How do engines know which cylinder is misfiring?

The sensor used to determine a misfire is primarily the crankshaft position sensor. If it takes too long for the CPS to register a sweep to the next cylinder in line as the engine rotates then the ECM will usually issue a flashing check engine lamp indicating a misfire has occurred.

During open loop operation the ECM is allowed to ignore misfires (full throttle and anything happening before engine warmup).

Chris

Two ways.

1) when the cylinder doesn't fire, the voltage across the spark plug follows a slightly different profile (believe it or not), and that is reflected back onto the primary side of the coil and can be detected by the computer. This method (I think) doesn't meet OBD-II requirements because its an indirect method of observing a misfire and if its set sensitive enough to catch all occasional misfires (rather than just the gross misfire that happens, for example, with a fouled plug or burned valve) then it is prone to too many false alarms.

2) When a cylinder misfires, the crankshaft slows perceptibly, and the crankshaft position sensor detects the deceleration and which cylinder it correlates with. That's a direct measurement of misfire and is what is generally used in OBD-II engines.

Wow, thank you both for the enlightenment! Its hard to imagine that it's sensitive enough to notice how a crank shaft slows down though in such a small amount of time and angle, since it has all the momentum keeping it going, but it makes sense.

Cheers,

Dean

Saturn has measured capacitive discharge of the 1/4 coil pack on their venerable 1.9 engines since 1996 (OBDII) and they continue to use a similar system on the newer 2.2 in lieu of a physical cam sensor. For as long as they've used this method, I would *ass*ume that Saturn was/ is in compliance with OBDII requirements or they would have amassed quite the stockpile of EPA fines. I don't know the ins and outs of the law, but it seems like there is some leeway in design technology as long as it (somehow) performs the prescribed task. Systems on OBDII cars are often designed with multiple technologies as chosen by each manufacturer; for example, EGR position can be determined by O2 and/or MAP sensor change in activity, a physical position sensor, Ford's differential pressure feedback sensor, EGR temp sensor, and probably other methods I am not familiar with.

That's certainly the common method, and the one that makes the most sense in terms of redundency and reliability (just don't let Daimler Chrysler know cam sensors can be reliable).

Toyota MDT in MO

It's actually a surprising amount. Although there is a lot of momentum coupled to the wheels, there is also a good amount that the whole drive train can twist between the wheels and the crankshaft; the steel is not all completely rigid. And it doesn't take much movement to detect.

--scott

Its hard to imagine that

Yes, but think about other modern electronics. Your cell phone can tell when a 900 MHz or 1.3 GHz signal shifts phase by 1/4 cycle, and that is orders of magnitude smaller shift than the crankshaft's deceleration.

Secret technology that died with Chrysler I guess. My 1993 (way pre-Daimler) still has its original cam sensor with 240,000 miles. ;-)

Heck, never mind the wheels... the two ENDS of the cranshaft don't move perfectly in-phase! That's why engines have harmonic balancers, after all.

It does at first seem surprising they can measure it but think about it, YOU can feel it so it's more then just a little bit of slowing.

So that's what they are for. I never really knew what they were doing other then balancing something or other yet a balance like a wheel balance never made any sense, this does. That makes two things I've learned today.