You
just brought up channel sixteen on the fader, wheel or keypad and nothing
happened. Oh dear, here we go again! Is it the DBO switch, the scene
master, the grand master, the softpatch table, the desk, the control
cable, the DMX splitter, the missing DMX terminator, the dimmer rack,
the dimmer fuse or circuit-breaker, the mains to the dimmer, the supply
to the local distribution board, the patch panel, the multicore loom,
the outlet socket, the splitter board, the extension cable, the tail
on the luminaire, the lampholder, the lamp or the safety cut-out switch.
Are the shutters or barndoors open on the luminaire? Is the desk in
the right mode? Is the DMX cable in the right socket? Has the dimmer
rack got the right DMX start address? Is the rigger's control overriding
the desk? Has the luminaire even been rigged yet, has it been delivered
by the hire company or maybe it's just pointing straight into a black
drape? No doubt you can probably add at least a dozen similar questions
the list.
What
are we to do? We could start at one end of the system, let's say at
the luminaire and work our way back to the dimmers or the control desk.
If the problem is a blown lamp or a configuration switch set to expect
MIDI Show Control data from a Microbee computer then we'll hit pay dirt
straight away. On the other hand, if the desk has been switched to AMX
192 output, or channel sixteen has been erroneously allocated to an
inhibit group, it's going to take a long time to discover. If faults
occur randomly within our rig, then we will on average test half the
system in order to locate any fault, but statistical averages do not
take Murphy's law into account. Murphy correctly predicts: Any search
for a fault will always start from that part of the system which will
maximise the time required to locate it. Don't get too smart and
decide to start from the other end of the system instead, because Murphy's
prediction has already taken that it into account.
We
could think back to the last time we encountered a similar fault and
start testing from there, but of course Murphy has a prediction that
covers those bold enough to make assumptions: Apparently identical
faults invariably have different causes.
Fault
finding really is as simple as ABC:
Assume
nothing
Believe
no-one
There
is an approach to locating faults which is generally more efficient
than the end to end or linear search. It's the process known as the
binary search, which involves dividing a system up in to successively
smaller sub-systems until the fault is isolated. At each step in a binary
search we eliminate half the system from our suspicions until what we
have left must be the cause of the fault, even if we aren't really
prepared to believe what we've found.
A
key element to applying this divide and conquer technique to fault finding
is the process of substitution: replacing the sub-system under suspicion
with another which is known to be working correctly. If for example
you wish to eliminate the dimmer rack from your suspicions, there are
two steps to follow. First, plug a known working load (which must be
a resistive load like a lamp, and of least 100watts if your dimmers
are not digitally fired) in to the dimmer in question. If the known
load works then you have exonerated the dimmer, if not proceed. Next,
plug the load from the faulty circuit in to a known working dimmer.
If the load now works, you have exonerated the load. In the case of
our hypothetical problem with channel sixteen the approach would be
to verify that an adjacent dimmer and its load are working, then swap
the loads between channel sixteen and the adjacent dimmer and note the
result. It is of course imperative to swap the loads back again after
the test to prevent an occurrence of that paranormal phenomenon, the
mysterious moving fault.
If
both the load and the dimmer test as OK then you have a fault which
is possibly best dealt with by taking up Permaculture, macrame or wool-spinning,
as you have completely missed the bus with electrical technology.
The
most powerful piece of test equipment required for rapid and successful
fault finding is an inquiring, methodical and above all, sceptical mind.
Very few if any tools are required. If the DMX data lights are glowing
on three racks, but the first one in the chain isn't working then you
don't need a $1,200 DMX tester, or even one of those natty little tester
plugs to verify the DMX stream. There's a pretty fair bet that the power
to the racks is on as well, so you can also leave your mains-rated multimeter
with its finger guards and shrouded connectors, resting in its high
impact holster. If you can't believe the evidence of your eyes then
don't: assume nothing, believe no-one, check everything.
