Does Teensy fix the problem of IBM's KRO?

[Shayderov]

That's the question.
Sorry if it is wrong board.

[fohat.digs]

What IBM keyboard and rollover are you referring to?

[Shayderov]

   
IBM Model M PS/2 52G9700 92F0332 and IBM Model M 52G9658 92F0332 1/24/96
I think they have 2KRO, no?

[tipo33]

No,  I'm pretty sure all the IBM BS boards are limited to 2KRO.  There was a project somewhere here to make a new membrane and controler, but I don' t know where it went.

[Shayderov]

No,  I'm pretty sure all the IBM BS boards are limited to 2KRO.  There was a project somewhere here to make a new membrane and controler, but I don' t know where it went.

Is not the controller is responsible for this? :'O
Then it is very very bad. Also, what is price for this two boards, don't you know?

[1391406]

No,  I'm pretty sure all the IBM BS boards are limited to 2KRO.  There was a project somewhere here to make a new membrane and controler, but I don' t know where it went.

Maybe for Model M's(I haven't checked), but my AT Model F has 5KRO.

[alaricljs]

Is not the controller is responsible for this? :'O
Then it is very very bad. Also, what is price for this two boards, don't you know?

The controller is a major point, but if the switch matrix isn't capable of signalling correctly, there's not much the controller can do.

[Shayderov]

No,  I'm pretty sure all the IBM BS boards are limited to 2KRO.  There was a project somewhere here to make a new membrane and controler, but I don' t know where it went.

Maybe for Model M's(I haven't checked), but my AT Model F has 5KRO.

How to check KRO?

[1391406]

No,  I'm pretty sure all the IBM BS boards are limited to 2KRO.  There was a project somewhere here to make a new membrane and controler, but I don' t know where it went.

Maybe for Model M's(I haven't checked), but my AT Model F has 5KRO.

How to check KRO?

Aqua Key Test

[jwaz]

No,  I'm pretty sure all the IBM BS boards are limited to 2KRO.  There was a project somewhere here to make a new membrane and controler, but I don' t know where it went.

Maybe for Model M's(I haven't checked), but my AT Model F has 5KRO.

How to check KRO?

Are you sure it's actually a problem and that you understand how KRO works? The rating comes from the shortest combo of keys and ISN'T a general rule. If someone can type past 150+ wpm on a model M then it shouldn't be a problem, even for high APM SC players.

Have you noticed keys getting missed because of the combos you're pressing? If not I wouldn't worry too much about it. NKRO, like sensors and polling rates, are rapidly becoming often meaningless gaming industry marketing terms.

[Shayderov]

wtf?

[Shayderov]

wtf?

Ex:  I play Call of Duty: I go with the press of "W", then I want to switch to run, If you want to run, you need to   hold "shift", then while I am running, I want to use a grenade
??
I can't

[wcass]

It doesn't work because Shift + W +(whatever grenade throw is) are 3 corners of a square. it might work if you use the other shift key.

As stated before, 2KRO means that the minimum number of keys that are guarenteed to work together is two. All 2KRO keyboards fail to register some key combination > 2, but will also properly register some key combinations > 2 - depending on exactly what keys are pressed at the same time. The reason for this (and how to predict where key presses will be suppressed) is the switch matrix.

For example, look at a very small 2x2 matrix (imagine tic-tac-toe playing field or #) with two “signal” lines (A and B) and two “sense” lines (X and Y). The controller sends signals down two of the lines and listens for the signals on the other two lines. If the controller senses signal-A on sense-X then the key at the intersection of signal-A and sense-X is pressed. But with that switch closed, you must understand that now all parts of sense-X are flooded with signal-A.

While holding that key down, let’s add another key – the one at the intersection of signal-B and sense-X. The controller reports that there are no signals on sense-Y, but sense-X has both signal-A and signal-B (two keys pressed). What is not reported, but very true is that now signal A is being carried on the signal-B line; and signal-B is carried on the signal-A line.

Now, let’s add a third key. Because the signal-A line is also carrying signal-B now and vice-versa, pressing either key on the sense-Y line will register both A-Y and B-Y. The extra key (the one without a finger on it) is what is known as a “ghost” key.

To prevent this, controllers are programmed to ignore keys that will generate a ghost key (and the ghost key with it). Any combination of keys that does not produce a ghost could be reported by the controller (up to the limits of the controller).

To achieve NKRO, each switch (sense/signal intersection) must be isolated with a one-way check valve. Conductive switch keyboards typically use a diode for this, but it is an inherent feature of the IBM Model F capacitive switch (XT, AT, 122). The “M” in Model M stands for membrane (a conductive type switch) and AFAIK no one has made a membrane with diodes because the membrane sandwich must be perfectly flat and diodes aren’t flat. I have managed to put high density diode arrays outside of the sandwich area. You can read about that here:
http://geekhack.org/index.php?topic=40111.0

[alaricljs]

I think MS's x4 has diodes on the membrane... I remember there being a pic.  But that's modern esoteric tech.

[dorkvader]

I thought the microsoft sidewinder used something resistive.
Nevertheless, there does exist at least one early rubberdome keyboard with diodes. Almost all membrane keyboards are limited to 2KRO.

There are two different kinds of KRO to consider. The KRO of the keyboard matrix and the KRO of the controller. In hte case of IBM buckling spring keyboards, only the capacitive ones are capable of NKRO. The membrane ones are capable of 2KRO. This is before the controller decodes the keypress.

So the controller can only have up to as high of KRO as the keyboard matrix can handle.

Now, a lot of modern USB keyboards will support 6 KRO, which is a limitation of the USB technology bandwidth. It can't send any more to the computer, so it doesn't. If it supports PS2, it can achieve NKRO over that, but to maintain compatability with USB1.1 the controller artificially limits KRO to 6.

Another case of the controller lowering a keyboard's potential KRO is with some of the old capacitive buckling spring keyboards. Some of them will only send upcodes for some keys (so many keys can not be held down). In this case, the keyboard matrix supports NKRO, but the controller only has a practical KRO of less than that (depending on the model, and how many keys are designated as "repeat action")

Specifically in your case, your model M's are limited to 2KRO by the keyboard matrix, so changing out the controller won't help that.

In the case of a model F keyboard, a teensy could be used to fix the KRO, since the capacitive matrix supports it.
For more information, you may want to consider reading the following article I found on the subject.
http://www.dribin.org/dave/keyboard/one_html/

[mkawa]

the sidewinder was resistive. basically, they solved the global resistor uniqueness problem for their particular matrix (and/or jimmied with the matrix until it was solvable): each key carries a resistance such that every combination of keys has a unique resistance. the problem with this is actually longevity/stability of the screen printing on the FCB. it's a bit like a capacitive switch in that it's quite difficult to keep stable capacitance values for each key without high manufacturing tolerances. the solvability of this problem also likely requires a slightly higher dimensional matrix than matrices which carry per switch diodes, or only maintain 2KRO, although i'm far too tired to crunch the (relatively simple) combinatorics right now.

anyway, you can't yet print a diode on an fcb for any reasonable amount of money. silicon is our most common, easiest to engineering semiconductor and it does not at all like to be printed or to have its substrate bent. i have thought of some more creative but naive solutions that require minor changes to the fcb production process, and use standard diodes packages, but the tooling for these is prohibitive (heck, the tooling for FCBs in general is prohibitive... grumble).

[Parak]

Yes, any Model M is 2KRO limited by the matrix, and any Model F is NKRO limited by whatever USB controller/adapter you might be using for it. In case of a Soarer adapter, a compatible Model F will give you NKRO over USB.