From:
http://www.worldcarfans.com/2070322.001/new-bmw-m3-v8-engine-in-detail
-8<-
Previously, i think the methods to detect knock have been piezo
detectors on cylinder/rotor head that listen to 6000-8000 Hz knock
signature, EGT/CHT trend monitoring, and pressure sensors routed to
chamber eg. thru spark plug body.
This new method seems to be quite reliable and can be implemented
easily
to ECU boxes that have ignition control.
This leads to think..
How apex seals will hold together when engine is on dyno and repeatedly
run against knock limit - to create clear figure where the limits
(schlimits!) are....
If seal has got cracks, it is likely that it's resonant frequency will
change - just like it happens with eg. wine glass..
Testing might be done - without engine breakdown - by fabricating a
probe that attach to spark plug hole and contact seal when it is
rotated
on position. Then probe will test it's integrity by exciting it with
sound sweep, and monitoring response when sweeping through it's normal
resonant freq.
Small surface cracks don't necessarily create big difference, but
depending on which mode the seal is vibrating the change in part
stiffness might be detectable.
Cheers,
Petri
Saab was fooling around with this quite a few years ago as I recall.
Also you want to monitor the intake air temperature and keep it below
about 140 F. When the intake air/fuel mixture is compressed 10 to 1
by the engine
and or a turbo charger the temperature rises. It can self ignite
before top
dead center by the increase in heat due to the compression. The
piezo knock
detectors are after the fact.
One should be looking for a knock or self ignition predictor. The
intake
temperature is the best predictor I have found so far. This was
known prior
to WW II and P38 pilots were trained to monitor the intake air temp.
Paul Lamar
Peak chamber pressure timing (measured by ion sensing or with a direct
pressure sensor (less practical)) should dictate ignition timing -
regardless of temperature. This will keep the engine from detonating
(or even pre-igniting) if the detonation is at all controllable by
ignition timing. (If the boost pressure*temperature gets high enough,
the ignition timing won't have any effect at controlling detonation -
the engine just diesels above a certain P*T. )
That's the beauty of measuring the timing of the pressure peak. It
keeps you from getting to a place where the engine knocks.
Fortuitously, the timing that leads to peak horsepower (torque at a
given RPM) is one which is cooler (more retarded) than one which causes
the onset of pre-ignition. This means that while on the dyno, you rock
the timing to get the best horsepower and measure the timing of the
associated pressure peak. You don't have to drive the engine into
knock/pre-ignition. Amazingly, as pressure and temperature vary, the
best horsepower chamber pressure timing doesn't vary, though the
ignition timing angle required to achieve that does..
Peak boost pressure should be limited by intake temperature - separate
Picaxe on the wastegate from the Picaxe controlling timing. This will
keep from getting to an operating condition (P*T) where retarding the
spark doesn't have any control over detonation.
An absolute pressure pop-off valve is useful in case the throttle is
rapidly closed.
Regards,
Matt-
Some of the 1998 SAAB work.
http://www.fs.isy.liu.se/~larer/Projects/main.html
http://www.swri.org/3pubs/ird2003/synopses/039305.htm
http://www.theautochannel.com/news/2001/04/20/018974.html
http://patft.uspto.gov/netacgi/nph-Parser?Sect1=PTO2&Sect2=HITOFF&p=1&u=%2Fnetahtml%2FPTO%2Fsearch
Paul Lamar
Thanks for input, Matt and Paul,
Let me further explain my point..
This means that while on the dyno, you rock
the timing to get the best horsepower and measure the timing of the
associated pressure peak. You don't have to drive the engine into
knock/pre-ignition.
Correct, when we already know boost/rpm limits.
But to create this map, we have to do it by "party animal" method..
"You don't know where the line is until you've crossed it" :)
With turbo on sea level, we have ample boost reserve to run our engine
to knock on almost any rpm.
Textbooks are quite dry on this issue, but common wisdom seems to be
that 200 psi (~14 bar) BMEP is about the maximum attainable with premium
pump gas and conventional engines.
Rotary engine should get past this figure, since their Friction MEP is
lower due to absent valvetrain and piston skirt friction - so when IMEP
remains the same, lower FMEP results higher BMEP. Though, regarding
knock limit, it might be somewhat compensated by non-optimal combustion
chamber geometry.
So, how much eg. 15 bar BMEP gives, horsepower-wise?
rpm,k 5000 6000 7000 8000 9000 10 000
hp 110 132 154 175 197 219
total 219 263 307 351 395 439
Formula: P(hp)=[V(ltr)*RPM*BMEP(bar)]/448
(this for 2-stroke and rotary. For 4-stroke, divide RPM by 2. For psi,
divide by 31 instead of 448)
Knock onset limits can be estimated by making an educated guess - but
when seeking to get absolutely max power out, we have to run our engine
against knock limit. Then we know where our 110% is, and we can set 100%
mark to our chart for that rpm and fuel (and altitude, and amb. temp).
Regarding spark advance, as Matt noted, there's regions. When doing test
runs on dyno with variable brake power, we set up a closed loop where
dyno increases brake force when we open the throttle - to keep rpm
constant.
Initially, we seek and set up the best power advance curve for lower
boost/loads, and run against knock onset somewhere higher. When reached,
we protect test engine by setting our system to cut injection
immediately - to back it off from situation before pre-ignition occurs.
Then briefly run on lower setting to let CHT stabilize to normal, and
run up again, now with somewhat retarded spark. Record where's limit
now, and redo with even more retarded spark, until we are on "insanely
retarded region"... X) - like greddy RX-8 turbo kit on maximum boost.
From this data, we now know our max boost limit, and can make a timing
curve that allows us to have max. power out on this rpm.
Redo for several rpm's, and after done, we have our data to make a
boost/timing map, and we know what's 100% power on every particular rpm,
and when it's time to open the wastegate - and bypass valve, if needed.
Then we switch to another fuel with different MON (premium(89),
100LL(102) or air racing(120)) , and redo the whole thing from start :)
This sounds the perfect way to waste a pile of money and time to engine
rebuilds, but knock in itself doesn't necessarily ruin the engine -
pre-ignition does, without a question. So, preventing prolonged
detonation should allow the engine to last, maybe thru the whole test set.
If seals crack and spit out, it is essential to prevent them from adding
insult to injury by hitting the turbine. That could be arranged by using
"branched wastegate piping" the way that was on Paul's drawing couple of
weeks ago. That exhaust arrangement could make sure that anything
heavier than gas goes straight to wastegate pipe and hits a bypass valve
there, not a turbine inlet. Though it should be dimensioned right, when
closed it is a standing wave resonator.
Altitude tests, if/when needed, are more difficult. To have really
useful results, we should bring the whole orchestra - instrumentation
battery, oil & water "radiators" and inter/aftercoolers up with the
engine. Likely a job for cessna 337.
The thing is, if we want really good altitude performance, charge air
cooling on thin air becomes an issue. Thermal load increases to abt. 3x
on FL300 and 4X on FL400.
If we are creating a "hotrod engine" for a brief climb or speed record,
we can bypass some of this by just gradually adding nitrous gas
injection when single turbo starts to give up, and by spraying water to
charge cooler inlet on peak compression to get away with a smaller unit.
Sounds like a plan?
I asked our faculty's engine workshop staff if it would be ok if i get
into some testing with their bench dyno... They were not too excited,
and resorted to silly excuses. They claimed that it will fly off the
handle after 300hp, it is used for contract work, fixing it takes
forever, first you need permission from here and there, etc. etc.
I'm a taxpayer and they're government funded, so basically that dyno is
mine! (at least 5E-6'th of it ;) - they don't seem to agree..
So, i need to find another that i can use - or have my own, somehow.
Fixed "Ballast" prop would be ok otherwise, but it cannot be adjusted
during run. Constant speed would be just fine, but it costs a lot and
requires a proper gearbox and hydraulic governor. Optimum would be a
dyno which sustains 10.000 rpm and is adjustable on the fly, so no
worries about gearbox and all tests could be done as planned.
Money (or lack of it) seems to creep in... i should find a "fundraising
artist" :) Now,wait, Mika Häkkinen (F1 champ -99) seems to have dumped
his ol' lady - maybe he could join our noble cause and lend some pocket
money. Maybe even contribute himself, now that there's no-one nagging
about polishing the crankshaft on living room :)
Cheers,
Petri
The Rotary Engine NewsLetter. Powered by Linux.
ACRE NL web site.
http://www.rotaryeng.net
Copyright 1998-2007 All world wide rights reserved.