Read Latex

Tuesday, June 02, 2015

Light Unit Testing and the Barely Bloggable


From the annals of the barely bloggable comes today's adventure, the repair of the light that refused to be repaired. Context: I am a unit testing nut. I have a deeply held and mathematically-decidable aesthetic about testing things before they are assembled and then testing them after they are assembled to make sure that they work, the whole work, and nothing but the work so help me God. This little light of mine, I was going to let it shine. It also magnifies and I was going to let it do that too.


Since this is barely bloggable I will hit the high points.

My desktop magnifier light has been on the blink for some time. By "on the blink" if you haven't gotten the fluorescent "air quotes" pun I mean, it would work, then it wouldn't, then it would, in a decidably undecidable way.

Setting it up after a recent move relegated it permanently to the "then it wouldn't" state, so I set out, for the second time in as many years to drill down and get to the dark root of the problem.

There are three main concerns in such a light of mine:
1) the bulb
2) the starter
3) the ballast

I replaced the bulb, assuming that like me, it had "burned out".

I flicked the switch and then "it wouldn't", or rather, "it didn't".

I took the bulb back to the local hardware store. I thought that the floppy plastic thing that completes the donut bulb held fragile wires. I thought these wires had given up the ghost, meaning, that one of them had broken or could conceivably break during reassembly, replacement or testing.

Much to my surprise the local hardware guy fished out a "flourescent bulb and starter tester" and proceeded to demonstrate that both my bulb and ballast were good. No Home Depot and Lowe's that I have ever frequented has ever hauled out such a tester. This alone saved me two trips to the hardware store, possibly saving my life if I had been killed in traffic on either one, but I digress.

I brought home the now-proven bulb and starter. The hardware guy said and I quote, "If it's not those then its the ballast", refering to the small black autotransformer that parks on the wall outlet.

After reassembly, it didn't, so I disassembled the ballast, un-tape-ing it and spilling out its transformer contents as I had done the during the last spelunking adventure with this light. I proved to myself that the transformer and connections inside were good by measuring the resistance of the one coil at 22 Georg Simon Ohms and confirming that the wiring connected to the blades of the plug by setting the voltmeter to the, "sing to me if its okay" setting, a setting which all VOM users will instantly grok.

So now I have a real problem. My three main concerns, bulb, ballast and starter work.

Enter into the, "take apart things that weren't intended to be taken apart" cave of what-have-we-here.

Buried, and I do mean hidden in the starter base were wires just lying there, never soldered by some worker who had to pee, or go on their lunch break or got a call on their phone. So I soldered the wires and reassembled that which isn't to be broken into and behold, my little light shines. I am going to let it shine before you because it is an excellent example of things not working when they actually should work, because some tiny but critical detail is hiding an assumption on which all illumination hinges. Hallelujah.

No comments: