deskthority

Hi guys!

I have some questions understanding keyboards but many websites makes my head not see things very well.

Basically I want the most compact, tenkeyless, mechanical, most light weight pressable, most symmetric and technical best keyboard as possible.

Today, I have tried Keycool and KBT race. The 75% layout is a must, I must have function keys or other keys in top.

I recently saw DSI modular mac keyboard, but not work for windows.

My goal with this thread is to make comparison of the technical specifications of:

Keycool 84

Noppoo Choc mini 84

KBT race keyboard.

I have both the ABS and PBT keycaps selected with various silencer for use and testing.

First:

All 3 keyboards claims to have NRKO, but the KBT race keyboard had another function. When I press and hold 2 keys, e.g "f" and "g", it would repeat them like "fgfgfgfgfgfgfgfgfgfg", but if I use keycool 84, it would repeat like "fggggggggg" Does noppoo choc mini have "NRKO", which can repeat multiple keys when held down? Also, the KBT race only has the ability to register 6 keys at a time, whereas keycool has unlimited. However that is not useful to me becuase I use only 5 fingers on keyboard when I work.

I currently use keycool 84 because I like the placement of the function keys, linear, vertically directly above the number keys, instead of slightly diagonally.

This brings me to the inevitable conclusion that the noppoo choc mini is almost identical to keycool 84, but might have a better NRKO (for example repeating) as well as build quality. I do like if the keyboard is a bit more soundy so I have a better feeling of how much force i put into keys, and I would also prefer to know if there are stabilizing/build differneces between keycool and noppoo choc in terms of the pressing effect: how fast and soundy it gets.

Thank you.

True NKRO won't cause "fgfgfg". If you hold two keys at once, the computer will always see one of them first, the other one second. Typematic repeat will always repeat the last key to be pressed, so you'll get "fggggg". This behaviour is by design and this is what you'll see on any computer running any OS and with any keyboard, 2KRO, NKRO or otherwise.

USB on the other hand, is by convention 6KRO — you can go higher (I think the new Tactile Pro is 10KRO) but convention and compatibility dictate 6KRO, as one USB data block only has room for six keys plus modifiers.

"fgfgfg" is what happens if you plug in two keyboards, and hold f on one, and g on the other. What you're seeing is a keyboard that is pretending to be two or more keyboards, and it's sending 'f' from one of its virtual keyboards, and 'g' from the other. Two virtual keyboards gives you 12KRO (2 × 6KRO), three gives you 18KRO and so forth. For whatever reason, when you connect two keyboards and hold one key on each, typematic repeat doesn't work properly: it repeats the last key held on each keyboard. I guess Microsoft never wasted time figuring out how to handle something that, to most hardware and software developers, would be a ridiculous scenario, but there is a genuine reason for it, and thus you'll see unexpected behaviour as a result.

Ouch! That's what happens when manufacturers try a dirty hack.

Two things about NKRO.

First: if you don't need it, Select, you don't need it. 6KRO may sound lame but that's actually a real mess of keys when typing. I never need as many as that either. Some gamers say they do, but I doubt they're getting much extra control mashing so many keys all at once! Sounds more like keyboard Twister. NKRO is a nice idea in theory, but rarely if ever required in practice. And the lack of NKRO is no hint at all that a keyboard is worse in any other way.

Second: full NKRO over USB is quite possible without reverting to multiple virtual keyboard hacks. Soarer pulls it off with his converter. My Model F sends more keys than I could ever hope to contrive a purpose for, and on a Mac, too; typically uninvited to the NKRO hack party. He does it using "the bitmap technique", which I only understand as a separate USB HID mode entirely. It works perfectly with just one exception: not recognised by every BIOS out there. Not an issue on Macs, but I did trip into it once while trying to boot a PC from a USB drive. Sometimes that PC sees my Soarer keyboard at boot, sometimes not, but it's 100% as soon as an OS is loaded. Macs, iPads, etc. are all fine with it, period.

Daniel:

Is multiple virtual keyboards worse/better than standard NRKO? I mean, what is the difference?

muirium: is there a way I can fake/artificialize a way to get multiple virtual keyboards with full NKey rollover?

Daniel Beardsmore wrote: "fgfgfg" is what happens if you plug in two keyboards, and hold f on one, and g on the other.

Hmm.. That doesn't happen under Linux. Only the last pressed key is counted.

BTW, A curious thing, though, is that when I press and hold "f" on one keyboard, and then press and release "f" (the same key) on the other keyboard, the repeat stops.

So what is the rollover difference between noppoo choc mini and keycool 84?

To get NKRO, first of all the keyboard's matrix has to support it in hardware. No magic can bring it back if it's not there at the start.

Then it's down to the controller (the chip inside the keyboard that reads the matrix and sends it out over USB or PS/2 or whatever else). It can typically send along full NKRO if it's using PS/2. USB is the puzzling part. USB is meant to be 6KRO. There are two ways around it that I know.

The multiple virtual keyboard method is pretty obvious. The keyboard manufacturer builds a hub into the keyboard and connects several fake controllers to it. (I assume this is all just done in the one chip.) Your computer sees several "keyboards" as soon as you plug it in. You have seen the downside: the computer gets easily confused about "which" virtual keyboard is telling it what keys are pressed. I consider this a dirty hack.

The other way is the "bitmap" technique Soarer uses, and Hasu developed I think. I don't know any keyboards that use it. But our own inventions do: custom controllers and converters.

So the USB controller chip in the keyboard is in the driving seat. That's where the decision is made how to handle NKRO over USB. I use older keyboards with a converter which acts as the USB controller instead. That's how I get NKRO without the hassle of virtual keyboards.

I read a comment here about Noppoo choc mini (comment number 2):

http://www.amazon.com/Noppoo-Mechanical ... B00A45WK4U

It states:

"Most keyboards don't have usb n-key rollover because the protocol was shortsightedly made without it. This keyboard gets around it by having a virtual usb hub with multiple virtual keyboards. This doesn't work very well with macs because they treat keyboard layout individually per board rather than system wide. There is either a driver online or you could use a less than ten dollar usb > ps2 > usb adapter. I don't see this as a problem against the rating because they never implied full macintosh compatibility."

Does this mean if I hold multiple keys simultaneously its gonna do "dsgfdsgfdsgfdsgfdsgfdsgfdsgf" instaed of "dsgffffffffffffffffffffffffffffffffffff"?

Thank you.

It's using the virtual keyboard trick, so yes. Unless you use PS/2.

Findecanor wrote:Hmm.. That doesn't happen under Linux. Only the last pressed key is counted.

As I suggested, it's likely a Windows design issue. It's not Microsoft's fault — it's not something any system designer is likely to anticipate (why would anyone be trying to type on two keyboards at once anyway?) As for pressing F on a second keyboard, with PS/2, it will be because the first keyboard sent F keydown followed by F repeat, and the second one sent F keydown then F keyup, terminating the repeat sequence (I guess the continued F repeats from the first keyboard get ignored). I don't know how USB deals with keyup/down, though.

I remember that Lemmings on the Amiga let you connect two mice at once and each player would get their own mouse cursor in two-player mode. Not possible on any other system I've used.

Select wrote:Daniel:

Is multiple virtual keyboards worse/better than standard NRKO? I mean, what is the difference?

It's an ugly hack, and to what end? Gamers who have one hand on the mouse don't need NKRO unless their genetics have afforded them a whole seven digits on their other hand and they're not using any modifier keys, which don't count towards the 6-key limit anyway. Considering that even the IBM Model M is only 2KRO, then NKRO is of dubious value to typists.

I never really understood all the NKRO fuss. The only time it was ever a problem was on my 486 PC when we used to play two-player action games where both players shared the same keyboard. That used to cause ghosting, which was confusing and annoying.

Ironically, UEFI on my OptiPlex 9010 hangs my Filco MJ1 (USB with standard 6KRO) — the keyboard freezes mid-POST (with all the LEDs lit) until Windows completes loading. No such problems with a cheapo Acer 6312-TW 2KRO membrane keyboard. I've heard of other similar problems with Filcos, so I do suspect there's a fault with the controller on them.

The reason Im asking about NRKO is cuz i like the virtual keyboard repeat function.

I might wanna get my hands on a noppoo choc instead of a keycool 84 then. Any major differences I should be aware off?

- backlighting doesnt matter
- sound difference doesnt matter much (can always dampen spacebar
- I got fitting keycaps

Select wrote:The reason Im asking about NRKO is cuz i like the virtual keyboard repeat function.

You whhahhhahaahhaaahahahahahhahhhahhhaaaahahahahhhahahahahahhahahhhhhaaat?

Not sure i follow

The key repeat behaviour is an artefact of how Windows handles multiple keyboards. It's not a feature of the keyboard. FWIW, Windows 8 still has the same behaviour.

Muirium wrote:The other way is the "bitmap" technique Soarer uses, and Hasu developed I think. I don't know any keyboards that use it. But our own inventions do: custom controllers and converters.

dfj has (low-speed?) NKRO, 9 June 2010
dfj has NKRO, 20 July 2010
Soarer has NKRO, 1 December 2010
hasu has NKRO, 2 December 2010
Soarer works around bad BIOSs, 5 December 2010

Objection sustained!