Thread Tools
Jan 06, 2010, 08:27 PM
Tree Antagonist
BobOD's Avatar
Here's a problem I see that I'm not sure has been brought up. If something were placed across, or fell across, the two cases, there is a possability of conducting current (at 12v) and generating sparks and heat....maybe a fire. It could be a strand of wire, a carbon rod...who knows. This isn't very much different than a car battery sitting there but it's something to be aware of. Sometimes years later, people forget about the danger.

And, btw, gfi protection works by disconnecting the power if there is current flow detected in the ground prong of the recepticle. If you disconnect the case from this ground point, and the 3 - 400 volts dc typically found in switching power supplies shorts to this case, the GFI will NOT trip. It might with the capacitor (AC grounded) setup shown above....but it might not too.

BTW, 120v is childsplay as suggested above. DC is far more dangerous...has a better chance of stopping the heart.
Sign up now
to remove ads between posts
Jan 06, 2010, 08:39 PM
Tree Antagonist
BobOD's Avatar
By the way, my suggestion, assumming no liability of course, is to get yourself a used PC case. Mount the 0-12v one grounded and the second (or third...) on insulated standoffs. Then everything is inside a grounded case.
Jan 06, 2010, 09:33 PM
ancora imparo
jj604's Avatar
Thread OP
My understanding is that the original gfi/ELCB's worked this way but that more modern ones work by detecting the imbalance between the Active and Neutral lines. If the current isn't the same in each it must be going somewhere else and they trip. Could be wrong about that though.

Some of us live where 240V AC is the norm and it seems to stop hearts pretty effectively. It's the current that kills you of course so it will depend greatly on the source impedance, how you are connected, the current path and your body resistance at the time but I always thought anything above 60V DC or AC needed to be treated as potentially fatal unless it was intrinsically current limited.
Quote:
Originally Posted by BobOD
And, btw, gfi protection works by disconnecting the power if there is current flow detected in the ground prong of the recepticle. If you disconnect the case from this ground point, and the 3 - 400 volts dc typically found in switching power supplies shorts to this case, the GFI will NOT trip. It might with the capacitor (AC grounded) setup shown above....but it might not too.

BTW, 120v is childsplay as suggested above. DC is far more dangerous...has a better chance of stopping the heart.
Jan 06, 2010, 09:37 PM
ancora imparo
jj604's Avatar
Thread OP
Interesting idea. What about my third point - the Charger case may be connected to the -ve lead and hence to the fault?
Quote:
Originally Posted by BobOD
By the way, my suggestion, assumming no liability of course, is to get yourself a used PC case. Mount the 0-12v one grounded and the second (or third...) on insulated standoffs. Then everything is inside a grounded case.
Jan 06, 2010, 10:30 PM
Tree Antagonist
BobOD's Avatar
Quote:
Originally Posted by jj604
Interesting idea. What about my third point - the Charger case may be connected to the -ve lead and hence to the fault?
Well, the -ve of of the ungrounded supply would not be accessible outside the case. The +ve of the ungrounded supply is but that is no different than than the original supply and is unavoidable. I don't design power supplies or anything but I would guess that they are designed such that the least likely failure to occur would be one that puts internal, high voltages on +ve. In other words, designed so if a failure in insulation were to occur, it'll go to -ve ground instead. Putting both inside a grounded case maintains the safety measures quite completely as best I can tell.

Regarding the GFI, a switching supply isolates input from output. There will be no imbalance for the gfi protection to sense while you cook. As far as it would be concerned, you would be the load.
Jan 07, 2010, 01:14 AM
ancora imparo
jj604's Avatar
Thread OP
Well, no to both of those I think.

1) The obvious failure that concerns me is if the Active (Line) Mains input connects to the ungrounded case. It happens. Possibly through a component failure or broken wire. In that situation the case is directly connected to the active Mains. But it is not grounded so the mains fuse will not blow. Any connection to the case which provides a path to ground (such as you) could be lethal. I would love to be convinced that it can't happen because of other current paths through the low voltage side but I'm not yet. My Hyperion Duo does not appear to have the case connected to either input or output but I wouldn't want to assume that for every charger. If the charger case is connected to -ve OUT from the supply, then it becomes live also.

2) "There will be no imbalance for the gfi protection to sense while you cook. As far as it would be concerned, you would be the load"

Well that's a trivial case and I agree that if you put yourself across the active and neutral then there is nothing that will protect you. But the failure mode we are protecting against is between active and ground. Then there IS an imbalance and that is exactly what ELCB type devices are designed to detect.

"a switching supply isolates input from output." So what? It's connection of the mains INPUT to case that is the danger.

This is turning out to be an interesting discussion. Thanks to everyone. Next opinion???

John

Quote:
Originally Posted by BobOD
Well, the -ve of of the ungrounded supply would not be accessible outside the case. The +ve of the ungrounded supply is but that is no different than than the original supply and is unavoidable. I don't design power supplies or anything but I would guess that they are designed such that the least likely failure to occur would be one that puts internal, high voltages on +ve. In other words, designed so if a failure in insulation were to occur, it'll go to -ve ground instead. Putting both inside a grounded case maintains the safety measures quite completely as best I can tell.

Regarding the GFI, a switching supply isolates input from output. There will be no imbalance for the gfi protection to sense while you cook. As far as it would be concerned, you would be the load.
Jan 07, 2010, 01:15 AM
Registered User
On some chargers the +ve is connected to the case, on others the -ve. Switched-mode PS isolate the input from the output to a degree, but if the insulation on the coil breaks down, a common failure mode, then the isolation can be violated.

I think Bob's suggestion of enclosing the pair of PS in a case which is itself grounded, such that only the main -ve and the main +ve are exposed, should bring you back to a safe setup so long as the floating PS is isolated from the case as Bob said.
Jan 07, 2010, 01:20 AM
ancora imparo
jj604's Avatar
Thread OP
This sounds like a key part of a safe solution - I agree. Whether it is enough I'm not yet convinced.

In mathematical proof terms it is necessary but is it sufficient?
Quote:
Originally Posted by kgfly
I think Bob's suggestion of enclosing the pair of PS in a case which is itself grounded, such that only the main -ve and the main +ve are exposed, should bring you back to a safe setup so long as the floating PS is isolated from the case as Bob said.
Jan 07, 2010, 02:10 AM
Registered User
When I had a quick look my first impression was an isolation transformer on the floating power supply might be benificial. Will have to think about this a little more. Something is not making me happy about what I again see posted. So far what it is it has not totally jelled. I think it has to do with a transformer in the floating supply becoming defective and going to ground. There is no longer a ground there so 120 volts is going to appear. The isolation transformer on the floating supply may reduce any risk of that occurance being dangerous. Again I will have to think about this. All I think is the event of that likelyhood is small but must be protected from. A ground fault device may be able to be wired into the floating unit instead.
Last edited by barry wilson; Jan 07, 2010 at 02:23 AM.
Jan 07, 2010, 03:20 AM
Registered User
Stupid idea of the day:
How about a zener diode/SCR crowbar between ungrounded supply and ground, set at 15-20 volts or so? In normal use it would be off and isolation between supplies is maintained, but when high voltage leaks into ungrounded case, the crowbar is triggered into conduction, creating a short and tripping the elcb. Of course it would also short the output of the supply and possibly take damage itself, but that would be only collateral damage necessary to protect the more important pieces of the system (aka. you).
Jan 07, 2010, 11:14 AM
Tree Antagonist
BobOD's Avatar
Quote:
Originally Posted by jj604
If the charger case is connected to -ve OUT from the supply, then it becomes live also.
No, not really. If the -ve of the charger is connected to its case and to the -ve of the first grounded power supply, then the charger case is grounded and will not be live.


Quote:
Originally Posted by jj604

2) "There will be no imbalance for the gfi protection to sense while you cook. As far as it would be concerned, you would be the load"

Well that's a trivial case and I agree that if you put yourself across the active and neutral then there is nothing that will protect you. But the failure mode we are protecting against is between active and ground. Then there IS an imbalance and that is exactly what ELCB type devices are designed to detect.
You are assuming that the ground source for any particular situation is the GFI protected one. There are other ground sources...the ground is one of them and it is not GFI protected.


Quote:
Originally Posted by jj604
"a switching supply isolates input from output." So what? It's connection of the mains INPUT to case that is the danger.
No it's not...well, it's not the only danger anyway...nor even the major concern. Switching power supplies often generate high voltage DC as a first step. They do this as a way to regulate high power without having to deal with such high current. I have one here that runs a 375vDC buss and stores a good amount of this energy BTW. This is so much more dangerous than even 220vAC.It is clearly designed to send any faults into the ground circuit. The grounded case is the most relied upon safety barrier. Eliminating it avoids this safety measure.

Putting both inside a grounded case, in my mind...not an expert by any means...maintains this safety measure quite well. Probably not perfect but a whole lot better than removing a primary safety measure and leaving it exposed. Even forgetting about an internal failure, I go back to my first mention. You have, by design, a large metal case at 12v with a lot of current capacity. It may not electrocute you but it is a fire hazard. For someone who is conscious of it, probably be OK as long as we pay attention all the time.
Jan 07, 2010, 11:18 AM
Tree Antagonist
BobOD's Avatar
Quote:
Originally Posted by kgfly
On some chargers the +ve is connected to the case,
I really wish they would not do that.

Do they still make pos grounded autos these days. That's probably what that's all about.
Jan 07, 2010, 12:34 PM
Adam
xStatiCa's Avatar
Quote:
Originally Posted by BobOD
You have, by design, a large metal case at 12v with a lot of current capacity. It may not electrocute you but it is a fire hazard. For someone who is conscious of it, probably be OK as long as we pay attention all the time.
I am confused by this. I am no electronics expert but I do remember from electronics class many moons ago that current is the killer not voltage. I have heard 1 amp can be lethel but on the other hand extremely hight voltage at very low amperage(a few millivolts for example) would not. Vandergraf generator experiments with hair standing on your head is a good example. Can you clarify what you mean?
Jan 07, 2010, 01:08 PM
7000mw of raw power!
rich smith's Avatar
Current is the killer but it requires voltage to get through the skin. Static is high voltage (i.e. half million) but harmless because low current. 12 high current DC (car battery) is pretty harmless too because not enough voltage to penetrate. 110/220vac is the most dangerous and accounts for most electrocutions because it has both.

Note that you can never be hurt unless the potential is across your chest like one hand to the other or head to foot. Getting hit finger to elbow or finger to palm is painful but can never electrocute.
Jan 07, 2010, 01:44 PM
Tree Antagonist
BobOD's Avatar
Quote:
Originally Posted by xStatiCa
I am confused by this. I am no electronics expert but I do remember from electronics class many moons ago that current is the killer not voltage. I have heard 1 amp can be lethel but on the other hand extremely hight voltage at very low amperage(a few millivolts for example) would not. Vandergraf generator experiments with hair standing on your head is a good example. Can you clarify what you mean?
I never understood the notion that "current is what kills". Power is what does the damage and power is a funcion of voltage AND current.

Anyway,
Your body is a resistor and has a rather high resistance.
Current (I) = Voltage (v) / Resistance (Ohms)

If the voltage is only 12v, then the current will be 12 / 200000 ohms (just to pick a number) = 0.06 milliamps. Not enough to harm.
At 300v, it's 300 / 200000 = 1.5 milliamps. Enough to do harm.


Quick Reply
Message:

Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
For Sale Used PC Power supplies converted for 12V DC chargers ddierking Aircraft - Electric - Batteries & Chargers (FS/W) 3 Jan 13, 2010 09:14 AM
Question Power supplies for lipo chargers? steelfinger Batteries and Chargers 4 Sep 04, 2009 04:28 AM
For Sale 12V Power supplies 30A $40 and 42A $50 for your chargers feathermerchant Aircraft - Electric - Batteries & Chargers (FS/W) 1 Sep 03, 2009 08:07 PM
Discussion Using PC power supplies for chargers? shurcooL Batteries and Chargers 9 Jul 25, 2008 01:44 PM