gettechnical

I have a 1999 Forester with a manual transmission. Last winter and again this winter, after driving on wet roads that have been treated with salt, the speedometer stops working. That is, the indicated speed immediately drops to zero. The effect is temporary, though. After a few hours of drying out, the speedometer starts working again. Once exposed to salt, driving in merely wet conditions with no additional salt will also cause the speedometer to fail until the salt is washed out or off of whatever it is affecting. So during this last summer, wetness caused no problems, the salt having been washed away, but the speedometer intermittency returned after the first ice storm this season. Because of this clear dependency on enviromental conditions, I don't suspect the usual '99 model year speedometer head problems as being the culprit. I was thinking it was saltwater getting into the plug on the speed sensor, which is located fairly low on the transmission, so this fall I blocked all the openings in the plug with putty, but it hasn't prevented the problem. The voltages delivered to the speed sensor are always right on spec, so far. One further clue: when the speedometer is in failure mode, stepping on the brake causes it to work again--but only as long as the brake pedal is depressed. Apparently there's some interaction with the ABS, but at least the ABS is still fully functional, since the speedometer is always working when braking . Occasionally the failure turns on the MIL with a speed sensor failure code, but the car is not functionally affected otherwise. If it's not the plug getting salted, what is it? The speed sensor itself is inside the transmission, and ought not to be affected by moisture or salt.

Well, I found the problem: the end of the vent pipe after the drain filter that sticks up into a hollow part of the body was blocked off by a spider's egg sac. I found this by removing the canister and blowing into the various hoses to find where the resistance was coming from. The key to finding the cause was the diagram showing how the parts of the evap system were supposed to work (instead of me guessing how it worked), as well as that link to the tips page for refueling problems. Since I apparently can't make attachments, here's a link to what it looked like: http://fotothing.com/tony/photo/c3116e6b437ac2a7e12cbfef149d77ef/ (The small hole in the blockage may have been caused by my blowing through the pipe--it was pretty tough to force air through!) Thanks again to all the folks who offered such great information!

Extremely helpful PDFs! Thanks very much! I was all over the Subaru technical docs site and couldn't find anything like this. Now I finally know what's going on. Thanks again!

That's very interesting, because it indicates that the displaced air during refeuling SHOULD go through the canister, condradicting what I just wrote to Joe. (Now I wonder what that vent hose is for, if not to return vapor to the nozzle...hmmm). I shall investigate tomorrow. Thanks very much!

The guage seems accurate. That is, when the guage shows nearly empty, the tank will accept at least 13 gallons of its 60 liter capacity as long as you fill it V-E-R-Y slowly, and the only way that can happen is when the attendant gives up in in disgust end tells me to fill my OWN goddamn tank, which I then proceed to do (breaking the law in the process). There is indeed such a line, and it's not pinched that I can see. Nor is there any obvious physical damage. I haven't ruled out the EVAP system, but as near as I can tell, the displaced air when refuelling is supposed to go directly to the top of the filler pipe through the aforementioned vent line to be sucked up by the vapor recover bellows on the nozzle. It doesn't appear that it takes a trip through any solenoid valves or the cannister, but one thing I have NOT yet found is a complete description of just how the whole fuel/evaporative system is supposed to work in detail. Apparently, mechanics just know this by osmosis. I suppose I could also remove the fuel pump assembly and try to fiddle with the vent valve from the inside. Haven't tried that yet. Thanks for your advice!

As if the rising price of gas weren't enough, it's nearly impossible to get the fuel into the tank at any cost. What happens is that the gas backs up into the fill pipe, shutting off the nozzle prematurely. Once the tank is about half full, it only takes about a half a gallon at a time unless you let it in very slowly. Unfortunately, I live in New Jersey, which is one of the two states which outlaw self-serve, and there's no way the attendant is going to nurse the gas in like this. The problem, as the astute observer has probably discerned, is that the vapor/air displaced by the entering fuel is being restricted in its passage out of the fuel tank. I suspect a problem with the vent valve. Replacing this item requires removing the fuel tank--a major operation I cannot do by myself. So my question is: does anyone know if this is a common problem, and is there a simpler work-around to fixing it (say, forcing a solvent down the vent line to clean out the vent valve or something)?