Norman Dunbar makes some magical things to make me read } } Hi Jerome, } } >> No, it never leak! } That bit was a joke - honest ! } } >> The signal propagates along the line, and when it reach the end of the } line, } >> it partially reflect (just like waves do on harbour) and then propagate } back, } >> crossing the new incoming signal... } This is what confuses my simple brain, in order for a reflection there } should be 'something' at the end to reflect back off of. If there is nothing } to reflect from then there shouldn't be a reflection. Using your harbour } analogy, there is a harbour wall to reflect back off whereas in an } unterminated cable, there is nothing. } } (Did I mention that I wasn't all that good with hardware ?)
But there IS something at the end of the unterminated cable. It's 'AIR', or VOID, and it's an isolant. The water is made of copper, the wall of the harbour are made of isolant. Sometimes, even nothing is an important thing. } } <SNIP> } } >> Single wire are not so nice, hence the need to correctly terminate the } line } >> with device or a resistor. } So how does having a resistor stop reflections then - and how do you pick a } correct value, Tony mentioned 330R what's so special aboput that value ? (My } knowlege here extents to V=IR). } Because a resistor will absorb the energy of the signal (transforming it into heat, thus reducing strongly the amount of energy which can reflect). The 'correct' value is related to the specification of the cable/bus. If you put a too small resistor (which close the loop between the 'signal' line and the 'ground line'), you will request too much current (I) from the generator to get the correct wave... overheating, melding the hardware... BAD. If you put a too big resistor, you still have some reflection. Different specifications (QL net, IDE, SCSI, ...) leads to different magic values.
