[Diy_efi] boosted compression ratio

[Mike]

At 08:17 PM 22/12/2002 +0800, you wrote:
>> Well why the heck dont you add a water injection setup and get the
>> humidty artificially - and you could chill it as well :-)
>
>Expense and water is scarce. :-)

Yeah methanol might soon be cheaper, the way the rains are going ;)

>Still trying to get a cheap inlet fogger together; that should do
>the trick. In winter, I'd have to heat the fog so that the air
>temperature rises as well. At 8 degrees C, air is saturated with
>only about 8 g of water per kg of air. Heating the air to about 25
>degrees increases that to about 20 grams. That's worth 3 Octane
>points; but the temperature rise reduces the effect by 1 to 2
>points. <sigh>

But hang on, you dont want water vapour and you dont want too
small a droplet size otherwise it will become vapour by the time
it gets to the cylinders. I dont recall what boost you were
running but over a wide range where the droplets will displace
a little air - you still want to introduce quite a large amount
of water so the cylinder compression can dump its heat in the
latent heat of vapourisation. SOmeone called 'Pete' on the
nissans13 user group on Yahoo did a nice setup and managed to
run straight ULP at around 15psi or so boost, went up to
almost 20psi on optimax,  using off the
shelf items and an atomiser head from a agri company, I think
it cost him around $200 and most of that was the pump which
I recall was one the best most reliable marine ones around
(I'm not on that group anymore - 9 lists are enough <sigh>)

>> The time is not the issue, its the fact that at higher revs there
>> are all sorts of other things that may 'sum' to way too much
>> noise floor for effective ping detection...
>
>Unless you know the noise floor and can compensate; which is what
>the (defunct) Harris chip did; measure noise floor outside the knock
>window and "subtract" that from the knock window levels. Well; it
>was rather more complicated but that's the gist. Hopefully, the
>valvetrain isn't clattering away just inside the knock window;
>though it's conceivable that some valves might be opening or closing
>at the same time.

mmmm, Bit difficult due to the random nature of noise but fair point
that 'noise' at different times but at the same rpms should be close
enough that a signal might get through, especially integrating over
the window and doing some convolution etc Its all those other
nasties like occasional valve bounce (minor I know) but changes as
the valves rotate, that and ringing through the rods, eccentric
motion around the big ends, twisting and reflections along the crank
etc - which contribute to making it unreliable to just sub the last
known noise - then the air con kicks in - another set of varying
noise or a change in gear propogates back through to change the
instantaneous torque fluctuations which make everything else 
twitch in frequency and amplitude ad infinitum etc etc

I really like the idea of correlating one knock sensor per
cylinder with those fibre optic pressure sensors (in spark plugs)
with individual EGT per chamber that might make it good
for high rpms and high loads - not quite synonomous with my
general use as a road car, compliments the twin tires I have
on it at moment - I must be probably the last person in Australia
with them if not the Southern Hemisphere <sigh>

rgds

mike


_______________________________________________
Diy_efi mailing list
Diy_efi at diy-efi.org
http://www.diy-efi.org/mailman/listinfo/diy_efi