How much will a 300w power supply run?
Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee, Devonavar
Hey, put some more crud in my box, and my 300w Zalman is still chuggin' away perfectly - additions are bold:
XP1800+
1.5gb DDR 2100
Radeon 9700 Pro (w/ Zalman ZM-80hp passive HS)
CD-ROM
CD-R/W
320Gb of storage
- 2 x Western Digital 100Gb SE (RAID 1, in SmartDrive enclosures)
- 1 x Western Digital 80Gb (in IDE pullout cartridge, not always in)
- 1 x Barracuda IV 40Gb
2 PCI cards (USB2/Firewire, Hercules Fortissimo - negligible draw)
5 fans (80mm PapstNGL - negligible draw)
2 powered USB devices (tablet, scanner - negligible draw)
Dual 12" cold cathode - two bulbs off of one inverter
I play UT2k3 at 1600x1200 resolution, hours at at time, no crashes (I say that because that game is noteworthy for stressing hardware). Quake 3 Arena is a thing of beauty now. Prime95, 72 hours at a time, no crashes.
Who needs a 550w PS again? Maybe 1/10th the people who have 'em.
XP1800+
1.5gb DDR 2100
Radeon 9700 Pro (w/ Zalman ZM-80hp passive HS)
CD-ROM
CD-R/W
320Gb of storage
- 2 x Western Digital 100Gb SE (RAID 1, in SmartDrive enclosures)
- 1 x Western Digital 80Gb (in IDE pullout cartridge, not always in)
- 1 x Barracuda IV 40Gb
2 PCI cards (USB2/Firewire, Hercules Fortissimo - negligible draw)
5 fans (80mm PapstNGL - negligible draw)
2 powered USB devices (tablet, scanner - negligible draw)
Dual 12" cold cathode - two bulbs off of one inverter
I play UT2k3 at 1600x1200 resolution, hours at at time, no crashes (I say that because that game is noteworthy for stressing hardware). Quake 3 Arena is a thing of beauty now. Prime95, 72 hours at a time, no crashes.
Who needs a 550w PS again? Maybe 1/10th the people who have 'em.
System:
3.06 GHz Pentium 4 with hyperthreading
Asus P4G8X motherboard
1.5 Gb PC2100 RAM
Barracuda V 120 Gb HD
CDRW/DVD drive
DVD+RW drive
ATI 9800 Pro 128 Mb graphics card
Audigy2 soundcard
2 x Papst 80mm fans
PSU: Fortron FSP300-60PN(PF)
RESULT: No stability problems - voltages seem to be stable under load.
HOWEVER, when CPU was at full load and the hard disk was thrashing, the PSU did start squealing in time to the hard disk activity. This was cured by replacing the PSU with a 350W version.
3.06 GHz Pentium 4 with hyperthreading
Asus P4G8X motherboard
1.5 Gb PC2100 RAM
Barracuda V 120 Gb HD
CDRW/DVD drive
DVD+RW drive
ATI 9800 Pro 128 Mb graphics card
Audigy2 soundcard
2 x Papst 80mm fans
PSU: Fortron FSP300-60PN(PF)
RESULT: No stability problems - voltages seem to be stable under load.
HOWEVER, when CPU was at full load and the hard disk was thrashing, the PSU did start squealing in time to the hard disk activity. This was cured by replacing the PSU with a 350W version.
P/S FORTRON 300W P300-60PN RETAIL
The above 300W is powering the following:
- P4 2533Mhz @ 2850Mhz w/ Zalman CNPS-7000-cu
- Asus 54B533-E motherboard
- 1GB (2x512MB) of PC2100 RAM
- 3 Barracuda V 80GB hard disks, 2 in RAID
- Geforce4 ti 4600
- 3 Panaflo 80L case fans (undervolted)
- Firewire PCI card
- V92 PCI modem
- 1 DVD+R/RW burner (Sony DRU-500A)
- 1 DVD player (Toshiba SD-M1402)
- Mitsumi floppy
Great power supply!
- P4 2533Mhz @ 2850Mhz w/ Zalman CNPS-7000-cu
- Asus 54B533-E motherboard
- 1GB (2x512MB) of PC2100 RAM
- 3 Barracuda V 80GB hard disks, 2 in RAID
- Geforce4 ti 4600
- 3 Panaflo 80L case fans (undervolted)
- Firewire PCI card
- V92 PCI modem
- 1 DVD+R/RW burner (Sony DRU-500A)
- 1 DVD player (Toshiba SD-M1402)
- Mitsumi floppy
Great power supply!
Enlight 300W PSU
-Gigabyte GA-8PE800 motherboard
-3.06 GHZ P4 (with Zalman CNPS7000-AlCu)
-2 x 512 MB PC2700 DDR
-Geforce 2 MX400 (with passive Zalman ZM50-HP)
-Samsung 60 GB HD
-Pioneer DVD-RW drive
-Philips CD-RW drive
-Samsung floppy drive
-Firewire PCI card
-USB2 PCI card
-network card
-Trust pci modem
-80mm case fan
-Gigabyte GA-8PE800 motherboard
-3.06 GHZ P4 (with Zalman CNPS7000-AlCu)
-2 x 512 MB PC2700 DDR
-Geforce 2 MX400 (with passive Zalman ZM50-HP)
-Samsung 60 GB HD
-Pioneer DVD-RW drive
-Philips CD-RW drive
-Samsung floppy drive
-Firewire PCI card
-USB2 PCI card
-network card
-Trust pci modem
-80mm case fan
-
- Posts: 23
- Joined: Mon Aug 11, 2003 7:09 pm
You powerhogs! You're the ones depleting the ozone and killing the fish in the seas!
</sarcasm>
IBM 185W PSU
MSI MS-6318 motherboard
1.3GHz Tualatin Celeron
1x256MB PC133
Radeon 9000 (passive)
Seagate Barracuda IV 40GB
Samsung 48X CD
NEC 4x4 CD-changer
3com NIC
USB KB/mouse/FF gamepad (powered off system)
2x12" CCFL occasionally
Edit - No, there are no case fans. There is a single 92mm in the PSU. That is all.
300W say what?
- M4H
</sarcasm>
IBM 185W PSU
MSI MS-6318 motherboard
1.3GHz Tualatin Celeron
1x256MB PC133
Radeon 9000 (passive)
Seagate Barracuda IV 40GB
Samsung 48X CD
NEC 4x4 CD-changer
3com NIC
USB KB/mouse/FF gamepad (powered off system)
2x12" CCFL occasionally
Edit - No, there are no case fans. There is a single 92mm in the PSU. That is all.
300W say what?
- M4H
I'll top that:
No-Name 180 watt SFX PSU (modded w/ 7volted L1A) powering:
Asus A7N266-VM Nforce mobo
XP 1900+
2x256MB PC2100
Maxtor D740x 40G HDD
GeForce2Ti (with big passive Celeron HS glued to it)
Sony 16x DVD drive.
92mm NMB B40 (the $0.39 special) w/ Fanmate
60mm no name exhaust fan
Anyone care to go lower than that?
(It's like a PSU limbo contest!)
No-Name 180 watt SFX PSU (modded w/ 7volted L1A) powering:
Asus A7N266-VM Nforce mobo
XP 1900+
2x256MB PC2100
Maxtor D740x 40G HDD
GeForce2Ti (with big passive Celeron HS glued to it)
Sony 16x DVD drive.
92mm NMB B40 (the $0.39 special) w/ Fanmate
60mm no name exhaust fan
Anyone care to go lower than that?
(It's like a PSU limbo contest!)
-
- Posts: 23
- Joined: Mon Aug 11, 2003 7:09 pm
-
- Posts: 23
- Joined: Mon Aug 11, 2003 7:09 pm
-
- Posts: 23
- Joined: Mon Aug 11, 2003 7:09 pm
-
- SPCR Reviewer
- Posts: 8636
- Joined: Sat Nov 23, 2002 6:33 am
- Location: Sunny SoCal
My old Bigfoot is still chugging away in daily use in my Sister's classroom. She's been using it daily for at least 5-6 years now. Pentium 90 (or 100?) + 24MB RAM + Bigfoot + Win98 Lite. Even her third grade students complain about how slow it is. They've probably all got better systems at home!Rusty075 wrote:Although I do own a lovely Quantum Bigfoot.
You remember bigfoots right? The harddrive that is everything a barracuda isn't. Slow, Loud, and Big. (Big only in physical size, not capacity)
-
- SPCR Reviewer
- Posts: 8636
- Joined: Sat Nov 23, 2002 6:33 am
- Location: Sunny SoCal
i dont have a high end rig any more but
a generic 320 watt psu
i have a shuttle ak31 3.1 kt 266a motherboard
maxtor diamondmax 9 80gb 8mb cache fdb baby
wd 120gb 8mb cache 1200jb uses about 10 wattts
liteon 52x burner
pioneer 115 dvd drive
sb live value
ultra 100 promise card
geforce 3 ti 200
3com 3c905c nic
pci scsi-2 card for an old ass scanner
any way my computer uses 130 watts idle it goes up to 140 durring cpu intensive aps.
i had a tv tuner card took taht out cause it used 20 watts even when i wasnt using it.
i have a watt meter thats how i know all this
a generic 320 watt psu
i have a shuttle ak31 3.1 kt 266a motherboard
maxtor diamondmax 9 80gb 8mb cache fdb baby
wd 120gb 8mb cache 1200jb uses about 10 wattts
liteon 52x burner
pioneer 115 dvd drive
sb live value
ultra 100 promise card
geforce 3 ti 200
3com 3c905c nic
pci scsi-2 card for an old ass scanner
any way my computer uses 130 watts idle it goes up to 140 durring cpu intensive aps.
i had a tv tuner card took taht out cause it used 20 watts even when i wasnt using it.
i have a watt meter thats how i know all this
-
- Patron of SPCR
- Posts: 946
- Joined: Sat Sep 06, 2003 9:57 pm
- Location: Berkeley, CA, USA
- Contact:
Just FYI, I recently purchased a kill-a-watt power draw measuring device..
http://www.laser.com/dhouston/killwatt.htm
This is a very cool real-time device that is also used in power supply reviews here at silentpcreview. Recommended.
Interestingly, my overvolted P4 3.2ghz system-- with a single hard drive-- draws a LOT more power than my Dual Athlon server with 4 hard drives and 3 SCSI CD-Rs.
Still nowhere near 300w though:
P4 system (overvolted northwood C-series);
idle: 150w
full load (3d gaming): 250w
Dual Athlon XP server (undervolted bartons):
idle: 150w
full load (prime95): 180w
There is a Radeon 9800 pro in the P4 system, but the vast majority of the power draw SEEMS to be the CPU, because I can get the P4 up near 250w by simply firing up Prime95..
It is fun to run around the house and measure how much power everything uses. Maybe I'm nuts...
For example did you know a 21" CRT uses ~100watts, 17" LCD uses ~30 watts, and 15" LCD uses 27 watts? And also LCD and CRT power save mode is really zero watts as advertised?
Another weird thing, the PCs use about 5 watts even when powered completely off.
http://www.laser.com/dhouston/killwatt.htm
This is a very cool real-time device that is also used in power supply reviews here at silentpcreview. Recommended.
Interestingly, my overvolted P4 3.2ghz system-- with a single hard drive-- draws a LOT more power than my Dual Athlon server with 4 hard drives and 3 SCSI CD-Rs.
Still nowhere near 300w though:
P4 system (overvolted northwood C-series);
idle: 150w
full load (3d gaming): 250w
Dual Athlon XP server (undervolted bartons):
idle: 150w
full load (prime95): 180w
There is a Radeon 9800 pro in the P4 system, but the vast majority of the power draw SEEMS to be the CPU, because I can get the P4 up near 250w by simply firing up Prime95..
It is fun to run around the house and measure how much power everything uses. Maybe I'm nuts...
For example did you know a 21" CRT uses ~100watts, 17" LCD uses ~30 watts, and 15" LCD uses 27 watts? And also LCD and CRT power save mode is really zero watts as advertised?
Another weird thing, the PCs use about 5 watts even when powered completely off.
My Nokia 445Pro is spec'd for 160W max.wumpus wrote:For example did you know a 21" CRT uses ~100watts, 17" LCD uses ~30 watts, and 15" LCD uses 27 watts? And also LCD and CRT power save mode is really zero watts as advertised?
Another weird thing, the PCs use about 5 watts even when powered completely off.
PCs use a bit of power when they are off to enable soft-power-on, WOL, WOR, etc. That's why it isn't wise to plug/unplug the 20-pin ATX cable if the power is still attached to the wall plug.
Cheers,
Jan
and crt's use more power with a white screen as opposed to black and if you make the brightness higher more power as well lcd's always use the same amount of power no matter what you are viewing unless they are off or standby.
i have a samsung syncmaster 955df 19" crt it uses about 75 watts on a black screen, 80-90 watts with white and the more brightness i add i can get it up to 100-120 watts and off i think it uses 7 watts or 5 i dont remember what. what sucks is it has no real standby it uses 75 watts in standby thats why you dont hear the static dissipate when it goes into standby in other words turn this monitor off otherwise its wasting power and creating heat even in standby.
but htis is off topic so i am sure i will hear it for this but i didnt start the off topic ness it was the guy talking about his kill-a-watt
i have a samsung syncmaster 955df 19" crt it uses about 75 watts on a black screen, 80-90 watts with white and the more brightness i add i can get it up to 100-120 watts and off i think it uses 7 watts or 5 i dont remember what. what sucks is it has no real standby it uses 75 watts in standby thats why you dont hear the static dissipate when it goes into standby in other words turn this monitor off otherwise its wasting power and creating heat even in standby.
but htis is off topic so i am sure i will hear it for this but i didnt start the off topic ness it was the guy talking about his kill-a-watt
-
- Patron of SPCR
- Posts: 946
- Joined: Sat Sep 06, 2003 9:57 pm
- Location: Berkeley, CA, USA
- Contact:
I think measuring exactly how much power your computer is really using is precisely what this topic IS about. Which is where the kill-a-watt comes in handy. Do you want to listen to a bunch of guys rattling off lists of items in their computer and what they believe to be true, or do you want to read actual measured values? It's the difference between science and voodoo..off topic ness it was the guy talking about his kill-a-watt
I think it would be possible to get over 300w load if you had a ~3.4-3.6ghz P4c.. that is, overvolted and overclocked, and a bleeding edge video card. Or perhaps an Athlon running at 2.3-2.4ghz also overvolted and (obviously) overclocked. But I think that's the only way.
My testing on about 5-6 machines around my house shows that hard drives, scsi cards, memory, video card, etc really don't matter since the CPU is the largest power consumer in your system BY FAR.. in my worst case (most powerful computer), adding 100w to the power load all by itself-- when going from 0% to 100% prime95 load.
At the other extreme, my low-end mini-itx box (I think the C3 533?) with laptop HDD and laptop DVD, uses a measly 13 watts at idle and 18watts under 100% load. Heh.
Yeah, wumpus is right. Actually knowing exactly how much power a system is drawing is better than just guessing at it.
Of course even the Kill-a-Watt is only partially useful, since its only measuring what the entire system is drawing at the wall. Even your P4 system isn't actually using 250watts. Your PSU is maybe 75% efficient at best (probably less) so the total system is actually only using 187.5watts peak.
Wumpus I dare you to go post that at Overclockers.com or OCAddiction the next time they say that they need a 500watt PSU. I hope you're wearing your asbestos undies.
What would be the most useful would be a way as simple as the Kill-a-watt to measure how many amps each component is drawing, and on what voltage line. Your P4 may only be using 180watts, but 160 of that may be on the 3.3 and 5volt lines. That's where an el-cheapo 300PSU wouldn't cut it. But a quality 300W PSU would work just fine.
The problems with sizing PSU's are:
1. You have no real way of knowing what you need.
2. The PSU makers market their product by only the BIG number, which in real use is meaningless.
I think PSU's should be marketed like fertilizer, where instead of listing only the big number, you list the numbers for each of the component lines.
So, for example the 300W Seasonic MikeC just reviewed would be a 92-150-216, while the 400W'er would be a 92-150-264. With the 3 numbers representing the wattage of the 3.3v, 5v, and 12v lines respectively. You break them down that way and suddenly you can see the real differences between the PSU's
(and Ryan200, I rarely do this, but man please learn some punctuation. I had to read your post like three times before it made sense)
Of course even the Kill-a-Watt is only partially useful, since its only measuring what the entire system is drawing at the wall. Even your P4 system isn't actually using 250watts. Your PSU is maybe 75% efficient at best (probably less) so the total system is actually only using 187.5watts peak.
Wumpus I dare you to go post that at Overclockers.com or OCAddiction the next time they say that they need a 500watt PSU. I hope you're wearing your asbestos undies.
What would be the most useful would be a way as simple as the Kill-a-watt to measure how many amps each component is drawing, and on what voltage line. Your P4 may only be using 180watts, but 160 of that may be on the 3.3 and 5volt lines. That's where an el-cheapo 300PSU wouldn't cut it. But a quality 300W PSU would work just fine.
The problems with sizing PSU's are:
1. You have no real way of knowing what you need.
2. The PSU makers market their product by only the BIG number, which in real use is meaningless.
I think PSU's should be marketed like fertilizer, where instead of listing only the big number, you list the numbers for each of the component lines.
So, for example the 300W Seasonic MikeC just reviewed would be a 92-150-216, while the 400W'er would be a 92-150-264. With the 3 numbers representing the wattage of the 3.3v, 5v, and 12v lines respectively. You break them down that way and suddenly you can see the real differences between the PSU's
(and Ryan200, I rarely do this, but man please learn some punctuation. I had to read your post like three times before it made sense)
-
- Patron of SPCR
- Posts: 946
- Joined: Sat Sep 06, 2003 9:57 pm
- Location: Berkeley, CA, USA
- Contact:
Right, I forgot about that. I did order a seasonic tornado though, so maybe we can get that efficiency up on my server which is on 24/7 and could benefit from it.. I think I have a power supply fetish now, thanks to silentpcreview.. I didn't know anything about efficiency ratings until I came here.Even your P4 system isn't actually using 250watts. Your PSU is maybe 75% efficient at best (probably less) so the total system is actually only using 187.5watts peak.
Good point. Power stability is a totally different (but important) issue, as is amperage. The kill-a-watt does have an amps mode but I haven't been looking at that so far..What would be the most useful would be a way as simple as the Kill-a-watt to measure how many amps each component is drawing, and on what voltage line. Your P4 may only be using 180watts, but 160 of that may be on the 3.3 and 5volt lines. That's where an el-cheapo 300PSU wouldn't cut it. But a quality 300W PSU would work just fine.
You know what else is a good idea: always run your system through a quality external uninterruptible power supply. Not just for the obvious power backup coverage, but because a good UPS conditions power and removes any sags, spikes, etc that could also cause instability. And believe me, it definitely happens (for some people), depending on the quality of your building's wiring and power feeds.
Very true. The old "garbage in garbage out" rule applies to power as well.wumpus wrote:You know what else is a good idea: always run your system through a quality external uninterruptible power supply.
An anecdote: Quite a while ago I had a friend ask me to help diagnose a workstation that was having random reboots. We swapped every conceivable component, but the symptoms remained the same. He lived in a fairly old mid-highrise building, and one night while over there for a Super-bowl party we noticed that the computer would reboot about 30 seconds before a guest would arrive, and again about 30 seconds after they left the apartment. Like clockwork. Well, after having a good thorough exorcism performed, to no avail, we discovered what was causing it: Every time anyone in the building used the 1920's vintage elevator, the machine would reboot. Apparently the old wiring was transmitting a surge or a sag throughout the entire building. A UPS unit cured it.
-
- Patron of SPCR
- Posts: 946
- Joined: Sat Sep 06, 2003 9:57 pm
- Location: Berkeley, CA, USA
- Contact:
lol, exactly!
My example was a system that would run fine at my friend's bench through every conceivable test.. but he'd take it over to a customer's house and it would invariably lock up or hang.
I asked him to rule out heat (people love "hiding" computers in totally closed cabinets), and once we did, it had to be a power supply issue at her home. It was the only remaining variable-- the system was identical. So I urged him to get a bench UPS to take with him on these kinds of problems.. lo and behold, the problem went away when he brought the UPS over.
My example was a system that would run fine at my friend's bench through every conceivable test.. but he'd take it over to a customer's house and it would invariably lock up or hang.
I asked him to rule out heat (people love "hiding" computers in totally closed cabinets), and once we did, it had to be a power supply issue at her home. It was the only remaining variable-- the system was identical. So I urged him to get a bench UPS to take with him on these kinds of problems.. lo and behold, the problem went away when he brought the UPS over.
Abit nf7-s 2.0 + Barton 2700xp @ 217fsb x 10.5 1.8 vcore
(backed off from 220x11 so i could lower my vcore and run quieter)
512megs of corsair xms 3500 (single channel) 7-3-3-2
Sound plaster Audigy 2
Sapphire 9700pro running at 350/330 (conservative ocing)
Western Digital 80GIG Special Edition HDD 8meg cache
Matshita CDRW
Zip 100 Drive
USB Wacom Tablet, Keyboard and Logitech mx700 Mouse
Cpu cooled by the Zalman 7000series alcu heatsink @ 2250 rpm - 48-55c
Couple of cooler master 80mm fans on a rheobus.
everything is running on a seasonic super tornado 300 watt ps.
Overall the voltage is rock solid, except the vcore which fluctuates wildly (this is observed by many nf7 owners even with 500watt psu, doesnt seem to be problematic though - im thinking this is more of a sensor issue)
i didnt plugin that secondary power supply cable (the 4 pin one) could that be the issue? CAn you pass me a tissue?
(backed off from 220x11 so i could lower my vcore and run quieter)
512megs of corsair xms 3500 (single channel) 7-3-3-2
Sound plaster Audigy 2
Sapphire 9700pro running at 350/330 (conservative ocing)
Western Digital 80GIG Special Edition HDD 8meg cache
Matshita CDRW
Zip 100 Drive
USB Wacom Tablet, Keyboard and Logitech mx700 Mouse
Cpu cooled by the Zalman 7000series alcu heatsink @ 2250 rpm - 48-55c
Couple of cooler master 80mm fans on a rheobus.
everything is running on a seasonic super tornado 300 watt ps.
Overall the voltage is rock solid, except the vcore which fluctuates wildly (this is observed by many nf7 owners even with 500watt psu, doesnt seem to be problematic though - im thinking this is more of a sensor issue)
i didnt plugin that secondary power supply cable (the 4 pin one) could that be the issue? CAn you pass me a tissue?
-
- Posts: 6
- Joined: Fri Sep 19, 2003 1:25 am
Fortron make some of the best PSU's available, but they are often bundled into cheap cases.
My AOpen case came with a FSP300-60ATV and it runs the following with no problems (I've tripped it a couple of times, but that was my fault)
- AXP 2100 Palomino @ stock speeds
- 2x512MB sticks
- CD-RW drive
- DVD/CD-ROM drive
- WD800JB 7200rpm HD
- up to 6 fans on a 6x6w fanbus
My AOpen case came with a FSP300-60ATV and it runs the following with no problems (I've tripped it a couple of times, but that was my fault)
- AXP 2100 Palomino @ stock speeds
- 2x512MB sticks
- CD-RW drive
- DVD/CD-ROM drive
- WD800JB 7200rpm HD
- up to 6 fans on a 6x6w fanbus