peritus said:
DON'T DAISY CHAIN IF YOU CAN HELP IT! You are much better off using an additional Firewire card for any non-temporary firewire connections. The problem, so I've read, is that a Firewire card's bandwith pushes the limit of the PCI bus (maybe not such an issue with PCI Express). Therefore, daisy-chaining divides any speed available to the devices in the chain (BOO!)...
Shouldn't.... FW800 still requires less bandwidth than the original 33 MHz PCI can provide.... Maybe your motherboard chipset sucks.
peritus said:
1) Buy a separate Firewire card (No motherboard firewire connections, no
video card firewire)
Wouldn't surprise me if some motherboard FireWire connections were lousy. Depends on the motherboard and how badly they cut corners in the design....
peritus said:
3) No daisy-chaining (explained hitherto).
4) No hot-plugging (just to be cautious)... I've read stories about periphials being lost from plugging/unplugging Firewire connections while the computer and periphial is on....
I've seen M-Audio's warning about hardware failures on hot plug. They don't say whether it hurt their gear or the computer, but they sort-of implied the latter, if memory serves. I wouldn't expect it to hurt the device; the FireWire spec says that you're supposed to be able to handle the surge current from plugging in up to a 33V power connection. The highest voltage anybody ships is probably 25 volts or so, so there's a -huge- safety margin....
About the only way I could imagine something going wrong with either the computer or the device is if it didn't properly have its signal lines protected from DC and the wires got crossed by severely torquing the plug somehow. Ideally, you should be able to flat short out a FireWire connection without causing any hardware damage. Yes, I've done this before, and no, it didn't cause any damage. Powered the machine off, though.
That said, I could see how some of those defective motherboard power capacitors from a few years ago might get pulled over the edge by that sort of hit. If so, the machine was failing already, and it was just bad luck that it happened on hot plug instead of when somebody turned on the power switch, IMHO.
Reminds me of a story. A few years ago, I talked to some folks who had been to the USB/FireWire plugfest at Apple's developer conference (a few years before that). They told of one "special" test machine that "complied with the FireWire spec"... and by complied, I mean that it ran at or near the maximum allowable voltage.
Anyway, at one plugfest, they reportedly pulled that thing out. Of course they warned the manufacturer reps in advance. Somehow, they were all expecting 12V on the bus (or maybe 18V). That was about the highest voltage Apple had used up to that point on any hardware. However, the letter of the FireWire spec allows for a maximum of 30V unregulated +/- 10%. That would be 33V regulated. You see where this is leading.
Needless to say, that little demonstration got their attention, and I haven't heard of any devices that fell over dead like that since then. The current Mac desktop hardware runs the FireWire ports at around 24-25V (thank you, Google), so I guess that's a good indication that sometimes being bastards about standards compliance results in better product consistency.... Oh, to convince somebody to build a test machine that feeds raw 110VAC to devices that don't comply with the USB Mass Storage spec....