64K MAP vs TPS table.

John Dammeyer johnd at autoartisans.com
Sun Aug 8 20:41:48 GMT 1999


Phil wrote
>
>John wrote;
>
>> MAP sensor vent for the Honda engine is located about 5 mm
>behind the
>> throttle plate.  Once the throttle is opened wide my MAP goes
>to
>> atmospheric.
>
>Sounds like you found a convenient place to stick a vacuum hose
>but given the shading of the port at part throttle very unlikely
>to be
>a "true" reflection of load. Try moving the hose to the plenum.


I'm really not trying to be all that clever.  I leave that to the Honda
Engineers some of whom have a bit of Formula 1 experience.  ;-)  I'm
using the stock Honda MAP sensor located on the stock throttle body on
the stock intake manifold.  I cannot use the stock ECU because the
distributor no longer exists at the end of the CAMShaft.  Also the
ambient pressure sensor is located under the steering column in the
autos and isn't always available so instead my ECU has its own ambient
air pressure sensor.

I probably didn't phrase my question well enough here.

Why would Honda put the MAP sensor just behind the throttle plate?


>
>All throttle excursions need to be treated with respect.
>More than X translation in an increasing load direction should
>cause the
>ecu to default from closed loop to open.

This I do.

>At this point the
>fueling should be 95% correct as it was established by steady
>state testing and calibrating.


Yes it is except that we haven't been able to set up and hold, say,
2300RPM at all throttle positions to establish proper mixture so I have
to extrapolate certain 'calibrated' values.

>
>It seems that your system may be "hanging on" to the closed loop
>algorithm a little too long. Instead of a base operating
>condition of closed loop try setting the base fueling to the max
>power table and then apply time/throttle deg/rate of change of
>rpm conditions before electing closed
>loop.

I thought that was what I actually suggested as a means to handle
acceleration.  Sorry.  Sometimes I don't clearly say what I mean.

>
>Just as you worried about the TECHONOGICAL miracle of valve
>timing
>vs injection timing you are trying too hard to keep closed loop
>operation
>as the basis of engine fueling.  Plain old dumb mapping is very
>efficient.


Are you making fun of me now?


>
>8 bit rpm X 8 bit load = 65,000 carby  jets
>
>By this I'm refering to a 4 point map with interpolation using an
>8 bit A/D converter.
>
>There must be a value for every load / rpm possible not just WOT.
>
>The accel enrichment is used to mask the lean out caused by
>system lag
>(your code plays a part here) If you can successfully calculate a
>new open loop injection value every rpm then accel enrichment is
>used for less than 0.5 of a second, any longer and it's just more
>smoke.


For a race car calculating a new injector value PW every stroke may be a
requirement but is that really needed for marine or aircraft
applications?

Opening the throttle quickly will certainly require more fuel because
there is less restriction to the airflow so obviously more air is drawn
into the cylinder requiring more fuel to maintain combustion/power ratio
but, because the rate of RPM increase is somewhat limited by the load,
is a pure table entry for every combination really needed?

At the moment, I calculate the Pulse Width (PW) every time I get new A/D
values from the sensors.  While the calculation for a new PW is taking
place I use the previous PW if the need should arise.   I am using an
8bit processor so I don't have room for a 64K table so there would be a
compromise at some point anyway.

But I see what you mean.  If I could look up PW for every possible MAP
and TPS to achieve 14.7:1 then I can either just change a constant to
request 12:1 (or whatever is desired) or  I'd need a different table for
14:1 and yet another for 13.5:1 etc.

>
>re bailing wire see page 194
>
>"Stainless-steel spring wire DIN no. 17 224
>
>I kid you not !


ROFL.

Thanks for you input.

John





More information about the Diy_efi mailing list