Introduction:

The Saab 9-3 is a great car, but there are a few issues that are common to them.  One of them is the recirculation motor/actuator fails when a plastic gear strips and prevents movement of the lever arm.  This actuator normally switches the A/C system between using outside/inside air.  Typically it's fine to always use outside air, but during the dead of winter, you need to recirculate the air to warm the cabin and prevent fogging.

My local Saab dealer wanted $440 to replace this motor with a new one!  This is one expensive plastic gear, and no, there isn’t a lot of labor involved (it took me about a half hour to replace the motor, not including gear-repair time).  I figured this was a great opportunity to solve a common problem and save some real money.

I experimented with a few different methods.  I'll explain the successful method here, which took about a half hour to do, and consumed maybe $0.50 worth of supplies (clay and plastic).

Summary:

-         Imprint a mold of good teeth.

-         Use the clay mold, along with original gear, as a 2-part mold to create the tooth.

-         Use Smooth-Cast polyurethane for the cast.

-         CAREFULLY trim excess polyurethane from gear.

-         Super glue the new tooth onto the original gear.

Part 1: Create mold of good teeth

You do not need to make a mold of the whole gear.  You need to create an imprint of a small fraction of the teeth on the gear.  This will enable you to simply slide the gear in and out of the clay, because the gear and clay, together, will be the mold for the new tooth.

Modeling clay works very well, since it is more durable, and less likely to break apart than Play-doh.  Feel free to expirement.  Bake-able modeling clay would probably work even better, as long as there is minimal shrinking of the mold as you bake it.  Then you don’t have to worry about disturbing the mold when casting the new tooth.

Part 2: Cast new tooth

I created the new tooth using “Smooth Cast 300”, which is a 2-part polyurethane plastic kit from Smooth-On: http://www.smooth-on.com/Urethane-Plastic-a/c5_1120_1209/index.html

This kit is easy to work with, and lasts a long time in the bottles.  It's been over 2 years since I used these bottles, and they're still good.  Stronger plastics are available on their website, but this is the kind I had.

To cast the tooth, mix a small amount of  "Part A"  and "Part B" together.  I used about 5 drops from each bottle.  Drip the mixture near the clay mold, and slide the original gear section with the missing tooth, into the clay mold.  The resin will be forced into the void for the new tooth.

Wait for the plastic to cure, this kit took about 15minutes.

Part 3: Trim/Superglue new tooth

First of all, it is unlikely that the polyurethane gear you created will unlikely adhere to your original gear.  Keep this in mind so you’re careful while trimming excess plastic from the gear.

At some point, the new tooth will start to flake off the original gear.  At this time, apply the superglue between the new tooth and old gear, and wait for it to cure.

Trim more plastic with a razor as needed, until it properly mates with the other gear(s).  Then test it :)

Figure 1: Overall Plot

Figure 1 depicts all of the raw data I collected. The black line is simply the freezing point. Since the heater turns on/off frequently, it creates a thick band of measurements, Figure 2 is a closeup:

Figure 2: Closeup

Here you can see the changes in the setpoint of our thermostat. My house has a programmable thermostat so it's turned down when we're gone, and up when we're home. The unit also tries to predict thermal properties of the house, so it'll be a certain temperature at the time you want. However, I think this feature is inaccurate and flaky. Anyway, this plot really shows how often the furnace is cycling. Figure 3 shows a plot of one-hour's worth of data:

Figure 3: 1 hour

So why did I go through the trouble of all this? 1.) Just out of curiosity, and 2.) Because I the numbers to know when I've done enough work, without having to wait for ice dams again. Once the attic is done, I'll replot these numbers to see if the difference between attic temperature and outside temperature is "close enough" so ice dams won't form, see below:

Figure 4: Difference between outside and attic temp.

On average, the attic is 26 degrees hotter than outside, but I've recorded a maximum differential of 43 degrees F. According to some thermo calculations, if 6 inches of snow covers the roof, the surface underneath the snow will be about 18 degrees hotter than the outside temperature. So, if it's 20 degrees outside, the roof surface is 38 degrees, more than enough to thaw the snow, only to refreeze at the soffits.

The magic temperature when ice dams will NOT form, is 32-18 = 14 degrees. According to my estimates, my roof surface is melting snow 36% of the time. If I can reduce the temperature differential (between outside and attic temperature) to 10-15 degrees, I'll be melting snow only 1-8% of the time. Although this is a very tall order, it will keep my roof in good shape, prevent future leaks, and prevent me from having to rake the roof all the time. This 36% number seems pretty high though, and assumes 6 inches of constant snow (unrealistic) which would imaginarily be replaced as it thaws.

Doing some more calculations, this means my average attic temperature must go down to 20 degrees, not 38 degrees. Wish me luck :)

Of course, all these calculations are based on a lot of assumptions, for this short week in January. Your mileage may vary.

Before I tear up my attic, let me give you a brief tour: Here is a sample view of my attic, as you can see, there is no insulation:

Since my house is on a slab foundation, all of the ductwork runs up through the attic. So, they're nice and hot up there for the winter (making icedams worse) and cold in the summer (possibly causing condensation). Fortunately, I haven't seen evidence water dripping from the ductwork.

My house was also built using the latest 1940's technology of wrapping newspaper around ducts, then "securing" it with rope tied around the ducts:

Ok, so there is no rope securing the newspaper in this picture.

Anyway, I bought some temperature loggers to do some qualitative thermal analysis, which I'll provide later. I've a lot of work to do, but will provide updates on this project when it makes sense.

• Thermal analysis of my attic to solve the ice-dam problem.
• Aluminum anodization.
• Car stuff
• Cooking.
• Other project type stuff.
• Other hopefully informative things.