5% is way too much! if it is higher than 0,5%, I would be surprised. the HC analysis is not that easy. you need to make sure, that the emssion-analyse is faster than the switching time of a si passenger car engine.
I wouldn't do it with gas analysis. OBDII does it with cranckshaft angular velocity changes, I believe. It could also be done with an in-cylinder pressure transducer accurately. I just don't know how often it happens. Lets say at high speed, 5500-6000rpm.
I used to work with US OBDII systems in the mid 90's for one of the Big Three. Yes, misfire is detected by deviant crankshaft acceleration. The OBDII system required a code for a "Type B" misfire which would cause exhaust emissions to exceed 1.5 times the standard. It required cylinder deactivation and flashing MIL for "Type A" misfire which could damage the catalyst. The threshold for those failures varied by engine, but Type B was usually set by 1-2% misfire and Type A would be 1-4% misfire.
In short, properly running engines are not supposed to misfire at all.
Wow, not at all? When I studied small engines (lawn mowers and such)in grad school I remember that they had an estimate of misfire percentage. I just don't remember what that estimate was. I know it depends on the engine design, but I was just looking for a very general number. Zero suprises me.
Well, you've got to consider the relative importance of good combustion to a lawn mower engine and a modern automotive engine. Misfire produces unburned HC in the exhaust which will put you over emission certification limits real fast (if OBD-II triggers a fault for 1.5 times the limit at 1-2% misfire, it'll prevent you from making original cert standards at .75-1.5%). The increased exhaust temps will destroy cats (last I saw, PGMs run $800-$1000/troy ounce, burning a cat is expensive) which lawn mowers don't have. The HC dilutes the oil, which can lead to engine damage. And that unburned HC is just a waste of fuel.
So automakers go to great lengths to get consistent combustion, along with low emissions, etc. Yep, 0 is the design target.
"not supposed to missfire" Problem is the typical missfire is not a fault of a properly running engine. As it is the fault of an improperly manufactured or handled spark plug, which I have seen alot of. It's pretty bad when you remove some high time nonmissfiring plugs to install some nice shiny new ones, just to have them missfire like crazy.
