EFI questions and answers

[Todd Knighton]

ehernan3 at ford.com wrote.

> PS: I don't have access to the Web, but I would like to know the list's experience and recommendations on aftermarket O2 sensors, 
especially the ones that use

	We have been using a Motec Air/Fuel ratio monitor for a couple 
of years, it gives us accurate air/fuel measurements using a 4 wire 
Bosch probe from 11.0/1 to 23.5/1.  Contact Motec Systems USA or JGM 
Enterprises, one in the same at (714) 897-6804 at 5692 Buckingham Dr. 
Huntington Beach, CA 92649.  The unit cost is $2,800.00 with the sensor 
and has outputs for 0-5v vs lamda or lamda/10 (14.7 = 1.47v etc.) and 
many other uses. It will display in lamda or air/fuel ratio for about 7 
different fuels.  The only problem we have with it is the price.  The 
sensors last about 200 hours on power testing and 50 hours on power with 
leaded fuel.  We've experienced higher hours due to most of the testing 
and setup being at part throttle and light.  Full throttle dials in 
pretty easy.
	The next best one seems to be utilizing the UEGO sensor but 
these units all get up over $7,000.00 plus and are probably not for the 
DIY'er.
	Cheap units come from Haltech, they use a standard 3 wire or 1 
wire Bosch type probe and use 30 led's for display.  If you had 
something to calibrate this unit against, it might work o.k.


	On to us,  I've converted an 1988 Bosch Motronics from Air Flow 
to Speed Density system and have some problems I need help on.
1.	Air Temp.  I'm altering the injector duty according to the old 
PV=nRT formula where T is in Degrees Kelvin.  This overcompensates and I 
don't know why.  The factory Air Temp function that was used on the Air 
Flow Correction was about 25% of this compensation for air temp.  I've 
tried calculating a function using the SAE correction factors for 
horsepower, because they use air temp as a function and these come alot 
closer to the factory compensations, but are still no the same.  
	Air temp compensation at idle doesn't seem to be the same as at 
part throttle or full throttle conditions either, why and how do you 
compensate for this.
2.	Idle compensation.  The closed loop idle compensation function 
Porsche had used on this unit worked pretty good for the Air Flow type 
system, but surges like crazy with the speed density system.  Does 
anyone have any good examples of closed loop idle comp. functions or PID 
loop stuff I could try, or maybe a fix for the Motronics one.
3.	Barometric pressure compensation.  Why do you need to compensate 
for BP when if you are using an absolute BP sensor.  I can understand 
the stoichiometric thing about high altitude vs low altitude and if it 
is part or full throttle, but who cares on a performance application, or 
let closed loop figure that one out.  Will the calc's be correct on a 
system that was designed to run, say, 13.0 / 1 air fuel ratio all the 
time, no matter what the condition.  Would the air fuel ratio change at 
different altitudes?


	We've used all of the above sensors mentioned in the letters, 
the Mass air flow sensor, Throttle Position Sensor, Manifold air 
pressure sensor, as well as the Flapper door air flow meter, they all 
certainly have their draw backs.
	The flapper door is restrictive and only see's CFM (cubic feet 
per minute), not Mass. As well as it uses a resistor element that's 
prone to wear out.
	The Mass Air Flow Sensor, sees mass but on the high power 
systems we do, when you have a sensor that will read 800 horsepower, the 
idle and low speed stuff is horrible.  The resolution is not according 
to load at all rpm ranges, but according to overall range of air flow.  
It also need to be trimmed for throttle position in our application 
because it's about 6 inches from the throttle body.  It gets so messed 
up from air flow disturbance and turbulance that the original calc's 
aren't even close.  Also on turbo engines it gets real interesting when 
you shut the throttle down from a boost condition and release all the 
air that thing just measured and is trying to inject the fuel even 
though the air is not going in the engine!  Resonance messes with the 
MAF's big time, especially in racing cam type applications, for get 
trying a straight calc.  The MAF's measure equally in both directions 
and don't know which way it's going so they tend to jump their reading 
up at high resonance periods or turbo pressure blow back.  The Bosch Air 
Flow meter is the best one I've used so far off the new Porsche 993's 
but it runs out of range at about 460 hp.  The Ford Hot wire is the 
stupidest thing I've ever seen, or the aftermarket ones anyway.  They 
use a 3.5" O.D. tube with a .5" hot wire mounted on the I.D. of the 
tube. (that's Outer Diameter and Inner Diameter by the way.)  And that 
thing if it's rotated about 5 degrees in our application is metering a 
whole new air flow stream.  It might work if you had room to put a 3 
foot pipe between the engine and the air flow meter with now curves 
whatsoever to laminarize the flow, but that's kind of hard to do in an 
engine compartment. 
	The Manifold air pressure sensor measures load at independent 
rpm ranges but certainly not mass and transients are definately 
interesting.  They seem to be able to be taken care of with a throttle 
position sensor, you do have to utilize acceleration enrichment as well 
as deceleration enleanment.  The Map sensor alone cannot compensate for 
enrichment without some other device.  There is a delay in the pressure 
change versus air flow.  The sensor doesn't know until it's too late 
that the motor needs enrichment.  I've also tried just straight 
calculating the injector duty cycle from engine displacement, pressure, 
temp, injector size, air fuel ratio, and volumetric efficiency, and was 
never able to get it to work everywhere.  I ended up making a trim map 
to trim all the possible situations, and by the time I did that I threw 
the whole thing away and just made a map that uses straight injector 
openings then compensated for the other things later, much less 
headaches.

Sorry for rambling on, but I've just joined this newsletter and I guess 
I had a bit to say.

Thanks for the help in advance

Todd Knighton, Protomotive Engineering
Email knighton at concentric.net