17-year-old Dead after Tweeting in the Bathtub



  • @morbiuswilters said:

    @ender said:

    I've got 2 electrical outlets in my bathroom, one is pretty much permanently used by the hairdryer, while the other one gets used to recharge my shaver, but they're on the opposite corner from the bathtub. I guess it would be possible to plug in a laptop to use in the bath, but the power brick would probably hang in the air.

    Are they GFCIs?  Are GFCIs common in Europe?

     

     

    Yes, they are called RCD (residual current device). They compair the current flowing on the Live wire (Hot in US) to the current on the nuteral (sometimes Cold in US).If the difference is more then 5 to 30 mA (depending on device) then is disconnects the supply, since the difference is 'leakage current'.  If it does trip then it means that somewhere downstream of the RCD there reason for current to leak away, be it to earch wire via wiring fault or to ground via a 17 year old. 20 or 30mA leakage limit reduces the chance of death.

     

    Sometimes in IT we have 500mA trip RCDs and they are more to prevent fire then electrocution.  Electronics such as PC leak a tiny bit of current to earth because the switch mode power supply are fitted with a mains side filter to stop the high speed switching noise getting on the mains supply and vise versa (the noise is filtered to earth). This means that with 50-100 or so PC a 25mA RCDs tends to trip randomly.

     

     



  •  I have an ability to stop threads with my know-it-all answers

     

    Ah well i am off to wash my stuffed cat



  • @Helix said:

     I have an ability to stop threads with my know-it-all answers

    You didn't say much that I didn't already know, except for the part about Europe having RCDs (and, of course, that they are known as RCDs on that side of the Atlantic).  I figure the discussion has just reached its end.



  • @morbiuswilters said:

    @Helix said:

     I have an ability to stop threads with my know-it-all answers

    You didn't say much that I didn't already know, except for the part about Europe having RCDs (and, of course, that they are known as RCDs on that side of the Atlantic).  I figure the discussion has just reached its end.

     

     

    Quoted above, I didn't ask what you didn't know.

     



  • Garbage Person

    @Helix said:

    This means that with 50-100 or so PC a 25mA RCDs tends to trip randomly.
    ...

    Those must be some really tiny computers you're runnin' there. Most I've ever gotten onto one 20A circuit was 50 - without monitors. And that circuit's favorite activity was tripping, because in order to pull that off, each machine had to stay under 50w total draw - which is no small feat even with brand new modern equipment. Most CPUs pull down more than that, nevermind the rest of the system and the atrociously inefficient power supplies.



  • @Weng said:

    @Helix said:

    This means that with 50-100 or so PC a 25mA RCDs tends to trip randomly.
    ...

    Those must be some really tiny computers you're runnin' there. Most I've ever gotten onto one 20A circuit was 50 - without monitors. And that circuit's favorite activity was tripping, because in order to pull that off, each machine had to stay under 50w total draw - which is no small feat even with brand new modern equipment. Most CPUs pull down more than that, nevermind the rest of the system and the atrociously inefficient power supplies.

    That was my same reaction to the comment, except my peak of connected units is merely 20 (damn you P4s)

    That and I've seen PC power supplies cheap enough to trip the GFCI protection on one P/S, let alone 20....



  • @Weng said:

    @Helix said:

    This means that with 50-100 or so PC a 25mA RCDs tends to trip randomly.
    ...

    Those must be some really tiny computers you're runnin' there. Most I've ever gotten onto one 20A circuit was 50 - without monitors. And that circuit's favorite activity was tripping, because in order to pull that off, each machine had to stay under 50w total draw - which is no small feat even with brand new modern equipment. Most CPUs pull down more than that, nevermind the rest of the system and the atrociously inefficient power supplies.

     

     

    Shame that you did not really understand my post, 20A current breaker has nothing to do with the leakage current.  Additionaly since I am in the UK, labs of PCs have approx half the current draw then for a same setup in the US, so this is less of an issue.

     



  • ♿ (Parody)

    @Helix said:

    @Weng said:
    @Helix said:
    This means that with 50-100 or so PC a 25mA RCDs tends to trip randomly.
    ...

    Those must be some really tiny computers you're runnin' there. Most I've ever gotten onto one 20A circuit was 50 - without monitors. And that circuit's favorite activity was tripping, because in order to pull that off, each machine had to stay under 50w total draw - which is no small feat even with brand new modern equipment. Most CPUs pull down more than that, nevermind the rest of the system and the atrociously inefficient power supplies.

    Shame that you did not really understand my post, 20A current breaker has nothing to do with the leakage current.  Additionaly since I am in the UK, labs of PCs have approx half the current draw then for a same setup in the US, so this is less of an issue.

    Did you even read what he wrote?  He wasn't talking about current leakage.  He was amazed that you even got to the point of worrying about the RCD.  As was I. 

    Why does the same setup draw fewer amps in the UK?  Is this like having bigger gallons?




  • @Helix said:

    @Weng said:
    @Helix said:
    This means that with 50-100 or so PC a 25mA RCDs tends to trip randomly.
    Those must be some really tiny computers you're runnin' there. Most I've ever gotten onto one 20A circuit was 50 - without monitors. And that circuit's favorite activity was tripping, because in order to pull that off, each machine had to stay under 50w total draw - which is no small feat even with brand new modern equipment. Most CPUs pull down more than that, nevermind the rest of the system and the atrociously inefficient power supplies.
    Shame that you did not really understand my post, 20A current breaker has nothing to do with the leakage current.  Additionaly since I am in the UK, labs of PCs have approx half the current draw then for a same setup in the US, so this is less of an issue.
    Ok, fair enough that we both forgot the 120/60 <->240/50 conversion means about twice the amperage we're working with, that still means that you're putting more units per circuit than we are, on average. Let's ignore that for a moment, and just chalk it up to "on the internet everyone lives where you live" syndrome, since we've all done it in the past.

    Instead, I would prefer to discuss how Shame that you did not really understand my post, 20A current breaker has nothing to do with the leakage current has nothing to do with the efficiency of the power supplies of 50 connected PCs. With 50 fluctuating PCs on a circuit at one time, of course the leakage has some play in things, no?

    Tell you what, before I try to keep going and continue to mangle an already (apparently) convoluted thought process, why don't you go back and explain how it has nothing to do with the leakage, for those of us just getting our caffeine fix for the day. Hate to be a shame on us when it might just be a shame on forgetting units of conversion.



  • Doesn't matter on the efficiency of the power supply.

     Let me just google that for you:

     http://www.coilws.com/AppNotes/Line10.html

     

    Or if your maths is not very good and need a more hands-on approach:

     http://www.epanorama.net/documents/pc/ungrounded_pc.html

     



  • @Helix said:

    Doesn't matter on the efficiency of the power supply.
    A less efficient power supply is more often a cheap power supply, which is more often going to use lower quality components, which is going to more often cause line problems up-stream, which is exactly what an RCD|GFCI is supposed to stop from happening, especially a finicky one. QED, if you have 50 cheap power supplies on one circuit, the chances of one throwing the breaker are still the same, but the odds of n number of those working in concert are much higher.

    Hence, what the hell are you talking about man? I get that efficiency just means you have to pull more wattage, but once again, that's as important, as I've not seen many recent computers drawing less than 70A/PC. Eventually you just get past 20A and it doesn't matter about the RCD|GFCI being in the way.



  •  @drachenstern said:

    @Helix said:

    Doesn't matter on the efficiency of the power supply.
    A less efficient power supply is more often a cheap power supply, which is more often going to use lower quality components, which is going to more often cause line problems up-stream, which is exactly what an RCD|GFCI is supposed to stop from happening, especially a finicky one. QED, if you have 50 cheap power supplies on one circuit, the chances of one throwing the breaker are still the same, but the odds of n number of those working in concert are much higher.

    Hence, what the hell are you talking about man? I get that efficiency just means you have to pull more wattage, but once again, that's as important, as I've not seen many recent computers drawing less than 70A/PC. Eventually you just get past 20A and it doesn't matter about the RCD|GFCI being in the way.

     

    You still did not read my post, efficient, or high quality components still does not matter in this case.  I wish you would try to think.

     

    Right, they must have taught you ohmslaw, V=IR at school????????

     

    All computers have filters in the PSU, Safety standards now impose limits on leakage current. 5 nF is the largest value cap in the filter.

    Rearrange ohmslaw to find current: I=V/R

    Replace R with reactance of a cap gives: I=V / (1 / 2.Pi.C) 

    Therefore leakage current to earth I = V.2.Pi.C

    So you have half of US mains voltage and assume maximum legal sized cap of 5nF:

    I=55x2xPix(5x10^-9)

    I = 1.7mA

     

    Therefore at 110volt on a 30mA GFCI you can have 17power supplies.  However here i calculated the largest legal cap size of 5nF, vendors what to keep this low as it will cause customers problems, but if they make this too small it will also cause problems getting the PC through FCC tests.  A common value is 1nF, which is 1/5th so we can assume labs of 85 psu will cause issues.

    Note in UK (and EU) at 235v this leakage doubles as the voltage doubles.  

    You mention the current usage of that many PCs will be a limiting factor anyway, we sometimes run higher circuits (30 amp breaker).  Further note that this kind of CLC mains filter is common in all PSUs, even for a small brick PSU for a router which could be far less then 50w.

     

    BTW for your information you seem to treat efficiency as some kind of subjective quality.  This is very wrong in engineering and you should take the systematic approach the efficiency is  (Power out / Power in) * 100.  Yes, some efficiency is lost via the leakage current but this is less then 0.0005% in a 75watt PSU and gets smaller as the PSU power rating increases, the rest of the power loss (efficiency loss) is as heat.




  • @drachenstern said:

    ...... I've not seen many recent computers drawing less than 70A/PC.

     

     

    70 Amps per PC !!!!!!!

    I guess you have not used a computer since ENIAC

     

    hahaha ROFL




  •  @drachenstern said:

    @Helix said:

    Doesn't matter on the efficiency of the power supply.
    A less efficient power supply is more often a cheap power supply, which is more often going to use lower quality components, which is going to more often cause line problems up-stream, which is exactly what an RCD|GFCI is supposed to stop from happening, especially a finicky one. QED, if you have 50 cheap power supplies on one circuit, the chances of one throwing the breaker are still the same, but the odds of n number of those working in concert are much higher.

     Interesting that you call that a less efficient PSU, i guess that is a managers approch.  In the UK we call that a faulty PSU.




  • I accidentally stuck a screwdriver into the guts of a desktop power supply while I was touching the metal part of the screwdriver.  And the power supply was plugged in, and the computer was running.  It gave me a big shock and hurt a lot, and I yelled.  The tip of my finger was slightly numb for a couple of hours.

    Don't ask me why I did such a stupid thing.  At least I wasn't standing in water (or sitting in a bathtub).  And my other hand wasn't grounded.



  • @DWalker59 said:

    Too bad I wasn't standing in water (or sitting in a bathtub).  And my other hand should have been grounded.

     

    FTFY

    We need more Darwin Award entries, and people stupid enough to do this in the first place deserve a spot on the wall of shame.



  • @Weng said:

    @Helix said:

    This means that with 50-100 or so PC a 25mA RCDs tends to trip randomly.
    ...

    Those must be some really tiny computers you're runnin' there. Most I've ever gotten onto one 20A circuit was 50 - without monitors. And that circuit's favorite activity was tripping, because in order to pull that off, each machine had to stay under 50w total draw - which is no small feat even with brand new modern equipment. Most CPUs pull down more than that, nevermind the rest of the system and the atrociously inefficient power supplies.


    He's in the UK: twice the voltage means each computer draws half the amps, so you can have twice as many computers for the same thickness of wire. Additionally, UK buildings tend to be wired using ring topology rather than the radial topology the rest of the world uses, which (in theory) again doubles how many computers you can put on a single circuit. Topping it off, UK circuits tend to be fused for 30 amps rather than the 20 amps of a US circuit.


  • Garbage Person

    @Carnildo said:

    Topping it off, UK circuits tend to be fused for 30 amps rather than the 20 amps of a US circuit.
    WTF. No wonder you guys have plugs the same size I use to plug in goddamn ovens, air handling units and forklift chargers. The voltage difference I knew - but I assumed you were using circuits with half the amperage (if only because electricians make a hell of a lot more money that way) - and lets not even talk about the topology issue. I'm still trying to wrap my head around that.



  • @Helix said:

    @drachenstern said:

    ...... I've not seen many recent computers drawing less than 70A/PC.

     

     

    70 Amps per PC !!!!!!!

    I guess you have not used a computer since ENIAC

     

    hahaha ROFL

     

     Lol, yeah, I'd hate to see his electricity bill running a 16.8kW computer. Or a lab of them.



  •  (8.4kW if it's in the US)



  • @Kemp said:

    Lol, yeah, I'd hate to see his electricity bill running a 16.8kW computer. Or a lab of them.

    I love rezzing old threads just to show I can multiply 70 times 240 and 70 times 120!



  • @morbiuswilters said:

    @Kemp said:

    Lol, yeah, I'd hate to see his electricity bill running a 16.8kW computer. Or a lab of them.

    I love rezzing old threads just to show I can multiply 70 times 240 and 70 times 120!

    Show a little respect for the skills: he also divided both results by 1000, jerk.


  •  Must be caught in some kind of time distortion, I could have sworn it was first page.


Log in to reply