Intel's Ronler Acres Plant

Silicon Forest

Tuesday, January 21, 2003

Mechanical Failures

Sump Pumps

We have two sump pumps in the crawl space under our basement floor. One was put in shortly after we built the house. The other was put in when we finished the basement. They both sit in perforated five gallon buckets that have been sunk into gravel lined holes about two feet deep and two feet in diameter. When we finished the basement, a perforated drain pipe was installed across the uphill side of the basement. It drains into the newer sump. Both pumps are controlled by float activated switches. When the water reaches a certain depth, the pump turns on and runs until the sump is empty, or nearly so. The switch on the first pump is inside a float that is attached to the pump by a power cord. It is free to flop around. The newer pump has a float that runs up and down a guide attached to the side of the pump. The five gallon bucket it sits in is just barely big enough to contain it, and sometimes the float would become stuck. When this happens the pump runs continuously, it never shuts off. I would periodically go down and check it, but it wasn't enough. After a couple of years of this abuse, it finally failed. I think it cooked itself. I pulled it out thinking I might be able to fix it. However when I took the top off the motor, I discovered that the motor housing was full of oil. Even if I could have fixed it, I wasn't prepared to deal with oil. So I put the lid back, set it in the garage, and went to home depot and bought a smaller pump with a floppy type of float switch. So far it seems to be working okay. And what I am going to do with old pump? It will probably sit in the garage till I retire, and then I'll throw it out. I called a couple of dealers about getting it fixed, but no one was interested. All they wanted to do was sell me a new one.

Radio Buttons

Saturday Ross and I fixed the radio in Mom's (Anne's) van. The volume on this radio is controlled by push buttons. One button increases the volume, another button turns it down. Every so often the down button gets stuck, and you can't turn down the volume anymore. You can still turn it up, but you can't turn it down. As this is the third or forth time this has happened, I have the repair procedure down to a science. Basically it involves removing the radio from the dash, taking the faceplate off and take it apart. The push button action is implemented with a flexible rubber sheet with bumps in it for each button. The problem is the bump for the down volume control gets stuck. Taking the faceplate apart allows it to pop back up. So now it's fixed, and it's just a matter of reassembling the faceplate and reinstalling the radio.

Taking the radio out is normally a onerous chore. On the Ford it is relatively easy, however, you do need a special tool. The special tool is a couple of pieces of wire, about the thickness of old wire coat hangers, bent in the shape of a "U", about three inches tall and an inch or two across. The ends of the "U" are something in the shape of crochet hooks. The ends of the U's go through holes in the front of the radio and disengage latches that hold it place. Theoretically speaking, you simple slide the special tools into the holes until the latches disengage and then pull to remove the radio. In practice, there is a good deal of pushing, pulling, yanking and muttering, but it eventually works, and it is much easier than having to take it out from underneath and behind, which used to be standard practice.

Tiny Screws

Sunday Ross and went shooting with a friend of mine in the Coast Range, about thirty miles west of our house. We found a gravel pit to shoot in, actually a hillside. A few years ago I bought a high-powered rifle with a scope in the expectation of going deer hunting. This is first year I have even shot it. It is very loud. It is at least twice as powerful as any of my other weapons. When I shoot it, I can fell the shock wave it produces in my entire body. Anyway, before you attempt to use a weapon like this, it is a good idea to "sight it in", i.e. to verify that the sights correspond with the trajectory of the bullet. So my friend and I tried a couple of shots each (Ross didn't want to have anything to do with it. The 22 caliber rifle is more to his liking) at some targets set up about 75 yards away, with inconclusive results. So we set up a larger, closer target, a sheet of particle board about four feet square about 25 yards away. We marked a spot
on the board and I fired about four shots at it. When we went to look at the board, it was apparently unmarked! But then we looked at the back side of board, and it was obvious that the board had been hit. There were large, ragged bunches of splinters sticking out. But even when we knew where the bullets hit, it was difficult to see where they had struck on the front side. And these were relatively large bullets, 30 caliber, almost 5/16" in diameter.

Anyway the four hits on the board were in a nearly vertical line, but they were spread out vertically over nearly two feet! What could cause this horrible inaccuracy? And then we look at the gun and notice that one of the scope mounts had come detached. The scope mount had been secured to the barrel of the gun with tiny flat head, Allen screws. Allen screws have a hex shaped recess to accommodate the driver. The cone shape of the screw head combined with the hex shaped recess combined to reduce the attachment point of the head to screw body to almost nothing, which made the screw very weak, and given the smallest task, it failed.

Ovenized Oscillators

At work I am working on a new radio transmitter. This radio uses a 10 MHz (ten megahertz, or ten million cycles per second) crystal controlled oscillator. Crystals are used in oscillators because they are very stable. If a crystal is generating a 10 MHz signal today, then it will be generating a 10 MHz signal ten years from now. However, the frequency generated is temperature dependent. If the temperature goes up, the frequency goes up. Not much, but for our purposes even the smallest deviation is unacceptable. So we use ovenized crystals. This means these oscillators contain a heating element that raises the temperature to the highest temperature that the crystal is expected to operate at, in this case about 140 degrees Fahrenheit (most Microsoft fonts don't have a degree symbol). I am working with on old prototype board and the crystal was attached to the board with double sided mounting tape (the kind used for hanging pictures on walls). This tape is made of foam to accommodate surface imperfections. This board had been used so much that the heated crystal had cooked the tape, which caused it to fail, which left the crystal dangling by its' wires, one of which broke. No big deal, we glued it back down using super glue and soldered new wires on.

Super Glue

A few weeks ago I tried to repair a die cast model tractor using super glue. It didn't stick the first time, so I filed the mating surfaces flat and tried again. Still no luck. Ended up using epoxy. Now the part I glued on has itself come apart, so I need more epoxy. Super glue is funny stuff. It will definitely stick to skin and to most other surfaces, but occasionally I will come across something that it won't stick to. Die cast metal for instance. A friend of mine speculated that it was because the metal is porous.