How to electrocute an elephant
Note: this article is about a passed event.
| How to electrocute an elephant |
|---|
|
| From 2011/04/09 10:00:00 to 2011/04/09 23:59:59
|
| Where: |
| Whitespace
Blekerijstraat 75, Gent, Belgium Loading map...
Costs 0 |
|
|
«Fixing the electricity, getting electrocuted»
We seem to be having an electricity problem, lets try to fix it!
Contents |
[edit] Bug report
The ground circuit is having some voltage on it. Touching the ground together with something grounded (e.g. the cable cage) gives a decent pulse through your arm. We already did a round of testing (unplugging things and measuring with a voltage meter from ground to body).
- Conductivity of human body is not very constant. Sweatiness, type of shoe's and other conductors you are touching, all influence the reading.
- For reference I measured these values at Void Warranties between GND and myself. (critique listed above applies)
- 4,6V AC
- 0,01V DC
[edit] We learned:
- our measuring technique is not always returning good (correct) results, resulting in false problem indications
- we need a proper setup for measuring, good ground, good connection to multimeter
- we should most probably also measure amps next to voltage with another multimeter
- the power is continuous: connecting the ground with a wire to e.g. the cages gives nice sparks, and it keeps on giving sparks, thus not a buildup of power, but continuous
- seems to have some punch at least, today we measured again 130 volts, and the previous time it went to 230 volts even, so this seems it could actually be hazardous. Watch out with the plugs everyone!!
[edit] Also some questions:
- Is our ground actually connected to "the ground" since it seems to be able to carry energy. Answer: not it was not - fixing this solved all problems
- Do we have a differential switch or some other device that should detect power going to ground?
- We do it's in the fusebox behind the mailboxes. (See pictures)
- It's a 300mA version, current regulation requires 30mA for new installations.
- a differential switch doesn't detect power going to the ground, it detects the difference between the L & N wires. So it should still work if flows somewhere else then GND in a faulty installation.
- Is the circuit box in our space ok? It shows some black area's from way back, and it uses old-fashioned plugs. Do we have backup plugs btw - they are use once only.
- Is the electrical cabling in our space ok? If it is, why do we have such problem locating the source of the power?
[edit] Attacking the problem:
We do this on a Saturday because:
- It should be daylight (hence weekend) to be able to turn of lights
- Shops should be open
- Our electrical specialists should be able to be present (hence not a weekend earlier)
- need a test-elephant ? Edison fries an elephant to prove his point, jan 4th 1903
[edit] References
Documents explaining the AREI (Algemeen Reglement op de Elektrische Installaties):
- Wikipedia article
- Huishoudelijke Elektrische Installaties (Vinçotte)
- praktische fiche Installatie (legrand)
- praktische fiche Beveiliging (legrand)
- samenvatting (O.C.B.)
Testing:
[edit] Result
- Mac Mini leaks voltage (around 18V, very low current)
- We added grounds to both roof top machines (they were not connected)
- Found out the space had no ground, so we connected to the ground: Our ground went outside, but was just a piece of wire not connected to the 'main' ground. We stripped the main ground and attached ours.
After connecting the ground, voltage measurement on ground dropped to absolute zero, and no more electricity shocks when touching both a ground and a cage, so we think it is solved.
Also, due to the fact that our ground was not connected to the ground of the rest of the building, this proves that whatever leakages were found, they were coming from our space and our space only.
Subpages
Images
| Has cost | 0 + |
| Has description | Fixing the electricity, getting electrocuted |
| Has end date | 9 April 2011 23:59:59 + |
| Has event type | |
| Has location | Whitespace + |
| Has start date | 9 April 2011 10:00:00 + |
