So if the direction of deceleration is towards Earth, then you might be able to consider the mystery solved and blame it on the process of collecting the data. But if the deceleration is towards the sun or another direction, we have an observation of an unknown effect in physics. If the latter is the case, I think the mystery is just starting to be understood--with a long ways to go and many more observations before we can consider it solved.
No, we may have an observation of unknown physics. We may also have an observation of well-known physics which nobody has been able to quantify. The acceleration is sufficiently small and the two Pioneer craft so similar (virtually identical, in fact) that nobody has conclusively ruled out such trivialities as the colour of the paint having changed or a thicker-than-expected layer of dust having formed - effects like these could be sufficient to cause the observed anomalous acceleration (think 'solar sails' for similar physics, but with the Pioneers emitting radiation as well as reflecting it). It really is a tiny effect.
The problem with ruling these kinds of effects out is that the Pioneers are the only suitable craft we have which have been going far enough for long enough to provide good data. The Voyager craft have gone a long way too, but they aren't spin-stabilised (like a bullet from a rifled firearm) like the Pioneers; they are stabilized by thrusters and firing the thrusters causes sufficient uncertainty that the anomalous acceleration is lost in noise. I recall a paper which looked at other spin-stabilised craft during their cruise phase to other planets which also reported the anomaly, but again the data was too noisy to be entirely convincing.
We really need a new mission to test this anomaly properly before we can say there is any new physics.
As far as the shuttle software is concerned, it's very small in size and given that there have been numerous computer problems on the shuttle, it's not a foregone conclusion that none of these errors are due to software bugs - signed off or not.
What problems have there been with the DPS (flight control) software? There's the year-end rollover thing, but I believe that's a bug in the specification, not the software. The BFS failed to initialise before STS-1 and there was a bug discovered in the fuel dumping code just before STS-2 - they were real software bugs. Some thrusters briefly fired inadvertently during STS-63, though really that was an electronics problem (a power-on transient), rather than a software problem. What else has gone wrong with the DPS?
I've only worked on my campus two years, and I've already seen that come up twice, where the logs were pulled for e-mail to say "did this student send a message like they said they did", or "did this student receive the message like they claim they didn't?"
So you agree with huge restrictions on your users because it's useful in a specific eventually which happens on average once per year? Are you completely fucking insane?
I could be entirely off base here and readily admit my lack of knowledge of physics involved. But if the earth is basically a cooler, with the outer layers of earth acting as insulation then the sun heating the earth is really no different than the sun heating a cooler of beer. The insulation stops heat from escaping and from getting in but there is some loss both ways. If the cooler is heated faster than it loses heat then the heat will build up inside the cooler and the internal temperature will continue to rise beyond the ambient temperature while the surface of the cooler will not. You could use other examples like a car.
The second law of thermodynamics disagrees. Heat moves from hot things to cold things, not the other way. In order for heat to flow from the outside of the beer cooler to the inside, the outside would have to be hotter than the inside. If the inside was hotter, heat would flow out, not in.
The part where the ambulance goes by is pretty funny.
At first I thought that was a lawyer-lawyer jibe and he was about to ask if the other lawyer wanted to chase it. But it was almost that good anyway:
MR. GABRIEL: Why don't you wait until the ambulance passes.
MR. BECKERMAN: I don't think we --
MR. GABRIEL: It may take a while.
MR. BECKERMAN: This is New York, Richard. This isn't Denver. We could be here all day.
MR. GABRIEL: Just try to keep your voice up.
As a computer engineer myself, I can say that it is *EXTREMELY* rare for a computer engineer to be a licensed PE.
Way off-topic, but programming desperately needs the kind of accountability and professionalism that 'real' engineering has. We're around where engineering was 100 years ago just now, with a hundred different screw threads and steam engines which explode in your face. 'software engineering' may be an academic discipline, but 'professional' (in their execution) software engineers are few and far between and professionally engineered software is rarer still. The lawyer is making a valid point.
Before you ask, I am a professional (it's my job) programmer. I'd love to be an engineer. I'd love to work somewhere where those kind of standards were applied. I'd get a CS degree (mine is in Physics), but those programmers I've worked with who have CS degrees don't seem much more engineer-like in their application than those without. Too much hacking, not enough engineering. Perhaps civil engineers would be the same if every bridge had "this bridge comes with no warranty, either express or implied" written into the contract.
Ben Franklin said he was a old widow with kids because saying his real identity would have distorted what he was trying to say. And I am sure he felt that way, otherwise he would have described himself at least as a man.
There is a massive difference between writing a letter to your brother's newspaper and writing for an encyclopedia. Few would take a letter in a newspaper as more than a single example or an opinion, if they believed it at all. An encyclopedia is supposed to consist of a higher grade of information. Passing yourself off as an authority in that arena is correspondingly a far more serious matter.
He spent an entire page talking about the plus80 certification and why it's important and how he chose a SeaSonic M12 modular powersupply that was plus80 certified.
An enire page which wasn't even wrong. They author obviously hasn't got the first clue about what they are writing about.
From TFA: "If a PSU meets the certification, it will use only the power it needs at a given load: In other words, it won't use more power than it needs. For example, if your PC requires only 20 percent of the total power of a 500-watt PSU, the system will consume no more than 100 watts"
That's just meaningless bullshit. If he's including the power used by the inefficiency of the PSU, then all computers only use as much power as they require. If he's not, then he's implying a 100% efficient power supply, which is not possible. 80plus means the computer as a whole will consume no more than 125% of the power the componenets exclusing the PSU require. In other words, that the PSU is at least 80% efficient. Can you guess where they got the name?
What sort of power conditioning? Would galvanic isolation, rectification into a substantial input capacitance and a regulated ouput suffice? That's what you get in a computer PSU.
The last survey that popped up here said that if SMART says your drive will fail, it probably will, but if SMART doesn't say it will fail, it doesn't mean much.
Actually, other than scan errors, the Google data shows that even with SMART errors, you drive probably won't fail soon (they all fail eventually, of course). It is much more likely to fail than a drive without any errors, but still more likely to last another year than fail.
Perhaps you misinterpreted the label on the y-axis of that figure. It is not in percent, it is a multiplier. So 0.5 means 50%.
Oops, indeed I did. It only scales my interpretation, rather than contradicting it though. It still indicates that wear is highly significant (which I expected, but previously erroneously asssumed was accounted for by MTBFs applying to power-on time rather than calendar time).
The quality of consumer hard drives nowadays is way, WAY low, and the manufacturers should do something about it.
Point, counterpoint...
I've never had a single one of my own hard drives fail. Not a single one, ever. I've had a dozen or so that I can remember, from the 20MiB drive in my Amiga to the 250GiB that now hangs off my NSLU2. They are all either still functioning or became obsolete before failing. Many of them have been run 24/7 for significant chunks of their lives and I don't replace them unless they become too small to be useful (they don't shrink of course, but my needs grow).
Sure, they will have reached "thermal equilibrium" after a short period of time. See Figure 9 in this paper " Reliability reduction with increased power on hours, ranging from a few hours per day to 24 x 7 operation " to see how I'm not sure that merely being hot is the problem.
The graph mostly seems to indicate that drives wear out when they are spinning. It's not all that far from a straight line (if you ignore the very low hours), which you would expect if wear was a significant component in the risk of failure. To a rough approximation, that graph shows a 0.5% risk of failure independent of usage level, then an additional 0.5% risk per 3000 hours/year of usage.
I always assumed quoted MTBF was for power-on hours and framed my discussion in those terms, but it seems I should have read the small-print.
It's hard to take someone seriously when they claim that their drives have a 100+ year MTBF, especially since precious few are still functional after 1/10th of that much use.
You're misinterpreting MTBF. A 100 year MTBF does not mean the drive will last 100 years, it means that 1/100 drives will fail each year. There will be another spec somewhere which specifies the design lifetime. For the Fujitsu MHT2060ATdrive which was in my laptop the MTBF is 300 000 hours, but the component life is a crappy 20 000 hours or 3 years - 93% of drives should make it that far given the MTBF. After the end of the design lifetime, all bets are off.
Do you seriously think a drive won't have reached thermal equilibrium after an hour, let alone after several hours? Mine seem to get up to their 'normal' temperatures in 30 minutes or less. And according to the Google study, heat doesn't lead to a significantly increased risk of failure till you get above 45 C or so.
TFA seems surprised by SATA drives lasting as long as Fibre...why one earth would your data interface have any consequences on the drive internals?
Fibre Channel drives, like SCSI drives, are assumed to be "enterprise" drives and therefore better built than "consumer" SATA and PATA drives. It's nothing inherent to the interface, but a consequence of the environment in which that interface is expected to be used. At least, that's the idea.
It's not how far it falls, it's how much shock it takes at the end. Doesn't take much high-density foam to trim the impact shock to something the drive can tolerate.
How much is not much?
Let's say you hit the ground at 100m/s (that's just shy of 200kts). A fairly arbitrary number, but in the right ballpark for a light aircraft crash I'd guess. I don't know what drives they use in iPods, but I bet the specs are pretty similar to the 1.8" Hitachi C3K80 drives I could easily find specs for. That specifies maximum shock of 1500G for 1ms, as a half sine-wave. Let us (with no justification beyond keeping the estimate simple) assume it can take 1500G for as long as it takes to stop and assume it's a square wave, not a sine. Taking g as 10m/s^2, we can handle an acceleration of 15000m/s^2. 100/15000 = 0.0067, so it'll take 6.7ms to stop from 100m/s. How far do we travel during that deceleration? Assuming a constant rate of deceleration, average speed would be half the initial speed, or 50m/s, so we'd go 50*0.0067 = 0.33m. About a foot of foam. And that's assuming constant deceleration from our foam (no chance), that we can compress the foam to zero thickness (no chance), that the drive is not operating (unlikely) and we're really can exceed the specified shock load in duration by a factor of 6 (unlikely). Two feet of high-density foam would seem closer to the mark and that's actually quite a lot of foam if you ask me.
No, it remains false until the user (ther person designing the op-amp circuit) creates the virtual short by introducing negative feedback. If you don't have negative feedback (which is not built-in to general-purpose op-amps), there is nothing to defeat. Try it: Take an op-amp, hook up the power then put 1V on one input and 1.5V on the other (via say 1k resistors, to give whatever magic you think drives the inputs in the absence of feedback a chance). There will be a 0.5V difference between the inputs. Where is your 'virtual short'? Now, introduce negative feedback by connecting the output to the inverting input. The output can now drive the inverting input to the same voltage as the noninverting input, which is exactly what will happen - there's your virtual short. No feedback = no virtual short.
There is a "virtual short circuit" across the input terminals. This is to say that the voltage difference between them is very close to zero at all times.
This is only true of op-amps used with negative feedback. They are usually used with negative feedback of course, but it's still a property of op-amp circuits, not of op-amps themselves.
Dude - you spent 3 paragraphs giving the "operation", but still didn't say what it DOES. What function does it perform?
It was in response to somebody saying they knew what it did, but not how it did it...
But anyway, operational amplifiers amplify the difference in voltage between their two (inverting and on-inverting) inputs. They're largely useless used open-loop, as they have voltage gains of tens of thousands to millions, so even slght noise sends the ouptut swinging about wildly. They are virtually always used with negative feedback (some connection between the output and inverting input). An op-amp with negative feedback drives the output till the two inputs are at the same voltage. With various simple circuits around them they can be made to amplify, add or subtract voltages, form the heart of filters (high-, low- or band-pass), buffer signals, integrate or differentiae signals, drive high-power loads or many other things I can't remember just now. Check out National Semicondustor Application Note AN-31 for a whole bunch of circuits you can build around an op-amp or two with a few other basic components. AN-4 and AN-20 give a written introduction to the theory and applications of op-amps.
The corporate English was much better than the self-proclaimed journalist's English:
Know this, while I disagree with this decision and think it is a monumental mistake, it will not effect our continuing coverage of Sony and the gaming software and hardware your company makes and supports.
The first comma should be a colon and "effect" should be "affect". I love the way so many bloggers rant about being journalists without bothering to learn the most basic tool of a journalist's trade - the language.
Slashdot Mirror
No, we may have an observation of unknown physics. We may also have an observation of well-known physics which nobody has been able to quantify. The acceleration is sufficiently small and the two Pioneer craft so similar (virtually identical, in fact) that nobody has conclusively ruled out such trivialities as the colour of the paint having changed or a thicker-than-expected layer of dust having formed - effects like these could be sufficient to cause the observed anomalous acceleration (think 'solar sails' for similar physics, but with the Pioneers emitting radiation as well as reflecting it). It really is a tiny effect.
The problem with ruling these kinds of effects out is that the Pioneers are the only suitable craft we have which have been going far enough for long enough to provide good data. The Voyager craft have gone a long way too, but they aren't spin-stabilised (like a bullet from a rifled firearm) like the Pioneers; they are stabilized by thrusters and firing the thrusters causes sufficient uncertainty that the anomalous acceleration is lost in noise. I recall a paper which looked at other spin-stabilised craft during their cruise phase to other planets which also reported the anomaly, but again the data was too noisy to be entirely convincing. We really need a new mission to test this anomaly properly before we can say there is any new physics.What problems have there been with the DPS (flight control) software? There's the year-end rollover thing, but I believe that's a bug in the specification, not the software. The BFS failed to initialise before STS-1 and there was a bug discovered in the fuel dumping code just before STS-2 - they were real software bugs. Some thrusters briefly fired inadvertently during STS-63, though really that was an electronics problem (a power-on transient), rather than a software problem. What else has gone wrong with the DPS?
So you agree with huge restrictions on your users because it's useful in a specific eventually which happens on average once per year? Are you completely fucking insane?
The second law of thermodynamics disagrees. Heat moves from hot things to cold things, not the other way. In order for heat to flow from the outside of the beer cooler to the inside, the outside would have to be hotter than the inside. If the inside was hotter, heat would flow out, not in.
At first I thought that was a lawyer-lawyer jibe and he was about to ask if the other lawyer wanted to chase it. But it was almost that good anyway:
MR. GABRIEL: Why don't you wait until the ambulance passes.MR. BECKERMAN: I don't think we --
MR. GABRIEL: It may take a while.
MR. BECKERMAN: This is New York, Richard. This isn't Denver. We could be here all day.
MR. GABRIEL: Just try to keep your voice up.
Way off-topic, but programming desperately needs the kind of accountability and professionalism that 'real' engineering has. We're around where engineering was 100 years ago just now, with a hundred different screw threads and steam engines which explode in your face. 'software engineering' may be an academic discipline, but 'professional' (in their execution) software engineers are few and far between and professionally engineered software is rarer still. The lawyer is making a valid point.
Before you ask, I am a professional (it's my job) programmer. I'd love to be an engineer. I'd love to work somewhere where those kind of standards were applied. I'd get a CS degree (mine is in Physics), but those programmers I've worked with who have CS degrees don't seem much more engineer-like in their application than those without. Too much hacking, not enough engineering. Perhaps civil engineers would be the same if every bridge had "this bridge comes with no warranty, either express or implied" written into the contract.
There is a massive difference between writing a letter to your brother's newspaper and writing for an encyclopedia. Few would take a letter in a newspaper as more than a single example or an opinion, if they believed it at all. An encyclopedia is supposed to consist of a higher grade of information. Passing yourself off as an authority in that arena is correspondingly a far more serious matter.
RoHS is about removing nasty stuff from devices, but it says nothing about recycling. It's WEEE which deals with the full life-cycle.
An enire page which wasn't even wrong. They author obviously hasn't got the first clue about what they are writing about.
From TFA: "If a PSU meets the certification, it will use only the power it needs at a given load: In other words, it won't use more power than it needs. For example, if your PC requires only 20 percent of the total power of a 500-watt PSU, the system will consume no more than 100 watts"
That's just meaningless bullshit. If he's including the power used by the inefficiency of the PSU, then all computers only use as much power as they require. If he's not, then he's implying a 100% efficient power supply, which is not possible. 80plus means the computer as a whole will consume no more than 125% of the power the componenets exclusing the PSU require. In other words, that the PSU is at least 80% efficient. Can you guess where they got the name?
What sort of power conditioning? Would galvanic isolation, rectification into a substantial input capacitance and a regulated ouput suffice? That's what you get in a computer PSU.
Actually, other than scan errors, the Google data shows that even with SMART errors, you drive probably won't fail soon (they all fail eventually, of course). It is much more likely to fail than a drive without any errors, but still more likely to last another year than fail.
Source?
Oops, indeed I did. It only scales my interpretation, rather than contradicting it though. It still indicates that wear is highly significant (which I expected, but previously erroneously asssumed was accounted for by MTBFs applying to power-on time rather than calendar time).
Point, counterpoint...
I've never had a single one of my own hard drives fail. Not a single one, ever. I've had a dozen or so that I can remember, from the 20MiB drive in my Amiga to the 250GiB that now hangs off my NSLU2. They are all either still functioning or became obsolete before failing. Many of them have been run 24/7 for significant chunks of their lives and I don't replace them unless they become too small to be useful (they don't shrink of course, but my needs grow).
The graph mostly seems to indicate that drives wear out when they are spinning. It's not all that far from a straight line (if you ignore the very low hours), which you would expect if wear was a significant component in the risk of failure. To a rough approximation, that graph shows a 0.5% risk of failure independent of usage level, then an additional 0.5% risk per 3000 hours/year of usage.
I always assumed quoted MTBF was for power-on hours and framed my discussion in those terms, but it seems I should have read the small-print.
You're misinterpreting MTBF. A 100 year MTBF does not mean the drive will last 100 years, it means that 1/100 drives will fail each year. There will be another spec somewhere which specifies the design lifetime. For the Fujitsu MHT2060ATdrive which was in my laptop the MTBF is 300 000 hours, but the component life is a crappy 20 000 hours or 3 years - 93% of drives should make it that far given the MTBF. After the end of the design lifetime, all bets are off.
Do you seriously think a drive won't have reached thermal equilibrium after an hour, let alone after several hours? Mine seem to get up to their 'normal' temperatures in 30 minutes or less. And according to the Google study, heat doesn't lead to a significantly increased risk of failure till you get above 45 C or so.
Fibre Channel drives, like SCSI drives, are assumed to be "enterprise" drives and therefore better built than "consumer" SATA and PATA drives. It's nothing inherent to the interface, but a consequence of the environment in which that interface is expected to be used. At least, that's the idea.
iHopeTheyDon'tMakeACrappyTVShowAboutTheSurvivorsOf ThisCrashBeingStuckOnAnIslandWhereWierdShitHappens
How much is not much?
Let's say you hit the ground at 100m/s (that's just shy of 200kts). A fairly arbitrary number, but in the right ballpark for a light aircraft crash I'd guess. I don't know what drives they use in iPods, but I bet the specs are pretty similar to the 1.8" Hitachi C3K80 drives I could easily find specs for. That specifies maximum shock of 1500G for 1ms, as a half sine-wave. Let us (with no justification beyond keeping the estimate simple) assume it can take 1500G for as long as it takes to stop and assume it's a square wave, not a sine. Taking g as 10m/s^2, we can handle an acceleration of 15000m/s^2. 100/15000 = 0.0067, so it'll take 6.7ms to stop from 100m/s. How far do we travel during that deceleration? Assuming a constant rate of deceleration, average speed would be half the initial speed, or 50m/s, so we'd go 50*0.0067 = 0.33m. About a foot of foam. And that's assuming constant deceleration from our foam (no chance), that we can compress the foam to zero thickness (no chance), that the drive is not operating (unlikely) and we're really can exceed the specified shock load in duration by a factor of 6 (unlikely). Two feet of high-density foam would seem closer to the mark and that's actually quite a lot of foam if you ask me.
No, it remains false until the user (ther person designing the op-amp circuit) creates the virtual short by introducing negative feedback. If you don't have negative feedback (which is not built-in to general-purpose op-amps), there is nothing to defeat. Try it: Take an op-amp, hook up the power then put 1V on one input and 1.5V on the other (via say 1k resistors, to give whatever magic you think drives the inputs in the absence of feedback a chance). There will be a 0.5V difference between the inputs. Where is your 'virtual short'? Now, introduce negative feedback by connecting the output to the inverting input. The output can now drive the inverting input to the same voltage as the noninverting input, which is exactly what will happen - there's your virtual short. No feedback = no virtual short.
This is only true of op-amps used with negative feedback. They are usually used with negative feedback of course, but it's still a property of op-amp circuits, not of op-amps themselves.
It was in response to somebody saying they knew what it did, but not how it did it...
But anyway, operational amplifiers amplify the difference in voltage between their two (inverting and on-inverting) inputs. They're largely useless used open-loop, as they have voltage gains of tens of thousands to millions, so even slght noise sends the ouptut swinging about wildly. They are virtually always used with negative feedback (some connection between the output and inverting input). An op-amp with negative feedback drives the output till the two inputs are at the same voltage. With various simple circuits around them they can be made to amplify, add or subtract voltages, form the heart of filters (high-, low- or band-pass), buffer signals, integrate or differentiae signals, drive high-power loads or many other things I can't remember just now. Check out National Semicondustor Application Note AN-31 for a whole bunch of circuits you can build around an op-amp or two with a few other basic components. AN-4 and AN-20 give a written introduction to the theory and applications of op-amps.
95% of people buy computers, not operating systems.
The corporate English was much better than the self-proclaimed journalist's English:
The first comma should be a colon and "effect" should be "affect". I love the way so many bloggers rant about being journalists without bothering to learn the most basic tool of a journalist's trade - the language.