Are there any USB NKRO keyboards on the market yet?

[Twinky]

Hello,

I've been searching for a commerical NKRO USB keyboard and haven't been able to find one yet. Is there a list of USB keyboards with NKRO?

Also, is there a comprehensive spreadsheet of keyboards with information like what kind of switches it uses, #KRO, and so on?

Cheers!
Twinky

[jabar]

I'd say the most readily available NKRO over USB keyboards are those of the Ducky brand; not sure if that encompasses all their models. Other brands such as Noppoo feature NKRO as well but don't have real US distribution.

[TotalChaos]

All the NKRO USB keyboards I have read real reviews of have said they are all flaky depending on what OS you use.

So basically you can forget about it for anything in the universe except Windoze XP, 7, 8.  It might not work on all 3 of those but at least one of them should work.

Plug your keyboard into any other computer and it will usually just go bonkers.

If you can wait 10 more years they might make us USB3.0 keyboards which could support NKRO without doing illegal tricks that only work by haxxing one OS and not the other.

In the meantime if you want 100% reliable NKRO just plug a PS/2 keyboard into a PS/2 port on your computer.

[damorgue]

They seem to have issues though. Choc uses some dual-device method and appears as two devices. If I remember correctly, this makes it behave strangely in certain OS.

[Soarer]

Even the dirty trick methods work fine except on Macs. Even so, they are NOT "illegal tricks".

No need to wait for USB 3.0 since full-speed USB 1.1 or 2.0 is sufficient to implement a good NKRO method.

The Noppoo and most other NKRO over USB keyboards are low-speed, which is a bit slow for it really. Using PS/2 would be better.

Full-speed is pretty rare on keyboards - the Cooler Master CM Storm Quick Fire Pro and Quick Fire TK are about the only ones I know of (and even for those I don't yet know the details of their implementation).

[davkol]

Tt Meka (that weird 96-key-like keyboard) should be 16KRO over USB. (So is MS SideWinder X4, but it's not mechanical.)

[TotalChaos]

Even the dirty trick methods work fine except on Macs. Even so, they are NOT "illegal tricks".

I assumed it was was illegal since it doesn't work on all devices and the whole concept of 1 device lying and saying its really 2 or 3 devices sounds dodgy. 

So it is 100% legal for 1 USB device to lie and say that it is really 2 or 3 devices?

And it is just that MacOS is lame and has a badly written or buggy USB Stack?

But MacOS is based on Linux and I assume uses lots of linux drivers for things, so does linux have these same bugs when using NKRO USB devices?

How about PS3 or other game consoles?  Does NKRO USB work on them?

[Soarer]

Tt Meka (that weird 96-key-like keyboard) should be 16KRO over USB. (So is MS SideWinder X4, but it's not mechanical.)

I have one guess where you get 16KRO from, and it's not a reliable source!

TT Meka says it uses full-speed too, so there's no reason to assume anything other than full NKRO.

The X4 used full-speed, but it used a very strange configuration for its reports.

[Soarer]

Even the dirty trick methods work fine except on Macs. Even so, they are NOT "illegal tricks".

I assumed it was was illegal since it doesn't work on all devices and the whole concept of 1 device lying and saying its really 2 or 3 devices sounds dodgy. 

So it is 100% legal for 1 USB device to lie and say that it is really 2 or 3 devices?

And it is just that MacOS is lame and has a badly written or buggy USB Stack?

But MacOS is based on Linux and I assume uses lots of linux drivers for things, so does linux have these same bugs when using NKRO USB devices?

How about PS3 or other game consoles?  Does NKRO USB work on them?

I've corrected you on whether it's illegal before

MacOS is not based on Linux )

Lots of USB devices appear as multiple devices, even bog standard 6KRO keyboards if they also have media keys and/or power keys, or an inbuilt trackpad etc. Strictly speaking in USB terminology one would say that one device supports multiple interfaces, but in Windows etc. those interfaces show up as devices.

The difference between Macs and the rest is simply an arbitrary choice in the OS about whether multiple keyboards are treated as completely independant or not. I'd say that the Mac gets it wrong here - there's no point treating them as independant if you can't also have independant app focus for each keyboard (i.e. multi-user capable). For example, if you plugged in both a USB keyboard and a USB numpad to a Mac, you wouldn't be able to use the main keyboard's modifiers with the numpad's keys!

I dunno about PS3 etc, but probably yes, since the Mac is an oddball.

There are theoretically some ways of doing NKRO over low-speed that don't even use multiple devices. Can't recall if any keyboards use them.

Bottom line is that NKRO doesn't have any such issues if using full-speed and done right (and it's easier to do it right than wrong).

[TotalChaos]

Even the dirty trick methods work fine except on Macs. Even so, they are NOT "illegal tricks".

I assumed it was was illegal since it doesn't work on all devices and the whole concept of 1 device lying and saying its really 2 or 3 devices sounds dodgy. 

So it is 100% legal for 1 USB device to lie and say that it is really 2 or 3 devices?

And it is just that MacOS is lame and has a badly written or buggy USB Stack?

But MacOS is based on Linux and I assume uses lots of linux drivers for things, so does linux have these same bugs when using NKRO USB devices?

How about PS3 or other game consoles?  Does NKRO USB work on them?

I've corrected you on whether it's illegal before

If so, I sincerely apologize.  Either I totally forgot or I prob just never actually went back to the thread to read the msg.  I have been wondering for a long time what the real story is.

To summarize:
Using USB 1.1 it is totally LEGAL for a keyboard to represent itself using multiple "interfaces" ("devices" under M$)

Then the trouble is that it is a matter of opinion whether the modifiers of one "interface/device" should modify the keys of another "interface/device".  Some OSes implement it one way and other OSes implement it the other way.

Do u concur?


 

[davkol]

But MacOS is based on Linux and I assume uses lots of linux drivers for things, so does linux have these same bugs when using NKRO USB devices?

It's not, never was. Original MacOS had nothing to do with Linux/Unix/whatever, only OS X is partly based on NeXTSTEP and Mach kernel, it's also a certified UNIX (Linux is not).

Tt Meka (that weird 96-key-like keyboard) should be 16KRO over USB. (So is MS SideWinder X4, but it's not mechanical.)

I have one guess where you get 16KRO from, and it's not a reliable source!

TT Meka says it uses full-speed too, so there's no reason to assume anything other than full NKRO.

I guess they claimed something like that on their website, but it appears to be only 6KRO@USB on my system.

[Soarer]

Even the dirty trick methods work fine except on Macs. Even so, they are NOT "illegal tricks".

I assumed it was was illegal since it doesn't work on all devices and the whole concept of 1 device lying and saying its really 2 or 3 devices sounds dodgy. 

So it is 100% legal for 1 USB device to lie and say that it is really 2 or 3 devices?

And it is just that MacOS is lame and has a badly written or buggy USB Stack?

But MacOS is based on Linux and I assume uses lots of linux drivers for things, so does linux have these same bugs when using NKRO USB devices?

How about PS3 or other game consoles?  Does NKRO USB work on them?

I've corrected you on whether it's illegal before

If so, I sincerely apologize.  Either I totally forgot or I prob just never actually went back to the thread to read the msg.  I have been wondering for a long time what the real story is.

No worries, I think it was just a small part of a discussion that was more about CPU loads etc.

To summarize:
Using USB 1.1 it is totally LEGAL for a keyboard to represent itself using multiple "interfaces" ("devices" under M$)

Then the trouble is that it is a matter of opinion whether the modifiers of one "interface/device" should modify the keys of another "interface/device".  Some OSes implement it one way and other OSes implement it the other way.

Do u concur?

That's pretty much it, yeah. The USB & HID standards say nothing on the matter of opinion.

I suppose with the example of the numpad, you might not want it merged with other keyboards - an example of that would be a IBM Space Saver + seperate numpad, where you wouldn't want the num-lock state to apply to the Space Saver.

The only truly correct way to handle it would be with options in the OS, so you could choose to build a virtual keyboard out of multiple devices.

HID did a great job of sorting out the mess of extensions for mice, joysticks, and other controllers, but apart from standardising media and power keys it really didn't move keyboards forward at all (with respect to keycodes etc).

[TotalChaos]

@Soarer

Do u know, or provide a link that gets the Soarer seal of quality approval, for explaining multimedia keycodes over PS/2 ?

A friend of mine was coding some sort of util for using multimedia keys and he was wondering about and I had no idea what to say.  I assume they just emit a keycode like any other key?

[Soarer]

Tt Meka (that weird 96-key-like keyboard) should be 16KRO over USB. (So is MS SideWinder X4, but it's not mechanical.)

I have one guess where you get 16KRO from, and it's not a reliable source!

TT Meka says it uses full-speed too, so there's no reason to assume anything other than full NKRO.

I guess they claimed something like that on their website, but it appears to be only 6KRO@USB on my system.

Oops, you're right, they do, sort of...

Anti-Ghosting Keys    16 (USB)

Strange way to write it, but I'd assume they mean KRO. After all, you don't need anti-ghosting if you've built the thing with diodes - it simply doesn't have ghosting!

[tufty]

Oh?  Can the holy and ubiquitous PS/2 "standard" not handle multimedia keys in a consistent way?

[damorgue]

I don't know about "standard", but I have some fairly old Cherry boards which are able to output media buttons over PS/2.

[Soarer]

@Soarer

Do u know, or provide a link that gets the Soarer seal of quality approval, for explaining multimedia keycodes over PS/2 ?

A friend of mine was coding some sort of util for using multimedia keys and he was wondering about and I had no idea what to say.  I assume they just emit a keycode like any other key?

Trouble is, a lot of PS/2 keyboards would've used non-standard codes and made you load a driver

Try Microsoft's translate.pdf. That shows mappings between USB HID and scan code sets 1 and 2 (set 2 used on the wire, but set 1 used inside the OS). Whether PS/2 keyboards actually put out those codes, I don't know.

[JPG]

I read that Max keyboards offer 20KRO over USB. If you type 20 keys at the same time, your typing with your ass, literally

Dunno if these keyboards are good, just saying they offer the 20KRO over USB.

[Soarer]

I read that Max keyboards offer 20KRO over USB. If you type 20 keys at the same time, your typing with your ass, literally

Dunno if these keyboards are good, just saying they offer the 20KRO over USB.

Smartass. Well their FAQ claims full NKRO. It does matter which since for anything that is >6KRO, full NKRO has a neater (and hence usually more reliable) implementation (even when using low-speed USB).

[urbanus]

full NKRO has a neater (and hence usually more reliable) implementation (even when using low-speed USB).

Could you please explain?

[Terrorsathan]

NKRO is nice and all, but I'm a pretty avid gamer, and have never found myself in a situation where usb 6 key roller wasn't enough...  I even have a PS2 port and adapter I can use but I just don't bother.


I think its more of an e-peen thing

[razorsharpgears]

Yeah I hope in the future they add an NKRO option for USB, because when I bought my motherboard I decided to buy one without a PS/2 option.
And now I am engulfed in this KeyBoard hobby, unable to get full NKRO.    :cries: 

[Soarer]

full NKRO has a neater (and hence usually more reliable) implementation (even when using low-speed USB).

Could you please explain?

It's neater because once you get up to 13 keys + modifiers, you can fit 112 keys (i.e. plenty!) in the same space by using one bit per key instead of a byte. But typically, if you increase the data sent from 8 bytes, you may as well double it for 128 keys worth.

It's usually more reliable because the code handling it is far less likely to have any bugs (both in the keyboard and the OS), just since it's so much simpler. Modifiers no longer need to handled separately, for one thing. And notably, there was a bug in all versions of Windows up to Win7 SP1 where if you press the maximum number of keys plus a couple of modifiers, the last key will get kinda stuck (as in, 6 keys plus modifiers if using a 6KRO keyboard, 12 keys for a 12KRO keyboard etc).

[missalaire]

The Cooler Master QuickFire TK has USB NKRO.

http://www.cmstorm.com/en/products/keyboards/quickfiretk/

[inteli722]

The Cooler Master QuickFire TK has USB NKRO.

http://www.cmstorm.com/en/products/keyboards/quickfiretk/

As does the QuickFire Pro (I don't know about the Rapid)

[TotalChaos]

The Cooler Master QuickFire TK has USB NKRO.

http://www.cmstorm.com/en/products/keyboards/quickfiretk/

They claim it uses USB 2.0 but then they say it doesn't work on the MAC.

So either they are lying about the USB 2.0 thing or Soarer is just plain wrong when he says that using USB 2.0 solves all the NKRO problems of USB keyboards.

This whole thing makes no sense. 

[inteli722]

The Cooler Master QuickFire TK has USB NKRO.

http://www.cmstorm.com/en/products/keyboards/quickfiretk/

They claim it uses USB 2.0 but then they say it doesn't work on the MAC.

So either they are lying about the USB 2.0 thing or Soarer is just plain wrong when he says that using USB 2.0 solves all the NKRO problems of USB keyboards.

This whole thing makes no sense. 

When they say it doesn't work on Macs, it's a limitation of the OS, not of the hardware.

[Soarer]

No no no...

USB 2.0 / full-speed doesn't magically make any implementation using it good, it just makes it possible. And relatively easy. So it's really not a good sign if they haven't managed to be Mac compatible.

[TotalChaos]

@Soarer

So u agree that it really has USB 2.0 and the MAC incompatibility is just a bug in MAC OS?

[Soarer]

I'm assuming it has USB 2.0; such a claim is nearly always correct. (However the claim that the Rapid has 1000Hz USB polling is incorrect, since it only uses low-speed).

I wouldn't call it a bug in OS-X. It's perfectly possible to write NKRO implementations that work with it, and using full-speed means there's no downsides (with low-speed you'd have to go down to a 62.5Hz polling rate, I think).

[TotalChaos]

Is the methodology of making a USB 2.0 keyboard work on OSX publicly available information for free?

Or do keyboard companies have to pay a bribe to Apple for the top secret infos?

[Soarer]

It's free; it's all in the HID spec. They just have to implement the main keyboard as a single interface (media and power are always separate).

Who knows - maybe they just say that Mac is incompatible because they couldn't be bothered to test, or would have to pay Apple for conformance testing or somesuch. Possibly it works fine anyway.

Someone should dump the report descriptors so we can find out what it's doing.

[metalliqaz]

My Leopold FC200R has N-key rollover with USB, verified at http://www.microsoft.com/appliedsciences/content/projects/AntiGhostingExplained.aspx

[hasu]

Shame. Manual says Quck Fire TK does NOT support NKRO on Mac.
http://www.coolermaster.com/cmstorm_files/QuickFireTK_User_Guide_EN.pdf

Cooler Master did failure of 'Noppoo choc mini' again? Why didn't they just use plain HID descriptor!? I believe plain HID descriptor can  circumvent  said Mac 'bug'.

Soarer, I'd call it a bug or laziness of Apple but I don't blame them on this harshly.
In most case we won't need that complicate descriptor, I guess.

[Soarer]

Also, "NKRO ... prevents some PCs from booting" - they need my trick of running both 6KRO and NKRO at the same time

It's good that the NKRO can be switched on and off, I suppose.

My guess then is that they started with something like Noppoo code, running at low-speed, and only really used full-speed to add the '1000Hz polling' to the spec sheets. Shame indeed.

[inteli722]

Also, "NKRO ... prevents some PCs from booting"

Maybe that's why the Quickfire Pro I bought my dad isn't getting recognized. Anyone know if it remembers what mode it's in?

[Oobly]

Also, "NKRO ... prevents some PCs from booting" - they need my trick of running both 6KRO and NKRO at the same time

It's good that the NKRO can be switched on and off, I suppose.

My guess then is that they started with something like Noppoo code, running at low-speed, and only really used full-speed to add the '1000Hz polling' to the spec sheets. Shame indeed.

I am interested in this running both 6KRO and NKRO at the same time. How would you implement it? As far as I know, for a keyboard to support boot protocol (so you can enter BIOS with it, etc.) it must implement 6KRO over low-speed. After boot you can use full-speed and NKRO, or if you don't care about the keyboard being usable at boot, just implement full speed and NKRO. But most keyboards will need to be recognised at boot for PC's with BIOS. Otherwise if your CMOS gets wiped from a power issue for instance, you can't go into BIOS to fix it unless you have another KB available that supports boot protocol.

Only solution I can see is having it switchable.

[Soarer]

BIOSs are happy with 6KRO over full-speed. If they didn't then you'd be right; the only way to work would be to switch completely from one speed to the other.

BIOSs should issue a SetProtocol command to make the keyboard use 6KRO, but some don't. That's key. If they all complied with the spec, a firmware writer could simply follow the guidelines in the spec to implement 6KRO for the BIOS, and then whatever they liked for the OS.

The 'trick' is to keep a 6KRO report running always, and have its descriptor tell the OS that it contains no keys, so that the OS ignores it. BIOSs see that it has the 'boot keyboard' flag set, so they see the keys in it fine (they ignore descriptors). A second endpoint provides the NKRO keyboard which doesn't have the 'boot keyboard' flag set, so the BIOS ignores it. Where a BIOS does issue SetProtocol, I suppress output on all but the 6KRO endpoint. It's all perfectly compliant with the spec

[C5Allroad]

All the NKRO USB keyboards I have read real reviews of have said they are all flaky depending on what OS you use.

So basically you can forget about it for anything in the universe except Windoze XP, 7, 8.  It might not work on all 3 of those but at least one of them should work.

Plug your keyboard into any other computer and it will usually just go bonkers.

If you can wait 10 more years they might make us USB3.0 keyboards which could support NKRO without doing illegal tricks that only work by haxxing one OS and not the other.

In the meantime if you want 100% reliable NKRO just plug a PS/2 keyboard into a PS/2 port on your computer.

My ducky works fine both on my PC and Mac.
Ducky seems to have the most reliable one.

[Oobly]

That's genius!

Great solution! I wish some hardware vendors would start to do this. What boards have you implemented this on? I am building a custom firmware for my own DIY board using a Teensy. Do you mind if I try out some of your concepts on it?

[Soarer]

It's in my converter, of course. And will be in my soon-to-be-released controller firmware, which has been tried on 4 keyboards so far (Retrete's Kevex, my mini, an IBM with Alps, and IBM bigfoot).

Sure, feel free to try these things out, they're not secrets

[Lu_e]

WASD keyboards also has their; 'Modified 6-Key Rollover over USB allows additional key presses after 6 keys to be recognized'

The V2's firmware will still only allow 6 keys to be recognized as being held down at the same time, but pressing up to 12 keys simultaneously will allow every key single to be sent, but only the last 6 keys will still be seen as being held down. Holding down and pressing additional keys one at a time will continue to allow new keys to be recognized.

In both the V1 and V2, Ctrl, Alt, and Shift keys are not counted in the key limit. So you can hold down Ctrl+Alt+Shift+6 other keys.

What say you on their implementation? Presses vs holding down? Is press vs holding a factor on older or custom controllers of today? Since there are 2 actions to the signal I think? switch on, switch off

[Melchior]

I think what you mean by NKRO USB is the USB HID large report size.

Where each report of key-presses was represented by ~16 (not the exact number...) byte reports.

There was thread about this here not long ago... my how time flies and knowledge is easily forgotten.

Also USB NKRO with a Coolermaster Quick-fire causes all kinds of wierd inexplicable issues (Windows 7 no less) - Inverted Insert Key State Anyone?

Also - I have seen the whole BIOS freak-out issue with multiple endpoint-devices; strangely enough with a mouse (that is also a keyboard), It just an temporary annoying quirk...

[Soarer]

WASD keyboards also has their; 'Modified 6-Key Rollover over USB allows additional key presses after 6 keys to be recognized'

The V2's firmware will still only allow 6 keys to be recognized as being held down at the same time, but pressing up to 12 keys simultaneously will allow every key single to be sent, but only the last 6 keys will still be seen as being held down. Holding down and pressing additional keys one at a time will continue to allow new keys to be recognized.

In both the V1 and V2, Ctrl, Alt, and Shift keys are not counted in the key limit. So you can hold down Ctrl+Alt+Shift+6 other keys.

What say you on their implementation? Presses vs holding down? Is press vs holding a factor on older or custom controllers of today? Since there are 2 actions to the signal I think? switch on, switch off

First thought - I'm not sure it provides any benefit. Maybe I'll think of one at some point, but if you want to hold more than 6 keys you're not getting what you want whichever keys aren't held!

For the implementation, it's a simple choice what to do when another is pressed when 6 keys are already held down - either block the new key down, or release the key that's been pressed the longest. The code is pretty simple for either case.

(You have to remember that the 6-key USB report is trying to convey the state of the entire keyboard, not events).

[terran5992]

Quick fire tk is nkro

[Linkbane]

All the NKRO USB keyboards I have read real reviews of have said they are all flaky depending on what OS you use.

So basically you can forget about it for anything in the universe except Windoze XP, 7, 8.  It might not work on all 3 of those but at least one of them should work.

Plug your keyboard into any other computer and it will usually just go bonkers.

If you can wait 10 more years they might make us USB3.0 keyboards which could support NKRO without doing illegal tricks that only work by haxxing one OS and not the other.

In the meantime if you want 100% reliable NKRO just plug a PS/2 keyboard into a PS/2 port on your computer.

This might be true several years ago, but you're really misrepresenting the facts. Several keyboards these days support USB over NKRO, including the CM QF TK, QFR Stealth, Noppoo series, and several others. Both the CM keyboards that I have used have worked with NKRO on any Windows computer, and they apparently don't work on Mac. But good luck finding a modern Mac with a PS/2 port still on it. The most commonly available ones would be the Cooler Master keyboards, but I'm sure that there are plenty of others which I do not mention.

[Findecanor]

The 'trick' is to keep a 6KRO report running always, and have its descriptor tell the OS that it contains no keys, so that the OS ignores it.

I suppose that a BIOS that won't SET_PROTOCOL() won't request any Report Descriptors either.

Couldn't you just set the protocol to BOOT_PROTOCOL from the beginning?
When the OS starts up after the BIOS, it couldn't expect the BIOS to have set it to one or the other, so it would always have to set the protocol anyway.

Except Macintosh doesn't have a BIOS ...

[Soarer]

The 'trick' is to keep a 6KRO report running always, and have its descriptor tell the OS that it contains no keys, so that the OS ignores it.

I suppose that a BIOS that won't SET_PROTOCOL() won't request any Report Descriptors either.

I have vague memories of attempting to set the protocol based on whether the report descriptor got read, but I can't remember now if it simply didn't work (i.e. some BIOS read the descriptors regardless), or if I decided it was simply too much of a hack!

Couldn't you just set the protocol to BOOT_PROTOCOL from the beginning?
When the OS starts up after the BIOS, it couldn't expect the BIOS to have set it to one or the other, so it would always have to set the protocol anyway.

Except Macintosh doesn't have a BIOS ...

Unfortunately you can't, because the OS doesn't SetProtocol - it only resets the device and runs through the rest of the initialization. SetProtocol is only ever used to set boot mode (in practice, as far as I know...).

[C5Allroad]

Ducky, Maxkeyboard, I think keycool, Noppoo, And I know some other boards also but I don't the name of them.

[Findecanor]

Unfortunately you can't, because the OS doesn't SetProtocol - it only resets the device and runs through the rest of the initialization. SetProtocol is only ever used to set boot mode (in practice, as far as I know...).

Awww. Thanks

The 'trick' is to keep a 6KRO report running always, and have its descriptor tell the OS that it contains no keys, so that the OS ignores it. BIOSs see that it has the 'boot keyboard' flag set, so they see the keys in it fine (they ignore descriptors). A second endpoint provides the NKRO keyboard which doesn't have the 'boot keyboard' flag set, so the BIOS ignores it. Where a BIOS does issue SetProtocol, I suppress output on all but the 6KRO endpoint. It's all perfectly compliant with the spec

I came across this snippet on Enhanced Keyboards and Windows on Microsoft.com:

However, there is a HID firmware solution that will work with a USB HID-aware BIOS that does not issue the HID Set_Protocol request. The solution is to implement two HID interfaces on separate endpoints in a USB composite-device fashion. One HID interface implements a standard HID keyboard with identical report and boot protocols; the other HID interface implements all the new enhanced keyboard features--for example, the multimedia and web browsing keys, the power keys, and so on. If done this way, the standard HID keyboard on the first interface will always work, regardless of whether the BIOS issues a Set_Protocol request.

But would that work with all BIOS or OS'es? Wouldn't an operating system choose the protocol on the first interface/endpoint and ignore the second? ...