Date: Wed, 19 Jan 2005 19:07:52 +0000 From: Colin Stamp <colinnospamp.plus.com> Subject: Re: Opened DI Casette - For those interested
On 18 Jan 2005 22:55:53 GMT, Dave Hinz <DaveHinznospamcop.net> wrote: >On Tue, 18 Jan 2005 22:42:51 +0000, Colin Stamp <colinnospamp.plus.com> wrote: >> On 18 Jan 2005 20:17:59 GMT, Dave Hinz <DaveHinznospamcop.net> wrote: >> >>>Check out this article: >>> >>>http://www.vehicular.isy.liu.se/~larer/Projects/main.html >>> >>>> , but the Saab setup does do the job neatly and I >>>> suspect more cheaply than using extra sensors. >>> >>>I'd venture to say much of what they're using the DI for can't >>>be done in other ways, since the spark plugs are the only thing >>>that can be used as sensors in the combustion chambers. There's >>>just no other way to get that information. >>> >> >> That's an interesting article. Cheers. >> >> It looks like the basic parameter that they are trying to sense is >> combustion chamber pressure. > >Well, conductivity, which is a function of a few things, but yes. The way I read it, they only measure conductivity because that's all they can measure using the spark plugs. They can get a rough idea of the pressure from the conductivity but, like you say, they get a few other variables mixed in. I guess that's a bad thing, but not bad enough to make the system non-viable. > >> The goals of sensing knock, misfire, ppp >> etc. are all achieved by interpreting the pressure waveform. The trick >> then, is to find a fast-response pressure sensor that's reliable and >> rugged enough to be used inside the combustion chamber. > >Right, which would be the "combined sensor/actuator" which looks >very like a spark plug. Yep. That's their method alright - and very neat too. > >> Using the spark plugs as ion sensors is really compelling since the >> plugs are in there anyway, but it's not the only way and I bet it has >> a few problems. For a start, the plugs have to be designed primarily >> to ignite the charge. That must make them far from ideal as ion >> sensors. Then they have to break off their role as sensors every cycle >> to produce a spark. > >Well, the timing of the cycle helps a lot. When they're measuring, they're >not firing the spark, and the other way around. They know _about_ when >these things will happen, just not _exactly_ when. Measuring the >conductivitiy of the air:fuel mixture gives them that. I think you're right. What they seem to be most interested in is how many degrees after TDC the pressure reaches it's peak - the PPP. Thankfully, that's well out of the way of the spark. > >This is the reason that using a different flavor of spark plugs in a >modern Saab is a bad idea. > >> One way round this would be to fit a second plug in each chamber. The >> second plug doesn't need to do any sparking so it can be designed to >> be a much better pressure sensor than a spark plug could ever be, and >> it can operate over the whole cycle. It doesn't have to measure >> ionization either. I've just done a quick Google for "combustion >> chamber pressure sensor" and got quite a few interesting hits. > >I'm not convinced that pressure is the only variable that they care about. >You could have the same pressure with wildly different air:fuel mixtures, >just by changing charge air temperature, humidity, and probably another >handful of things. The actual value of the pressure at it's peak doesn't seem to be important, just the crank angle at which the peak occurs. > A second sensor, of any type, would also change >the head considerably, and possibly screw up the combustion chamber >geometry, flow patterns, and a bunch of other things I can't even imagine. Well yes. It would have to be properly designed. > >> Most >> seem to be peizo-resistive but there's also fibre-optic ones. Some are >> even built into spark plugs or diesel glow plugs so you don't even >> need an extra hole in the head, so to speak :o). > >Yeah, 'cuz we'd need that like we'd need a hole in the head. >(had to do it; you understand I hope). And there was me, trying to keep it down to a subtle smiley ;o) > >> Using a dedicated pressure sensor would probably allow better control >> than Saab currently get with their spark-plug ion sensing system, but >> I bet they're more expensive and the emissions regulations mean we >> don't need to resort to them yet, so I reckon Saab have the right >> idea, at least for now... > >Goes a long way to explaining why they're not using a dizzy and coil >any more, eh? Ah well probably not actually. coil-on-plug ignition seems to pre-date spark plug ion sensing. When it first started to come in (can't remember by which manufacturer), the elimination of unreliable dizzys and leads was given as the reason. Of course, it makes ion sensing a lot easier, but that seems to be accidental. Cheers, Colin.