I start the engines the next day to do some VFR work. As the engines fire up, the engine start checklist directs my eyes to the tachometer, then the oil pressure, and to the suction gauge. The suction gauge indicates that the right vacuum pump has died. Okay! That sort of explains the zombie heading indicator, and means that I've snagged the wrong instrument. I should have not written it up while it was questionable. The laws don't allow me to write, "never mind, it's not broken after all," which is why there are so many airplanes flying around with sort of broken stuff. As soon as it's written down, it grounds or restricts the airplane. As I understand it, only a maintenance engineer is allowed to correct a pilot's mistaken snags.
Fortunately, the effect of the unserviceability is the same, as I'm not allowed to conduct IFR flights with only one vacuum pump working, and we're in the north in the summer, so "night" is virtually non-existent. What's weird, though, is that the reason there are two vacuum pumps, is that one is supposed to be able to serve as a back up. Either suction pump on its own is supposed to be sufficient to run both the attitude indicator and heading indicator. I'm supposed to be able to safely complete an IFR flight with one suction pump failed. Come to think of it, the autopilot is not supposed to engage if only one vacuum pump is operating, to ensure that there is sufficient suction to run the essential instruments. The autopilot is still working, although a little half-heartedly, as the heading indicator is still sluggish. So is the attitude indicator. The shuttle valve that is supposed to disable the autopilot must have failed, so the autopilot is remaining on line, leaving not enough suction for the heading indicator and attitude indicator. I turn off the autopilot, but it doesn't help much.
After a few hours the heading indicator and attitude indicator are both completely useless. Weirdly, the autopilot will still engage. (I keep forgetting and snapping it on when I want to look something up or open a Clif Bar to eat). But it will attempt to follow the slowly toppling attitude indicator. Pretty scary, really. How far would it roll the airplane? That's not in the manual.
In the old days, there were no heading indicators or attitude indicators or autopilots. The turn and bank instrument and the compass were all you had for staying right side up, and if the former died, it was just the compass, which has all kinds of errors or lag and lead. The old bush pilot trick for letting down through an overcast was to do so on a heading of south. On south, if you bank at all, the compass will immediately swing around, even before the heading has changed, especially at high latitudes. I try this at the end of the flight, because my destination is pretty close to due south of my position. It's pretty hard. Try it sometime.
As I taxi in at the end of the flight the suction gauge indicates that the second vacuum pump has failed. Well this would make a fair amount of sense: the pressure differential between the working and failed vacuum pumps is what is supposed to throw the shuttle valve to disconnect the autopilot. And obviously if they have both failed I can't expect the suction instruments to work properly.
Company has already made arrangements to ship a vacuum pump to a maintenance shop a few hours flight away, so now they will ships two vacuum pumps, and we will head there on the next flight.