Hard Light Productions Forums
Off-Topic Discussion => General Discussion => Topic started by: Topgun on September 10, 2007, 08:29:42 pm
-
so I get this new computer, assemble it and turn this sucker on.
first, the fans go on and... nothing. the keyboard doesn't receive any power and the monitor shows no signal.
so then I take out the pcie grfx card and replace it with an old 4mb pci card. everything works now... at least for the first few secands, then the computer turns itself off.
any help guys?
-
sounds like you dont have enough power
but id be able to make a better judgement if you posted your specs.
-
450 watt, the grfx card says I need a 400 or greater.
-
well yea but thats how much your card needs, how much does your cpu and drives suck up?
add up all the wattages for every part in your system, you psu should be 100 or more watts above that (and more if you intend to upgrae in the future).
-
basically I have this computer
http://www.cheap-n-easy-pcguide.com/build_your_own_cheap_computer.html
but with an 450watt PSU.
if he got it to work with 400 I think I can do.
-
my old computer has about the same specs. it burned out its 420 watt psu and i put a 550 watt in and it runs fine.
-
I just got the PSU today.... :sigh:
-
The wattage ratings (and requirements, for that matter) don't mean squat. What actual model do you have?
-
thermaltake tr2-430
my bad it's a 430 no 450.
-
Disconnect/remove everything except for a video card, CPU, keyboard, and memory. Tell us what happens then.
-
did that already... nothing.
EDIT:
Maybe it tunes it self off if it's idle? not likely...
EDIT2: I understand the behavior better now. the first time I startup, after it checks dmi data, I get "invalid boot disk" yata yata... if I ctrl-alt-del to restart it goes normaly until it's time for a boot disk... then it just turns off.
EDIT3: if I let it run the first time (when it's asking for a boot disk) it will turn it's self off in about 25 sec.
-
D'oh.. edit..
Just in case, but have you already re-seated the CPU and RAM?
Can't say I'm enthused about the brands, but eh. I honestly don't think the PSU is the problem. I think the mobo is the one saying 'nuh uh.'
Case in point to what CP said, I'm running a 7900GT with a dual core proc, two 7.2K rpm harddrives, SB-XFi, AND yanking power for LEDs via USB for my keyboard mouse and joystick. All this from a supposedly 250 watt PSU. (Dude, I can't kill this Dell!) Eh, still relevant.
Do you have a Windows CD in the drive when it's doing this boot disk yammering?
-
I havn't put in a cd drive yet.
-
Er, and it's a new hard drive too? Might be a bios that doesn't want to do anything without a bootable drive so it turns off...? Kinda guessing now.
Do you actually get the option to enter the BIOS? (F2, F10 whatever). I'm at least thinking the pci-E slot may have to be enabled first before that'll work...
-
It says I can but when I try it, it just goes on like nothing.
my motherboard manual says if this hapens it is because my system overheated :wtf: ?
-
Yikes! Yeah... Make sure you have the CPU heatsink on properly, for sure. Try taking off the heatsink and checking to see if there's an obvious 'bubble' showing no contact between the CPU heat spreader and the HS. What kind of heat-sink are you using btw? Or did you get a retail CPU (as opposed to OEM as the site mentioned)
-
my temp reads at 50c, I don't think that is the problem.
-
But... where are you seeing that if it isn't letting you in the bios? And if it is reading the temp correctly (which it should, considering the 3500s have the Cool & Quiet™ including an integrated thermometer in the die) then yes that'd be ok... running out of ideas.
-
it gives me the CPU temp at start up.
my guess is it has some sorta case detection thingie that turns the computer off when the case is open. but that does not explain the pcie thing...
EDIT: I am going to bed... keep posting ideas if you are on a different timezone, for those of you that are on est, then... I suggest you go to bed.
-
Eh... You should be able to have the motherboard sitting bare on a desk with all necessary components plugged in and still have it run without a hitch. Only 'open case detection' I've ever dealt with has been on enterprise servers.
And yeah, the PCI-E vid card issue is still odd.
I just read on the comments for that mobo that the ram slots were numbered incorrectly... might try swapping ram from white to green or vice versa. I also read another comment that described the EXACT PCI-E problem...
-
The wattage ratings (and requirements, for that matter) don't mean squat. What actual model do you have?
thats because people dont know how to read them. most people only see the single number rating, which at best is a gross approximation of what the power supply total is (at peak). its more there for advertising and really isnt an accurate rating.
the actual power rating is that grid of numbers which give you the maximum amppage draw on each voltage line, overloading one or more of those is generally bad for the long term health of the system. each device should have a power rating for each hot pin. if you sum up the amppage draw for all the 5 volt pins in the system, it should be about 75% of the maximum amp draw.
now most devices tend not to make their per pin current draw apparent perfering a simplified and highly innaccurate approximation. infact ive only seen proper detailed ratings listed in only about half of the computer partds ive ever bought.
the reason companies like dell can get away with much lower power ratings is simply that they apply the hard numbers. for example a psu might need more power for the 12v cpu connector, for example when using amd over intel, yet need much lower 12 and 5 volt lines because youre only running a couple drives. now you can tilt the aproxamate total power by ramping up the power on more useless lines on the psu. dell gets away with it because they have the power to order custom production lines specifically designed for their power needs. in other words when they want to produce a new model, they get some engineers together. when we want to build something, we buy parts from newegg and hope the manufacturer's didnt lie about their stats.
-
how can I be sure it is the PSU? this is the only one I have that can plug into this motherboard.
-
Are there no stores nearby that sell some generic ATX PSUs? Nab one, test it and return it... better if that's a big chain though since most smaller shops are more unforgiving on the restock fees... maybe if you blame some moronic friend who said he needed an "I dunno, what's the usual?" PSU and turns out he's got an iMac. Dumb friend. Stupid stupid mac user friend! :nervous: What?
If there are any small computer shops (and there are bound to be, unless you're in the Mohave) you could also see if they'll test the mobo with a PSU they have sitting around. Same at the big store, though I'm sure they'll be more inclined to charge a testing fee.
Anyway, from the comments on TG I read about that particular motherboard, I'm still more inclined to blame that. ThermalTake has supposedly cleaned up (so I've read) it's act on overstating their PSU capabilities, and in my experience, the symptoms described are not like any power issue I've come across. No self-respecting 150watt PSU would fail to power a 3500+, a HD and PCI video card.
-
tech suport says the thing is overheating :doubt:. it might also be the fan. the tech guy said if it does not have a cpu fan conected to the motherboard it will read as an overheat. I have a cpu fan but it is not conected to the motherboard, just the PSU. I guess I need to buy a cpu fan... :sigh:
-
the actual power rating is that grid of numbers which give you the maximum amppage draw on each voltage line, overloading one or more of those is generally bad for the long term health of the system. each device should have a power rating for each hot pin. if you sum up the amppage draw for all the 5 volt pins in the system, it should be about 75% of the maximum amp draw.
Those numbers are pretty useless too. There are many power supplies that essentially have nonsense written on those stickers and will explode at half their rated current loads. You have to look at individual models (or brands, at the most) and what their reputations are.
The only voltage lines that are relevant these days are the 12V ones though. All of the power hungry components (processors, video cards, hard drives and optical drives) run almost exclusively off 12V, and even crappy power supplies put out enough on 3.3V and 5V.
how can I be sure it is the PSU? this is the only one I have that can plug into this motherboard.
The symptoms you describe seemed to indicate a PSU problem. That power supply should be quite okay given your system though. The midrange (400-650W) TT power supplies are fairly middle of the pack. They aren't great, but there is much worse out there.
tech suport says the thing is overheating :doubt:. it might also be the fan. the tech guy said if it does not have a cpu fan conected to the motherboard it will read as an overheat. I have a cpu fan but it is not conected to the motherboard, just the PSU. I guess I need to buy a cpu fan... :sigh:
That's possible. I got a new board last weekend and it kept making an alarm sound from the PC speaker until I moved the fan to the CPU fan slot. It's usually possible to disable a warning system like that from the BIOS though.
If your fan has a 3 pin connector and is not a very high speed model (with a power draw less than about 1.5A), you can run it off the motherboard. There are also converter cables available for a few dollars.
-
Note to motherboard designers: The mobo should register the temp from the CPU, and not worry about the CPU fan! :mad:
Good thing you didn't try to use water cooling.
What'd they say about the PCI-E problem, though?
-
he said it causes overheat so bad that it won't even turn on. I am starting to believe this overheat thing since I tested the CPU heat continually. the temp usually says 45c at startup but after awhile it goes up to 70c, which is the max temp before the computer will turn itself off.
-
WOAH! Erm... Sure about that heatsink placement? That really sounds like a HS/CPU contact issue again...
-
45c at startup? That's horrible. No CPU, even a Prescott, should ever get that high after a cold boot. Something is seriously wrong with your thermal conductivity.
-
I will be geting this cpu fan.
http://www.tigerdirect.com/applications/searchtools/item-details.asp?EdpNo=1862753&body=REVIEWS
I think my heat sink is broken.
-
45c at startup? That's horrible. No CPU, even a Prescott, should ever get that high after a cold boot.
Some old ones do. The 1.4ghz Athlon C in my retro game rig idles at 58C on startup and loads at 75C or so. It works fine though.
But yeah, that's way too hot for a 3500, assuming the board is just not misreporting it (which is not unheard of).
I will be geting this cpu fan.
http://www.tigerdirect.com/applications/searchtools/item-details.asp?EdpNo=1862753&body=REVIEWS
I think my heat sink is broken.
What heatsink do you currently have? They can't really break like that in most cases.
-
(Dude, I can't kill this Dell!)
That is a hobby of my friend's. He changes the admin passwords on the school PCs, adds network connections, deletes network files, changes the BIOS, etc, etc. (of course he doesn't do this on the new HPs)
-
My XP system died, same sort of err messages....
...simple solution, smell your hard drive (C:\) if that smells like melted solder, your comps turned to crap and its time to build a new one.
(dont buy Vista its rubish)
-
checked the heay sink and it is too small for my 939. this new one should fix it.
-
I will be geting this cpu fan.
http://www.tigerdirect.com/applications/searchtools/item-details.asp?EdpNo=1862753&body=REVIEWS
I think my heat sink is broken.
Chances are that you can get it cheaper at NCIX.com (http://ncix.com)
-
I once had a 3500+, and it worked perfectly. Until...
After a few months of using it, the cpu temperatures started to rise. And rise. And rise. And rise. When it used to idle at 40c, it would go past 60c and into 70-80c when I ran anything even remotely intensive. After a while, I decided to replace the fan, because that would be the only way to fix it. (all other options exhausted). Lo and behold, when I removed the fan, the fan was literally baked onto the cpu via the conductive gel. NOTHING would be able to get it off without annihilating my CPU. Unfortunately, a... er... slight accident happened with the motherboard when the cpu was removed, damaging the clamps that kept the cpu in the socket. My computer never booted again.
To this day, I use my Macintosh. It's a dual-cpu G4 mac. The heatsinks literally belong in servers. It sounds like a mother****ing leafblower. Leafblowers are totally useless machines, btw.
-
unless you need to blow leaves :D
rather than getting one of those made in china power supplies. i got a made in germany power supply. i know im overpowering the thing. e6700, 8800 gts, 2 satas, xfi, and im using all 8 usb ports. yet my psu is only rated for 600 watts. but it just chuggs along fine and all my voltages are all within their tolerance range. now i doubt i could sli on this thing, but i think i got my moneys worth. german engineering ftw!
-
To this day, I use my Macintosh. It's a dual-cpu G4 mac. The heatsinks literally belong in servers. It sounds like a mother****ing leafblower. Leafblowers are totally useless machines, btw.
That would suck. :p I thought Macs are usually quiet. I had a fan like that at one point (a 6000rpm 60mm screamer that you could hear a few rooms away) and got used to it over time, but I later switched to a larger heatsink with a 2700rpm 92mm fan and the sound felt that much more comfortable.
You can get very powerful heatsinks that are also quiet. I have used a Big Typhoon for the last year or two and the stock fan is inaudible, even if I put my ear up to it. The Tuniq Tower is another good option that is bit more powerful, although also slightly louder.
rather than getting one of those made in china power supplies. i got a made in germany power supply. i know im overpowering the thing. e6700, 8800 gts, 2 satas, xfi, and im using all 8 usb ports. yet my psu is only rated for 600 watts. but it just chuggs along fine and all my voltages are all within their tolerance range. now i doubt i could sli on this thing, but i think i got my moneys worth. german engineering ftw!
Actually, if it's a good quality 600W, it's more than enough and will easily power another video card. For reference, my current setup consists of a 3.6ghz E6750, X1900XTX at 700/800mhz (this old card is actually more power hungry than any of the 8800s), two hard drives, X-fi, 10 fans and some other stuff. It uses 400W on load (Orthos and rthdribl running simultaneously) measured at the wall outlet. If I subtract 90W for the monitor and take the power supply's 75% efficiency into account, that comes out to a measly 230W drawn by the actual PC components.
-
my case really only has 3 fans, and the smallest is a 120mm. the 2 250mm case fans really give me a ****load of cfm for not that much noise. the psu fan is a 140mm i think. with fans heat spreaders and heat pipes on all the major stuff, i dont think i have a heat problem. it sedom goes over 40c.
how would one go about measuring the actual power useage for the whole machine?
-
There is a power meter on my UPS so I use that, but you can buy a standalone one for $20 or so to measure it. The Kill-a-Watt is a popular one. Make sure to factor in the monitor's power usage and PSU's efficiency rating.
-
I once had a 3500+, and it worked perfectly.
dude, I have a 3800+ x2. why is everyone saying I have less? did I make a typo?
-
The link you posted about the 'build a cheap PC' listed a 3500+ ;)
-
the new fan seems to fix everything, havn't tried the graphics card yet though.