Windows keyboard repeat-rate gremlin
April 4, 2006 11:15 PM

My Windows 2000 system has started spontaneously reducing the key-repeat rate and the rate at which windows scroll, slowing each to a crawl. Other weirdness accompanies this.

This is always preceded by the NumLock, CapsLock and ScrollLock lights flashing once then going dark. The num keys still work as numbers, though. Pressing Num Lock once does nothing, pressing it again relights the NumLock indicator.

In Control Panel, my key repeat rate indicates that it's still set at max. Moving it down, then back to max fixes the speed problem - until the next time the indicator lights flash, just minutes later.

I have a Netgear wireless router, ZoneAlarm, & PestPatrol.
I've run AdAware, Spyware Blaster, Spybot S&D, and CCleaner, none of which found anything to explain this.

Any ideas what's happening here?
posted by Tubes to Computers & Internet (9 answers total)
What brand of a keyboard and what interface are you using? I assume it's not wireless, since you have LED indicators?
posted by disillusioned at 11:25 PM on April 4, 2006


It's my trusty old NMB keyboard, with PS/2 interface. Never had any problem with it.
posted by Tubes at 12:03 AM on April 5, 2006


I would have guessed PS/2. It sounds like there could be a short, and I'm not sure how the key-repeat functionality works, but I wonder if you were to unplug and plug the keyboard back in (outside of this problem occuring) you'd perhaps see the same behavior?

That said, try another, USB-based keyboard? Try this keyboard on another system?

It might just be time to upgrade past a spec unveiled in 1987. It's almost of drinking age.
posted by disillusioned at 12:25 AM on April 5, 2006


To find out if it's the keyboard, buy,find, or borrow another PS2 keyboard and plug it in. This should tell you if it's the keyboard or the PS2 controller on the motherboard.

I'm almost certain that this is an electrical problem- as in, the cable to the keyboard has been run over by an office chair and isn't making good contact, internally or where it connects to the computer. PS2 devices really hate to be plugged/unplugged, and will often behave in bizarre ways, such as you describe.

Whatever you do, when you're testing, don't plug and unplug the keyboard while your computer is on. I've seen motherboards killed this way.
posted by fake at 1:10 AM on April 5, 2006


Fake brings a good point correcting my shortsightedness. Older motherboards were somewhat susceptible to voltage issues re: PS/2 reconnections, and you could permanently damage them. On the other hand, the PS/2 spec doesn't call for it to work again after being plugged back in, even, though I've seen cases where it does and doesn't.

Seriously, consider a $5 USB keyboard from a neighborhood Goodwill, or a $2.50 PS/2 replacement.
posted by disillusioned at 1:40 AM on April 5, 2006


You're thinking of AT keyboards, which should never be unplugged from a running machine. I blew up a $400 motherboard that way, years ago. PS/2 devices can be plugged and unplugged with impunity, it's part of the device spec. (On the other hand, Microsoft's PS/2 mouse drivers are extremely poor at coping with a mouse that has been unplugged. Perhaps that's where this misconception arose.)
posted by majick at 7:33 AM on April 5, 2006


Majick—that's right. But I stand by my assertation that I've experienced troubles getting the keyboard working again after a reconnect on PS/2 models.
posted by disillusioned at 8:22 AM on April 5, 2006


G'morning! Thanks, all -- I'll try swapping keyboards around when I get home tonight. It sounds plausible.

(By the way -- I've always assumed that an unplugged PS/2 device would require a reboot upon plugging in. I've never had any PC pick up a reconnected PS/2 keyboard or mouse without rebooting...)
posted by Tubes at 9:02 AM on April 5, 2006


I'd put a caution on PS/2 hotplugging. I've not seen it offically supported, and a MS article at
http://support.microsoft.com/?kbid=262798 notes "Because PS/2 hotplug is not in the PS/2 specification, some PS/2 controllers may not correctly handle "hotplugging" of a device, and some may short out devices."
posted by bitmage at 9:03 AM on April 5, 2006


« Older Framing Prints   |   The limits of logic Newer »
This thread is closed to new comments.