This information applies specifically to 2003-2008 'SG' model Foresters. Although it illustrates the dual filament H4 bulbs used in the earlier years of the SG model, the wiring in the later models with separate high/low beam bulbs is pretty much identical right up to the final point. So everything should work the same throughout the entire SG line. These concepts should also apply to the earlier SF models, since Subaru employed a similar setup there.
A quick review of the diagram for the 2009+ SH models indicates a different approach to this wiring, so that this information is probably not applicable.
A while ago Ferret gave us a great analysis of the peculiar situation that occurs when one of the headlight fuses blows on a Forester that uses H4 combined high/low beam bulbs. Briefly summarized, the light on the side with the good fuse appears to come on as normal, while the light on the side with the bad fuse glows dimly. Yes: bad fuse, no apparent electricity, but still glowing dimly! This is due to what engineers call a "sneak path," which is an unintended complete circuit that can produce unexpected results.
You can see Ferret's full post at: Subaru Forester Owners Forum - View Single Post - Dim head light
. Ferret is good at this stuff. He can find a sneak path almost as fast as the electrons do!
This situation isn't limited to blown fuses. It will result from anything that interrupts the flow from the battery to the common terminal on one of the headlights. A broken wire or a terminal on the headlight connector that has been destroyed by overheating are two other possibilities, as is a specific type of failure within the bulb itself. It's a result of Subaru's choice to power the bulbs on their common terminals from separately fused circuits and to switch them in parallel on the low side. Note that the National Electric Code forbids such approaches in house wiring, in part for many of these same reasons. But for automotive engineers, the word is "If it works, do it!" (Maybe they did it for safety? Fuse has blown but both lights are sort of "on" anyway? As we say in the IT profession, "Not a bug, a feature.")
Since the symptoms of this situation are so bizarre and don't intuitively lead one to suspect a blown fuse, the problem keeps coming up here and on other fora. I thought that illustrating it with a diagram might be useful, so here it is. In this example assume that the low beams are turned on and the connection from 12 volts to the right side bulb has been interrupted for some reason—blown fuse, broken wire, broken terminal on the connector at the back of the light, or a broken connection inside the light. But the situation is fully symmetric and will occur for any combination of high/low, right/left. You can see the sneak path starting down the unselected
high beam filament on the left, then going up the high beam filament of the right side bulb—traveling the wrong way on a one-way street. Normally it couldn't do this because of the higher voltage that would be present at the top, but since that voltage is no longer there the electricity is able to use the high beam filament as part of its path to ground, completing the path down through the right side low beam filament. Three filaments are now in series, sharing the 12 volts among them and all glowing dimly. The right bulb looks dim, and the left one is actually a bit brighter than usual, since its low beam filament is illuminated at full strength and its high beam is on dimly. Depending on the exact situation, the blue high beam indicator bulb might also be glowing dimly.
For clarity, I've simplified the switching mechanism by just showing an ordinary switch. In reality it involves relays, DRL resistor, etc.
The fuse numbers are those shown in the shop manual for my 2003 XS. Fuse numbers for different models, years, or national markets might be different, but the principle should be the same. A general caution: I have noticed a few minor discrepencies between the shop manual and the actual wiring, so always check things out with a meter before doing anything significant to your electrical system. The meter doesn't lie.
Just to clarify things, I arranged the various parts of the bulbs in my drawing in order to illustrate the various flows. The drawing is not intended to represent the arrangement of the actual terminals on the bulbs, though it would seem like the obvious thing for the bulb's designers to have done. Here's the actual pinout, as seen looking into the connector
. The wire colors are as specified in the 2003 wiring diagram. YMMV.