New PSU = 30% faster boot! Why?

[Breeze]

This is odd. I was using an Antec Truepower 2.0 550W on an older A7N8X-E mobo (that Mobile Barton's still chugging away at 2.4 gHz...) and sometimes it wouldn't start, just dying after the fan started spinning; but if I turned off the power supply, and then started it would be fine. It took about 3.5 minutes to get from cold boot to complete desktop and I sometimes got a message from my 6600GT drivers that the card wasn't getting enough power on bootup. So I figured it was time for a new power supply.

Then today I got and plugged in a brand new OCZ GameXstrem 600 watts and my boot time went down to under 2.5 minutes! The system also seems to respond better but that's probably because video card is now getting enough power.

I guess it goes to show how important good power is. But I still don't understand why the bootup is so much faster now. Would bad or low power make bootup routines fail and retry, making code re-execute and extending the boot? Seems to me none of the subsystems speeds could have been affected without causing errors and crashes. Thanks for any light.

 

[ratbuddy]

Wait, the system takes over 2 minutes to boot and you're ok with that?? What OS ????

 

[OldSkool]

Personally, and this is relating to a build yet to be finished, if my system takes more than 10 seconds to fully boot into windows, i'm going to realize that something is wrong.

 

[Jolly-Swagman]

Wait, the system takes over 2 minutes to boot and you're ok with that?? What OS ????

Yeah that still seems too long for boot up, you should have put the new PSU in your new Rig in your SIG and that one in your old rig

 

[funnyperson1]

The TP probably has bad caps at this point, but as stated, Windows should start up within 30 seconds or so.

 

[JamesXP]

my server (XP 2500+, 320mb ram, 5400rpm hd + server 03 is faster to boot than that... somethings a little wrong.

 

[Oklahoma Wolf]

Not to be a downer, but the A7N8X is a 5V oriented board - you're crossloading that GXS terribly, and that design doesn't handle such situations well to begin with. If it's done long enough, the GXS will eventually fail. The TP2 was also crossloaded powering that board, though the capacitor issues they often have won't help anything either.

My best suggestion is to find a PSU better able to handle the 5v/3.3V demands of that board like an older FSP350-60PN with a higher combined 3.3V/5V rating (about 200W should do it) and single 12V rail.

 

[Breeze]

When I talk of "bootup" I'm talking from the time I turn on the computer to the time EVERYTHING is up and loaded and drive activity stops, which is usually after my Yahoo widgets load up. I've got tons of (useless?) crap loading on this system (over 60 processes running after boot is finished!). I agree it's a long time, and I'm sure if I took out a lot of these services and applets it would be faster. The A7N8X system is still using an ATA boot drive and I tried moving my OS to SATA but found it impossible to do without reinstalling the whole OS (and all its software...) from scratch.

FYI, the C2D in my sig boots up in about 30 secs. It's my work system and I keep it lean and mean: there's less than 25 processes running after boot is complete.

I find Windows XP desperately tries to be "fast" at bootup by bringing up the desktop way before every applet and service is up and running, so I don't count boot up as time to the desktop.

At any rate, none of this relates to the question of why, all other things being equal, a boot would be faster with a better power supply...

 

[Breeze]

Not to be a downer, but the A7N8X is a 5V oriented board - you're crossloading that GXS terribly, and that design doesn't handle such situations well to begin with. If it's done long enough, the GXS will eventually fail. The TP2 was also crossloaded powering that board, though the capacitor issues they often have won't help anything either.

My best suggestion is to find a PSU better able to handle the 5v/3.3V demands of that board like an older FSP350-60PN with a higher combined 3.3V/5V rating (about 200W should do it) and single 12V rail.

Thanks Oklahoma Wolf: I just looked up "crossloading" as it's the first time I heard of it. This could very well have been the downfall of the TP2. Is there any software that can monitor and analyze a system to detect this? Sounds like the only way to not worry about it is to get a monster power supply that just can't be a victim of it.

Maybe Jolly-Swagman is right: I should swap this powersupply for the one in my C2D case, which is an Antec Sonata 2 case with the supplied 450 watt power supply. I could try putting the Antec supply in the A7N8X case, though IT probably will then be subject to the same crossloading issue. The GXS is a much better PS than the Antec, right? At this point, buying a used power supply for this system that's still in good shape seems like a tall order.

And THANKS funnyperson: I just looked through the fan grill of the TP2: THE CAPS ARE BULGING!!! WTF? Was this power supply made with the crap-caps that appeared on a host of motherboards several years ago? So I might be able to save this PS by changing the bad caps... Is this a known issue?

Thanks for all the feedback!

 

[Oklahoma Wolf]

Is there any software that can monitor and analyze a system to detect this?

Nope, the only thing you can do is recognize a mainboard that pulls CPU power from 5V. If it's missing the 4 pin 12V connector every new board these days comes with, it's a 5V based board. Accordingly, you need to pick a PSU that can handle that. As mentioned, look on the label for a combined 3.3V/5V rating over 200W - anything lower than that is a crapshoot.

You'll have your work cut out for you on that one - PSU's that can handle those old 5V based boards have been disappearing at a steady rate. Zippy still makes server grade units that can do it, but for the most part such units are few and far between. FSP/Sparkle and Enermax are about the only two left worth mentioning anymore.

The Silverstone Zeus ST56ZF will also do it, but this too is a server grade unit, not cheap, and very much overkill because it can power a modern heavy 12V based load too.

I just looked through the fan grill of the TP2: THE CAPS ARE BULGING!!! WTF? Was this power supply made with the crap-caps that appeared on a host of motherboards several years ago? So I might be able to save this PS by changing the bad caps... Is this a known issue?

A well known issue - Fuhjyyu capacitors don't like heat and CWT built Antecs like the TP2 run the fan too slow. Lifespan is usually 3 years at best to less than one year at worst depending how hot they run for how long.

The GXS is a better PSU than those older Antecs, but still not designed for your board. You need to find something that has, as I mentioned, a higher combined 3.3V/5V rating.

 

[theELVISCERATOR]

I got a one 1ghz slot one PIII right now booting in under one minute with onl 512 ram..

 

[Mr.Guvernment]

go to start / run , type in msconfig

go to startup, uncheck everything but AV and spyware apps, reboot

now some apps will load anyways and re-enable themselves, unless you get like spyware search and destroy which will block the apps from starting.

 

[hokiealumnus]

2.5 minutes is less than preferable, but not too awfully bad! My horrible, horrible IBM PC at work takes a full six minutes to start up Win2k. It's a 2.6G P4 w/ 512M RAM. I can't WAIT until we get new ones...have to show up ten minutes early to get logged on to the system on time. It's really a burden being a moderately tech-savvy OC"er and having to put up with a company with computers still running Win2k.

/completely off topic/
Nice to see another Hokie on here! GO HOKIES!!!
/return to the regularly scheduled discussion/

 

[Mr.Guvernment]

^^ i would come in on the time you start and if someone complains say, hey sorry , computers too slow, still booting..

 

[sp00nix]

yikes that is a long boot indeed. My laptop is up and running and idle under a minute! (specs below) My desktop is about 1.5 minutes because of some specialized hardware drivers.

 

[vt420]

win2k may load a little slower (and shut down slower too) than XP and vista, but once it's running it's very fast and stable. I personally skipped XP on all my personally hardware as it seemed pointless to sidegrade to it from win2k (especially at launch when it was just a pretty(ugly) face on almost the same OS with a little bit of random crap thrown in)

my last build before this one ran win2k, current one runs vista. I've had to professionally support win2k, XP, server2k3, and vista machines (not to mention fixing older win versions at my old job) and I've never really liked XP, win2k was about the most stable of all of those well, tied with 2k3.

Jeff

 

[[Spectre]]

New PSU = 30% faster boot! Why?

 

[wingless]

If the voltage is too low across the motherboard then this could cause signaling errors between the North Bridge and CPU/RAM. The packets trying to get sent back and forth would face frequent time outs and have to resend. This is kind of like increasing the drive strength in your RAM settings or increasing the LDT Voltage on a Hyper Transport system to increase HT bus speeds.

This is the only reasonable explanation I can think of that would cause such an issue. It has to do with the signaling strength back and forth between components. We need an engineer to explain in more detail and make corrections to my explanation though.

 

[Mr.Guvernment]

^^ as well if the HD isnt getting full powert to start up it can take some time for it to finally kick in, recent issue with a friends system, psu crapping out so you can hear the HD stutter on boot up for almost a minute - new psu, problem gone.

 

[Drew@PSU]

It's a mobile barton, yes? So wouldn't it throttle itself if it were low on power, say in a battery low situation ( I'm just hypothosizing, or however you spell it )? That's what I would jump to at first.

In other news, [Spectre] wins the thread for the Fast and the Furious style pic.

<-- Hates John Walker
</0.02 dollars>