Computer Hard Drives

I haven't used Citect but I know that they sell software not hardware. I'm guessing that you're using an ordinary desktop PC for your application. We also use them for our HMI but the PCs are inside airconditioned control rooms so it's not an issue.

If your PCs are exposed to the atmosphere of your plant, then you do have a problem. You could do one or a combination of several things: buy hardened industrial PCs; put your PCs in a controlled environment (perhaps even vibration proof environments), or relocate your PCs to a safe area and use an Ethernet KVM module so that your mouse, keyboard, and monitor can be located away from your PC.

Another possibility is that there is something wrong with your PC that's damaging the hard drives. Try replacing them.

Good luck.

The three PCs are kept in a cabinet in an air conditioned control room and there is no vibration whatsoever and yet the hard drives for the plant scada system keeps failing. We tried using silicone sealant on the cards but that didnt help either.

Currently what i'm experimenting is using the Anti static bag that comes with the hard drive to see if that will help.

Hello fbompeh,

for an environmental measuring project I once packed the electronic in metal cases bought for little money from a military surplus dealer. You can get those boxes with protection after IP68 (pressure proof, dont know just what it is in NEMA). Dont forget that the boxes must have (or get) a membrane for balancing atmospheric pressure differences made by temperature cycling, otherwise you`ll have the moisture in the box after some days and don`t get it out again - because it`s sealed ...
Good luck + regards Uwe

If all your PCs are in the same room and only the three SCADA PCs are affected, then we can rule out corrosion, vibration and the UPS. Greg G suggested replacing the power supplies and that might solve your problem. Unfortunately, that's only one possibility. What about your grounding system? Is it correctly done? Are the signals going in and out of the PCs clean (no noise or voltage spikes)?

The again, it might just be that your Pcs are failing. If you're not a computer technician, then I'd suggest you call one or send your PCs off to a service center for a complete physical, and maybe a couple of transplants.

As an alternative to rotating HDD technology, you may want to look into this fast high-volume Flash solution:

http://www.adtron.com/products/

its very interesting. do you mean hard driver a fixed disk? or other equipment? if its disk. you should keep it away from acid. cause the sulfur acid can be corrosive to your circuit. the circuit is explosive. I suggest you'd better pack the disk in a seal box. for example plastic box to keep it away from acid. if the whole system no problem only but the disk, you can change the disk and its link cable as well. I sugeest you change the disk as the computer work in that circusmstance is use to use industrial level computer. so it must be there something worong with your disk. you cna also check out the virus as well. v

As posted by cnpower, HDD's are prone to fail if excessive vibration or corrosive environment are present, not just because of an acid damaging the circuit, but also because of acid damaging the Vacuum-Seal integrity.

Besides the remote possibility of a "power-loss while writing" type of crash, his advice is excellent: Isolate it in a vibration and acid free environment, and it should be fine.

A portable computer completely encased does not resolve the problem. The frequency of failure indicates one of two concerns:

The slighest of vibration destroys rotating hard drives about once per month. The drive will refuse to turn except spordiacly. One rare startup may allow transfer of information but don't expect more than three tries until total destruction.

An isolated ups typically protects input power from electrical surges but for the computer of concern each and every one of the input signals are not properly isolated

every time an invisible cloud of static electricity passes overheat (cloudy day monthly even in clear weather, invisible local sulfuric cloud, other local static charged cloud)

tall stacks and good grounds attract electrical surges that play havoc with magnetic memory. In this case, the drive may not be damaged but rendered rather dumb.

Electrically isolate all physical inputs.

If you think that the printed circuit on the drive gets damaged by acid vapors, you could apply a "printed circuit lacquer" to protect the circuits. Be sure not to spray on bearings and connectors. More thin layers give a better result than 1 thick layer.

If the other computers are in the same room, and their hard drives are not failing, I would replace the power supply in the computer housing the failing drives, since this would be a likely cause of consistent hard drive failure. Another likely cause would be the data/signal lines of the SCADA system. Make certain they are all properly filtered and signal conditioned and the computer case is grounded by means of a grounding strap or braided conductor. Check for ground loops.

As far as the sulfur fumes as a possible cause, unless there are visible signs of corrosion problems in the room, I would not think that has anything to do with your problem. If it could cause only hard drives to fail in a month or so, there would be plenty of other failures over time, and visible corrosion problems.

Regards, Greg

Once upon a time I mounted two HDDs, one on top of another. One of them quit pretty quickly. You mentioned the drive is in an antistatic bag. I'm guessing it's hot in there. My suggestion is to make sure there is plenty of air flow about the drive.

Good luck

There are several things that worry me here and they have to been covered to a certain extent by others but I will reiterate.

Heat Dissipation.. You mentioned that the computers were on a shelf and that you had placed anti static bags over the disk drives. I can just about guarantee that by placing the bags over the disk drives you are going to cause them to fail prematurely so get rid of the bags pronto. After this check how hot the drives are getting during normal operation. The cooler the drives are the better and the longer they will last. I have an 870 MB drive that has its own cooling fan and operates at slightly over room temperature. This drive has been operating continuously for over ten years now and I have never had a problem with it. Also check that there is sufficient ventilation around your computers, they pull a fair amount of power and that ends up as heat so if there is insufficient ventilation you are going to cook your machines and damage them.

Earthing Check that you have a good earth on the secondary side of your UPS and that the neutral line is tied to ground. Some UPS systems completely isolate the output and there is no reference to ground. Use a multimeter to check the voltage to the three pins at the UPS output. One pin should have mains voltage on in ant the other two should be close to zero when compared to a known good earth point.

Vibration. The heads on a hard disk drive do not actually touch the disk surface like they do in a floppy disk drive. If the heads touch the disk both the heads and disk will be damaged and you will loose data. Vibrations can cause a head crash or skip and this is usually followed by a cascading failure and the total failure of the drive. If you have blasting operations of things that can cause sudden jolts like this then install some sort of shock absorbing material under the computers.

Corrosive Environment Certain environmental contaminants can indeed damage the disk drives and cause them to fail prematurely. Ozone O₃ is on such substance and damages the surface of the disk. This damage ultimately causes a head crash and a catastrophic failure of the disk drive. This could certainly be your problem and if it is the solutions are limited. If you can find some way of isolating the computers from the environment and filtering and conditioning the air that they use for cooling.

That gives you a few thing to check and try before you start trying to do the expensive one of isolating the computers form the environment or purchasing ruggedized systems.

One other thing you may wish to look at is replacing the disk drives with solid state disk drives that us flash memory instead of a rotating magnetic media to store the data. To give you an idea how reliable these are, one was being used in a high altitude experiment and the descent parachute failed. The disk drive fell several kilometers to the ground landed on a highway an was run over several times, at least once by a truck, before it was retrieved. The memory module was still intact and they managed to retrieve all the data. These are considerably more expensive that the usual disk drives but considering you are loosing them each month and the amount of lost productive time they might be worth serious consideration.

That covers just about everything I can think of. I would start with looking at the cooling first as I think that is the most likely cause of you problems. I hope this give you something to work with and wish you luck.

Masu as usual some good advice.

If the drives are in bags, they will fail soon with overheating. Any drive that is significantly warmer than your hand needs cooling. I would guess that this is your problem....

Also, what I have not seen from anyone is if you have a source of clean, uncontaminated air, then feed this into the PCs so that they are slightly pressurised above the room pressure, a little is good, this will keep out any gases that maybe are attacking the drive seals....though I find such contamination thoughts to be a very long shot....for me, the problems are most probably high temperature and inadequate cooling.....and/or cheap drives.....

The hard drives were not in bags when they failed. He is only experimenting with that bad idea now.

Since the original post, the OP signed on as "fbompeh" (post #7) and clarified a few things.

I think the SCADA connections are a likely culprit, but the power supply is still suspect. I have had similar problems over the years from both.

Greg

I have some experiance on such PCs which run continuously HDD @ Hi speed. I looked into the problem of a Communication set-up having the same problem. The firm suppled Equipment was asked for the problem & told that the HDDs run continuosly, 24x7x365, so they have to be replaced @ specified period. As the HDDs were of Industial-Std, they charged much higher price.

If the PCs or HDDs are supplied by the Acid Plant manufacturer then ask the firm for this behavior.

Anti-static bags are conductive. They'll short out the circuit on the drive and computer. As other had suggested heat is a problem. Are the case sealed air tight to prevent acid frame? What kind of HD are they? 1.5" microdrive? 2.5" notebook drive? Or 3.5"? What's the capacity? Could you substitute with solidstate flash drive? There are IDE replacement which use flash or even SDRAM/DDR. Capacity is below 10GB.

I agree that the overheating is the most likely cause of the problem and that sticking the drives in anti-static bags is a really bad idea.

I did however have a case a long time ago with CDC 300 MB removable platter disk drives at a mine that were installed in a computer room with filtered and pressurized air. There were two drives and they would crash about once a month each and sometimes at the same time. It required the replacement of all the heads and the removable media needed to be junked. With 20 heads at $500 each per drive, new filter at $300, a new disk pack at $300 and a days labor to rebuild the drive there wasn't much change from $12,000 for each crash.

Initially we thought that the blasting in the mine could have been causing the problem and discounted the contamination as the room was specifically designed to keep any contaminants out. In the end the culprit turned out to be O₃ (ozone) from an electrostatic air filter that they were using to clean the air for the room. Ozone can damage the coating on the disk platters which in turn caused a head crash. Once the filter was replaced they never had a disk crash ever again.

As Guest said in post #12 if the disk drives feel anything more than slightly warm they need additional cooling. As the computers are on a shelf I will bet they are crammed together and there is insufficient free space around them to allow adequate ventilation.. If this is the case then moving them so there is more space and better ventilation may fix your problem.

There are a host of things that could be causing this problem but like any problem it is more than likely the simplest solution that will work. Overheating is easy to check and relatively easy to remedy so it is definitely the thing to tackle first. One thing I believe everybody is in agreement with though is that the anti-static bags around the disk drives are a really bad idea so get rid of them pronto.

Allow me to mention once more the issue of vacuum-seal integrity, which usually is hardly more than a stick-on patch, with a weak and "soft" adhesive, prone to fail in a corrosive atmosphere, and all-that besides, heat, vibration, and power discontinuity on-the-fly.

What vacuum seal?

Disk drives are sealed but not completely and they are full of air. They need air or they won't work because the heads actually use the air to separate them from the rotating disk platter.

They also have small bleed holes to allow air in and out as it expands with temperature. The bleed hoses have filters on them to keep out particulate matter but this will not stop anything that is dissolved in the air like water vapor and corrosive gasses.

Disk drives are sealed but they are not vacuum sealed and are not completely isolated from the atmosphere.

...What vacuum seal?...

Correct me if I'm wrong but: On the thin side of each hard drive, you may notice a chrome-coloured patch, with attached warning, never to fiddle with this "sticker" to avoid the loss of warranty.

This is a Vacuum seal, meant to protect the mechanics of the sucker from air friction and particle contamination, and keep the mechanics separated from the PCB, as these rotate at 7200 rpm and heads tackle at tens of iterations per sec.

As it is located near the mounting screw bores, it caused lots of agony to inexperienced self-taught technicians, as they tried to cage-mount a hard-drive, and accidentally compromised this seal's integrity, by pushing a screw right into it. As the drive naturally croaked, they turned to the shop, only to be advised of the above mentioned warranty warning. So there.

Hi Yuvalmate,

It is a seal and is intended to show if anybody has tampered with the drive but believe me the inside of the drive is filled with air, there is no vacuum inside them. Some manufacturers also use this to stop you from swapping the drive electronics. This is because the defects in the magnetic media on the platters need to be mapped to the electronics and this makes the electronics non-interchangeable.

If you look carefully you will find a bleed hole somewhere on the drive that allows air to pass in and out as it expands with temperature or the atmospheric pressure changes.

The heads in hard disks do not actually touch the disk platters and literally fly on a layer of air that is rotating with the disks. If you remove the air the heads will crash into the disk surface and damage both.

The reason that they place a seal on it is to make sure that nobody opens up the disk chamber. The insides are assembled in a clean room environment and the slightest contamination will result in a catastrophic failure.

So there is definitely no vacuum inside a hard disk drive, it's full of air and there is a vent hole to allow the air pressure to equalize with the outside.

Hi Masu,

I agree fully with all you say about causes of HDD failure.

The reason I suspect a bad power supply or the SCADA system I/O wiring/cabling is that if I understand him correctly, ONLY the hard drives in that one computer out of the three are failing every month or so, while the other computer's hard drives are fine.

Therefore I wonder what could be different about the SCADA computer, since they all seem to be in a similar environment.

If it is loaded with I/O cards and the like, which SCADA computers typically are, that would seem to be a clue. Power supply issues, transients on one or more of the I/O lines, heat from the extra hardware etc. Data and I/O lines connected to the SCADA computer and running out to various locations in the plant can easily result in transients, or other problems depending on the shielding, filtering, signal conditioning, proper grounding, etc. Extra cards, also consume more power and add additional heat, which gets us back to the heat issue which you have mentioned.

Something is apparently different in that one computer. Of course, even in the same room, or cabinet, conditions can vary substantially (as in computers being exposed to different surrounding temperatures, having restricted air flow, and so on), but I am assuming conditions for all the computers in the room are approximately equal.

What else might be different about the SCADA computer?

Perhaps fbompeh can enlighten us further.

Regards, Greg

Hi Greg_G

It's a good point that only the single PC is failing but I have never seen a dodgy power supply that consistently destroys only same sub assembly within a system. If the power supply was going far enough out of specs to damage the electronics in the disk drive then there is a pretty good chance that it would damage something else as well. At leas I would expect to see other devices, like the CD/DVD, to fail as well at the hard drive failures.

There is one power supply problem that disk drives are uniquely susceptible to and that is write errors due to voltage transients. The thing with this type of fault though is that it usually only destroys data and not the drive itself. It depends on when and what is corrupted but it can vary from a simple data corruption that shows up later as a disk read error or complete loss of all the data. The thing is though if you reformat the drive it will work ok again afterwards. I don't know if Francis has tried reformatting the drive prior to replacing them but it is something that I would try first.

I have seen this sort of problem due to poor earthing and no earth neutral MEN link on the secondary side of transformers and UPSs. A lot of electricians don't realize that if you install an isolation transformer you need to add a new earth neutral link on the secondary side and consequently omit it. The result is that even though you have the correct voltage between the terminals they are not referenced to ground and float all over the place. You can easily check this with a multimeter by looking at the voltage between all three pins. If everything is connected correctly one pin should be at mains potential above the other two and there should be little to no potential between the others. If there is no MEN you will have a mains voltage between two pins and some strange voltage between these and the remaining pin. If this is the case then you need to get the person that installed the UPS to come back and install the MEN link.

I still have the feeling that it is more than likely a heat problem because a power supply problem would more than likely manifest itself in various fault types rather than the same one over and over.

Hi Masu,

You're absolutely right about the UPS grounding, the heat problem and your feelings about a defective or overtaxed power supply.

I can't explain why, but I have worked on several computers where the power supply only affected the hard drives. In the worst case, it blew a chip on the hard drive controller board which I didn't realize until after I replaced the drive, turned on the computer and smelled a whiff of burnt electronics, had a new but dead drive which upon examination had a fried chip. Upon examining the original drive, I observed that it too had a fried chip and although the drives were different, the chips appeared similar in placement. I replaced the power supply, put in another new drive, and all was well: There was no damage to anything else at all, not the motherboard, memory, CD drive or floppy ... nothing. As we know, these anecdotal things are a poor basis for extrapolation, yet they stick in our minds.

You mentioned something which I had considered in suspecting the power supply, and that is corruption of the data rather than damage to the drive itself. The OP never mentioned mode of failure or any forensics he may have performed. If the boot sector gets screwed up the computer can fail to properly recognize the drive and this is often taken as a drive failure if no examination is done. A SCADA computer is typically performing reads and writes constantly, while other computers only do them intermittently. Therefore any glitches are more likely to result in a bad write of data or to the allocation table.

Many "SCADA" computers have I/O cards for data acquisition and control, with multiple I/O lines or multiple communication cables connecting to other remote devices and these lines can be a source of noise, transients and voltage and/or ground differentials. While damage tends to be to the cards themselves, or the motherboard I have also had occasions where the hard drive was damaged or, more frequently was corrupted, generating a drive failure error message of some kind.

The first thing the OP should have done was try to determine what actually failed on the drive to the extent possible. The next thing I would likely have done would be to replace the whole computer if the drives were actually being damaged and I didn't know why, because there isn't much difference in effort between replacing the primary drive and reinstalling everything and replacing the whole computer ..... in fact the computer replacement is probably quicker if the operating system is already installed.

Trying things like applying sealant around the cards or putting an anti-static bag over the drive is certainly not the way either of us would proceed.

If in fact the issue is heat, shame on the OP. Fans are cheap, and even temperature controlled fans are under $20.

Regards, Greg

Hi Greg,

I can't explain why, but I have worked on several computers where the power supply only affected the hard drives. In the worst case, it blew a chip on the hard drive controller board

Ah the cheap power supply syndrome. This sort of problem is nearly always due to over voltage and can cause massive damage, you are lucky it was only the hard drive that suffered. It relates to a property of switch mode power supplies where a certain type of failure causes them to supply an over specification output voltage. It is usually a hard or consistent fault but can be intermittent and occur only at start up or for short periods of time during operation. Series pass power supplies rarely fail like this so are considerably less prone to the problem so if power consumption isn't a problem I always prefer a series pass regulator. In your case it is likely to be the +12 V supply that was the problem and you got away with only loosing the disk drive as there wasn't anything else that sensitive using the +12 V

Any switch mode power supply, worth even thinking about using, should have some sort of secondary over voltage protection that protects the electronics it is driving. It's not difficult to do all you need is an appropriate zener diode and some sort of fusible link at the output of the power supply. If the voltage goes above the zener breakdown it shorts out the output and blows the fusible link thus protecting the electronics. I have also used diodes as the fusible link because they will burn out faster than a fuse and protect the down stream electronics faster. The only hassle with using a diode is the additional voltage drop, but this can easily be allowed for.

Speaking of fuses I used to repair PLC output cards that protected a 50¢ transistor with a $5.00 fuse. Even though the fuse was rated to blow sooner than the transistor the time delay always meant that both would blow. I ended up replacing the fuse with an appropriately rated diode that cost 10¢ and never blew the transistor up ever again.

Hi Masu,

"In your case it is likely to be the +12 V supply that was the problem and you got away with only loosing the disk drive as there wasn't anything else that sensitive using the +12 V"

I fully agree that is the most likely culprit.

Unfortunately, as you point out there are many poor power supplies out there, and in my experience at least, the worst offenders are the low end computers from the well known name brands, which is why I myself would never buy one. They do so many things to save pennies here and there that these low end, attractively priced machines tend to be unreliable, but in any case never live up to their processor clock speed potential because they burden it down performing tasks that should be done by dedicated chip-sets or cards. I have often been appalled by some of the power supplies I have seen in these machines.

I always build my own so I can select each component separately, and it all starts with an appropriate case, a good power supply and a good motherboard and processor. Then, if I am building it on a budget, I'll look to save money on the other hardware wherever appropriate, considering the final primary use of the computer. Even when the name brand low end machines use name brand components such as motherboards or video cards, they typically are not compatible driver-wise with the "same" but separately sold versions of the motherboards, video cards etc.

I have done similar things with PLC (especially sinking) output circuits, but off the card on a separate board, since some of the environments were extremely "noisy", and/or it wasn't possible to modify the card.

It would have helped if "fbompeh" was more forthcoming in details originally. Now, since it is all three computers, but in the same room, the power supply scenario is out the window, and the SCADA I/O now seems far less likely also, so it is most likely one of the things you mentioned.

Regards, Greg

Hi Greg,

I always build my own so I can select each component separately, and it all starts with an appropriate case, a good power supply and a good motherboard and processor.

Back in 1993 I build a computer system for monitoring my house, from scratch, based on a 386 motherboard. I had a half height 19" rack so I in an appropriate box on rails then saturated it with fans that blew filtered air over everything to keep it cool. The result was that the inside remained spotlessly clean and nothing got much warmer than room temperature.

I was finally shut down last Wednesday. We are moving house and since we plan to rent the property I felt that the average tenant would not appreciate it. The only time that it was unexpectedly off line during that period was when the power was out too long for the UPS to keep it going. All up it was off line only a handful of time, over the period, with a total down time of about 3 hours. Nearly all of the 3 hours was due to a power pole, down the street, being knocked down by a car and the time it took them to replace the pole.

It was shut down occasionally to add new functionality and upgrade small bits and pieces but the mother board, power supply, memory, most of the IO cards remained the same. I did replace the disk drive but that was to get more capacity and that was over 10 years ago now.

It shows you how reliable a PC, running Windows 3.1, can actually be if you use quality components and keep it as cool as possible. I believe the single biggest factor in the reliability, however, is the fact that everything is keep as cool as possible and only marginally above room temperature.

I have done similar things with PLC (especially sinking) output circuits, but off the card on a separate board, since some of the environments were extremely "noisy", and/or it wasn't possible to modify the card.

Back in the early 1980s I was national product support specialist for the Struthers-Dunn line of PLCs. One of the things S-D did was with their boards was to use sockets for all ICs, the idea being to make any board easily repairable. This was a huge mistake, the sockets cause nearly all the faults and the only ICs I ever replaced were PROMS, when I upgraded the operating system. For super critical applications I would rebuild the boards and remove all the sockets soldering everything, except the proms, directly to the board. It was extremely time consuming but it increased the reliability immeasurably.

Another things I had trouble with was the triac outputs. Noise was getting onto the mother board from transients, when inductive loads were switched on at the peak of the AC cycle. They used a small transformer to isolate the mains from the electronics and this would allow any spikes to get back into the system. The introduction of the opto-coupled triac though has been a huge improvement and if available at the time would have cured this problem.

I ran across a site that is still selling second had Director-4001 PLC for roughly the same price that they were new in the early 1980s. I nearly fell of my chair when I saw this as I thought, after a quarter of a century, there wouldn't be any left in existence, let alone a market for them. I don't know if this is because they were that good or that nobody knows how to replace them with more up to date technology so whenever they fail a direct one for one swap is required.

I am retired but whenever I get the chance I keep an eye on what is going on. It makes me wonder though because I see the same mistakes, as were being made 25 years ago, with triacs switching mains voltages. It costs little more and in some cases less, to do it properly yet designers seem to not learn from past mistakes.

Oh well I don't need to really worry about it any more so I shouldn't let it get to me, however, if we don't learn from our mistakes we are doomed to repeat then and will never get anywhere.

Hi Masu,

Like you, I have been around awhile. Not only do I see the same mistakes repeated over, I also see many instances of lost knowledge which is probably more distressing.

One instance that immediately comes to mind is the use of drip loops on cabling entering a building through the wall. This was standard procedure years back, and worked 100% of the time. Nowadays, under the misguided spell of modern silicone sealants, it is rare to see a drip loop on a new installation. And as good as the newer sealants are, they are far from perfect depending on the surface they are applied to, and how much time passes.

Bottom line: water entering the structure and causing damage, something that had been corrected many years ago has returned due to the lost knowledge of the simple drip loop.

Computer home and even office furniture was two or three steps backward when it first came out, and only relatively recently has it begun to even approach the usefulness of older writing or typewriter related furniture. I consider my self fortunate to have acquired some Steelcase brand desks for my home office many years back, just as offices were replacing them with "modern" modular workspaces (a funny term for something that offered precious little space to do anything but type on a keyboard and move a mouse). My desks have a large 60" wide by 30" deep top, conveniently located large drawers, and that most useful of bygone accessories, the pullout side leafs on each side. I use the right one for my mouse, allowing me to comfortably position my arm, and the left one whenever I need extra space for a pad, some papers, or an open book. These desks are head and shoulders above any of the modern office furniture.

Very few modern electric fans approach the high air volume and quiet operation of the fans made in the days before AC was common.

I'll stop here, but I could go on at length .....

Regards, Greg

Those so called computer work stations are a complete waste of space. At the university I worked at for a while they purchased a batch that had a shelf for the printer directly above the shelf for the screen. The users kept complaining that the screens were faulty and were giving everybody headaches. The problem was that the transformers and motors in the printers were causing a 50 Hz wobble on the screen. Some people just couldn't be convinced that the problem wasn't with the screen and that it was due to them purchasing a poorly designed piece of crap furniture.

My home office until recently was a complete office that I got for free from a previous employer that was going out of business. When I got the golden handshake from them the boss let me clear out my office and when I say clear I mean clear everything, desk char, shelves, filing cabinets, rug from the floor, the lot and keep everything. The only thing he asked me to return was the company car, he even forgot to ask me to return the corporate Amex card. Unfortunately I don't have the room for it any more so it got thrown out last week. One thing that I do use is one of those arms that you could get to put your screen on. I use it for my laptop and have a keyboard and trackball that I use with a big comfortable reclining chair to do all my CR4 work form. I find trackballs considerably easer to use than a mouse because you don't need to obligatory mouse pad to sit it on and you don't get the same sort of wrist strain problems with them.

Greg,

The original post hardly ever mentioned causality. The quoted "we have to replace the hard drive and do a set up" hardly even hint to their proper investigation, not even of "negate-to-find" replacement method.

One may even have hard time understanding if the replaced and "set-up" is of corrupt boot-sector mapping, or a hardware failure, as they never mentioned any attempt to recover the so-called failed unit.

I follow (with interest) this discussion, and everybody is saying "missing detail here" and "missing detail there". Has the "Guest" re-appeared to answer any of your questions? Yuval

Hi yuvalmate,

The guest has registered, since filing the initial post, as fbompeh and given some more details in post #24. You are however correct and as yet there is no information how they diagnosed that the drive was faulty and what steps, if any, were taken to try and restore the drive to service. I have asked fbompeh in post #26, the answers to which, will hopefully shed some light on what is going wrong.

Yuval,

As Masu already stated, the OP signed in as "fbompeh" and has provided some additional information, but as is so often the case here, the devil is in the details and the details are slow in coming.

Regards, Greg

We do try formatting the drives but always comes up as disk error.

No, not just formatting, because if the boot-sector is corrupt you cannot re-format. Why not try to re-assign the disk-space anew, which practically mean something like "re-write a new boot-sector" as follows:

If it is for Win NT, Win 2000 or Win XP, start the setup program and following the F8 press, and prior to format, there is a stop where you can re-define partition size. Start by deleting any existing partitions (press D then L). At this time with no partitions at all, re-boot the machine, and enter the same stop again. Now, re-assign the partitions you want, and then format.

If it starts working again, it means that yours was probably not a hardware problem in the disk itself, and it may have been caused by noisy (spiked) power-line, or in the rare case, the power was cut or interrupted while the heads were during write mode, which may lead to the corruption of the boot-sector.

I once witnessed a disk formatted with Win 2000 NTFS gone corrupt, right after being scan-disked or de-fragged (I can't recall which) by an XP SP2 system. I later found out that Win2000 NTFS can only "see" (define) 132 Gb, while XP SP2 does not have this limitation, so the lesson was clear: Those NTFS versions are not compatible.

If re-assigning new partitions does not help, especially if after you re-boot and hear "crackle" noise from the disk, it usually means that the disk hardware is faulty, to any extent not worthy of re-assignment.

If you work with Win9x or so, the partitioning program is called "fdisk" and can be run from within Setup Disk Prompt. The procedure should be the same: delete all existing partitioned, re-boot, then reassign new partition(s), then format.

Maybe it was the other way around: XP SP2 NTFS disk was corrupt by Win2k NTFS system.

We are using window XP Professional.

Are your PC's configured in a workgroup having more than one version of windows? - maybe one file-system does scan-disk or de-frag to another, in maintenance routines, thus corrupting the disk?

Is the problematic PC having more than one physical disk partitioned by different file-system versions - maybe the fault is as above

Is the BIOS having trouble auto-recognising the disk? - maybe the disk is not compatible in some way to the specific motherboard

Are there more than one PC feeding off a common UPS? - check it's reccomended working load, maybe the total output of the UPS is overtaxed (as hinted by Greg)

How long does it take for a working disk to croak? Is it too hot by then? Does it make strange ticking noise after booting it as "dead"? - maybe it's power supply not stable enough for this specific disk

Did you try to feed the problematic PC with an alternative power supply? - maybe it's being fed with a spiky of fluctuating line

By eliminating all possible causes of fault - one by one - you may reach a working configuration (There is a saying here in Israel: If it's there, and working, why fix it?):

1. Try to re-assign the disk-space (re-partition) as described before. Not working? - then:

2. Try to replace the power supply. Not working? - then:

3. Try to mount one of these disks on a PC with different motherboard. Not working? - then:

4. Try the problematic PC with a different type of disk. - Not working? - then ... etc. etc.

Ok that means that the drives are indeed failing catastrophically and that it is not some sort of transient causing the corruption of the data. That pretty much excludes any external effect as we would expect to see the same problems on the other machines.

Therefore it must be something to do with the specific machines that have the drives that are failing and top of the list has to be overheating. This can cause numerous problems with the electronics as well as the mechanics and disk media.

We really need to know how hot the disk drives are getting during operation.

You might like to think about taking the cover or enclosure off one of the systems and letting it run, for a couple of months, open to the air. This should reduce the temperature that the drives are operating at and if it fixes the problem will confirm that the problem is due to overheating. This may not be possible so if it isn't don't worry if you can't do it.

Another thing that you may like to try is purchasing a small 12 v DC fan and positioning it inside the system so that it circulates the air around the disk drive.

Finally if you still have one of the faulty drives dismantle it and see if there is any scoring on the disk media or particulate matter inside the disk chamber. This will give us and idea if the failures are associated with the electronics or something that is damaging the media and causing the drives to crash.

Have a think about these options and if you can get back to me with what you find. In the meantime I will have a further think about your problem and see if anything else comes to mind.

Not that only one computer hard drive keep failing, we are using three computers for our plant citect scada, and the supply for these computers comes from a UPS which supplies power to other computers for the offices. The problem is only the three citect computers hard drives keeps failing atleast every month we have to replace a hard in either the number one,two or three computer, so the problem is not on one particular computer.

The citect computers are always on for 24hrs we dont turn them off unless there is a problem.

Hi fbompeh,

Ok the problem is occurring in multiple machines but these are all one particular brand of computer and other computers in the same vicinity don't exhibit the problem. I have a few questions and thins that would be helpful if you could test.

When the drives fail have you tried reformatting them prior to replacement or are you just replacing them with new drive straight away?

Have you ever dismantled one of the faulty drives? If you carefully dismantle them and get to the heads and platters it may give you some idea of what is going wrong. If they are indeed crashing for some reason you will see scoring in the surface of the disks and particulate matter inside the disk chamber.

If you can get hold of a multimeter check the voltage coming from the UPS. You will need to check the voltage between each of the terminals in a power socket to each of the other terminals. If they are labeled A, B and C you should see something like this:

A – B mains voltage

B – C mains voltage

C – A close to zero volts

If you don't get something like this post what you do get and I will see if it means anything. By the way what voltage is the mains supply where you are?

Finally how hot are these disk drives getting during operation? They shouldn't be much warmer than the atmosphere and only feel slightly warm to the touch.

Don't despair there is a solution to your problem and I am sure with a little perseverance it can be found.

One thing that was not mentioned as a possible cause.......the posiblity of a BIOS Trojan Worm. Someone could have possibly implanted a BIOS Trojan ROOTBOT.

I would actually change to the ASUS Motherboard and use the Kingston RAM, a dual processor setup. Next, I have worked in several types of plants, mils, and mines, all full of vibration, and all sorts of dust and soot particles in the air. Our control systems had heavy filtration on the environmental control boxes that the CPU's were located. Each had its' own a/c climate control, most of the control rooms had more vibration then a fish out of water, this however never interupted the opperations of any of these computers. Your scada sytems are bunk......build your own PC platforms and don't rely on some other sorce. I once worked at a phosfate production plant, our main plant opperations control room was located atop a platform in the main process area, which had more then 10 conveyor systems, with electric motors on average of 50hp each. The vibration and dust was like being in the canistor of a tornado vacuum. We never had any real cpu failure do to these harsh environmental elements.

So, this would only lead to one thing that has not been even concidered, SABOTAG!

Sorry for the ugly dark cloud...... :o :eek: Maximo

I have never seen a virus damage a drive in a way that stops it from being reformatted. Corrupt the data and destroy all the data on it yes, but never destroy the drive itself.

That doesn't mean it doesn't happen but I have never seen it, has anybody seen a virus or malicious program actually render a hard disk completely unusable and un-formatable?

...has anybody seen a virus or malicious program actually render a hard disk completely unusable and un-formatable...

In my nightmares, but never in concrete reality. It's a piece of software. A virus may however drive the drive crazy by forcing it's heads to re-iterate endlessly, or drive it into irregular rotation to an extent beyond it's designed specification. But this is highly unlikely, and I can only assume the disk's firmware is written on ROM, to prevent such occurrence.

There has been some very interesting replies to this question. The one thing missing from the replies is a focus on anything other then the PC platform. I would be looking "outside the box" on this one.

I would be looking at your SCADA application to find the root cause of the failure of these drives.

I have had experience in the past (during Y2K remediation efforts) with a logic programming conflict between a SCADA system and one of its down stream PLC processors that could easily have accounted for the failure rate you are seeing on the SCADA system's hardware.

What whas happening was...

The default state for a bit in a bit array in one of the downstream PLCs was false. When the I/O daemon on the SCADA box read the bit on the PLC, it wrote the state of the bit to a variable on the SCADA box. The last thing the I/O daemon did was trip the "changed data" flag on the SCADA platform. When the SCADA saw the bit state change, a routine in the SCADA box said "crap... this bit needs to be true" and promply issued a write request to the I/O daemon to set bit XYZ123 on PLC ABC true. The I/O daemon did as requested and the PLC promptly executed the request.

On the next scan of the PLC, the bit was promply set false once again... so off we went to the races.

The impact of this on the whole system was 10 second screen update times on the SCADA HMI terminals and log files that went from 0 bytes to multi megabyte size in literally 5 minutes.

We found it by throwing an Ethernet protocol analyzer on the traffic into and out of the box and by looking at the packets to see if they made sense. When we saw the repeated read and write requests to the same spot on the PLC we new we had an issue in the logic. Fixing the code fixed all of the problems and network traffic dropped to anticipated levels and system response time also went to normal.

If left alone... I could well see how the situation we had could have easily smoked a hardrive or two in a month. The heads were constantly flying all over the disk platter reading and writing huge amounts of data that they were not designed for.

So... FWIW... check your code. If it all appears as it should, maybe a RAM based HD is your answer.

Rick...

The only time I have had a drive die completely the cause was one of the heads came detached from the arm. I am going to repair it one day. I'll let you know when I get it back together. I have no idea why it did this, I had replaced another that seemed to be in trouble but that is still going. SOD's law again. Are they just too cheep?

Unless you have a clean room and the head became detached wile the drive was not spinning you have Buckley's chance of getting it to work again.

Time will tell. the funny well not so funny part was that it was a replacement drive for one that is still working all though at first said to be dying. Just can't tell. Yes I have the required work space. The faulty drive lasted all of a few days. 80Gb 7200 rpm. I have yet another just in case. My ultimate is still working 20years later in my Amstrad 1640 only 20Mb but it does run DOS 5.2.

...My ultimate is still working 20years later...

Funny you should mention this: I have an old 40 meg Seagate with DOS 6.22 and win 3.1 still working well (installed as a push-in tray) on a 3 Ghz Athlon, no special drivers required.

On it, I have Norton Commander, a beautiful F22 3D simulator, and a beautiful 3d Pool simulator, both still in use under DOS.

My old Pentium (32 Mb RAM) and 486 (16 Mb RAM) are still operational, and waiting to be sold for billions of dollars after my own demise. Any starter bids?