ECM Fault?

[Akselrud, Boris]

I did better.

I installed a cut out switch and a precision multi turn resistor which
allows to dial an exact voltage of a  MAP sensor at the time (by voltage
dividing between 0V and 5V). Then using the switch I substitute the resistor
instead of MAP sensor.
Idle smoothens for a sec or two, then it roughens again and the engine
always stalls after 10 sec or so.
I think this is because ECM program is designed to continuously lean the
mixture and watch MAP and O2 reaction. As soon as it senses lean condition,
It would compensate for it and add more fuel. Since It can sense no change
in MAP while leaning up the mixture, it continues to lean it, until the
engine stalls. That's why I need to complete my own fuel injection device
and replace ECM with it. Then I can figure out if there is a fuel mode which
is making the engine happy.

By the way this problem with the negative feed back loop(engine-ECM) and not
immediate engine reaction time is a common problem in negative feed back
control systems. Rarely it is possible to implement it without causing the
system to oscillate for at least a shortest time.

-----Original Message-----
From: Programmer [mailto:nwester at eidnet.org]
Sent: Monday, December 06, 1999 7:16 PM
To: gmecm at efi332.eng.ohio-state.edu
Subject: Re: ECM Fault?


Boris,

Get a vacuum pump and hook it up to your MAP sensor. Slowly apply 15 in to
22 in of vacuum to the MAP and note if the idle smoothens out. Let me
know...

Lyndon
-----Original Message-----
From: Akselrud, Boris <AkselruB at moodys.com>
To: gmecm at efi332.eng.ohio-state.edu <gmecm at efi332.eng.ohio-state.edu>
Date: December 6, 1999 7:54 AM
Subject: ECM Fault?


>That's me again.
>I am just one of those guys who don't take no for an answer.
>
>In short, I did find the anomaly with ECM fuel injector signals which is
>causing idle roughness and idle rpm fluctuation. As you all know, in TBI
>system two injectors are driven by pulses 1..1.2ms wide and spaced with the
>time interval required to make proper air/fuel ratio. What I found by
>looking at the signal with an oscilloscope is that there is an additional
>abnormal impulse approximately 0.1 .. 0.3 ms wide in between the normal
>impulses opening the injector. This abnormal impulse comes and goes,
>perfectly corresponding to idle rpm fluctuation. Even when I step at the
>accelerator pedal, up to a 2500 rpm I can see the abnormal pulses and
again,
>they are not always there, they come and go with 1..2 sec interval.
>I am also seeing about 70V spikes produced by inductance of the injector
>coil and when I put zener diodes to cut them, the engine would barely
start.
>
>I am also taking a second cut on the device I've designed to simulate
proper
>injector signals to the engine. My intent is to see if there is a fuel mode
>which would make the engine run smoothly. What I do is disconnect the
>injectors wires and connect them to my device with help of which I can
>provide accurate timing and play with duty cycle. However my current
version
>needs to be redesigned properly to absolutely prevent it from opening
>injectors for more then  1.2 ms, otherwise I have had a little too much
fuel
>in throttle body on occasion.
>
>What do you guys make of it?
>As you probably remember I am dealing with 1993 Cadillac fleetwood 5.7 TBI
>
>Boris.