Dont forget ally heads are shinier! i'll get me coat.....
As well as cooling, the water molecules apparently get in the way of the the actual flame front, controlling the burn speed an physically preventing it getting out of control.
The problem as i see it is unless you have a really accurate way of metering the water, you are getting the cooling effect when it isnt needed as well as when it is, as Dave says causing unwanted cooling and resultant loss of torque. My point was that a system which can target individual cylinders only when they pink and with just enough water to control the burn might be better? Maybe it would just do exactly the same job but in a different way!
Glad to see the oft-misunderstood 'higher compression with ally heads means more power' myth has been busted, a pet favourite of mine (yes I went with ported iron heads!). Having said that, the weight saving alone is worth having. On small blocks the iron W2's are still the best bang for the buck IMHO.
As far as the knock sensor is concerned, about time someone brought a kit out, we've been using them in the industry for years. The pricing page for the Hughes (wash my mouth out!) kit just shows truck applications, which already have an ECU (aka ECM, PCM, etc). The knock kit probably intercepts the stock signal and sends it's own instead. However, the main blurb does say suitable for distributors etc, wonder if they are going to price that up. Not cheap, but if it works well (there's the caveat!) it would be worth it.
Some of the difficulties with a knock sensor are that there is (literally) a lot of mechanical noise in the area, and it needs to filter out the regular stuff to find the high frequency knock signature. If it's calibrated right it should be OK. But with one knock sensor and eight cylinders, it will get a different strength signal from cylinders further away, so some cylinders will be more difficult to correctly 'hear' and adjust. Most modern multi-cylinder (6 and more) engines use two sensors.
One thing my company (amongst others) has come up with is an Ionisation system, which looks at the Ionisation current across the spark plug in each cylinder after the spark to determine knock. That way you are directly sensing in each cylinder, and effectivly looking at pressure (inferred) rather than listening for high freqency vibrations (a knock sensor is just an accelerometer). Mercedes for one use Ionisation coils.
Bring on the day when we get those as aftermarket kits!!
Gavin you have echoed my thoughts (worries?) about knock sensors in general. I agree with everything you said there but you lost me a bit on the ionisation front!
I guess you are saying you can tell by current flow (or voltage drop?) when or even how combustion occurs much like my old Jacobs ignition unit but in a more sensitive and analytical way?
Gavin, are these knock sensors susceptable to natural resonances in the engine? If so, how do they filter that out?
Could be a good uni project, I have a second year project coming up. I was thinking of settling the age old argument of whether rough or polished ports and chambers are best for power and/or flow. Initial lines of enquiry tell me that polished will be better for flow, but that doesn't equate much to power since the flow test is so far detached from real engine conditions. I think the major factor here will be atomisation of fuel/air mixture... but I digress.
That's the trouble James. Most flow test benches only flow air. Engines flow an air fuel mixture that is never evenly dense or at a constant velocity either in total or in cross section. Yes you do digress don't you?
Yes that's it Dave. Basically when the spark is done there is still a very small current flowing across the electrode due to the gas being ionised. If you measure the current it will change depending on pressure in the cylinder (due to normal or abnormal combustion - knock).
I think it's only a matter of time before this replaces knock sensors. It would enable individual cylinder A/F ratio control, etc.
When you think what % of the energy in petrol is actually used to do work (can't remember but it's low), the possibilities for future improvement are still huge. There is some good stuff on the way! Not all the technologies will be applicable to our old engines, but some will!
Hey James, I did my first degree thesis on modelling knock in an internal combustion engine, coincidentally!! Trying to correlate some of the theory with practice. Hmmm, I think I have forgotten more than I now know (not much to start with). The brain whithers away once you stop studying.
And on the water injection - the most difficult think about knock is detecting it interactively. Once you've got that cracked (by whatever sensor), it's pretty much free to retard the ignition to fix it, but expensive to do that via water injection.
Yes you can use water injection to prevent knock in an open loop system (fixed settings), but if you go closed loop the gains are higher, and that's when you need a realtime detection system.
