Hardware Reboot when launching Orbiter, serious issues at system boot

[N_Molson]

Well, I currently have very serious issues with my computer. I post them here, because I guess there are hardware-specialists over there. I have some experience with computers (more than 10 years), but I never had this kind of issue before.So, at this point, I take any ideas.
The last-resort option being, of course, buying a new computer...

Replacing the MB or any other component may prove difficult, because this is "old" hardware that stores do not sell anymore.

Core Hardware :

Motherboard : ASUS, A8N-SLI Premium (Socket 939)
CPU : AMD, Athlon 64 3700+
GPU : NVIDIA, GeForce 7600 GT, "Point of View", PCI Express, 256MB GDDR3
RAM : 2x 1024MB, total 2048MB.

Power Supply : HEDEN, 480W

HDDs :

n°1) MAXTOR STM3500320AS (IDE) e120GB
n°2) SAMSUNG SP1213N (SATA) e500GB

Software :

OS : Windows XP Pro SP2, installed on the n°1 IDE HDD
MB : supports NVIDIA nForce 4, installed (2006). BIOS up to date (2006), flashed with no improvements.
GPU : NVIDIA Drivers, up to date
HDD : NVIDIA IDE drivers

Antivirus : AVIRA, Antivir Personal, up to date.
Firewall : CHECK POINT, ZoneAlarm, free edition, up to date.
Maintenance : CCleaner, Spyware Terminator.

Other peripherals :

TV Tuner : ASUS MyCinema-P7131-Hybrid (PCI slot1)
CD/DVD ROM driver/writer : LG, super multi
Mouse : USB, Microsoft
Keyboard : "non-USB", Microsoft
External SoundCard : LINE6 Toneport UX1
Modem : SAGEM LiveBox-mini (ADSL), connected via Ethernet to the MB (Marvell-Yukon driver).

First time the issues appeared : a few seconds after launching Orbiter 2010 (Rel. ed), the case is described below.

SYMPTOMS :

- Issues when booting the computer. I get apparently random re-boots, which seem not related to the booting process itself. I can try to classify these issues in 4 great families :

Case 1) The BIOS screen briefly displays, then I get a reboot. Sometimes I can access the BIOS, but the re-boot occurs when I'm inside the BIOS menu. The reboot seems time-related rather than process-related. I can get a few seconds in the BIOS menu if I'm fast. Time seems limited.

Case 2) The booting process successfully passes the BIOS screen. Then I have the MS-DOS prompts screen. At this point, the process usually freezes a few seconds (2-3s) to allow me to enter the SATA-RAID utility, which seems to be present in the Motherboard. Same issue as above : I can briefly get into the utility, but a re-boot occurs anyway.

Case 3) The booting process reaches the memory check / sys.ini MS-DOS screen, which displays very briefly (1-2s) as usual. Once this step finished, a reboot can directly occur, or I can get to the Windows lowres "Pre-loading screen" ("Windows is loading"). But then it doesn't last more than a few seconds and the reboot occurs.

Case 4) The booting process ends and Windows takes command. This is the critical step : from there the OS is stable. I launched antivirus/spyware scans last night and Windows ran flawlessly during more than 8 hours, the scans were complete as expected (negative reports).

​

- Issues when loading a "power-3D" application. I took Orbiter 2010 (Rel. ed.) and Moonbase Alpha as exemples.

• Orbiter 2010 loads any scenario correctly, displays the scene a few seconds (controls work), and then I get a reboot, followed by the same issues as above (yeah, that really sucks !!)

• Moonbase Alpha fails to load : I get the Windows "Moonbase Alpha encountered a problem" screen 2 times, then I get back to the Desktop.

• Google Earth works perfectly though. But unlike the two applications above, it uses OpenGL.

- Other misc. issues :

• I try to reboot the computer after a Windows session. The Windows process seems to close correctly, and I fall under the booting problems described above. However, I noticed that Windows often displays the "Windows did not close correctly last session..." MS-DOS screen when the booting process gets this far.

• The probability to get a successful booting process seems again time-related. If I let the computer power off during more than 10 minutes, chances that the computers boots correctly are fair. However, if I power down the computer less than 30s., there is almost no hope to get a successful boot. Which makes me suspect capacitors (MB or Power Supply) : looks like things are better if they can fully discharge.

• The booting process loops when it fails. Usually, the process loops 3 times (with different results), then the computer gets in a frozen state, all systems powered and running but actually doing nothing. The only way to get out of it is the power-off button pressed for 4 secs. Sometimes however the computer might seem frozen, and suddenly loads windows, with no screen display before that (here I suspect the GPU).

Thank you for reading, I don't really know what to do (except backing up data while I still can ?).

:hail:

 

[Mantis]

If you're getting a crash/reboot on the BIOS screen or during the POST, you've got some serious issues with that computer. At that point your operating system hasn't even kicked in so its definitely a hardware issue. I'd start my troubleshooting with checking the BIOS - have you recently updated it - maybe with the wrong version? I'd also look at heat - are your fans working? What temperatre is your CPU operating at (often visible throug BIOS)? If that's all normal, I'd take a look at system memory - perhaps you have a defective memory module. Failing all that, perhaps you have a main board problem.

 

[shortstar]

If you hardware is a little aged like mine, have a quick glance at the capacitors in the mb. The tops could be anywhere from slightly swollen to leaking electrolyte on the mb or gpu board. These caps can be replaced easily. Also, if it is a laptop, just disconnect all power including battery, and hold the power button for about 30 secs, then insert bat and power plug and turn it on. Might work with a desktop as well, but I cant confirm this.
Best of luck,
shortstar

 

[N_Molson]

I flashed the BIOS, using the CD that was delivered with the Motherboard, but that didn't change anything.

Concerning the temps, they were (really) excessive 1 month ago, but since then I cleaned the computer. Maybe these temps put a lot of stress on the hardware ?

Temperatures (room at 25°C) :

CPU : 35°C, idle at 3500 rpm, max 6000 rpm at full load.
MB (passive cooling, no fan) : 41°C
GPU : 40°C, fan running at 31%

"Manual" feedback seem to correlate with those values, I feel no excessive heat. I adjusted the setting of the GPU fan to 100%, just to be sure, I can clearly feel a flow of 30-40° air.

The BIOS tells me I have 2048 MB of RAM as expected, I'll try to run the computer with only 1 module, that's worth a try, it's probably the easiest part to replace.

Edit : the PC is a desktop computer.

I'll take a serious look at those capacitors, but there are lots and lots of them, and I'm not sure to recognize a "swollen" one... I didn't even knew there was some fluid inside.

Edit2 : managed to run Orbiter 2010 in "RGB" mode (1-2 FPS, but no re-boot).

 

[cinder1992]

<snip>AMD, Athlon 64 3700+

oh dear. when did you buy that? there was a bad batch of those that had a decay problem, ended up replacing every part in my mother's computer (ram, motherboard, HDD, power supply, and we switched to windows 7) before we found out that the processor had more than a few ...erm... lets say major problems

 

[N_Molson]

Hmmm... Antivir detected a Trojan in "autorun.inf" after i unscrewed, unplugged, cleaned, replugged and rescrewed virtually each part of my PC !

Now that would explain a lot... But the war is not over !

 

[garyw]

oh dear. when did you buy that? there was a bad batch of those that had a decay problem, ended up replacing every part in my mother's computer (ram, motherboard, HDD, power supply, and we switched to windows 7) before we found out that the processor had more than a few ...erm... lets say major problems

I've got that motherboard and it's worked like a champ for me. Do you have any proof of this claim?

---------- Post added at 00:28 ---------- Previous post was at 00:27 ----------

Hmmm... Antivir detected a Trojan in "autorun.inf" after i unscrewed, unplugged, cleaned, replugged and rescrewed virtually each part of my PC !

Now that would explain a lot... But the war is not over !

Try a memtest. A dodgy stick of RAM can often cause all sorts of grief including random reboots. I had that on one of my servers last year.

 

[N_Molson]

I closely looked at the component and their capacitors : all the "trademark" stuff like ASUS, Nvidia, is in perfect shape once the dust removed. And when you closely see at these components, you can "feel" the quality of the assembly.

I looked into the power supply, and now that was a very different story. I'm sure that at least one of the 2 biggest capacitors (220V) is leaking. There are big glue spots, and even once cleaned, the circuits look almost... rusty...

 

[garyw]

Swapping out the power supply should be fairly easy - what wattage is it? I know that some places only provide a PSU that has just enough grunt to power up the machine.

 

[N_Molson]

Yeah, I'm gonna to change that power supply anyways. It's 480W, I'll take 500W or more to be safe.

Here are pics of the 2 main capacitors of the power supply, can someone confirm that the brown spill is a capacitor leak, as I strongly suspect ? That would be the simplest explanation, as everything else seems in order. When the computer is at full load, during startup (all systems check at the same time) and during power 3D application, the damaged supply could be unable to deliver enough Watts, causing a shutdown and then a reboot.

I tried to run the system on 1 RAM socket, and switched the 2 to test if one was failing. No changes : the 2 sockets are good, or they are both bad. That was quality stuff so I am rather confident.

 

[Mantis]

It could well be a mainboard/cpu problem. Even if you had a trojan it cannot cause a reboot during POST (Power on self-test) or within the BIOS settings itself. If the problem was the OS or a virus or something, it would only manifest itself after the POST and during the boot-up or operation of the operating system. Only hardware failures can do that. It's possible that the power supply could be the problem but I'd bet on RAM, CPU or mainboard.

 

[Xyon]

You'd be surprised how many computer issues stem from PSU problems. It's usually my first suspicion when trying to diagnose it. I think you're on the right lines, replace the PSU with a ~500W one and see what's what after that.

 

[shortstar]

power supplys are super cheap, so go that route, that cap is def having probs. For giggles, go to your local tv repair shop and grab you a pair of caps to replace your bad ones, just ensure you get the same voltage cap and preferably a bit more uf and brush up on some repair skills, think of it as a real life extension of keeping your "spacecraft" in service. Just ensure you inspect the traces after the caps are removed, if a trace is halved or seriously injured, you may have to just pony up for a new supply.

A good 600w supply should run about 25 bucks, and the caps to fix yours should be about a dollar.

 

[Xyon]

I'd seriously advise against mucking with the capacitors inside a power supply. It's much safer to just get a different unit. Since the power supply is connected to pretty much everything, directly and indirectly, you never quite know what it'll break when it dies.

 

[shortstar]

I'd seriously advise against mucking with the capacitors inside a power supply. It's much safer to just get a different unit. Since the power supply is connected to pretty much everything, directly and indirectly, you never quite know what it'll break when it dies.

Those little fellow dont have the leg up on the high voltage caps(32kv) that i muck around with, but a person should try and be safe when working around caps. just remove the supply from the computer, and cross the terminals of both large caps with a resistor(preferably) or a screwdriver. Once you have ensured the caps aren't powered up with the above its a simple soldier job. As far as what it will break when it dies, it will send (if the caps go bad) ripple current into the motherboard, and wreak havoc.

I'd def do something before too much of that happens.

 

[Xyon]

I appreciate what you're saying may be perfectly sound, but I personally believe it's bad advice for a mainstream user, who may not be as proficient in electronic repair as yourself. Meddling with the internals of a power supply is dangerous if you don't know what you're doing, which was the crux of my point.

 

[N_Molson]

I'm not an expert in electronics and I'm fully aware of it. In fact, I was almost worried to break the "do not open, hazardous voltages inside" seal of the power unit. My opinion is that electricity is a little like explosives : errors do not forgive !

So I ordered that Power Supply (the 600W one), which is modular and seems to be compatible with the SLI functions of my MB. It has received awards from a lot of gaming / hardware reviews, which is certainly not the case of the current one. The 3-years warranty is also a proof that's serious hardware.

OCZ 600W MODXTREAM PRO POWER SUPPLY

 

[shortstar]

A person should always operate within their limits, so do what you think is safe. Hopefully this will fix everything right up, as long as no damage has happened to the mb already.

 

[N_Molson]

Woohoo my new power supply is arrived ! I'm going to fetch it at the warehouse. :leaving:

 

[cinder1992]

power supplys are super cheap

umm, what? That only applies if you buy the cheesy 100w PSUs. you need at least 600w to run a modern computer, and a 700w minimum for good graphics cards (no, not the "off the shelf" ATI and NVIDIA cards, i mean real powerhouse cards, the sort that CFD guys use).