|
|
|
| We offer a Koi Rescue Service. |
Electrical Safety: This article by Dr. Johnson may be a little technical for some members, but your safety and that of you koi may be at stake. GFCI - the inside scoop from an Engineer - ...with thanks to Roark for the 'heads up' Also, how fish get electrocuted and why it does what it does. Let's tune into Roark. Knowledge is power, so here's the inside scoop on GFCI's: The current generation of garden-variety GFCI's do not have anything to do with the ground wire. No kidding. The ground goes into the unit, serves as a zero-potential reference to the internal monitoring circuitry, and then passes right through it *without* being monitored. (Exception: Medical grade units...explained later) Current-generation GFCI's employ two very sensitive bridge amplifiers. These amps each monitor a leg of the AC circuit. Amp #1 monitors the "hot" wire, and amp #2 monitors the neutral. The ground wire goes straight-through the unit 99% of the time (see below for exceptions on "medical" grade units). The outputs of amp 1 and amp 2 are directly proportional to the current flowing through each of these legs. Their outputs are routed to an analog comparator. As long as both amps show the same output level, the comparator's output is off. The comparator is connected to either an output driver stage (which is either a disconnect relay or a slow semiconductor like a triac). To trip the GFCI, all that is required is a very slight current imbalance between the hot and neutral wires. This difference gets amplified and the difference causes the comparator to saturate... which "trips" the output stage thereby disconnecting the power. This all happens in just a couple of milliseconds. Enough to zorch a fish in close proximity, but not enough to kill a healthy human. Depending on the intended use of the GFCI, some employ some smarts to prevent false tripping. Some equipment has internal circuitry (ie capacitors in the power supply) which will shunt a small amount of current from hot to ground when it is initially powered-up. This will cause a "trip" on some units unless the unit is smart enough to "cycle count" (ie, look for the trip condition on two successive cycles of the AC line. This trades raw response time for convenience). Now, for a variation on this basic GFCI theme we have the ultra-high-end medical grade" units. These employ a very high gain amp (to the tune of over a million to one) which monitors current on the ground lead. (The other 2 amps remain exactly as described) If *any* current whatever is flowing in the round lead, it trips the unit. (This type of unit is generally configurable as to its behaviour. There are times such as when a cautery is in use, where there will simply *be* a current flowing in the ground pin even though there is no actual "imbalance". This type of leakage causes false trips, and the last thing you want if you're a surgeon is to hit the juice when you've got a field full of bleeders... and the entire OR goes dark!). Medical units also have what is called an "output
crowbar". This is a nice, hefty and very fast semiconductor clamped directly across
the hot and neutral leads. When an imbalance occurs, this semiconductor saturates and
provides a So why the crowbar? Sheer response time as well as a certain degree of redundancy. When the output is crowbarred" (the name came from some wag who equated this action to dropping a metal crowbar across the output leads) no current can pass *out* of the GFCI.... although quite a lot is flowing *inside* it. Electricity takes the shortest path to ground... and the crowbar definitely provides this path in spades. This direct short is present for several milliseconds... until the output stage (which is much slower than that last semiconductor) can disconnect itself. This buys you time. The human body can take an utterly unbelieveable amount of both voltage and current for a few microseconds. (Basically, electrocution affects the muscles first. Muscles just can't physically get pissed-off in such a short time. Yeah... you'll be spastic in the morning, but 50,000 volts for a couple of microseconds will *not* kill you. Not enough time. Admittedly, neuro effects are another matter... You're alive, you're breathing fine, but you can't move your legs for a few weeks... hehehe) A Related Roark Tip: This brings-up another point which I'm about
to go harp-on for my web page. Folks with submersible pumps can dramatically reduce the
possibility of shocking fish in the event of a pump electrical failure by placing their
pump within a basket of aluminum window screen. I use nylon wire ties to secure the screen
into a "baggie which fits very loosely around the pump. This www.koivet.com Dr Johnson |
|