Engine Hunts at Start and Then Rough Idle When Warm

[MikeWard]

There is no getting around the PRV hunting when cold.

With respect David, that is total bollocks

[David T]

With respect David, that is total bollocks

I am guessing here but I think "total bollocks" loosely translates to "total bullsh-t" in American English. Must be a British saying I assume. Anyway addressing your point, if you disable the idle motor and the Lambda system you can get the PRV to run pretty smoothly, even cold. My point in saying the PRV hunts is that with all of the stock engine systems in place and operating as intended, hunting is normal. The more you open the air screws and bypass the idle motor, the smoother it may run but the idle motor will have less and less control. It needs that control so that when you are running the A/C it can keep the motor from slowing down as the compressor clutch engages. It also helps to prevent stalling as you release the clutch going into first from a stop. Hunting is to be considered "normal" even by the specifications. If you refer to the Workshop Manual, in several places you will find the idle spec to be plus or minus 50 RPM's in recognition of the fact that the motor will hunt normally. It is not that the speed is supposed to be 775 plus-minus, it is that the speed will vary above and below 775 RPM's and still be within spec. I refer you to D:05:02 System Notes #5.

[Bitsyncmaster]

The stock idle ECU has no fancy crystal controlled clock and fancy digital logic. The idle speed tolerance would be the tolerance of the idle ECU components and the variable of aging and heat changing those component values.

[Elvis]

... It is not that the speed is supposed to be 775 plus-minus, it is that the speed will vary above and below 775 RPM's and still be within spec. I refer you to D:05:02 System Notes #5.

That's just your interpretation David.




Funny that nobody seems to understand what a PID controller does with a dead time.

I learned this an a college in California...any my friend Jordan recognized the dead-time / slow
reaction...

[NckT]

That's just your interpretation David.




Funny that nobody seems to understand what a PID controller does with a dead time.

I learned this an a college in California...any my friend Jordan recognized the dead-time / slow
reaction...

I understand PID control. I tried to explain the control function of the engine control in post no.19 in:

http://dmctalk.org/showthread.php?15...umpiness/page2

[lazabby]

It appears I cannot get to closed loop. I unplugged the WOT and I'm still getting a dwell reading of about 53.8 on the dwell meter (4 cylinder). The RPM ran about 1200 but once I plugged the WOT back in the RPM settled back down to about 800. I unplugged the CO2 and got the same reading. Plugged it back in and got the same reading. The CO2 is new. Perhaps the wire on the CO2 is bad?

[NckT]

It appears I cannot get to closed loop. I unplugged the WOT and I'm still getting a dwell reading of about 53.8 on the dwell meter (4 cylinder). The RPM ran about 1200 but once I plugged the WOT back in the RPM settled back down to about 800. I unplugged the CO2 and got the same reading. Plugged it back in and got the same reading. The CO2 is new. Perhaps the wire on the CO2 is bad?

You'll get the figure of approx 54 deg if the switch on the wide open throttle is closed OR the lambda temp switch is closed.

When the coolant temp is above 80 c, the lambda switch will then open (with the wide open switch is open), before the ecu will operate in closed loop mode using the oxygen sensor voltage input as reference, or it will default to 45 deg. The design is that the fuel computer will only operate in closed loop mode once the engine is at this temperature.

You could remove the fuel computer electrical plug and remove pin 7 that will remove the lambda and wot switch inputs to the ecu and then see if you get a varying dwell with a hot engine. OR on a hot engine, disconnect the wot and check to see if you get continuity between pin 7 and pin 5. This will determine if the lambda switch is open or closed. It should be open circuit on a hot engine.

I hope I've explained this ok so far.

Good luck Shannon

[Bitsyncmaster]

I agree you should get the wire disconnected to trouble shoot if it's the temp switch or the lambda ECU bad. Most owners probably won't attempt to remove a pin from the lambda ECU connector (although it's not hard to do with experience). You could also disconnect the wire at the bulkhead connector but those female pins are hard to pull from the connector. Another option is to cut the wire and then join it back when your testing is done.

Since your holding that 54 degree dwell it must be the temp switch or lambda ECU bad.

[Elvis]

You'll get the figure of approx 54 deg if the switch on the wide open throttle is closed OR the lambda temp switch is closed.

When the coolant temp is above 80 c, the lambda switch will then open (with the wide open switch is open), before the ecu will operate in closed loop mode using the oxygen sensor voltage input as reference, or it will default to 45 deg. The design is that the fuel computer will only operate in closed loop mode once the engine is at this temperature.

...

Nick, please explain - 80°C ?

the manual says 15°C ?!? D:04:10

before 80°C the ecu operates in closed loop mode ?

Sorry I'm confused. What am I gettting wrong ?


@ Shannon - for the O2 sensor to operate it needs 300°C. It then will generate a "high" voltage (1V ?!?) that
will activate the ECU into closed loop mode. Before that - it will stay at the cold / enrich PWM.
This can take a while and sitting in the garage at idle in cold weather - will take a long time.
I've seen that some weeks ago when I measured my lambda ECU and O2 sensor. I was close to replace the
sensor but before that I talked to a friend who is a car genious and runs his own shop. He told me to wait
or drive the car to heat the sensor up above 300°C.

[NckT]

I made a mistake, it is 15 c not 80 c as I said. (That's the last time I try and convert c to Fahrenheit in my head as I got that completely wrong too as it's 59 not 80 !

The o2 sensor will give a sinusoidal voltage signal between 0.2 to 0.8 volts.