Comment 8 for bug 120258

I was obviously unclear. I don't mean that the battery time is only wrong after suspending. I mean that the battery profile (which you can see in the "profiled discharge time" graph of power history) is influenced by time while the laptop is suspended - and since the laptop is using hardly any power while suspended, this makes the battery profile vastly underestimate the normal power usage (and hence overestimate the time remaining).

This manifests (for me) as a large spike in the profiled discharge graph at around the 30% mark, around the level I usually resume from suspend, which I can reduce by allowing the battery to discharge through that level a couple of times, and not suspending.