Monday, October 26, 2009
Currently at work, I've been assigned to a project where I have to design a capbank (as in a collection of capacitors housed in one box). Now for those of you who are not familiar with capacitors, they are more or less like a battery. Actually, a lot of them do shape like a battery, but unlike any traditional batteries from the grocery stores, if you touch the terminals, the chances are you will get a good jolt depending on the capacitance. The general rule is you don't touch them unless you have touch the terminals with a screw driver (with an insulated handle of course) first to discharge it.
As my teammates and I are about done with designing this thing, of course safety people would have to review it before it gets built. In my previous experience, these meetings never go well. They have always caused me nothing but a great deal of frustrations and anguish. Now don't get me wrong, I'm an advocate for safety but I'm just not sure if these meetings does the job. I usually call them the "What If" meetings because that is pretty much what the safety guys do; proposing a list of "what if" scenarios with an increasing order of ridiculousness. Here is usually how they go.
Safety Guy (SG): What if the operator forgets to discharge the capacitors?
Design team (DT): We've designed it to discharge automatically.
SG: Then what if the discharge is not quick enough?
DT: We can design it to make it real labor intensive to dissemble so it will take longer.
SG: How about what if it blows up. . . spontaneously?
Me in my head: The box is freakn' Aluminum!!
SG: What if the operator is really good at unscrewing screws and pulling things apart?
Me in my head: Then he deserves to get fried!? Anyone would know better than to treat lightly of MULTIPLE capacitors that are each a size of a SODA CAN!
SG: What if some one forgets to turn off everything?
Me in my head: What if the operator licks the terminals?! Come on?!?
My dilemma is where exactly should this "what if" game stop? The whole idea is to protect people from doing dumb things and hurting themselves in the process, but is there a measurement to that? To what level of idiots am I suppose to protect? Is there a measurement of idiocy? I propose that there should be one. In my everyday design tasks, we often deal with the term "6 sigma". A sigma is a standard deviation. So what 6 sigma really means is whatever we design, the final distribution curve of this thing should fit within a 6 standard deviation window or roughly 99.99966% of the time. Or in another word, whatever it is that I'm designing, it better work 99.99966% of the time. Now take that same paradigm and apply it to measuring idiots. Given a normal distribution (or bell curve). The middle bulk part would represent people with average intelligence. People to the right and left edge of the curve would represent the extreme genius and idiots, respectively. With such a curve, I think the safety guy will have a better idea of where should the "what if" questions stop because now there is a demarcation of idiots that can be saved (near the left edge of the curve but within the 6 sigma window) and those that may just jump into a swimming pool with a live capbank, what I also call the "6 sigma idiots" or people outside the left side of the bell curve.
***Note: Knock on wood, knock on wood, rabbit's foot . . . that I won't become that 6 sigma idiot and fry myself.