On the verge of giving up. :(

[Chibiabos]

I thought I was hot $#!t really. Ah, the 1990s when people thought you were a GOD for being able to assemble your own PC, and having even minimal IT skills virtually guaranteed you a good paying job and whatnot.

I put together several of systems starting with the early days of Pentium. The last "good" system I put together was my 1 GHz Thunderbird. I bought with that a terrible heat sink ... it worked, really, as in the comp didn't overheat, but the noise was horrible, and I finally appreciated a good heat sink and replaced the one I had bought with a THG-recommended one that improved noise tremendously (that is, ran much quieter).

I finally tried to piece together a new system in Spring of 2005 or so and the result? Utter, complete disastrous failure. I was only working part time, so it had taken me all those years to save up to get the components I wanted. The hard drive failed, as did its replacement and its replacement before I finally gave up on SATA. I couldn't figure it out ... I went with Western Digital which had always been a solid brand for me (as opposed to Seagate and Maxtor, both of which had always been failures), but brand new drives would go bad just a few days after installing them.

Giving up on SATA and going back with IDE didn't solve all the problems, though. Crash after crash, hard lockups whenever I tried playing a more recent 3D game. In the case of Homeworld 2, within 10 minutes. I could usually manage in normal Windows stuff, and very old games (like Starcraft), but even then I'd suffer the occaisional crash or lockup.

I didn't overclock anything, just an Athlon64 3200+ and a Radeo 9600 AIW graphics card. Since I couldn't tell what was failing, I didn't know what I should try to get a warranty on and wound up just eating it. A couple months ago, I tried fixing it with a new heat sink and new computer case that was more spacious and had a lot more fans, but that did not do the trick. I gave up and decided to try building a new computer from the ground up, only re-using the monitor and hard drives from the original (so I could copy my old stuff; I ordered a new hard drive for a fresh OS install, of course).

Its a completely new mobo, a completely new graphics card (X1300 with DUAL SLOT type heatsink on it), new RAM, everything ..... yet I experience very similar same lockups and crashes. WHY!?

I'm a big TBS fan and love, for instance, Galactic Civilizations II ... generally just a few minute into playing it, the game locks up and the monitor darkens and flashes me a "Signal Out Of Range" on it. It couldn't be the graphics card, could it? This is a new one taking up dual slots with a heat sink, and I've confirmed the fan on it does work.

I've wasted around $1500 the past two months trying to get my old system to work and (when that failed) building a new one to repalce it. On part time income, that's been killer ..... I could have used that money toward getting a better car or something.

I really, truly want to give up at this point ... but I do not trust the "major manufacturers" (Gateway, Dell, etc.) I like the idea of having a good gaming system and don't think I should be expected to spend $4000 on a system whose components I know are less than $1,000. I know labor isn't free ... but does anyone have any recommendations on gaming PCs? I'd like to be able to pick my components (esp. graphics card) and have the system assembler build and test it and guarantee its not going to lock up on me. I've thought about maybe getting a notebook with everything built in, but those still seem to be twice the cost for half the power.

I would -love it- if Tomshardware made and sold systems. I know it might seemingly defeat the "here's what's killer, build it yourself!" approach ... but I'm obviously not up to the task of being able to build a stable system, even following THG's recommendations. It'll be awhile before I have the money to put up for a good rig, but I want a professional to do it who would overkill on cooling everything (I'd even spring for ... in fact would probably insist upon ... some kinda super water-cooling rig with heat sinks for every component including RAM, hard drive, etc., even though I would NOT want it overclocked).

So what kinda shops do you guys recommend? Or might I be better off asking one of my friends who have not had the problems I've had and paying them to build a full system for me?

Thanks ...
Chibibabos
beaten, battered and desperate

 

[fishmahn]

Same sort of thing that hit me a few years ago.

Power supply.

Ok, that may not be the problem (about a 1% chance of it not being the PSU, IMO), but recently power has become much more important than just 'have a power supply'. Its now, 'have a good quality power supply with lots of amps on the 12v line(s)'.

Based on your description of constantly crashing on doing almost anything, That's my first guess.

So, what brand, wattage and voltage caps. does your PSU you have? (12v,5v & 3.3v primarily)

Also, what are your CPU & chassis temps like?

Mike.

 

[Chibiabos]

I've run the gambit, several (I think 5 or 6) different power supplies between my failed system and the new attempt from the ground up. The one I got with the current system turned out to be the lowest one (for some reason I thought it was better than the one I got), its a 350W - http://www.newegg.com/product/product.asp?item=N82E16817154012


I've been up to 500W I think was the largest one I tried, but different power supplies didn't seem to make a difference.

Despite a large cooler, the previous system got up above 60C for the CPU. :/ If I leave this one on awhile, it sometimes gets up to 55C, but generally below that with a system (chassis) temperature generally below 35C. I wish I could get a temp reading on the graphics card. :/

 

[CHEEZball]

The Signal (or Frequency) out of Range error and crashes is more than likely your PSU as fishman said. I'm having the same problems, and lookin in the bios I can see that it's running low on juice.

Also, your monitor, mines old as snot and shuts off if you even dare touch it. You haven't replaced yours either right? It could be wearing out. I've read that that specific error "The Frequency is out of Range" means that your refresh rate is too high for your monitor. I'm not entirely positive that the monitor could be a culprit, but hell I'm replacing mine anyway just for a shiny new toy (mmm flat screen)

 

[CHEEZball]

Try a new monitor, further reading up on the issues, it does sound like it could be the culprit. You could try lowering your refresh rate and resolutions but in game it will continue to try to run at a higher resolution that is intended for the game, which can confuse your poor computer and crash you.

 

[Chibiabos]

It does it in the middle of the game. A nagging voice in the back of my mind suggested maybe the monitor is the culprit ... but if it was the monitor, the monitor couldn't give some kind of feedback that actually crashes the system, could it?

If the system isn't crashing, when I'm playing a game like Galactic Civilizations II, I can switch out of the game through Ctrl+Esc (to activate the Windows Start menu) and that should kill the music. That doesn't happen if the game experiences the "freq out of range" error and I hit Ctrl+Esc. If it was just the monitor, I should be able to control/run the system blind, right? I even tried blindly "Ctrl+Esc" and "R" for "Run...", "command" + enter to run the command promp, then "Alt+Enter" which (if my system isn't crashed) gives me a full screen text mode command prompt with the thought that forcing the system to change the display mode (especially to such an old standby as text mode) might show me something if the system weren't completely hung and it were just the monitor, but that never works.

A power supply isn't cheap, but a lot less expensive than a completely new system, so I guess it wouldn't kill me to try that before giving up completely. Even just Firefox crashes on me a couple times per day, though, and ... that shouldn't happen. :/ I frequently have my system set to run a scandisk on my boot drive because of all the crashes, else the crashes increase after successive hard-crash-and-reboot events.

 

[Chibiabos]

Try a new monitor, further reading up on the issues, it does sound like it could be the culprit. You could try lowering your refresh rate and resolutions but in game it will continue to try to run at a higher resolution that is intended for the game, which can confuse your poor computer and crash you.

I didn't think it would be possible for a monitor to cause a full system crash. Is it possible? Its just an IBM 19" CRT, nothing special.

I am 99.9% certain its a full system crash, not just a monitor display problem, as game music will continue to play on my speakers for awhile even if I hit Ctrl+Esc to try to activate the Windows Start menu (which normally switches out of the game screen and causes the game music to stoop playing). Is it possible for the monitor to actually cause a system crash?

 

[CHEEZball]

The age old question, are your drivers all shiny and new/up to date?

I can't explain the hard lock ups, but ..

I'm a big TBS fan and love, for instance, Galactic Civilizations II ... generally just a few minute into playing it, the game locks up and the monitor darkens and flashes me a "Signal Out Of Range" on it.

These are my only suggestions

Signal Out of Range Fixes

Neat explanation I stole from another forum

A computer screen (both CRT, the TV like monitors, and LCD, the flat panels) can only show so many pixels (the tiny dots that make up EVERYTHING on the screen). This would be the maximum resolution. 480x680, 800x600, 1024x768, and 1280x1024, are the "standard" resolutions. Depending on the physical size of the screen (17", 19", etc.) different resolutions may look better than others. In addition to the resolution the pixels are updated only so many times per second, with a higher setting looking more crisp and readable, this is called a refresh rate, and is measured in hertz (hz). A screen set to 75hz is updated 75 times per second.

When a game starts it will default to a preset resolution and refresh rate. If either the resolution or refresh rate is higher or lower than what the monitor can display you will recieve an "Signal Out of Range" message.

 

[fishmahn]

Crappy PSU. It may work fine in your 1Ghz t-bird, but not an A64 with a modern video card. Well, not reliably.

I had a PSU much like that one. Came with this case http://www.newegg.com/Product/Product.asp?Item=N82E16811154017. Didn't power my Athlon XP 2600+ with a 9600XT and 1 HDD very well. It worked, but not completely stable. Not as many lockups as you, but still unstable. Found out about PSU's here, and I'm now using a Fortron (FSP) 400w, and its much better.

FSP (Fortron), Sparkle, Enermax, OCZ, Antec (TruePower series is best), Seasonic and a few others make quality PSU's. FSP & Sparkle (FSP makes many of Sparkle's PSUs - pick one with an 'FSP' in the part #) make inexpensive ones that are pretty good. You want at least 400w so you have some headroom, more (450 or 500) if you plan to upgrade to better video, or have a lot of hard drives, or overclock. Nowadays, you need to see the amperage on the 12v line(s). It should probably total over 20A. I would post a question in the Power supply section to get a response from people more expert than me at PSUs.

Mike.

 

[Chibiabos]

The age old question, are your drivers all shiny and new/up to date?

Yuppers.

I can't explain the hard lock ups, but ..

I'm a big TBS fan and love, for instance, Galactic Civilizations II ... generally just a few minute into playing it, the game locks up and the monitor darkens and flashes me a "Signal Out Of Range" on it.

These are my only suggestions

Signal Out of Range Fixes

Neat explanation I stole from another forum

A computer screen (both CRT, the TV like monitors, and LCD, the flat panels) can only show so many pixels (the tiny dots that make up EVERYTHING on the screen). This would be the maximum resolution. 480x680, 800x600, 1024x768, and 1280x1024, are the "standard" resolutions. Depending on the physical size of the screen (17", 19", etc.) different resolutions may look better than others. In addition to the resolution the pixels are updated only so many times per second, with a higher setting looking more crisp and readable, this is called a refresh rate, and is measured in hertz (hz). A screen set to 75hz is updated 75 times per second.

When a game starts it will default to a preset resolution and refresh rate. If either the resolution or refresh rate is higher or lower than what the monitor can display you will recieve an "Signal Out of Range" message.

Except I'm not sure this is applicable. "When a game starts" refers to actually when the game starts, right? The game starts, and usually I can get at least a couple minutes of play and then I get the out of requency message on my monitor. Does it still apply?

 

[Chibiabos]

Crappy PSU. It may work fine in your 1Ghz t-bird, but not an A64 with a modern video card. Well, not reliably.

I had a PSU much like that one. Came with this case http://www.newegg.com/Product/Product.asp?Item=N82E16811154017. Didn't power my Athlon XP 2600+ with a 9600XT and 1 HDD very well. It worked, but not completely stable. Not as many lockups as you, but still unstable. Found out about PSU's here, and I'm now using a Fortron (FSP) 400w, and its much better.

FSP (Fortron), Sparkle, Enermax, OCZ, Antec (TruePower series is best), Seasonic and a few others make quality PSU's. FSP & Sparkle (FSP makes many of Sparkle's PSUs - pick one with an 'FSP' in the part #) make inexpensive ones that are pretty good. You want at least 400w so you have some headroom, more (450 or 500) if you plan to upgrade to better video, or have a lot of hard drives, or overclock. Nowadays, you need to see the amperage on the 12v line(s). It should probably total over 20A. I would post a question in the Power supply section to get a response from people more expert than me at PSUs.

Mike.

I will try again when I have the $$$, but as I stated, I've tried power supplies up to 500W and same problems. I couldn't name the brand names on them all, so I guess its possible even the powerful ones sux0red, but seems improbable so many would be bad (unless it takes a truly rarely supreme one to make the grade ... in which case, I would have expected to find an article on Tom's saying only a few brands of PSes are okay and all the others cause crashes like mine).

 

[fishmahn]

D'oh! Forgot something very important. Your CPU temp is too high for an A64. 60C is where I would put the shutdown temp. It should be below 50 if you have decent airflow, some people with good heatsinks get below 40 and you should probably be in the mid-upper 40's with the stock heatsink. Your chassis temp seems ok.

Soooo, what kind of case do you have? What kind of fans? What CPU HSF? Generally, you should have a case with a side panel fan blowing onto the CPU's HSF (or a duct to the HSF), an exhaust fan on the back panel below the PSU, and the PSU fan. Maybe an intake fan in front to cool the hard drives if you're worried about that.

The GPUs you have are all lower-end ones, so there shouldn't be any heat issues as long as you stick to ones with decent cooling (as it sounds like your current one is). Most lower-end GPUs don't have temp sensors, but higher end ones do.

Mike.

 

[fishmahn]

Were all the higher-watt PSUs on the old system with temps of 60c? The problems with the old system could easily have been heat. In which case a gigawatt PSU couldn't fix that.

Mike.

PS: This isn't really the section for your question - it really should go in the hardware section - whichever subsection seems closest to the source of he problem or general. You're not getting any of the true experts here - just some lunatics (and Cheezy) that escaped from the Other. (enter there at your own risk - but hey, you may like it... or not.)

 

[Chibiabos]

D'oh! Forgot something very important. Your CPU temp is too high for an A64. 60C is where I would put the shutdown temp. It should be below 50 if you have decent airflow, some people with good heatsinks get below 40 and you should probably be in the mid-upper 40's with the stock heatsink. Your chassis temp seems ok.

Soooo, what kind of case do you have? What kind of fans? What CPU HSF? Generally, you should have a case with a side panel fan blowing onto the CPU's HSF (or a duct to the HSF), an exhaust fan on the back panel below the PSU, and the PSU fan. Maybe an intake fan in front to cool the hard drives if you're worried about that.

The GPUs you have are all lower-end ones, so there shouldn't be any heat issues as long as you stick to ones with decent cooling (as it sounds like your current one is). Most lower-end GPUs don't have temp sensors, but higher end ones do.

Mike.

I have 60C as the shutdown temp on my current one. I do on my old one as well, but not before a couple lockups, reboots to discover my CPU was in the 70C range, which made me think I may have already fried my system and eventually led to my starting over!

This is my current case: http://www.newegg.com/product/product.asp?item=N82E16811129017

It doesn't have side panel fans like my last case attempt for my old system did, but it does have 3 large (120mm) fans. I am gonna get a 4th one for it.

Unfortunately, I don't have air conditioning in my actual room, and I think that may be a contributing problem ... my room itself really heats up with the computer in it, I generally try to have a box fan blowing air in from the hallway straight at my workstation, but unfortunately it fell over and smashed. :/

I don't like the front door on this case ..... it seems totally inane to have air blowing out the back when the front door seems to not allow air coming in the front. The case is spacious and compartmentalized, with the hard drives in one isolated compartment, the mobo/CPU/etc. in another, and the power supply in still another isolated compartment ... but seems stupid to me there doesn't seem to be a way for air to come in the front unless you keep the front bezel door propped open. I've wasted sooo much money on cases, its not even funny, I literally have a STACK of 3 cases I'm going to get rid of ... I would have been better off going with one of those water cooled cases off the bat. 🙁

 

[fishmahn]

You have an awesome case there. That one is highly recommended here, IIRC. That may be why your temps aren't over 60c this time around. Don't worry about not having a side panel fan on this case.

See the grill on each side of the front of the case? That's your air intake for the front.

Is that the same CPU as went up to 70+ previously? If so, I would try to torture-test it with Prime95 or some other similar program (see hardware forums for other options) for a few (24+?) hours to see if its still stable. That may be the current problem - a few circuits inside may have been hosed from hitting 70+. But don't quote me - I'm not an expert.

Anyways, if you have 3 fans already, I doubt a 4th will be necessary. First make sure you have airflow - think of how the air flows. The PSU is an exhaust, the one on the back panel should be an exhaust, and in front should be intake. That should be optimum airflow for that case. If you want to see if you need more case cooling, run the PC with the side panel off and a big fan blowing into the side. Run a 100% load test (aka torture test) for an hour or so and see what the temps read.

Per your room issue: Your chassis temp is not bad, so the room likely isn't a problem. Oh, are those temps while sitting at the windows desktop or while under 100% load? If not under full load, load up Prime95 or some other torture test for a while and check the temps then. The idle temps are pretty worthless because the CPU doesn't generate as much heat when its idle.

One more thing to check (Proof I'm not an expert - It's important and I didn't think of it at first) is that the HSF is firmly seated on the CPU. Also, did you use thermal compound? What type and how was it applied?

Mike.

 

[Chibiabos]

You have an awesome case there. That one is highly recommended here, IIRC. That may be why your temps aren't over 60c this time around. Don't worry about not having a side panel fan on this case.

See the grill on each side of the front of the case? That's your air intake for the front.

Is that the same CPU as went up to 70+ previously? If so, I would try to torture-test it with Prime95 or some other similar program (see hardware forums for other options) for a few (24+?) hours to see if its still stable. That may be the current problem - a few circuits inside may have been hosed from hitting 70+. But don't quote me - I'm not an expert.

Anyways, if you have 3 fans already, I doubt a 4th will be necessary. First make sure you have airflow - think of how the air flows. The PSU is an exhaust, the one on the back panel should be an exhaust, and in front should be intake. That should be optimum airflow for that case. If you want to see if you need more case cooling, run the PC with the side panel off and a big fan blowing into the side. Run a 100% load test (aka torture test) for an hour or so and see what the temps read.

Per your room issue: Your chassis temp is not bad, so the room likely isn't a problem. Oh, are those temps while sitting at the windows desktop or while under 100% load? If not under full load, load up Prime95 or some other torture test for a while and check the temps then. The idle temps are pretty worthless because the CPU doesn't generate as much heat when its idle.

One more thing to check (Proof I'm not an expert - It's important and I didn't think of it at first) is that the HSF is firmly seated on the CPU. Also, did you use thermal compound? What type and how was it applied?

Mike.

Oh, right at the edges? Grah, I'm blind, didn't notice those. 😛

The CPU and mobo are new (not the ones that went over 70C), as is the graphics card, RAM, and boot drive. The case I initially bought these new ones in though was quite a bit smaller ... it was so bad, in fact, I could barely squeeze the IDE cables around my graphics card which I presume killed the airflow (IDE cables being so fat). I definitely appreciate spacious case.

I actually did not apply the HSF to the CPU, as I no longer trust myself since I seem to have royally screwed up my last system, I had portatech do that (pre-assemble, burn in+test). I moved the mobo over to this case without touching the HSF+CPU (I moved the mobo over with them still mounted with the retention bar down)

I will try that torture test you mentioned ... is it freeware? (I hope!). Right now I'm very tight on cash from a combo of these expensive failed attempts to get my comp working, buying a new junk car, and putting a deposit down for a new place to live because I'm quitting my old job and getting one that's too far away for me to commute in a sane fashion. After a month or so though things should be smoother and in a few months, cash for a new system shouldn't be too bad of a problem. If I can get this one working, that would be fantastic ... doubly so if I can get my old one working as well (its sitting in a case ATM not as extravagant but similarly large-sized case ... I do need to remount the CPU+HS as I didn't do it properly on my last attempt, for some reason even though the fan spins, the CPU shoots up past 60 degrees within just a couple minutes of starting it up (not even long enough to make it through boot, I can watch the CPU rocket up just going immediately to BIOS/PC Health Status) ... has no hard drive or anything in it anyway. Dunno if its salvageable, as *that* is the one that went over 70C.

 

[fishmahn]

Ok, good so far.

What HSF do you have? Was it a stock one from a retail package, or an add-on one? I know you had them set it up for you, but maybe you know or have a receipt or something.

BTW - I had all kinds of trouble getting my 1ghz T-bird to work - not the same problems you're having but no bootup, etc... Had to pay a tech place to make it work for me. 😳 😳 😳 Embarrasing as hell to me since I'd been doing PC's for over 15 years back then. I know just how you feel. Get this one running stably, then maybe you can fire up the old one and if its still workable, sell it and recoup some of your expense.

Prime95 & Memtest86 (google them - they're both free) are my favorite 2 programs for this. Memtest to make sure your RAM is stable - run it overnight - and Prime95 to stability test. Prime95 is a windows program, Memtest boots itself, so you burn a CD or have it make a boot floppy (different downloads) and boot to it directly.

Mike.

 

[Chibiabos]

the receipt describes the cooling system as " Ultra Cool Package (UltraFan CPU Cooler - Artic Silver 5 Thermal Compound - 2 High Perf. Case Fans)"

(Of course the "Case Fans" are obsolete since I am in the new case with 3x120mm fans)

I dunno if "UltraFan CPU Cooler" is an actual brand name for the cooler or just their description of it.

 

[fishmahn]

Oh well, that doesn't say much. We'll assume for the moment that the HSF is properly attached... (might try reseating it later if temps are too high)

Lets see what the temps are during a stress test. That'll tell us if it overheats or not. Thinking about it, if it heats up to shutdown temp under load, that could be why it's not stable. And if 50C is idle, then 100% load could break 60c.

Mike.

 

[Chibiabos]

I'm running that Prime95 program, its at iteration 100,000 and running 0.107 - 0.116 per iteration. Its generally getting faster with the first 5 iterations at 0.111 or above and all except the second to last of the last 5 iterations at 0.108 or below. Grah, it took 2 hours, 58 minutes to reach 0.50% ... that means it'll take just shy of 600 hours to achieve 100% if I'm reading that right. :/

 

[fishmahn]

Whoops - I should have mentioned that one of the options is a torture test, and you don't have to download a work unit thingy. Sorry. 😳 Oh well. That works too.

After a half hour or so, check the CPU & chassis temps and see what they are. Watch them for a few mins and see if they've stabilized.

Oh, do you have Cool & Quiet turned on? If not, turn it on and see how your idle temps fall.

Mike.

 

[Chibiabos]

I downloaded a thing from AMD to monitor my temperature and whatnot, but the program just fails, I don't know why.

 

[fishmahn]

Figure out what motherboard you have and download the tool from the mobo mfg, or get speedfan or another system monitor. There are others too that are free, I just don't have their names handy.

Mike.

 

[Chibiabos]

I forgot to mark down the system temp, but the CPU temp was up to 58C upon reboot after running the Prime95 for several hours, and after about 10 minutes of just having Windows up and not doing much, it was at 50C CPU and 29C Chassis.

The mobo is confusing, as for some insane reason portatech lists a ECS nForce3-A939 (Highly Expandable / Excellent Value AGP Motherboard) on the receipt, but the board I am pretty sure is a Gigabyte K8 Triton (GA-K8U-939)

 

[Chibiabos]

Figure out what motherboard you have and download the tool from the mobo mfg, or get speedfan or another system monitor. There are others too that are free, I just don't have their names handy.

Mike.

Grah, the Gigabyte Easy Tune 5 tool seems useless for the hardware monitor ...... not sure where its getting the numbers, but it doesn't match the temperature that the BIOS Hardware Monitor reports.

The BIOS reported 52C / 28C for CPU / Chassis at the reboot I just completed, for instance, but Easy Tune 5 reports only 29C for the CPU though it matches 28C for the chassis, and it reports 0 RPM for both CPU and chassis fan speed. :/

I had even worse problems with the AMD CPU monitor, wouldn't give me any reading at all on the temperature. :/