Why does one need 64 KRO (Key Roll-Over)?

[Kamen Rider Blade]

http://www.tomshardware.com/news/cooler-master-trigger-z-keyboard,25884.html

Assuming the 64 KRO is true and not just some arbitrary number the marketing guys put on the box, what is the benefit?

Most people don't share 1 keyboard amongst multiple people.

Most people don't faceroll when gaming.

Humans have 10 fingers generally and are limited by that factor.

USB is limited to 6KRO for 1 instance of a Keyboard device in memory via driver.
The only way to break the 6KRO limit is to instantiate another instance of the same keyboard within driver.
So you really only need 2 instances of the same USB keyboard in driver memory via the keyboard drivers.

Wouldn't all that extra traces to gurantee anything past 12KRO be pointless and a waste of resources for the company?
Wouldn't the extra cost be past on to the customer and be a wasted feature?


I guess if you go by bigger = better marketing, some people will be fooled by the 64KRO, but in reality, if you implemented 12KRO, wouldn't that be more than enough to satisfy anybody that is logical and not fooled by marketing?


Any thoughts from the folks here?

[inteli722]

Because Faceroll

[jabar]

The only way to break the 6KRO limit is to instantiate another instance of the same keyboard within driver.

No. Full-speed USB can yield NKRO without the USB hub hack.

[JPG]

Well it's probably marketing, but NKRO over usb is possible. It's true that at some point, it's meaningless, but if we omit the usb difficulty, NKRO is not that hard to achieve. The model F is NKRO, it's 30+ years old. If we were not buying cheap crap, we would not even ask ourselves if our keyboard is NKRO, it would just be by default.

[osi]

These guys appreciate it.

[quickcrx702]

Why stop at 64KRO?  Why not 128KRO?  Or 256KRO?  Who cares if there aren't that many keys on a board, put some keys on the bottom!!!  LOL why do they make titanium sporks?...because they can, and people will buy that ****.

[damorgue]

The problem I see with 64KRO is that they might as well have gone for NKRO. You will need diodes to achieve this anyway, which means that the matrix supports NKRO, so why put in an arbitrary limit?

[Kamen Rider Blade]

The problem I see with 64KRO is that they might as well have gone for NKRO. You will need diodes to achieve this anyway, which means that the matrix supports NKRO, so why put in an arbitrary limit?

I concur

[TheSoulhunter]

Why stop at 64KRO?  Why not 128KRO?  Or 256KRO?  Who cares if there aren't that many keys on a board, put some keys on the bottom!!!  LOL why do they make titanium sporks?...because they can, and people will buy that ****.

Well, sporks are actually nice for camping/field-trips/etc. and titanium is lighter than stainless-steel but more rigid than aluminum.

[Soarer]

I need a titanium spork.

Correction: I'd like a titanium spork.

Clarification: I'd pay money for a titanium spork.

Conclusion: OP asked the wrong question.

[Hak Foo]

http://www.trekt.com/3414-Snow-Peak-Titanium-Spork.aspx?zmam=41843225&zmas=1&zmac=2&zmap=63267&gclid=COj6iqOOrrwCFdJhfgodbkoAmQ

There you go.  Nine dollars.

Seriously, I thought USB used a series of "this is which key is down" messages, while PS/2 used a "key pressed/key released" messages.

Maybe their controller firmware is only capable of queuing up enough messages to support 64 keys down at one time.

[Grimey]

http://www.trekt.com/3414-Snow-Peak-Titanium-Spork.aspx?zmam=41843225&zmas=1&zmac=2&zmap=63267&gclid=COj6iqOOrrwCFdJhfgodbkoAmQ

There you go.  Nine dollars.

Have two of those for my lunch kit at work, no regrets.  The cafeteria in my building does not have real plates/cutlery so I went with a camping set (2 for when the wife joins me) as to avoid disposable plasticky wares.  Also there is a whole flight of stairs between me and food, better get the lightest stuff possible, or perhaps just overpayed programmer... you decide.

[IvanIvanovich]

They are attempting to make it more futureproof...

for when upgrades to humans become available.

[Findecanor]

The only way to break the 6KRO limit is to instantiate another instance of the same keyboard within driver.

No. Full-speed USB can yield NKRO without the USB hub hack.

That's right. The keyboard declares the format of its report packets in a "Report Descriptor", and that format supports also true N-key rollover to be expressed.
I would like to see a dump of Trigger Z's 64KRO keyboards' report descriptor, so that I can see what is wrong with it, what is making it not work with Mac.

The problem with NKRO is not to express it, but the problem is with BIOS:es. They require that the keyboard supports a special "boot protocol", and it is this protocol that is 6KRO. The BIOS is supposed to tell the keyboard to switch protocols, but many common BIOS:es don't even do that.
Most keyboards talk only this protocol, and therefore they also define this protocol in their report descriptors.
Having to support two protocols adds complexity.
My guess is that CM Storm is trying to get around bad BIOS:es by talking the 6KRO boot protocol and also send additional report packets that extend them.

[Soarer]

My guess is that CM Storm is trying to get around bad BIOS:es by talking the 6KRO boot protocol and also send additional report packets that extend them.

Haha, I wouldn't go as far as ascribing intelligence in the design,  it is CM after all

[Linkbane]

They are attempting to make it more futureproof...

Show Image

for when upgrades to humans become available.

Finally, we can make use of the 'useless' 1 ms polling..

[hasu]

Just question.
Can low speed USB handle NKRO?
I mean, more than 64KRO. Does low speed packet size(8byte) limits to 64KRO?
Or we can achieve 128KRO/256KRO with 16/32 bytes report even on low speed?

[Soarer]

Just question.
Can low speed USB handle NKRO?
I mean, more than 64KRO. Does low speed packet size(8byte) limits to 64KRO?
Or we can achieve 128KRO/256KRO with 16/32 bytes report even on low speed?

Sort of, yes. You can have a report that spans multiple packets. But then you reduce the response rate to poor levels, even for typing.

Using 64 bits would represent 64 keys, which would be enough for the (base layer) keys of a Poker, for example. But, of course, the secondary layer counts as extra keys! So that doesn't really work unless you implement all the layer functions on the host computer.

64KRO is a really strange number for them to come up with. I can't think of a logical reason for it!

[Grendel]

Although USB specs limit the INT EP payload to 8B, you can do multiple transfers in a row until you you got everything across. It's slow as pointed out, you probably could do slightly more efficient NKRO w/ multiple 7B reports (+ 1B report ID.)

[dorkvader]

They are attempting to make it more futureproof...

Show Image

for when upgrades to humans become available.

Finally, we can make use of the 'useless' 1 ms polling..

Useless? I quite disagree! 1ms polling is quite useful and shows amazing forthought.

Despite the limitations of the cherry MX switch design (and most contact based switches) polling a matrix at 1ms or faster is quite easy and effective to do on hall effect switches (and possibly capacitive). Some hall sensors are even designed for a few hundred KHz scanrates, for much-fater-than-1ms poll rate.

Now they can support highly superior keyswitch designs with only minimal changes to the controller code.

[Linkbane]

They are attempting to make it more futureproof...

Show Image

for when upgrades to humans become available.

Finally, we can make use of the 'useless' 1 ms polling..

Useless? I quite disagree! 1ms polling is quite useful and shows amazing forthought.

Despite the limitations of the cherry MX switch design (and most contact based switches) polling a matrix at 1ms or faster is quite easy and effective to do on hall effect switches (and possibly capacitive). Some hall sensors are even designed for a few hundred KHz scanrates, for much-fater-than-1ms poll rate.

Now they can support highly superior keyswitch designs with only minimal changes to the controller code.

In the future, everyone goes back to PS/2 or hall effect, because ten fingers aren't good enough.  :p
I'm sure that some modification to the brain will be required to allow the use of three fingers and the ability to think quickly enough to use all of them, eh?

[hasu]

Hmm, responsiveness... I see. Report ID idea seems to be nice, though 1 byte out of 8 is not trivial.
I like to try this someday when I have spare time.

Thanks.

[BucklingSpring]

The problem I see with 64KRO is that they might as well have gone for NKRO. You will need diodes to achieve this anyway, which means that the matrix supports NKRO, so why put in an arbitrary limit?

The answer is easy. Once you sell NKRO, you're done.

But If you sell it by increment... You can stretch the sales by selling better products at each iteration.

Moore's law is not a prediction. It's a conspiracy. 

We consumers are getting technology enhancements at a pace much slower than science advancements simply because we are welling to pay big money just for a slight improvement.

[quickcrx702]

Why stop at 64KRO?  Why not 128KRO?  Or 256KRO?  Who cares if there aren't that many keys on a board, put some keys on the bottom!!!  LOL why do they make titanium sporks?...because they can, and people will buy that ****.

Well, sporks are actually nice for camping/field-trips/etc. and titanium is lighter than stainless-steel but more rigid than aluminum.

But in the grand scheme of things, it doesn't take significantly more space to pack both a spoon and a fork, and a spork sucks at both functions.  I'd rather have a spoon for soup and a fork for meat.  That said, I'm not going to lie... I thought it was an awesome product, but I can't justify buying one LOL.