So my computer starts to act up. I had to fix some files. More files damaged so I shut the thing off with plans to only turn on the thing one more time to get everything off of the hard drive. I take it to the local computer place and the guy suggests that maybe the computer is running hot. So I put the whole thing in the fridge and it worked fine. Taken out it lasts only about 20 minutes. I still figured it's the hard drive but it's not. The computer motherboard is failing. As long as the computer is in the fridge it runs great. What I don't understand is how files on the hard drive were damaged by the computer. But lots of files were damaged. It's not a virus most likely because I ran two updated virus programs and they detect nothing. And now that I have put the old drive into another machine it works fine. Nevertheless it will be destroyed after I get all the files I want. ERS
Unless you were doing disk writes to the files the files are likely fine. The bad thing is the files are comprised of many chunks that are found by file(s) that point to where at the minimum the beginning of the chunks are.
A computer that is on is doing a lot of file writing.
Are you sure it's the motherboard, or could it be a fan? Fans (even the PS fan) are easy to replace, and a lot cheaper than motherboards.
As you read and write the data, it could easily be getting corrupted by a failing CPU, motherboard chipset, or even cables. Either "dying motherboard" or "dead fan causing motherboard overheating" (or several other possibilities) could easily give these symptoms.
It doesn't sound like it was the problem... of course, at the cost of disk drives today, I guess replacing it on a "better to be safe" basis couldn't hurt.
Well, if the computer crashes while writing to disk, it can make a mess. If the file directories are in memory, are altered when you change a file, then are corrupted by bad CPU or memory operation, and then are written back to disk, you have a royal mess. I just had this on my kids' computer due to CPU overheating. It started getting the "blue screen". I got the vacuum and sucked a huge wad of dust out of the CPU heatsink, and the problems went away, but I had file corruption in the registry. I tried to fix it, but the recovery stuff wouldn't do it. I finally copied all the user files off, erased the disk and reloaded Windows, and put the user files back. So, there was nothing wrong with the hardware, just it was overheating.
There is so much woodoo going on in a PC that almost anything can cause serious problems. Many motherboard failures are related to CPUs overheating, commonly due to dust build-up in the heatsink fins and the fan blades themselves (stalling the CPU fan, particularly those small fans). I've seen paper stickers on the bottom of CPUs turn tan/brown in color, from clogged-failed CPU heatsink fans.
My new machine has 3 each, (80mm) 3" fans.. 1 in the PSU, 1 on the case and
1 on the CPU heatsink. I've seen CPU fan/heatsink combos in stores that utilize the 3" fans now, instead of the troublesome small ones (about
1-1/2"). The better grades of fans have 2 ball bearing assemblies in them.
Many motherboards have protective circuits under the physical location of the CPU to detect overheating, which should shut the machine down. Some PC manufacturers actually produce machines with that feature disabled. The temp sensing circuits on the motherboards and in PSUs are supposed to make the fans run faster as temps rise, and/or slow the CPU clock down to reduce heat generation.
The (separate) power supply/PSU may also have temp sensing circuits, but maybe not. I'm not aware of any heat sensing circuits within the voltage regulation circuitry area located on the motherboards, although they may exist, and this is the second or third area that can create significant heat. This was the area that the bad electrolytic caps were located in. I suspect that most of those caps have been used up or destroyed since that was quite a few years ago, that those were used. When you look at one of those boards, it's very obvious since the tops of 'em are domed outward.
My previous W98 machine had a front panel that was easily modified for a filter, and I added a 3" case fan to push/draw in filtered air into the case. No silly-assed HEPA paper, just 2 layers of spun poly and a charcoal layer, scissor cut from a big universal (20x25") furnace filter (actually left-over from cutting 12x14" filters for room air cleaner). When I've had that case open for changes, there was no significant dust residue anywhere (despite a nasty operating environment) that would create problems. The filter gets furry, not the heat-generating/dissipating components. I utilize the case fan pushing filtered air into the case, which keeps the media drives clean (as opposed to exaust fans sucking room air in thru the drives).
I've seen pics of liquid-cooled CPU heatsinks, machined (milled) from blocks of copper, for use by CPU overclockers that pump ice water thru the blocks. Other experimenters have used peltier coolers for CPU heat.
Regardless of the problem, the solution is the same.
Divide problem into 2 possible sources: Disk Rest of computer
How much is your data worth?
At today's computer costs, it worth your while to borrow/ rent / buy a similar era box that is a known good quantity. Add the drive as a slave to the existing good drive in the "new" computer. Copy your files over.
Buy a new hard drive every 3 years or so........ copy important files over.
I'm a true believer in more-air-is-better, but there are risks. The air flow in your PC has been designed. I.e., the size & location of the fan(s) and the size & location of vents are such for a certain speed and
*path* for the air flow. Adding a fan will change this, possibly for the worse! E.g., the additional fan could buck the designed air flow, resulting in less cooling. Or, it could change the path of the flow, so as to reduce the flow over some components.
The safest course for additional cooling is to simply boost the current fan. In many PCs, the PSU fan also cools the PC. Adding a fan in series with the PSU fan would be safe, direct, & easy. Attach it outside, over the PSU vent. If you need to be told about the importance of the direction of the fan's flow, you shouldn't be doing this.
Disclaimer: I have no design experience with PC's or fans.
Throwing a drive into another PC used to work fine, but I don't think that'll work with the Vista OS.
One would probably need to get an external HDD enclosure with a USB cable (and a separate PSU/AC adapter usually, and even then, old versions of program/user data are likely to be incompatible. It's probably best to stay away from Vista all together, but I wanted a new PC and it was on it. I' hoping this machine will transition smoothly to Linux (proper pronuncation lye-nukes).
Agreed. Another possible source of problems is the power supply connector. If the current is on the high side, it will eventually overheat a pin or two, and perhaps even turn the solder joint on the board into a cold-solder joint. Anyway -- glitches from the pin going high-resistance can cause corruption of disk sectors as they are written. I've seen this happen on an AT&T UnixPC (68010 CPU, not an
80x86). Depending on which pin you get different problems. One pin powers the system RAM, another the disk controller chips, another pair the disks themselves, and yet another the built-in monitor (in this case), resulting in strange dancing text on the screen. :-)
The Sun Fire 280R which I just got has three 5" fans (one for the PCI cards, one for the two CPU modules, and one for the DIMMs holding the (up to) 8 GB of RAM.
In addition to this, there are two hot-swappable power supplies with two 3" fans each. This is *not* a quiet machine to keep in the bedroom running 24/7. :-)
The machine in question above is designed to live in an air conditioned machine room, so it runs the fans full speed all the time. But it does have temperature monitoring. Currently 91 F in the box itself, and 131 F in each CPU. (And they don't start to complain until they reach just below the boiling point of water. :-)
[ ... ]
I've got a rack-mount chassis with a built in filter in the shop ready to control the CNC Bridgeport once I finish some conversions. (The original LSI-11 based controller has severe electronics Altzheimer's. :-)
And one machine which I have is quite interesting. It uses heat pipes to carry the heat away from the CPU to a heat sink and fan assembly high on the back panel. This is one of the "Shuttle" Intel boxes designed for the gamers who live by overclocking their systems. It is a "Shuttle" -- a nice compact case, but with only one PCI slot, and one for an AGX card.
I made a water cooled heat sink from some sheet copper. I used some oak to make a die to form one bit of copper into a hat shape and then soldered that with a couple of tubes onto a flat piece of copper. And used a cheap pump made for desktop fountains for circulating the water. No radiator. Just had the pump in a plastic jar that held about a half gallon of water. It was overkill for my CPU. But is did get all the CPU heat outside the computer case.
Your CPU cooling device sounds like a very practical approach Dan. One could toss ice or one of those lunch cooler gel-pac things into the water, too. Don't forget the corrosion inhibitors.
A friend and I had talked for countless hours about using various cooling methods several years ago. I determined that really low temps in a vented cabinet would create a problem of condensation and sweating/dripping droplets around a CPU, or otherwise inside the case.
The Pugetsystems project that Ned referred to (in this thread) seems to work very well.
A non-ice method would be to use a metallic tank (with massive fins) to dissipate the heat to floor-air temperature, too (just to make it a more metalworking intensive project).
As a note of a much simpler method, I saw a heatsink/fan assembly at Staples (Antec product maybe), that has a slab of copper inlaid into a shallow pocket in the bottom face of the aluminum heatsink. I don't know how the copper is attached/mated, and the idea may just be a gimmick, since oxidation of the materials could be a significant drawback to the efficiency of the heat transfer.
fluorinert, made by 3M. Pricey stuff, like $135 for 250 cc. As Gunner sez, peanut oil works quite well. Hams used to get a kilowatt out of metal 6L6's running upside down in about a gallon of peanut oil.
Oh, YEAH! I think I recall that three of them in parallel gave you a plate resistance of about 16 ohms -- and thus, the birth of the "gutless wonder" direct-coupled audio amp.
(built one! Used it as a sub-woofer driver long before sub-woofers were the vogue)
PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.