So how is shotgunning doing? Which parts do you know – without doubt – are not defective?
Your problem is typically found in the power supply ‘system’ - which is more than just a power supply. Especially useful would have been a signal from power supply controller that orders the power supply on and a signal from the supply telling the CPU it can now operate.
Things such as memory were never on the list of viable suspects. With 30 seconds of labor, by getting six voltage numbers using a multimeter, my next post could have defined which one component is defective. A less than $18 tool that would have identified the suspect immediately if not limited the problem to very few suspects.
What the meter measures: set to 20 VDC. Touch meter’s probe to each requested colored wire where power supply connects to motherboard. Measure the green, gray, and purple wires both before and when the power switch is pressed. Report numbers and power up behavior. Also measure any one of red, orange, and yellow wires when the switch is pressed. Report all six voltages here. So little information isolates the failure to one or very few suspects.
Not even disconnecting one wire or component is necessary. IOW no money to replace perfectly good parts. No replacing parts on wild speculation. No shotgunning.
Looks like I found the culprit. After plugging in my friend’s old geforce 7950, it posted.
How simple after all. It’s weird because boot behavior wasn’t consistent, my first board didn’t properly beep code (and why wouldn’t it for merely a bad video card?) and in a power surge, the graphics card is the last thing I’d expect to be damaged.
I’m going to put in my old motherboard to make sure both parts didn’t fry, then return the new board and buy a new video card I guess.
Is there anything I can try to resurrect my dead video card before I give up on it?
The only thing I can suggest is for cleaning the contacts. You can use a pencil eraser on the contacts of the video card to erase the oxidation. I really doubt it’s the problem. When I’ve found the problem on a video card I’ve found a missing resistor or such physical damage and you won’t be able to fix that.
You can see if the manufacturer offers money for a trade in. Sometimes they do.
Forgive me if this is ignorant, I haven’t bothered to look at the specs on those video cards: Are you using the same slot every time? I’ve seen one slot on a motherboard fry and leave others unaffected.
I was wrong. The old motherboard works - it refused to boot with ram in the 3 or 4 slots, but not the 1 and 2 slots. I don’t recall P35 boards having any necesary order in which you have to use the ram slots, so I’m not sure if that’s a design issue or a malfunction issue.
But… I’m running on the old system with the older, working video card. I’ll need to do some stress testing but everything looks good.
My old motherboard, the p35, won’t give me any error codes whatsoever. When I knew everything worked except the video card, no beep. When I had the ram in the wrong spots, no beep. When I had no ram at all, no beep. The speaker/controller works because it gives 1 beep on POST, but that’s it.
For example, a completely defective power supply can still boot a computer. And a good power supply in an otherwise perfectly good computer can fail to boot. Shotgunning would not reveal any of this. Is anything in your power system ‘good’ or ‘bad’? You still do not know.
You have assumed everything is either 100% good or 100% defective based only in shotgunning principles. Those who recommend shotgunning do not learn the many details that actaully determine good or bad. They assume it if appears to work, then it must be good. And so the defective power supply boots a computer.
List of reasons for your strange behavior could be 30 items. Most of which others do not even know exist. And still you do not yet know of even one component that is definitively good or definitively bad.
You did not first learn what the problem is. So you reset the CMOS and what else? How many problems did you ‘fix’ that were not defective? At this point, how many more things were fixed and are now defective? Fixing something without first learning what is defective - shotgunning. Nobody knows how many defects now exist until you start defining what is bad (definitively) and what is good (definitively).
Measure progress by what gets moved from the ‘maybe’ list to a ‘definitive’ list. Nothing gets there by shotgunning. You still have nothing on either list - the “definitively good” or “definitively bad”. No progress.
Welcome to why we fix things ourselves. To learn. And to learn from our mistakes. You are doing what a majority recommend rather than learn good diagnostic practice. You can learn from those who actually do this stuff. Or from an overwhelming majority who do not even know how electricity works. Who know only using speculation, swapping parts, and junk science reasoning - called shotgunning.
You had resources to ‘follow the evidence’ including a description of how to do that in 30 seconds. Instead you kept swapping parts. No responsible technician - not even a car mechanic - does that and stays employed.
So after doing so much work and posting so often, how many components do you know as “definitively good” or “definitively bad”. Provided was how to know using 30 seconds and resources that were always available.
I can appreciate your frustration. It is common and too routine. You saw and feared a simple procedure. Many routinely fear and avoid something they never did before. As a result, how many things have you “fixed” and still accomplished nothing? Even your provided conclusions are only speculation - some not even based in how computers work. They say in CSI, “Follow the evidence”. But when presented how to do that, people typically do not understand what that means. How to ‘follow the evidence’ means 30 seconds to get six numbers in http://boards.straightdope.com/sdmb/showpost.php?p=11448193&postcount=21
Your answers will only be as useful as the information provided which is why so many others only posted wild speculation – it could be this or could be memory or could be …
Motherboard works. CPU works. PSU works. Ram works. Video card did not work. Other video card works. What else is there to know?
My original motherboard did not output error codes, and I did not until yesterday have a second video card to swap in from the first. What is it that I could’ve done that would’ve isolated my problem?
I already said that the PSU was outputting the correct voltage to the old motherboard - what more was a multimeter going to tell me?
Are you saying my method of swapping out working parts (which correctly identified the problem) is inferior to your method of putting a multimeter to the PSU (which was not defective)?
Quite frankly, I don’t understand where you’re coming from. Swapping in known good parts to a system one by one, or swapping out suspected bad parts to a known good system are perfectly valid ways to diagnose computer problems. It would’ve solved my issue in an hour - if I had another working computer with the same type of architecture, I’d swap the parts out one by one and I’d have realized my video card is faulty quite fast. The problem in this case was that I did not have the spare parts, so I made an educated guess and bought the part I suspected was broken. I was wrong, so I found a way to acquired working hardware of the next most likely candidate, which turned out to be correct.
Again, the computer can boot just fine when the power supply is defective. How do you know that supply ‘system’ is OK? Computer booted with one video card? A classic mistake made by shotgunning.
One perfectly good video card can work in a computer with defective supply. Other perfectly good video card can fail in the same system because the supply ‘system’ is defective.
Did you know it is a ‘system’ or just assume the PSU is the entire supply ‘system’? What were those numbers? Was the supply under full load when measured? But again, those same numbers say to me much more than you will ever know. To have a useful reply, all those numbers (especially when the supply ‘system’ is under maximum load or in the defective machine) must be provided. Just too many facts in those numbers to say why. Your ‘voltages are good’ declaration says nothing useful. What exactly were numbers from all six wires and how was the load maximized?
Listed was how to tap knowledge from the few to determine what really is and is not defective. Currently, you have posted no reason to believe the power supply ‘system’ (and yes it is a system containing multiple components - not just a power supply) is good. Explains why you are so confused. Without numbers, those who really now this stuff remain mute. Your replies will only be as good as facts that you provide. Currently nothing says the power supply ‘system’ is good. Your symptoms (sometimes it boots; other times it does not) are classic of a supply ‘system’ that is defective. Using what you have posted, only wild speculation can identify a suspect - which is why you have so many posts that say, “Try this and replace that” - shotgunning.
Fixing something comes after the problem is located. Just another reason why voltages on six wires from a 3.5 digit multimeter, both before and when the power switch is pressed and when the system is under maximum load, are essential for a useful reply.
So, new board + GF7950 = POST
old board + GF7950 = dead
new board + GF8800 = dead
old board + gf8800 = dead
I would suggest that these results also indicate a possibly flaky power supply. Graphics cards such as these require +12V, via a 2x3 pin connector from the power supply. Some power supplies may generate +12V for the CPU and the PCI-E from the same regulator. The 8800 is more power hungry than the 7950, IIRC.
Tell you what, if this system can run games with complete stability, and do everything else I want a computer to do indefinitely, and yet somehow has some secretly defective power supply… I’ll learn to live with it somehow.
The voltages stay within 2% of rated capacity even when stressed with benchmarking utilities (like running prime95 and ATItool). CPU voltage suffers load vdroop at exactly the rate expected of the cpu and motherboard, and how they previously worked.
It never booted. I never got a POST until I replaced the video card. Sometimes it would power itself up and power itself down, which is different from intermittently booting. That seemed to be the results of an error checking system responding to different configurations. I’m told it’s normal for example if you don’t have ram installed for the system to power up, beep, and then power down (or reset).
You really have a hard on for the power supply.
I’m curious - let’s say I got a multimeter and the power supply checked out. What would be your next step in diagnosis? I sure hope it wouldn’t be replacing parts 1 by 1 to see when it did and didn’t work.
Incidentally, if you are really trying to help, I’m not trying to take an insulting tone. But you seem needlessly aggressive, admonishing me for getting the system running without your preferred diagnostics.
Which says nothing useful. Again, if those voltages are known, then post the numbers. Why short your best help of facts? Does not matter whether those numbers are in 2%. What matters is what each number is, as defined: before the power switch is pressed, when the power switch is pressed, and when the system is under maximum load. Why do fear to provide numbers when only the numbers (not your conclusions) provide useful replies?
Prime95 does not load the system. Learn what Prime95 does. But you have provided a fact. You never loaded the supply ‘system’. So that 2% number is even less definitive.
It is a symptom of a defective supply system. But we still know nothing but speculation until those numbers are provided.
What is this error checking system? If that error exists, a BSOD display exists. Which means you have critical numbers and error messages to post here. You have assumed a response by an error checking system that does not exist.
What tells a microprocessor to operate? Power supply ‘system’. Same system that remains undefined (neither definitively good nor definitively bad) because those critical numbers were not provided. You made conclusion from speculation and did not even know the supply ‘system’ permits the CPU to start and can shut it down. Is that ‘system’ defective? Nobody can say until those numbers are provided.
By your own admission, why some combinations work and why others do not, obviously will confuse you. But then you also assumed PSU, motherboard, and video controller are good when your own post says all three remain in a third condition - still undefined.
Nothing I have posted says your power supply or other parts of that supply ‘system’ are good or bad. Nothing you have posted says a supply ‘system’, video card, or motherboard are “definitively good”. So you remain confused. After all that work, nothing appears on a list of accomplishment. Everything, based upon what you have posted, is still undefined – not yet ‘definitively something’.
“Sometimes it would power itself up and power itself down” is a symptom of a defective power supply system AND a symptom of maybe 15 other problems. Why does intermittent booting occur? The only useful answer comes from “following the evidence”. That starts in a supply ‘system’ that can make all other components act strangely and confuse you. The only useful answer starts with 30 seconds and numbers from those six colored wires. Until the many components of the power system are definitively good, then confusion can exist.
What happens next? Wait a minute. You know supply voltages are within 2% and do not have a multimeter? What did you use to measure? You cannot know those voltages are in 2%.
What happens next? Remember the part where I keep saying those numbers contain information you do not even realize? Without specific numbers, too many ‘next steps’ exist to discuss. Nothing I have posted is insulting. This is how logic works, Only facts are posted. Any ‘tone’ is only your own biases. Of course, if you detect exasperation, well, again, look at the facts that I keep reposting. You said certain voltages were within 2%. Did not say what voltages. Did not properly load the system. Did not even have a multimeter which means you have no idea what those voltages are. You have done almost everything necessary to subvert any assistance. You shotgunned, and than do not understand why you were confused? No insult - just facts.
The fact that my computer runs, and passes stability tests, where before it wouldn’t even POST, is not useful?
Sir, I have to assume that you are not quite mentally all there. Why would I “fear” posting those numbers?
This is a happy ending. I had a hardware failure, I isolated and replaced the part, everything works again. I don’t get what your problem is.
But if you must know, the voltages under booted idle are CPU core 1.216v, 3.3v rail is 3.29, 5v rail is 5.03, 12v rail is 11.9, northbridge core is 1.34, vdimm is 1.93 (set in bios at 1.9), 5 volt standby is 4.98, battery is 3.29. All normal voltages.
Load voltages are CPU: 1.2v, 3.3: 3.28, 5v: 5.03, 12v:11.9, nb 1.34, vdimm 1.93, 5vsb 4.98, bat 3.31.
The CPU voltage drop at load is normal for core 2 CPUs, called “vdroop”
These are basically the same voltages I have recorded from when I built my system.
I’m well aware of what P95 does. The blended torture test will maximize the CPU’s power usage beyond anything any user will do with it, and not only that, but it double checks the result of any math in order to detect any potential errors the CPU puts out due to power or heat problems.
I’m also using ATITool, which does the same thing for the GPU. Between stressing the CPU and GPU beyond anything a normal user would ever do, error checking the whole way, the power supply will be taxed beyond anything else someone is going to do with it. What else could the power supply be doing that it’s not using when those programs are running the torture tests?
There is no evidence whatsoever that my power supply is defective.
If you remove everything from your motherboard except the CPU, what will happen is that it will power on, beep 3 times, and then power off. Is your PSU defective? Woops, better break out your multimeter - the fact that it works fine is no evidence that it works!
Uh, what? You know what POSTing is, and what beep codes are, right? The motherboard attempts to power up and boot the computer - if it encounters a critical error with the hardware during this process it’s designed to beep out the PC speaker a certain code that indicates the problem. My old motherboard didn’t, the replacement did.
You understand that a computer that cannot POST - which does not even reach the point where anything is displayed on screen, will not have a BSOD error, right?
The supply provides the system with stable power that allows it to pass stress test. The numbers from two seperate voltage meters in two seperate motherboards indicate that voltages are well within spec.
You don’t seem to fundamentally agree that something working proves that it works.
The parts that work now and stand up under stability testing clearly work.
Powering up and down is also the symptom of an aborted POST process due to non-power-related faulty hardware. You can prove this to yourself if you were so inclined by, as I said, removing everything but the CPU from your motherboard and powering up. It will power up, beep, and power down.
You do understand that there are devices within a motherboard that detect and regulate voltage, right? They are effectively a built in multimeter. This information is available to the computer BIOS and any program that reads those memory locations where this data is stored.
I have done “nothing but subvert any assistance”? Is fixing the problem nothing? That’s sort of a big deal, when, you know, the whole issue here is fixing the problem.
So if it’s your case that sure, your computer will stand up to stability tests, and sure, your computer can be used for whatever purpose you want it to be… BUT IT’S STILL SECRETLY BROKEN MAYBE!!! … well, I can live with that.
I also don’t think you’re using the term “shotgunned” correctly. Testing defective parts one by one does not seem like a shotgun approach. Now, changing 5 things at once, and having it work, and not knowing which of those 5 things fixed it sounds like a “shotgun approach” and it makes sense that that’s a bad plan. But I replaced parts with known good parts 1 by 1 until I isolated the hardware failure, and now everything works. How is that “shotgunning”?
There is no evidence that your supply is defective because you did not post the relevant numbers requested, because your numbers come from an unnamed source - not the necessary 3.5 digit multimeter which you fear, AND because those numbers are taken without load. Furthermore, you still did not provide the important numbers - that were requested. Nothing even in your last post provides those important numbers.
''Fear of learning" is widespread - especially among computer ‘experts’ who only learned shotgunning. Why did you fear to take 30 second. Still not provide the required numbers from six wires? If not fear, then why would you spend a hour posting accusations and denials rather than spend 30 seconds getting the useful numbers? If not fear, than what is it inside you that wants to keep you uneducated?
Shotgunning - what only the least informed must do due to insufficient technical knowledge. Same people would then attack rather than learn.
A power supply can still boot a computer - and be defective. And still you post “it boots, therefore it must work”. At what point do you learn even that simplest concept? Well if everything is so good, then you can now explain why some of four good video cards work here but not there. After all, proper diagnostics means you also know why those failures still exist. Shotgunning means you never learn.
You assumed Prime95 provides sufficient load. No it does not. Your explanation demonstrates why Prime95 provides insufficient loading.
Not that you will hear any of this. Rather than take 30 seconds to provide useful and requested numbers, you choose to shotgun, challenge, and accuse. The problem could have been isolated in the next post. Instead you said the computer boots, therefore it is fixed. Ok. Which part was defective and why? You shotgunned. Best you can do is speculate.
A problem increasingly apparent in America where more have less ability to ‘think through’ or diagnosis simple problems. Therefore a majority can only shotgun. You ‘worked harder - not smarter’. It explains why a majority of employees in the Silicon Valley must be Indian and Chinese immigrants (ICs). Too many so called ‘computer experts’ do not even have basic diagnostic skills.
If not fear, then what causes you to resort to shotgunning? Shotgunning - what one must do because of near zero technical knowledge and due to fear to “work smarter - not harder”. Even important signals from the supply controller to tell the CPU to work or stop working - you did not even measure them - did not even know they exist. Instead you ‘assumed’ some mythical error checking circuit must exist. Instead you challenged your help rather than learn.
This post is not about fixing anything. This post is about a problem that could have been isolated in a day. Instead, it took how many days, how many denials, how many ‘try this or it might be that’ speculations, how much replacing of perfectly good parts, and still you don’t know why some video cards would not work in some places. Shotgunning exemplified.
The reason for shotgunning? Well you still did not post some important numbers. And still never learned why those simple numbers were important. You choose to fight and accuse rather than learn. So America needs even more Indian and Chinese - people who actually want to learn. This post is about why America so desperately needs immigrants. This post is about the so many who have so little technical knowledge as to shotgun. So little grasp of good diagnostic procedures as to defend shotgunning.
Prime95 does not provide sufficient loading. When first told that, then why did you instead deny? Those with the least knowledge will deny the most. Will continue to shotgun long after they should have learned not to. Long after they should have learned how to “follow the evidence”.
And still you did not provide numbers number that discuss what the supply controller is doing. It boots, therefore it must be good? A defective supply can still boot a computer. So we pretend that reality does not exist - declare it good only because it boots. Shotgun reasoning - or why even more American jobs must go overseas.
This is the post where I stop trying to be helpful and say why being helpful is futile. So many ‘experts’ who would rather fight than learn.
The source was not unnamed. The two motherboards with voltage detector/regulator devices reported the output voltage of every major rail of the PSU.
Haha, what, I fear? Like, I’m going to hook up a multimeter and say “MY GOD THE BLUE WIRE… IT’S 1.7 VOLTS… MY ENTIRE LIFE IS A LIE!”?
The power available on the 3.3, 5, and 12 volt rails of a PSU are not the important numbers?
I don’t have a multimeter, and given the information available to me, one wasn’t necesary. Plus it sounds like a pain to read the wires under load - since an ATX power connector is sealed when connected, I would have to strip wires to get load voltages and why bother with all that when the motherboard has a device designed to report voltages on it?
I wasn’t shotgunning. I replaced one part at a time until I found the defective one. Shotgunnign would require me to apply more than one change at once, right?
A power supply that boots and every voltage remains stable and normal under load is very unlikely to be defective.
Prime95 and ATITool. That means that the CPU and GPU are running at maximum power draw. The motherboard and memory will be operating at near maximum power draw. The fans are all spinning at their max. The only thing not working at the time is the hard drives and optical drives, which are a small part of the overall drain. My computer, under normal use, would not draw any more load than it would under those tests.
The video card was almost certainly broken. The CPU is known good, the ram is known good, the motherboard is known good, the PSU is very likely to be good, and the video card - well, the system did not boot with the previous video card but boots with a known good one. There’s a very strong chance that it’s the defective part.
I never called myself a computer expert.
Your technical knowledge doesn’t seem very deep. The power supply does not “send signals to the PSU” - the voltage regulator module on the motherboard powers the CPU, and interprets any “signals” sent by the PSU.
How? Tell me, step by step, how I could’ve had the problem isolated in a day.
I don’t know what “some video cards would not work in some places” is supposed to mean. I corrected my earlier post in which the old motherboard and known good video card didn’t work together. So the known good card works, the other did not.
Haha, indeed, me fixing my computer with the “test suspected failed parts with known good parts” approach is definitely the downfall of America.
Why wouldn’t maxing out the CPU and GPU and motherboard and memory provide a sufficient loading?
Haha, now I’m at fault for American jobs lost. If the power supply puts out the voltages it always has, if it can stably run the computer under load, what reason is there to suspect that it’s bad? There’s no evidence that it is faulty in any way.
You weren’t trying to be helpful beyond the first post. You are for some reason angry at me for solving my problem without following your exact instructions. You seem like you may be a crazy person. I was initially grateful for your help, but since I fixed my system you’ve been nothing but angry and insulting.
Incidentally, your technical knowledge is so lacking that I doubt you’d have lead me to the correct conclusion eventually. After all, your next step to fixing a computer that does not POST - that does not boot, that does not even display any data to the monitor, was to look for BSOD codes.