Engine Troubleshooting: Lean when the fan turns on

[Nick H]

As our car is idling, when the fan turns on the Air/Fuel ratio becomes extremely lean, mis-fires and runs horrible. According to the PE ecu, the injector open times is staying constant, and fuel pressure is staying constant. Have any other teams experienced a problem like this? What could be causing it?

Also, when the fan turns on, we see a voltage drop of 0.7 Volts, which is reasonable.

Any ideas?

[Nick H]

As our car is idling, when the fan turns on the Air/Fuel ratio becomes extremely lean, mis-fires and runs horrible. According to the PE ecu, the injector open times is staying constant, and fuel pressure is staying constant. Have any other teams experienced a problem like this? What could be causing it?

Also, when the fan turns on, we see a voltage drop of 0.7 Volts, which is reasonable.

Any ideas?

[Dan Deussen @ Weber Motor]

Nick,

does the PE ECU have a compensation table for injector dead time based on battry voltage?

The problem here is that when the battery voltage goes down the injector dead time (time it takes from when the injector sees voltage until actually fuel passes through it) will increase. So when the voltage drops, the injection time needs to increase for the same A/F ratio.

This is especially important on semi-sequential=batch-fire systems where the injector is fired twice as often and with only about half the opening time as on a fully sequential system.

[B Lewis @ PE Engine Management]

Nick,

I agree with Dan's comments. There is a batterry voltage compensation table that you can adjust in the software that will take care of this.

Good luck!

[Pete M]

Yes, getting your dead time comp right is something worth spending time on. Ideally, you'd do it before spending much time tuning your base map, because changes in dead time comp will change that pretty majorly. The effect of the comp being wrong is most pronounced at low pulse widths, hence why you're noticing it at idle.

I can give you some tips if you want.

[VFR750R]

Make sure you have good grounds and no feedback from the fan. Otherwise probably related to dead time.

Could maybe also run ecm with electric motor attached to another flywheel and physically measure fuel injected over time with or without fan on. If engine will continue to run, could also just pull one injector off to measure flow.

One other thought, with no spark, O2's will read lean, maybe your coils aren't grounded well and don't get enough voltage when the fan comes on.

[Nick H]

in regards to our last question, for the most part, the battery voltage compensation took care of those problems.

Another problem we are encountering is at a constant TPS, constant RPM, 14.8V, general steady state conditions, we are experiencing oscillations of the injector open times which causes the engine to run extremely rich. The only compenstation that we have occuring at that instant is air temp (seen in picture 2). Why is the injector open times oscillating? At almost every other point in the map that we have looked at, the injector open times have been constant, or acting appropriately under certain compensations.

Pic 1 showing the oscillating injector open times (blue) and the red is air fuel ratio (0-5V, so 1.5 is 13:1 and 1 is 12:1)
http://img149.imageshack.us/img149/6826/cpp1pi7.png

Pic 2 showing a table of our compensations. The only one acting at this instant is Air Temp which should not be causing the open times to oscillate.
http://img99.imageshack.us/img99/5715/cpp4dz4.png

[Pete M]

What is the shape of the main fuel map in that area? Any sudden discontinuities (like the one i can see in the graph?). Are all your compensation tables smooth looking? Are the values coming from the sensors (eg air temp) sensible and relatively stable?

I'm not familiar with your ECU, we're a motec team. On the motec, you can separately log the result of the base fuel map, the results of all the compensation tables and the overall fuel pulse width. If your ECU can do that, it'd probably help find which one is fluctuating.

[Nick H]

Yes, there are some large discontinuities however we are not operating near them during this test. We have not taken the car to the dyno, so we are just working on the low-load portions of the map by revving the motor in neutral.

The ECU is sensing load by TPS only (not map) and for some reason the injector open times seem to be erratic.

[Pete M]

What does the output of the TPS look like? If the pot has crap in it, the signal can bounce around a fair bit. Loose connections can also do it. An intermittent break in the circuit or short to 5V or ground would briefly flick the TPS readings to another site. This may happen too quick for you to notice in the real time display. If you're running any form of acceleration enrich, this will exacerbate the problem.