Sunday, March 15, 2009

Useful Things I Learned This Weekend

1. World of Goo is a very cool game.

2. An hour playing with World of Goo will give you a better intuitive feel for stress analysis and the effects of resonance in a structure than ten hours reading a textbook will.

3. The ramshackle buildings you wind up with in World of Goo remind me of a lot of some source code I've seen.

4. The Wii Internet Channel's a pretty good browser (being Opera and all), but it doesn't have the horsepower to play an NPR podcast.

5. An electric crock pot turned on "high" can crack a Corian countertop.

6. You can make an effective trivet using a 12" piece of ceramic tile and five of those sticky-back felt tabs you put on chair legs. You really only need four tabs, at the corners, but the fifth in the middle is belt-and-suspenders.

7. If important work e-mail's going to arrive, the most likely time is right after you've mixed the catalyst with the self-etching primer and started spraying.

8. Just a hunch, but I'm guessing the "answer e-mail while wearing latex gloves half covered with catalyzed primer" scenario didn't figure prominently in Blackberry usability testing.

Sunday, March 08, 2009

Sealing the trailing edge

The stripper arrived about a week ago. I put the traling edge wedge from the rudder in the portable spray booth, now pressed into service as a fume hood. (As it turns out, the the stripper doesn't seem to give off a lot of fumes, but I was taking no chances with stuff capable of removing epoxy.) I didn't need to strip the rudder's skins because the epoxy adhered much better to the trailing edge than it did to the skins.

Stripping was a matter of brushing the epoxy on, using a disposable paintbrush, waiting a while for it to bubble or show signs of softening, and then cleaning it off. The stripper dissolves in water, but it took a while to find a suitable scraper. I eventually settled on a green Scotchbrite pad, which did a pretty good job of taking off epoxy without itself dissolving. I went through three coats of stripper. The epoxy came off just fine--in fact, it came off pretty well with the first coat--but I'd also primed the wedge. To make sure the fuel tank sealant adhered well, I also wanted to strip as much of the primer as possible, so the sealant would be gripping aluminum on both sides of the bond. Except for a couple small patches, almost all the primer came off with a combination of the stripper, scrubbing with the Scotchbrite pad, and a bit of wet/dry sandpaper.

The intervening week was a busy one with no opportunity for workshop time, and we were out of town this weekend. When we returned this evening, I mixed up the batch of sealant. If you've never seen the stuff, looks like sticky, somewhat runny black tar that gets into absolutely every place it shouldn't. You can remove it with MEK or acetone, at least before it cures, but it's a challenge to keep it under control. It won't stick to the non-sticky side of packing tape, though, so I put a piece of packing tape the length of the aluminum angle. After squirting sealant between wedge and skin on top and bottom, assembling the trailing edge, and clecoing it to the aluminum angle, I spent a fair bit of time with paper towels cleaning up all the sealant that oozed out and trying mighty hard to get the sealant off the aluminum instead of spreading it further.

Also, to ensure the trailing edge came together fairly tightly (without the wide gap I ended up with before), I decided to clamp a second piece of aluminum angle along the narrow space between the rivet holes and the very trailing edge of the rudder. You can see that here.

The next step is to wait. The sealant has a 30 minute working time, but it'll need to stay there at least 48 hours to cure properly before I can mess with trying to rivet the trailing edge.
Posted by Picasa

Wednesday, February 25, 2009

Disassemble, reassemble

I spent some time last night taking apart the rudder's trailing edge.

The process of drilling out rivets is pretty simple. If you haven't seen one before, a rivet has a broad head and a cylindrical body. You drill partway down the body, use a pin punch to pop off the head, and then push the rest of the body through the hole. The tricky part is you don't want to make the hole any larger, or the rivet you put in there when you reassemble the thing won't hold properly. I usually have a problem with the drill bit trying to walk a bit off center when I start drilling, so I use a center punch to make a starting dimple in the center of the head. It's not perfect, but it helps. I also check often while drilling.

After a lot of careful drilling, I eventually managed to drill out all the rivets without enlarging the holes. Here you can see the wedge that goes between the trailng edge skins. What you can't really see in the photo is that the wedge has dried T-88 epoxy on it. (So do the skins, but not as much.) The web site for the epoxy manufacturer says you can remove dried epoxy using a suitable paint remover. So there's now a quart of Certified Coatings Sure Strip paint remover on its way from Aircraft Spruce.

By the way, Aircraft Spruce was most helpful with this order. According to the epoxy maker, "suitable" means the paint remover must contain methylene chloride. I couldn't tell from the Aircraft Spruce web site whether Sure Strip contains methylene chloride, but the folks at the Aircraft Spruce order desk went into the warehouse, pulled out a can, checked the contents, and called me back to confirm that it does.
Posted by Picasa

Monday, February 23, 2009

Follow-up on the rudder

I heard back from tech support this morning. They recommend drilling out the rivets, cleaning it up, and redoing it with fuel tank sealant. It'll be a pain, but I agree with them.

So now I need to order some sealant. Proseal at Aircraft Spruce is tres expensive, $78 for the smallest quantity, but it turns out Vans has it, or a newer incarnation of it, in 3.5 oz quantities for around $16. Horray for Vans!

I'm still working out how to clean the epoxy off the aluminum parts.

Saturday, February 21, 2009

Rudder issues

I'm having rudder issues. One of the more ticklish bits of building this particular airplane is the trailing edge of the rudder, one of the last things to touch the air as the plane goes by. You have to get it straight, and ideally you need to have it nice and, well, together. The trailing edge consists of two rudder skins and a wedge-shaped piece of aluminum that all meet at a point. The instructions say to put fuel tank sealant or some kind of glue on the wedge, glue it all together, clamp it, leave it alone for 48 hours, and then go through the ticklish process of riveting it so it comes out nice and straight.

Now, this is the second time I've built this rudder. The first time around, I got it all finished and hung it on the wall of the workshop to get it out of the way. (I was in a rather small workshop at the time; space was precious.) A few weeks later, the string I'd used to hang it broke, the rudder took a catastrophic fall, and the trailing edge got crunched rather badly. After much wailing and gnashing of teeth, I ordered new parts.

I recently finished the rebuild, and it was time to do the trailing edge again. The last time I did this, I used fuel tank sealant and had a whole bunch of the stuff left over, that I wound up tossing because, let's face it, how often are you going to use fuel tank sealant around the house? (Fuel tank sealant comes in two parts. Once you mix them, it'll eventually set. I ordered a small quantity, but it came in a container where you had to mix the whole thing, and the whole thing was enough to do a whole fuel tank and quite a bit more than I needed. You can store mixed sealant in the freezer for up to about six months, but at the speed this project's progressing, six months won't cut it.)

So this time I tried T-88, an epoxy used in aircraft construction. Here you can see it clamped together so the glue will set. That little aluminum bit at the left is a test piece. All seemed to go reasonably well. I gave the glue a full week to set and the test piece seemed pretty solid, so I started working on it. That's where I ran into problems.

Here you can see a shot of the new trailing edge, on top, and the old one on the bottom. The epoxy separated as I set the rivets. It turns out T-88 on aluminum just doesn't have the peel strength fuel tank sealant does. When I tried peeling the test piece apart, I was able to pop it as well. I've written the tech support folks at the kit manufacturer to find out how serious a problem this is.

The trailing edge itself came out pretty straight. It wasn't perfect, but there weren't any major problems. I've contacted tech support about the width of the trailing edge. We'll see if I need to drill out the rivets and redo it.
Posted by Picasa

Friday, February 06, 2009

Driving in the Rain

This is not a rant. OK, it's sort-of a rant, but actually it's an attempt to turn a rant to productive use. In the interests of improving the general level of rain driving skills in Southern California, please consider a few thoughts from someone who's lived in places where serious, heavy rain is a regular occurrence.

1. Use your headlights. Seriously, it's not about you being able to see. It's about other people being able to see you and not run into you. By turning on your headlights, you're also turning on your tail lights, side lights, and all the other things that make your car a mobile Christmas tree, improving the chances that traffic coming your way will stop in time.

2 Be predictable. This is basic defensive driving stuff. The people around you may be hydroplaning, might have fogged up windshields, or might be totally freaking out because there's precipitation falling from the heavens. Use your turn signal when you change lanes. Don't weave. Don't make quick lane changes except in an emergency. Basically, pretend you're surrounded by a bunch of jumpy, freaked-out psychotics armed with two ton chunks of plastic and steel and don't do anything to startle them.

3. Drive in the wheel marks of the car in front of you. The wheels of the car in front of you are doing a bang-up job of clearing water off the roads. You can see those wheel marks on a wet road as a somewhat more dryish patch. The tires throw the water to the sides, and then the water gradually flows back where it was before. That means that, for a window of time, there's a somewhat more dry patch of road. By driving in that dry-ish patch, you reduce your chances of hydroplaning.

4. Drive towards the middle of the road. Most of the big puddles will be towards the edges of the road, because that's how we slope roads. Drive towards the middle to avoid them.

5. Allow extra stopping distance. Your brakes don't work as well when they're wet. And even if they do, your tires have less friction on a wet road. And even if they work well, you might find out, at the wrong time, that you're hydroplaning. Allow yourself some extra stopping distance.

6. Try not to hit the brakes in the middle of a curve. In a curve, your tires are already doing what they can to hold your car on the road. Hitting the brakes might be just what it takes to get them sliding instead of gripping. Try to slow down before you get to the curve. Once you're in it, if you need to slow down further, take your foot off the gas and coast. Use the brakes in the curve as a last resort.

7. If you're hydroplaning, try not to steer or brake. If you're hydroplaning, your car wants to go straight ahead. Your front tires aren't gripping well, if at all, so steering can send you into a spin. The best thing to do, if the road and traffic will allow it, is to take your foot off the gas and coast till the wheels grip again. (The faster the car, the more likley it is to hydroplane. By slowing down, you give the water more time to get out from under the tires, so the tires can grip the road again.) If you have antilock brakes, they should still work to slow the car, but not as well as on a dry road--see "allow extra stopping distance" above. If you don't have ABS, and you have to use the brakes, use them with caution and remember that you might have to pump them. Sometimes you'll hit a large puddle on only one side of the car. That'll tend to pull the car in the direction of the puddle, because of the friction from the water. Correct as you would for any other skid, steering away from the puddle to keep the car moving straight ahead.

Driving in Flood Waters

And now a word for the aspiring Captain Nemos out there--and I count myself a proud member of your ranks.

First off, let's talk about why driving in flood waters is usually a bad idea.

If it's not a bad flood, your car's likely to be hydroplaning a good bit of the time. Also, you can get water up the tailpipe, which will toast the engine pretty darn quickly. As you drive forward through the flood, a bow wave will build up ahead of the car, and water will start sloshing through the engine compartment. That sloshing water can short out the ignition wires, stalling the engine. Also, a car's door seals are designed to keep out rain, not flood water, so she'll start taking on water, which means, at a minimum, you're going to have to replace the jute that glues the carpet to the floor, you'll need to worry about rust, and you might soak the computer (which is often on the floor under one of the font seats).

If it's a bad flood, the tires might leave the road surface. A car will float, at least for a few minutes. During that time, the front wheels become rudders and the drive wheels act like paddle wheels, but minus the paddles. Congratulations, skipper: your vessel will doggy paddle along at maybe 5 miles per hour until she stalls, sinks, or gets swept away by the current to a nasty fate somewhere downstream.

So, it's usually a bad idea to drive through flood waters. However, if you should find yourself in this situation, here are a few things to keep in mind.

1. Consider whether bailing out is the better answer. If the car's in danger of getting swept away, you mght be better off abandoning it. After all, chances are good you can swim better than a car can.

OK, Cap'n Nemo, you've decided to hang with the car. Let's talk about driving in this new environment.

2. Downshift. You need to keep water out of the tailpipe, which means you need to keep lots of exhaust bubbling out. Hybrid drivers (and I'm one) don't have a lot of choices at this point, but folks with non-hybrids should downshift (that's "L1" if you're in an automatic) to keep the engine revved up.

3. Watch the speed. The last thing you want to do is slosh water up onto the ignition cables and stall the engine, because then water gets into the tailpipe and it's all over. (As an aside, if you do stall it, but you manage to coast to a dry-ish area, drying off the ignition cables may be all you need to do to get it started again.)

4. If the wheels leave the road, don't panic. You'll feel it if it happens. Keep the engine revved up and doggy paddle to the nearest relatively dry area. Remember, though, not to rev too much because once those wheels grab again, you don't want the car to suddenly launch itself into something. Once you're back on solid ground, make an appointment with your mechanic, cause you'll probably need it.

You can find more rain driving tips here.

Wednesday, January 21, 2009

Airlines "Hedging" Fuel Prices

During a lull at work, I was skimming the financial headlines and came across an opinion piece with this provocative headline: "The perils of plunging oil prices: UAL's effort to fight fuel costs backfires". The piece goes on to talk about how airlines' heading strategies are backfiring now that oil prices have dropped, so they're suffering financial losses and having to lay people off.

What's odd here is that hedging shouldn't backfire. Hedging is a method of reducing risk, kind of like insurance, but if you're a clever airline on a tight budget you can often do it for no out of pocket cost. For instance, an airline might decide it can afford jet fuel when oil's $100 per barrel but no more, so it might sign a forward contract, a contract with a supplier to pay that price for jet fuel--no more and no less--for the next two years. True, if oil drops below $100, the airline may be upset that it's locked into the $100 price, not the lower market price, but those are losses it's already figured into its business model. At least theoretically. Or if it wants to to limit the range of fuel prices, it could set up a collar (or, more accurately, a costless collar): pay someone for the right to buy fuel from them at the $110/barrel price (buy a "call" option), and finance it by selling someone else the right to sell fuel to the airline at the $90/barrel price (sell a "put" option). If oil goes above $110/barrel, the airline's covered, and it oil goes below $90/barrel, the airline has to buy it at $90, so the airline's locked into the $90-110 price range.

A look at United Airlines' quarterly report for the quarter that ended September 30, 2008 reveals a somewhat different story:

Aircraft Fuel Hedges. We have a risk management strategy to hedge a portion of our price risk related to projected jet fuel requirements primarily through collar options. The collars involve the simultaneous purchase and sale of call and put options with identical expiration dates. In order for the Company to obtain more favorable terms for a portion of its hedge positions, the Company also entered into collars with additional features. These hedge positions include extendable collars, referred to above, and collars that include twice the amount of put volume as call volume. Gains and losses derived from the Company’s hedge positions are not accounted for as cash flow or fair value hedges under SFAS 133. The Company’s hedges that are classified either in Mainline fuel expense or nonoperating income (expense), based on the nature of the hedge instruments.
(emphasis added) This looks more like speculation than hedging. Twice as many "put" options as "call" options? No wonder they're getting hit. Using the collar example above, for each person they've paid for the right to buy fuel at the higher price, they've sold two people the right to sell them fuel at the lower price. And the price of oil dropped way below that lower price. (In reality, probably no fuel is changing hands. Instead, they just settle up the amount of money each party would have gained or lost if it had traded fuel. Financially, the net result should be pretty close, give or take some taxes.)