Saturday, June 15, 2024

Are people with Asperger’s excellent in data mining because of their pattern recognition?

 Are people with Asperger’s excellent in data mining because of their pattern recognition?

I can’t speak for you or anyone else, only myself.

I think that characteristics that seem to be part of my Asperger’s have made me very good at troubleshooting and repairing electronics and electromechanical devices.

You must have seen the Tshirt that says:

There are two kinds of people in the world

  1. Those who can extrapolate from incomplete data

I’m definitely in category 1. I’ve worked on a lot of different kinds of electronics, and I’ve noticed that many techs will overly-generalize symptoms. It seems to follow the 90/10 rule. 90% of the jobs, they’ll be right, but then spend 90% of their time and parts on 10% of the repairs trying to figure out what is actually wrong after replacing the wrong parts. They often “shotgun” the repair.

When someone “shotguns” a repair, it means they just keeps replacing parts willy-nilly in an ever expanding pattern around the affected circuits. It is a waste of time and parts.

Whereas I’ll see slight differences and treat those 10% repairs differently, spending a little more time looking for the real cause and a LOT less time repairing.

Often, all I have to go on is a mental visual model (I have a vivid imagination) and subtle clues. In the repair world, you can’t build up a vast array of data and then process it. You must pay attention to little squeaks, pops, even smells, make up some hypothesis about what is going wrong, make predictions and test them, and see if it responds how it should if you are correct.

I will also consider what other parts may have been stressed when the obviously defective parts went bad, or what else may have caused those parts to go bad but not be obviously defective. As a result, it is very unusual for me to have a repair come back.

Remember when email spam really started taking off? I got really good at determining which emails were spam just from the subject line. There were no email filters at that time, or they were primitive and prone to failure. I can’t tell you exactly what I’d look for. Naturally many spam subject lines are very obvious, but not always, and I was very good at picking those out.

I worked for decades without a degree, but HR departments won’t even schedule an interview without one. So when I had an opportunity to get a 2yr EET degree tuition-free, I jumped at it.

Just a couple of weeks after I started, one of the other students brought in the controller for an in-floor heating system he and his father had installed. It had apparently burned up. He said this was the second time this had happened.

He brought it in to see if any of the electronics instructors could figure out why this had happened twice. They all clucked over it like brooding hens, and after a couple of hours of consulting with different instructors, they announced “it has a short in it.”

I was floored. That is what people who know nothing about electronics (no offense) say is wrong when their remote control buttons don’t work correctly, or the TV picture is rolling.

I got a look at it, finally. I asked this gentleman a series of questions and determined:

  1. The heating element has a positive temperature co-efficient and so even if you connect it without a thermostat, it will only get to around 90 to 100F in the room running full out.
  2. Both times this happened, the room temperature seemed normal, not warm.
  3. He and his father save and reuse old wire nuts.

This tells me first of all that the controller did NOT short, because the room would have become very warm, and under normal current limited by the heating element, the controller should not have gotten more than slightly warm.

Looking at the controller, wiring, and remains of the wire nuts, I determined that the heat and burning started at the wire nuts and proceeded down the wire. All of the burning and melting was on the outside of the plastic box, none came from inside. The controller board looked perfect.

Clearly, the old, previously used wire nuts had failed. A little heat, the plastic gets soft. The metal loses its spring and doesn’t hold the wires as tight, causing more resistance. This causes more heat. As the copper gets hot, it oxidizes, further increasing resistance and heat.

The insulation begins melting. The excess wire had just been left in a partly coiled heap on top of the control box. Eventually, two wires contacted enough to really get the insulation burning. Only then did the breaker finally blow, because up until this point, it had been drawing normal amounts of current.

This took me less than 5 minutes to suss out. And most of that was asking questions.

A year later, my instructor brought in a winch controller with burned wires and a button stuck down. See the little box with two buttons? Thanks to Zoro dot com:

It had two red and two black wires. He said he’d hooked it up to a car battery, hit one button and it let out cable, hit the other button and nothing happened. So he pressed and held the button, then the wires started smoking and the button stayed down when he let go. At that point he yanked the wires off the battery.

He offered a day off to anyone or any group who could tell him what went wrong. I said that I would not compete, as I felt it would be unfair of me and because I figured I’d give it away just asking questions and looking at it. So I did not even look at the winch controller.

After nearly an hour of students in groups working on this, he had everyone write down what they thought went wrong and he went through them.

I was gobsmacked - he chose two that simply said “It has a short in it”!

I couldn’t let this pass. I said “Now do you want to know what actually happened?”

I drew two DPDT symbols on the whiteboard with appropriate wires to show how one switch connects the motor B+ to M+ and B- to M- for Forward, and B+ to M- and B- to M+ for Reverse. And that what the instructor had done was mix up the wires so he had one motor wire on the battery, and one battery wire on the motor. Then when pressed one way, it would connect the motor to the battery but would actually run in the opposite direction to the marking on the button, and the other button would short across the battery. This would melt the plastic inside the housing and so the button stuck down.

I had still not looked at the winch controller. I said “If you go look at it, you’ll find that there will be some small way to tell which are for the battery, and which are for the motor, and that one battery wire and one motor wire will be burnt.”

That is exactly what they found.

No comments:

Post a Comment