down to the keyboard would look pretty horrible... see pic:

The issue with running the wire is NOT that it would look horrible, but that you THINK it would look horrible even though it wouldn't.

The whole idea that wires are ugly is ridiculous. Having tons of wires laid out stupidly is certainly ugly, but one or two USB wires that you won't trip on? The issue is with you, not the wires.

 

[Supercharged_Z06]

The issue with running the wire is NOT that it would look horrible, but that you THINK it would look horrible even though it wouldn't.

The whole idea that wires are ugly is ridiculous. Having tons of wires laid out stupidly is certainly ugly, but one or two USB wires that you won't trip on? The issue is with you, not the wires.

OK, a wire running down the wall and over to the keyboard would simply look great for this setup. I am definitely wrong about wanting a wireless mechanical keyboard for this and certainly must have issues. My apologies!