From a common sense standpoint, the 2% error you describe seems to be right. I mean, we're talking about an instrument designed to measure "something." For it to be off by XXppm, then you have to stop and say, if that's the case, what's the point of the meter?