Frickin' techie miracle

Mercutio

Mercutio

Fatwah on Western Digital
Joined
Jan 17, 2002
Messages
23,004
Location
I am omnipresent
Just now, on accident, I installed a Via USB2 PCI card into a Gateway E-4200 (P3-450) running Windows XP.

While it was on, and at an XP desktop.

This is of course something that makes even experienced techs cringe.

I didn't know it was on (these things are REALLY quiet and the power LED must be burned out or something), and being a lazy guy I didn't unplug it.

So here's the amazing part: The machine didn't belch flame and die. It didn't kill the card. It didn't even reboot or lock up. It remained fully responsive, and, this just kills me ran the plug and play install process instantly, on the spot. I turned its monitor on just in time to see it install a USB root hub.

I plugged in my DVD writer, it's working, and I'm typing on the machine right now. Haven't even rebooted yet.

Wow.
 
I'm guessing the card had some sort of delayed start feature where all of the bus is in a high impedance state for a set time after it gets power. When you first installed it, it took a fraction of a second after it detected power to start sending and receiving signals on the PCI bus. Luckily, you probably slide the card in really quickly(within the delay period). Also, perhaps there is no built-in delay, but it does take a finite time for the power capacitors on the card to charge(and by extension for the card to be powered up). You don't always necessarily kill the card or machine doing what you did, although I certainly wouldn't recommend it. BTW, I've actually changed components on live circuits, but it's never components connected directly to the power. Rather, it's things connected to ICs, most of which can take a momentary short on the output. Generally, though, I would never do that on a PC because shorting a signal can have all sorts of unforeseen and unintended consequences.

In any case, I would say you were very lucky here. I'll unplug a fan on a live PC, but that's about it.
 
'Course not, Free - the PATA spec included this capability from day one. Buck really intended to say that Merc was going to hot swap an XP1700+ overclocked to 2700+. Not that there's any problem with that either.
 
If an internal drive is actually "internal", and the case isn't open, I'd say swapping a drive would in fact be something of a miracle.

Of course, if I could do that I'd be walked out of Best Buy with a bunch of 200GB DM9s that I "hot swapped" out of their retail packaging.
 

  • > So here's the amazing part: The machine didn't belch flame and die. It didn't kill the card.
This sort of thing won't be particularly amazing once PCI Express is up and alive (¿2004?).

PCI Express, or simply The King Of All PCI (Intel's thoughts), supports hot swapping -- even with yer graphics card :o allegedly.
 
I once hot-swapped a Quantum hard drive of ~210MB that was inside a Compaq 486DX33 desktop. It didn't kill it instantly, but the drive eventually failed one week later in his new bed. That was in '95 IIRC, so the drive was already old.
 
Gahhhh... I dunno what all the fuss is about. Everybody knows dat.

PC's plug into the wall, right? The power that comes out of the wall is AC, yes? Runs at 50 cycles per second. (Possibly 60 if you are in the Good Ol US of A, where they drive on the wrong side and do all zorts of other weird ztuff.)

Az any fool knows, AC power follows a sine wave: at 50Hz, a 100V AC power plug delivers anywhere between about -200V and +200V. (I forget the exact figures. JTR would know. The 100V is a nominal figure, calculated by the root mean square (RMS) method. Or possibly Voodoo, I get them mixed up sometimes.) And exactly 100 times a second, it crosses the zero Volt line. Mercutio simply waited for the correct moment and then plugged the part in.

Nuffin to it.

Great reflexes. Merc.
 
Actually, the peak voltage is the square root of 2 times the RMS voltage, and the average voltage is pi/2 times the peak voltage. So for the good old USA(60 Hz, 120 V nominal) we have peaks of +- 169.7V and an average voltage of 108.0 volts. The RMS voltage is derived from the DC voltage that would give an equivalent power flow is a pure resistive load. For an inductive or capacitive load on an AC line the power input isn't the product of RMS volts time RMS amps as it would be for a pure resistive lopad due to the current leading(capacitive load) or lagging(inductive load) the voltage. A perfect example is my washing machine. It tested at about 7.5 amps RMS but the power input is only ~400 watts instead of the 900 that voltsxamps would give you. Ditto for all of our room air conditioners and fluorescent lights(the magnetic ballest is an inductive load).
 
Although I didn't expect this to work with PCI cards, I do routinely hotswap my CDRW drive from one machine to another without a hitch. Only thing needed to do was to "scan for hardware changes" in device manager.
 
You must log in or register to reply here.
Back
Top