Aw, heck. If God wants your house to burn down, it's gonna burn down.

When I said "safety" I wasn't thinking about fire (although, turning things off does reduce that risk).

IMO the biggest risk is your local utility, followed closely by acts of God (lightning).

We don't get much lightning here, but I did lose a disk controller to lightning once. Never figured out how the motherboard and hard drive survived when the jolt had to get through one or the other....

---

Aw, heck. If God wants your house to burn down, it's gonna burn down.

When I said "safety" I wasn't thinking about fire (although, turning things off does reduce that risk).

IMO the biggest risk is your local utility, followed closely by acts of God (lightning).

We don't get much lightning here, but I did lose a disk controller to lightning once. Never figured out how the motherboard and hard drive survived when the jolt had to get through one or the other....

We had lightning strike here that took out local sub-station transformer, just across the road. It only took out one computer. But that hit the MB, HDD, graphics card and monitor. The cpu and memory was found to work, and the monitor would switch on, so investigated internally and it had been struck through the data cable from the graphics card.

Andy

Aw, heck. If God wants your house to burn down, it's gonna burn down. Life is about taking risks. I am at risk every day when I get outta fricken bed. Just leave 'em on and don't leave any flammable in close proximity. I have 8 rigs crunchin' away 24/7, and my kitties keep an eye on them.
Seriously, what is the risk? Everything is supposed to be UL approved to sell in the US, which means the flammability of all materials used is suupposed to be 'self extinguishing' so unless there is something supporting a flame, it will go out on it's own.
I recently changed out my Commando mobo for a new P5K, and when I took the Commando out, much of the area around the PWM transformers had pretty much turned to charcoal. (Thoght I had smelled something burning).
They had been covered up by the insulation for the phase cooling evap, and some of that had turned to toast as well. Those babys run hot!!!
But, no flame, no foul. And even if there had been some combustion, it is doubtful that it could have spread. So little combustible material. May have made a bit of a smoke bomb, but not much else.
So I guess it all depends on whether you are a risk taker in life, or a paranoid schizophrenic.
(Not making fun of those who choose to take the conservative route).

I had a system catch fire last March. The fan on the PS died and I didn't know it. Smoke was all over the place and the smell was awful! The capacitors were exploding one by one until I found the fire extinguisher and gave it a blast through an open slot where I'd pulled a CD drive out. The only things salvageable were the hard drives (and then only to offload the data...don't trust them any more).

I had another system stacked on top of that one, so it suffered heat damage as well. Lucky I had a spare case to rebuild the second system into.

The insurance settlement was good enough to get an AMD X2 4600+ to replace the system that burned. I convinced them that since I build my own systems that it would be cheaper for them if they paid for a pre-built barebones system than if they replaced the individual parts at cost and I rebuilt it...and it was.

...
IMO the biggest risk is your local utility, followed closely by acts of God (lightning).

We don't get much lightning here, but I did lose a disk controller to lightning once. Never figured out how the motherboard and hard drive survived when the jolt had to get through one or the other....

We had lightning strike here that took out local sub-station transformer, just across the road. It only took out one computer. But that hit the MB, HDD, graphics card and monitor. The cpu and memory was found to work, and the monitor would switch on, so investigated internally and it had been struck through the data cable from the graphics card.

Andy

When I worked in dial-up ISP Tech Support, we would regularly get calls from people whose modem had stopped working. Usually the cheapest internal modems which would report "no dial tone" or similar.

After a while I got to ask the users a few pertinant questions, e.g. have you seen or heard lightning/thunder recently? It does not need to be a nearby stike to induce a pulse down a phone line.

---

Sir Arthur C Clarke 1917-2008

When I worked in dial-up ISP Tech Support, we would regularly get calls from people whose modem had stopped working. Usually the cheapest internal modems which would report "no dial tone" or similar.

After a while I got to ask the users a few pertinant questions, e.g. have you seen or heard lightning/thunder recently? It does not need to be a nearby stike to induce a pulse down a phone line.

I had a lightning strike take out my modem (and only my modem thankfully) once. It was a dang good modem from Zoom. Hardware based 16bit ISA modem that performed better than any other modem I had ever owned. Shortly after that, I switched over to SBC xDSL and did away with modems altogether.

After that, I went out and purchased a 10,000 joule surge suppressor with modem protection. I know it won't help in a direct lightning strike, but it'll help during most storms.

---

I will usually let my one and only rig run 15 hours straight during the day and then turn it off when im sleeping. I also leave my computer on during eletrical storms unless directly overhead.
My computer shuts down if it reaches a certain temperature, But i would certainly never leave my rig on 2 weeks straight.

---

- Luke.

I will usually let my one and only rig run 15 hours straight during the day and then turn it off when im sleeping. I also leave my computer on during eletrical storms unless directly overhead.
My computer shuts down if it reaches a certain temperature, But i would certainly never leave my rig on 2 weeks straight.

Repeatedly switching computers on and off over a few months cause socketed chips to crawl out of their sockets as the metal expands and contracts with some slight 'one way-ness' in the socket with the result that connections become unreliable and the machine starts crashing when the pins move a few nanometres too far to keep the current flowing.

You can tell that this happens by pressing down on a socketed chip and hearing a click. Anyway, nearly everything is surface mounted now, so this shouldn't happen anymore.

Power cycling also introduces an extra failure point for hard disks and adds extra stress to the system. A drive may work indefinitely when on, but fail to restart after shutdown, so make sure you have everything backed up. I wouldn't get on an aircraft if it had 1000:1 chance of crashing!

In a low dust environment there is absolutely nothing to fear from running a machine permanently - I have done so with several for 20 years now without problems, and even the 20 year old machines still work.
In a high dust environment, the processor overheats and slows down.

My webserver sits in a huge room shared with over a 1000 other servers. None of them blow up either, although I'm sure a hard disk fails occasionally due to MTBF.

If it doesn't blow up while you're sitting next to it, it ain't gonna blow up when you're not sitting next to it!

Well, if you turn the computer off when you sleep, then it would make sense to turn it off when you're not even there; otherwise, why turn it off while sleeping? :)

After watching the Live Earth concerts on TV, I'm re-thinking the whole value of running BOINC. Looking for a signal from outerspace has no real practical purpose and since I already believe the universe is too big for their not to be ETI out there... hmmm

Also running climate prediction.. here I am contributing to the pollution of the atmosphere by running a climate prediction model that is most likely trying to figure out what rate global warming is/would happen given different scenarios. A bit of irony/hypocracy there.

---

Well, if you turn the computer off when you sleep, then it would make sense to turn it off when you're not even there; otherwise, why turn it off while sleeping? :)

After watching the Live Earth concerts on TV, I'm re-thinking the whole value of running BOINC. Looking for a signal from outerspace has no real practical purpose and since I already believe the universe is too big for their not to be ETI out there... hmmm

Also running climate prediction.. here I am contributing to the pollution of the atmosphere by running a climate prediction model that is most likely trying to figure out what rate global warming is/would happen given different scenarios. A bit of irony/hypocracy there.

Well, if you feel strongly about not running cpdn and seti, then remember, you can always run Rosetta, Docking, Simap, and proteins to try to help find cures for Cancer, Aids, alzheimers, and other protein related diseases. You could run MalariaControl and try to help out with that. These projects might be of significant benefit within your lifetime.

[the previous message is my own opinion. The opinion of others may vary]

I will be gone for about 2 weeks visiting Washington D.C. My question is will my computer explode if I leave it on to crunch while I'm gone? I usually run BOINC all day but turn my computer off while I'm sleeping.

The fact that you have to ask tells me that you are uneasy and uncertain about the notion leaving your computer on for that length of time unattended.

Would you leave your television on continually for two weeks with nobody there?
Or your games console?
Or your stereo?
Perhaps your dvd player on repeat play for two weeks?

No, nor would I!
Switch it off.

Guess it's time to get my two cents into this debate. :)

First, I was an electronics repair technician (fancy word for tv repairman) for many years with a company that repaired all types of home electronic equipment and was an authorized factory service center for the Magnavox brand. I'm also a licensed amateur radio operator. I've been playing around with electrical/electronic gadgets ever since I was in the third grade in school. (I'm 52 years old now so that is a LONG time!) So I know what I've seen and experienced through the years.

First, "most" failures occur when a device is first turned on. Think of a light bulb. How many times have you had a light bulb be turned on and just suddenly blow out? Now think of how many times you flip a light switch and just get a flash then nothing? I've had the same experience with radios and tv's. Most people who contact us to repair something that didn't work properly say it happened when they first turned it on. Some typical complaints would go: "My tv was playing perfectly when I turned it off but when I turned it back on it just clicked (or popped or made a frying sound or numerous other complaints/fireworks) and then nothing." Or "My tv was playing perfect when I turned it off, and now when I turn it on there's nothing but a bright line across the screen (or no sound, or sound but no picture, or numerous other problems). The key point is nothing happened until they turned it off and back on.

This is due to the sudden stress of powering up a "cold" circuit. Capacitors loose their charge while off, and when the circuit is first powered up there's a surge of current while they are building back their charge. Parts are cold and in normal electrical and electronics circuits a cold element has less resistance than a warm one. Then there's actual heat stress on the components themselves. Everything in the world with just a few exceptions expands when heated. As Andy Lee Robinson pointed out, this can cause socketed chips to gradually creep loose, but it can also cause stresses to build up in surface mounted devices. This causes internal failures inside the chip.

A common symptom on a tv or radio is the set is playing but someone walks across the floor and something goes wrong with it, or the other way around, it won't play properly unless you walk across the floor or smack it. I've found a large part of the time that this happens is when a large power resistor is soldered onto a circuit board with it's only support being the leads. The continual expansion and contraction of the resistor flexes the leads eventually causing a break. This is truly a case to explain a repair bill my dad saw a fellow hand out... "One drop of solder, 0.25, Knowing where to put it, 20.00".

I leave two computers and a tv on all the time. They never get turned off unless the power goes off or I'm working on something that requires their power to be off. One of the computers has a dead cmos battery and I have no way of knowing how long it has been dead because it had been several months since I had the computer turned off!

Now I don't want anyone to do anything they're in any way uncomfortable with, so if it worries you to have a computer or other electronic device turned on while they are asleep or away from home, then don't do it. There's enough worries that the normal person has to cope with without adding additional worries. But if you feel safe doing it, then by all means do it. It is better on the equipment. It may run your power bill up some, but you can always rationalize that by having BOINC running 24/7 and contributing to science.

In the case of my tv, I rationalize that by the fact that if someone comes to my house and all the lights are off, no sound, and no one answers the door, then we're likely not at home. If the tv and lights are on then somebody just might be around close so if they were to break in they just might get caught! (We had been broken into several times before I started leaving it on 24/7.) And yes I have been away from home for several weeks at a time. Since the computers are connected to the network, I can always get on a friends computer and check on the condition of the ones at home.

---

Jim

Some people plan their life out and look back at the wealth they've had.
Others live life day by day and look back at the wealth of experiences and enjoyment they've had.

@Jim-R

My computers are never turned off either. I've got about a decade more than you in the electronic service industry. I started as a TV tech also. I couldn't agree more with your comments.

As for electronic equipment starting a fire, I found this to be extremely rare. There is very little inside electronic equipment that will actually catch on fire and todays cases are all fire resistant. Most times the smoke you see or the acrid smell come from the electrolytic of a blown capacitor or some other overheating component. I couldn't start to count how many piece of equipment I've had to be replace because of fire extinguisher damage when simply pulling the plug would have left the unit with only a minor repair job.

---

Boinc V7.2.42
Win7 i5 3.33G 4GB, GTX470

No, nor would I!
Switch it off.

You seem to miss the point - the uses you mention require the presence of a human to enjoy their functions. Of course it would be an obscene waste to leave them running for no reason.

One could argue that as these are unnatural human recreational activities and consume resources for the sole purpose of keeping a vegetating human brain entertained with intellectual chewing gum without actually do anything constructive for the environment is also an obscene waste.

However, computers can do useful work 24/7 without needing human presence, and I believe the end justifies the means.

You seem to miss the point - the uses you mention require the presence of a human to enjoy their functions. Of course it would be an obscene waste to leave them running for no reason.

A typical home pc fits this description perfectly...

One could argue that as these are unnatural human recreational activities and consume resources for the sole purpose of keeping a vegetating human brain entertained with intellectual chewing gum without actually do anything constructive for the environment is also an obscene waste.

Again, in what way does this not describe a home computer...

However, computers can do useful work 24/7 without needing human presence, and I believe the end justifies the means.

Just because something can be left on 24/7, it does not logically follow that it should. I'm not in the business of running a 'farm' with its associated costs. SETI get my 'spare' cpu cycles - if I'm not actually using a computer for a specific purpose, it's turned off. On its own, for me, running SETI is not sufficient reason for enlarging my energy bill.

You seem to miss the point - the uses you mention require the presence of a human to enjoy their functions. Of course it would be an obscene waste to leave them running for no reason.

A typical home pc fits this description perfectly...

The difference being as the previous poster pointed out that unless there is someone there there is no point in leaving a TV or stereo running.
With a computer it can be doing usefull work while you're not there, be it Seti, Einstien, video encoding, running database queries, running a web server, etc, etc.

---

Grant
Darwin NT

Well, if you turn the computer off when you sleep, then it would make sense to turn it off when you're not even there; otherwise, why turn it off while sleeping? :)

Because it makes noise while he's trying to sleep?

---

Indeed, the ol' computer "Keep it running vs turn it off" question is decades old. There has never been two greatly split camps since the Blue and Grey. I figured most people kept their comps running 24/7 these days though. I say surge protect, UPS, and don't worry. I have a dual P3 that has been power cycled less than the refrigerator since '99.

Has anybody noticed how many electronic devices don't actually switch off these days, unless you actually disconnect them from the power source, they only go into standby mode.

---

@Jim-R

My computers are never turned off either. I've got about a decade more than you in the electronic service industry. I started as a TV tech also. I couldn't agree more with your comments.

As for electronic equipment starting a fire, I found this to be extremely rare. There is very little inside electronic equipment that will actually catch on fire and todays cases are all fire resistant. Most times the smoke you see or the acrid smell come from the electrolytic of a blown capacitor or some other overheating component. I couldn't start to count how many piece of equipment I've had to be replace because of fire extinguisher damage when simply pulling the plug would have left the unit with only a minor repair job.

It is a little know fact that electronic equipment actually works on packaged smoke. This can be observed as after the smoke is released from the packages, the item in question usually stops functioning.

---

Boinc Button Abuser In Training >My Shrubbers<

Indeed, the ol' computer "Keep it running vs turn it off" question is decades old...

The "must keep it on" was especially important in the days of the old thermionic valve computers. Turning the machine back on after a power-down had a very good chance of blowing a few valves.

Another trick to try to improve the compute reliability was to have a maintenance schedule where banks of valves were replaced after so many hours of life before they failed.

For today's machines, the power-up/power-down cycle is likely not so bad now for semiconductor technology itself. What is still bad is the temperature cycling for the chip packaging (with the hundreds of chip pins/pads getting ripped off the motherboard), and for all the other mechanical parts such as HDDs and fans.


Google published a very good research paper on their HDD failures. Anyone know of any manufacturer figures showing the trade-off between HDD spin-up/spin-down cycles and continuous spin life? Is it better to have "power saving" enabled and let the disk spin down? Or better to keep it spinning?

Happy crunchin',
Martin

---

See new freedom: Mageia Linux
Take a look for yourself: Linux Format
The Future is what We all make IT (GPLv3)