Re: Power supply build quality pictorial. part 2
Late 90's sounds about right, especially with that 5VSB spec, as you mentioned. Also, I just noticed a Gateway logo on the second case shot, so this most likely came out of a Gateway PC.
Interesting.
I suppose that was a valid concern back in the day, as many motherboard ICs were directly connected to the PSU rails, including 5VSB. Actually, I think this went on pretty much until late-era socket 462 and 478 boards.
One of the reasons why many eMachines PCs died from the infamous Bestec ATX-250-12E PSU is because the socket 462 boards in those PCs typically had their LPC SIO, as well as some other TTL power control logic circuitry directly connected to the 5VSB. Thus, with the 5VSB from the Bestec overshooting to 10V or more, many of those ICs died. I actually have a board like that in my "for a rainy day" pile that I've attempted to fix. Most of the TTL ICs I mentioned were rated for 4.75V-5.25V. So no wonder they released the magic smoke.
Nowadays, the 5VSB on motherboards usually goes through a linear regulator of some sort (typically generating 3.3V stand-by) and then goes to the ICs. So many boards are somewhat more protected from overshooting and quick-rising 5VSB.
I wouldn't be surprised if that's the case. For many OEM PSUs, the maximum rating on the label meant maximum _continuous_ power the PSU can supply (and even higher peak, if needed). Hence the overbuilding. Quite a contrast from the cheapo PSUs with 250-300W ratings that could only hope to peak that high for a few seconds before blowing up.
And the 12V rail! Even a Core Solo won't be able to run with that PSU.
But for an old 5V-heavy PC, it should be fine - especially socket 462 boards, as those usually didn't support S3 sleep (only S1, which means 5VSB was used, at the most, for WOL, WOR, and power by [PS/2] keyboard/mouse.) Actually, most Pentium II and 3 boards supported S3 sleep, but even those were stated to require no more than 500 ma to 1 Amp on the 5VSB. So I have a good number of motherboards that will work with this PSU.
Originally posted by PeteS in CA
View Post
Originally posted by PeteS in CA
View Post
I suppose that was a valid concern back in the day, as many motherboard ICs were directly connected to the PSU rails, including 5VSB. Actually, I think this went on pretty much until late-era socket 462 and 478 boards.
One of the reasons why many eMachines PCs died from the infamous Bestec ATX-250-12E PSU is because the socket 462 boards in those PCs typically had their LPC SIO, as well as some other TTL power control logic circuitry directly connected to the 5VSB. Thus, with the 5VSB from the Bestec overshooting to 10V or more, many of those ICs died. I actually have a board like that in my "for a rainy day" pile that I've attempted to fix. Most of the TTL ICs I mentioned were rated for 4.75V-5.25V. So no wonder they released the magic smoke.
Nowadays, the 5VSB on motherboards usually goes through a linear regulator of some sort (typically generating 3.3V stand-by) and then goes to the ICs. So many boards are somewhat more protected from overshooting and quick-rising 5VSB.
Originally posted by PeteS in CA
View Post
Originally posted by PeteS in CA
View Post
But for an old 5V-heavy PC, it should be fine - especially socket 462 boards, as those usually didn't support S3 sleep (only S1, which means 5VSB was used, at the most, for WOL, WOR, and power by [PS/2] keyboard/mouse.) Actually, most Pentium II and 3 boards supported S3 sleep, but even those were stated to require no more than 500 ma to 1 Amp on the 5VSB. So I have a good number of motherboards that will work with this PSU.
Comment