 | |
Which PSU?
Hardware Overclocking Discussions about overclocking (alt.comp.hardware.overclocking)
 |
| | Thread Tools | Search this Thread | Display Modes |
#11
 
| |||
| |||
|
Re: Which PSU? -
04-28-2007
, 01:03 PM
|
Post some numbers, already! |
Quote:
|
I am especially interested in how well voltage/temperature/fan monitoring applets work, and what nMonitor and nTune do for your system. |
http://www.alcpu.com/CoreTemp
#12
| |||
| |||
|
|
'Ed Medlin' wrote, in part: | PC Power and Cooling seem to go another direction with one very | large12v rail. | http://www.pcpower.com/products/view...php?show=S75QB | is the PSU I have and you can read the info there on how they go | about doing it. 60a is a LOT for a single 12v rail. Let me know what | you think about it yourself. _____ I think you have a very good power supply, but that the engineering department of 'PC Power and Cooling' is subordinated to the marketing department. One single 60 Ampere + 12 VDC rail may be SLIGHTLY more convenient than three 20 Ampere rails, but at the expense of meaningful overload protection. My Antec Neo HE 550 supply has three + 12 VDC rails at 18 Amperes each (obviously not simultaneously); 216 Watts should be more than enough to run any currently available components. And 12 X 20 = 240 rails for a 750 Watt supply would not seem to be much of a limitation. Either way your overclocked E6600 / 2 X 8800 GTX overclocked system would be well taken care of. Post some numbers, already! I am especially interested in how well voltage/temperature/fan monitoring applets work, and what nMonitor and nTune do for your system. Phil Weldon |
run its three outputs.
http://www.hardwaresecrets.com/article/361/5
Between www.jonnyguru.com and hardwaresecrets.com , you'd be surprised
how many supplies are breaking the rules. PC Power and Cooling is
being honest about it, while the others practice deception. I wish
Jonnyguru would do more high current testing, to show which supplies
don't have current limiters. Or limit at what are basically thermal
limits (rectifier overheat).
I'm still waiting for a report of a spectacular failure... :-)
Paul
<<snip>>
#13
| |||
| |||
|
|
'Ed Medlin' wrote, in part: PC Power and Cooling seem to go another direction with one very large12v rail. http://www.pcpower.com/products/view...php?show=S75QB is the PSU I have and you can read the info there on how they go about doing it. 60a is a LOT for a single 12v rail. Let me know what you think about it yourself. _____ I think you have a very good power supply, but that the engineering department of 'PC Power and Cooling' is subordinated to the marketing department. One single 60 Ampere + 12 VDC rail may be SLIGHTLY more convenient than three 20 Ampere rails, but at the expense of meaningful overload protection. My Antec Neo HE 550 supply has three + 12 VDC rails at 18 Amperes each (obviously not simultaneously); 216 Watts should be more than enough to run any currently available components. And 12 X 20 = 240 rails for a 750 Watt supply would not seem to be much of a limitation. Either way your overclocked E6600 / 2 X 8800 GTX overclocked system would be well taken care of. |
PSU with 2 HD's, 1xDVDRW, 1x floppy, and a 6800GT. I'm guessing that my new
PSU was a good investment if I was going to even attempt a C2D OC.
Unfortunately, I think the one I ordered got lost in the mail so I might
have to get another.
--
Phil
#14
| |||
| |||
|
|
'Paul' wrote, in part: | This web site does anatomy checks. The NeoHE 550 has one 12V rail to | run its three outputs. | | http://www.hardwaresecrets.com/article/361/5 | | Between www.jonnyguru.com and hardwaresecrets.com , you'd be surprised | how many supplies are breaking the rules. PC Power and Cooling is | being honest about it, while the others practice deception. _____ What Hardware Secrets naively call 'jumpers' are actually shunts. Each rail has a shunt. The over-current sensor measures the voltage drop across the shunt. In the case of the three +12 VDC rails if any of the three sensors measures an overcurrent then the entire + 12 VDC is shut down or limited. Page 29 of http://www.formfactors.org/developer...public_br2.pdf explains the short circuit and current limit protection. The Antec Neo HE 550 provides this through an IC that, among many other functions, senses the voltage drop across the three + 12 VDC shunts ( + 12 VDC for the CPU, + 12V1DC and + 12V2DC. So it really is that the 650 Watt PC Power & Cooling supply does not adhere to the specifications while the Antec Neo HE 530 does. |
30A rail specifically for video, with its own overload protection.
#15
| |||
| |||
|
|
'Paul' wrote, in part: | This web site does anatomy checks. The NeoHE 550 has one 12V rail to | run its three outputs. | | http://www.hardwaresecrets.com/article/361/5 | | Between www.jonnyguru.com and hardwaresecrets.com , you'd be surprised | how many supplies are breaking the rules. PC Power and Cooling is | being honest about it, while the others practice deception. _____ What Hardware Secrets naively call 'jumpers' are actually shunts. Each rail has a shunt. The over-current sensor measures the voltage drop across the shunt. In the case of the three +12 VDC rails if any of the three sensors measures an overcurrent then the entire + 12 VDC is shut down or limited. Page 29 of http://www.formfactors.org/developer...public_br2.pdf explains the short circuit and current limit protection. The Antec Neo HE 550 provides this through an IC that, among many other functions, senses the voltage drop across the three + 12 VDC shunts ( + 12 VDC for the CPU, + 12V1DC and + 12V2DC. So it really is that the 650 Watt PC Power & Cooling supply does not adhere to the specifications while the Antec Neo HE 530 does. Phil Weldon |
engineering, so I missed that. I see they are labeled on the PCB as "R508"
and "R509", so they do look like they're intended to be resistors.
There are supplies that don't have a current limit, because on at least
one of them, they managed to draw 30A from the 12V rail, even though the
label on the supply was rated for a lot less.
Like I mentioned, I'd really prefer if the jonnyguru site would include
some overcurrent testing, just to see what will come out. Overcurrent
circuits have to be set a bit higher than the rating on the label (like
maybe 30% higher), and with a power supply tester, it would only take
a few minutes of testing, to see how far each rail can go, and
whether the current level is just 30% over, or a lot more.
And I don't understand how this 20A limit thing is being handled. For
example, the PC Power and Cooling Supply, has quite different agency
approval stamps on it, than another supply I looked at. Very minimal in
fact, on the Silencer 750.
But what is weird is, the claimed 240W secondary limit is supposed to be
part of IEC 60950 (and follow-on specs). At least, that is what a couple
of the ATX power supply specs (from formfactors.org) say.
So, when I look at this product:
http://www.newegg.com/Product/Produc...82E16817148008
the output is listed as 12V @ 35A, implying a single output, and
yet they have a number of agency approvals listed with 60950
in the title. (This is in the specification section of that page.)
What I'm missing here, is I've never seen a copy of 60950,
either the older ones (when it was called IEC 950), or the
more recent ones. I don't know if someone doesn't feel the
spec applies to them, or what.
Paul
#16
| |||
| |||
|
|
Ed Medlin wrote: I had to do a bit of case modifications to get it installed. I also moved the water pump and have the water block installed on the P5N32-SLI. I see there is a metal fan guard? It almost seems that it should be on the other side of the fan to keep the wires out! No fan guard, but there is also no problem routing and securing wires |
panel wiring.
Quote:
|
The 2 8800 GTXs are going to fit fine, of course I will have only one available PCI slot with those. That is great that they are going to fit, and good luck with the bottom drives! You know, I thought I would have problems with the top drive bay and |
ears on the removable sliders do touch the top 8800. The rest of the
bays are fine. I can't figure out the procedure for mounting a drive in
that sucker. Swiftech did not ship instructions for the Antec case, so I
am on my own there. The bottom drives are actually harder to get wiring
to than the top. I have to route the SATA and power cables through a
small hole just beside the 100mm center fan. I should have it up and
going sometime today and ready to install the OS. I am going with XP
Pro, although I did get Vista Ultimate OEM if I want to go that way
later. If I do, I will probably dual boot until the problems with Vista
and SLI video are resolved.
I have everything installed except for optical and HDDs. The case is
extremely quiet and the water cooling install and bleeding went
flawlessly. The 350 pump moves a ton of water and the water block is a
massive hunk of copper, aluminum and stainless. This has to be the
quietest running system I have ever built. The three 100mm fans are
silent as are the two 8800 GTXs. I apprecieate all the info on the case.
I saved a bit of time there.
Ed
Quote:
|
I ordered a new cooler, the Thermalright Ultra 120 Extreme, and am thoroughly disappointed! The E6400 is taking way too much voltage to keep it stable at 3.52 GHz under load (1.45v), and it's getting too hot. I got the Noctua NF-S12 1200 RPM fan for it, which was a mistake. The space at the ends of the blades leaks so much air rearward that I thought I had installed it backwards and even took it off. I feel hardly any air coming out the other side. The cooler appears to have a piece of thin copper sheet covering the bottom, which is folded-up and has holes from which the heat pipes protrude. WTF, an extra thermal interface? Perhaps it is soldered, but I thought solder was a lousy conductor of heat. It was so not-flat that I was compelled to lap it smooth for a half hour. I guess that's why they tell you to apply thermal compound to both surfaces. Of course the nickel plating is now gone-- another poor conductor, I understand, but relative non-reactive. The mounting system applies all the pressure to the CPU on the front and centers the force on the rear entirely on the same small area on the rear through foam tape. The Biostar board is so thin and flexible that this mounting system scares me. In retrospect, I wish I had bought the Noctua Cooler instead, for the better mounting system, but with a different fan. I have not extensively tested, but I can run it cool at 3.2 GHz at relatively low voltage, but must drop memory speed to 400. Both Superpi and Sandra show no gains over 440 x 7 = 3.08 GHz. Perhaps I should have left well enough alone, but I just had to know. I said to myself, "You're going to need that cooler for the Quad core in August." Next plan, lap the CPU's IHS and swap the Noctua for one of the Antec Tricool case fans... |
#17
| |||
| |||
|
|
'Ed Medlin' wrote, in part: | PC Power and Cooling seem to go another direction with one very | large12v rail. | http://www.pcpower.com/products/view...php?show=S75QB | is the PSU I have and you can read the info there on how they go | about doing it. 60a is a LOT for a single 12v rail. Let me know what | you think about it yourself. _____ I think you have a very good power supply, but that the engineering department of 'PC Power and Cooling' is subordinated to the marketing department. One single 60 Ampere + 12 VDC rail may be SLIGHTLY more convenient than three 20 Ampere rails, but at the expense of meaningful overload protection. My Antec Neo HE 550 supply has three + 12 VDC rails at 18 Amperes each (obviously not simultaneously); 216 Watts should be more than enough to run any currently available components. And 12 X 20 = 240 rails for a 750 Watt supply would not seem to be much of a limitation. Either way your overclocked E6600 / 2 X 8800 GTX overclocked system would be well taken care of. Post some numbers, already! I am especially interested in how well voltage/temperature/fan monitoring applets work, and what nMonitor and nTune do for your system. Phil Weldon LOL........patience, patience. I am going as quickly as I |
remove the bottom 8800 again....:-(. The way they designed the MB, the
SATA connectors are behind the bottom card. It won't interfere with it
at all, it is just impossible to get to them with the card installed.
Even my skinny fingers can't get to them at all. I am getting close to
finishing it up and getting XP installed.
Ed
#18
| |||
| |||
|
|
Phil Weldon wrote: Post some numbers, already! Yeah, we know you stayed up all night like you always do! I am especially interested in how well voltage/temperature/fan monitoring applets work, and what nMonitor and nTune do for your system. Try this for monitoring temperature: http://www.alcpu.com/CoreTemp Got it..........thanks. |
Ed
Quote:
#19
| |||
| |||
|
|
You know, I thought I would have problems with the top drive bay and the video cards. I am losing one (the top one) drive bay because of the ears on the removable sliders do touch the top 8800. The rest of the bays are fine. I can't figure out the procedure for mounting a drive in that sucker. Swiftech did not ship instructions for the Antec case, so I am on my own there. |
http://www.silentpcreview.com/article255-page4.html
I had the most trouble with the floppy and optical drives, knowing which
holes to use.
Here is the downloadable manual, if you'd call it that:
http://www.antec.com/us/productDetai...?ProdID=81802#
I assure you the one that came with it is no better and much smaller!
I found the online reviews of the case more helpful than the manual.
Quote:
|
The bottom drives are actually harder to get wiring to than the top. I have to route the SATA and power cables through a small hole just beside the 100mm center fan. |
though, right? http://www.silentpcreview.com/article255-page3.html
#20
| |||
| |||
|
|
Next plan, lap the CPU's IHS and swap the Noctua for one of the Antec Tricool case fans... |
layer of AS5 I had applied. I checked both mating surfaces with
a machinist's square. The IHS was virtually flat, but I lapped it a
little for good measure, but didn't even get down to copper. The
so mounting system makes it easy to turn the cooler slightly \ or /,
hopefully that won't effect the curing of the AS5.
The Antec fans only have a four pin Molex connector, so I used
a Nexus "Real Silent" 1200 RPM case fan I had ordered at the
same time. Unfortunately, it took some modification because
of the way the Thermalright fan clips attach requires that the
mounting holes don't go all the way through, which, of course,
they did. The Nexus doesn't move a lot of air either, but there
is much less perceived leakage. I have the system set for
426 x 8 = 3.4 GHz @ 1.382 vCore which is the minimum
required for stability. 1.375v failed Orthos testing after about
7 hours. Temps as reported by CoreTemp are upper 50s to
low 60s when running Orthos, depending upon FFT size, with
the case closed and all fans on low. The Biostar bundled
monitoring utility reads CPU temp as cooler by about 5° C. I
see no difference in temperatures after remounting. I'm having
trouble imagining this cooler working well on a quad core. I
guess I'd have to try a faster fan before making that judgement.
|
« Previous Thread
|
Next Thread »
| Thread Tools | Search this Thread |
| Display Modes | |
Linear Mode
| |
Powered by vBulletin Version 3.5.3
Copyright ©2000 - 2008, Jelsoft Enterprises Ltd.
Copyright ©2000 - 2008, Jelsoft Enterprises Ltd.