"practically all" <> "all". Now neither you nor me know exactly how many devices have been made out of specs, but the fact is current limiting device implementation has been mandated as of SCSI-II, so I bet you the figures are at best exceedingly low. Even if everyone set all devices to TP=ON, he'd run into device deaths of "natural causes" way before any TP related issues appeared, and that's assuming he did have badly designed devices on the bus in the first place.
In fact one could argue about having two devices supply termination power, the ones located at the two ends of the SCSI bus (and thus closest to the bus terminators), as being a better configuration, if lots of devices and a lesser quality cable of mismatched impedance come into play. There is voltage drop along the cable and dual TP can and will help in marginal cases.
While your recommendation is generally correct, you presented it as though multiple TP=ON is a death sentence. It's not.
C/P excerpt from a white paper of the SCSI Trade Association:
"How are terminators powered?
Terminators are powered from the "term power" line on the SCSI bus. Term power can come from any device on the bus, and is provided by either the host, a drive on the bus, the backplane, or any combination thereof. Term power is provide through a diode and fuse - historically - the drop accross the diode and cable allows for a term power range of 4.0 to 5.25 volts."
Regards
, no ?
