>>
>> Uhhhhuuumm. How do pull down resistors build up are charge?
>
>The terminating resistors put a load on the bus (I think it is 2 mA per
>bit). If you instantly remove this load, the voltage will increase because
>of the series inductance of the bus. Theoretically, with any inductance and
>any current, at t[0] the voltage will tend towards infinity. Practically
>speaking, built in ESD protection diodes will clamp this voltage to a
>safe level because the amount of energy one can store in the inductance
>of practical cables is very low.
Paired with the capacitance of the cable this would equal about nil.....
>The real getcha is the problem of messing up the bus phase of an active
>bus. Note that both the drives and the controller are capable of driving
>the bus. They can't do this at the same time. If one of the data bits
>is driven high by one perhipheral and low by another, something has to
>go.
But that something does not mean a spike, that can cause psychical damage.
It means the entire bus doesn't know what is going on and takes a massive
shit. Data that is in transit could be corrupted...that's it.
>Because modern SCSI busses are designed for high-speed, the output
>drivers are not current limited well enough to always survive. They
>can usually take a short-circuit to common (ground), but can't sink the
>current from a bit that's being driven high by another device.
When did we go to pure ground? A digital low and ground are two very
different things....
>Hot-swap devices have an interface transceiver. It has the NOT of all
>the bus bits ORed into its enable pin. It does not connect until the
>instant that all the bits are low. Upon disconnect, it sets the enable
>pin false before you can extract the device.
>
>This protects the bus and its devices. It remains for the drivers and
>the Operating System to restore the machine state because you probably
>disturbed something.
>
>Stop-A on the Sun will quiet the SCSI bus .... but .... Most people are
>lucky!
Again, if the bus is inactive, not dead/quited but just inactive, you stand
no chance of doing any physical damage, as long as the device is not
charged with static. If it is active the chances are still basicly nil,
though data does have a high probability of getting screwed up.
------------------------------------------------------------------
Elite MicroComputers 908-541-4214 http://www.psychosis.com/emc/