Sunday, 17 February 2008

diagnostic tech 92 ford tempo eeciv



Diagnostic Tech: '92 Ford Tempo EECIV

We're going to need to set the Wayback Machine for 1992. Ford was

using Electronic Engine Control System IV, or EECIV. By today's

standards EECIV lacks sophistication but in its day it certainly

performed its tasks very dependably.

On the plate today is a 1992 Ford Tempo w/2.3 engine and automatic

transmission. The complaint is an intermittent Service Engine Soon

light on and a high idle speed at that time. This system does store

trouble codes in memory but should only turn the SES light while a

code setting condition is present. In this case it just seemed to me

that a test drive to verify the complaint was in order.

Initially the car ran fine. No light came on and even though I was

driving in stop and go traffic there were no idle speed problems. I

decided to return to the shop and just run the EECIV self-tests. When

I pulled into the driveway the SES light came on. The idle speed was

fine until I stopped and put the car's shifter into park position. The

idle speed went to around 2000 rpm and stayed there for about three

seconds before slowing back to normal. I had just verified the

complaint. Time for the EECIV checks.

The car's test connector is at the left shock tower, near the battery.

Technically there are two connectors. One is a single wire called the

self-test input connector.

I'm going to mention that there are alternatives to retrieving codes.

You don't have to use a scan tool. That being said, you will find the

scan tool method more reliable with less chance of misreading a code

number or accidentally clearing memory codes.

It is important to understand EECIV self-tests before using them to

find a problem. Ford has the car's computer run specific tests and

report back areas that failed the tests. The failures are reported as

trouble codes. Running the tests improperly can give you misleading

results.

A quick warmup cycle for the engine and I'll get started.

After connecting the scan tool to the test connector and entering the

car data I want to start with "Quick Tests" and the Key On Engine Off

tests. If a sensor fails the KOEO test then it should be repaired.

You'll need a "Pass" or a no fault found condition in the KOEO

self-test before any KOER (Key On Engine Running) tests can be

reliably done.

The KOEO running test results are called "On Demand" codes. In this

case I get a Pass code 111 which means no problems were found. After

the On Demand codes are displayed the Keep Alive continuous memory

codes will be displayed. The Keep Alive, or continuous codes are codes

from the car's computer memory. Codes that have set in the past. This

car has a 332 in memory. A 332 indicates insufficient EGR flow. I

don't think a problem with the exhaust gas recirculation system it

going to be related to the high idle or even the SES on at that time.

An EGR not flowing should not cause the idle speed to be higher and

since the EGR is not supposed to be flowing at idle anyway I don't

think it would turn the light on at that time. I'm going to move on to

the KOER tests because that test should be done before getting too

involved with Continuous codes. The KOER test will reveal any sensor

failures that happen during the test.

The KOER test codes indicate a 332 again. This time I know the EGR

failed a test in real time and currently has a problem. Again though,

I don't think it is related to the idle speed problem. Two more codes

do indicate a problem with idle speed control, a 412 and 411. These

codes indicate the car's computer not being able to control the idle

speed. These codes do not store to memory so without running KOER

tests we would not see them. If you think that was a lot of trouble to

find that something is wrong with the idle speed you have a point.

What was learned though is that the several sensors that contribute to

the computer choosing an appropriate idle speed do pass self-tests.

That means I should start by concentrating my efforts in testing the

computer's ability to control the idle speed. In other words, I've

learned that the inputs to the computer are likely to be fine but

there is a problem with the idle control output circuit.

It's a good idea, I think, at this point to check the base idle and

minimum air adjustment. Before any throttle adjustments are made I

want to be sure I have a clean throttle plate. I'll remove the air

intake hose, the MAF and the air cleaner cover as a unit. The throttle

plate actually doesn't look too bad. It's pretty clean on the air

cleaner side. The air filter looks good and the air cleaner housing

was properly fastened.

When I open the throttle fully and look past the plate I can see major

carbon buildup. I can even see why the EGR flow is low. The EGR

passages are just behind the throttle body and fully plugged with

carbon. The best thing to do here is remove the throttle body, remove

the EGR valve and do some cleaning.

While we are in the neighborhood, this is a good time to remove the

Idle Air Control (or Air Bypass Valve) and clean that passage as well.

The engine computer uses this valve to control the amount of air that

bypasses the throttle plate. The amount of air allowed controls the

engine idle speed.

Once everything is clean and reinstalled it is a good idea to clean

the MAF sensor or at least see if it needs cleaning. Two tamper proof

screws hold the MAF to the MAF housing. A torx T20H will fit the screw

heads. Carefully remove the MAF from the housing. With a magnifying

glass and a good light you can look for dirt on the two tiny wires at

the bottom of the sensor. If you see dirt, carefully clean it. These

days there are spray MAF cleaners on the market. That wasn't always

the case so I still use the same Berryman Spray Carb Cleaner that I

always did. For stubborn spots I use a Q-tip as well.

This particular MAF looked pretty clean so after reinstalling it I let

the engine warm up again. I know that I now have a clean throttle, a

clean MAF, a clean IAC and a clean EGR passage but what I don't know

is if I've fixed any of the complaints. It's time to run the KOER test

again.

After getting the same results with the KOER self-test as before all

this work, I decided to see just what was happening at the Idle Air

Control as far as the computer control was concerned. The voltage at

the IAC signal wire should roughly match the position being commanded

by the computer. A command to open the valve 34% would have 3.4 volts.

As a load is placed on the engine the command would increase, so a

voltmeter could be used to check the command to see if the computer is

trying to slow the idle speed during the high idle problem. Another

method would be to disconnect the IAC wire connector from the valve

and see if the engine slows. You see, the valve is spring loaded to

return to minimum air position without a voltage command to the

contrary. In fact, it is by disconnecting the IAC and adjusting the

throttle position stop screw that minimum idle air is adjusted. All I

have to do is wait for the high idle problem to appear. After a few

minutes of running time the engine idle speed begins to race. When I

disconnect the Idle Air Control the rpms do not change. The IAC valve

is bad. I suspect a vacuum leak within the valve is the problem.

After replacing the IAC valve I now have no more racing idle and no

more idle control related fault codes. One thing very good about the

EECIV testing method is that I can rerun the test after a repair to

verify the problem is fixed. The idle problem is fixed! BUT the EGR

problem remains. I still have a 332 fault code in the KOER test.

During the KOER self-test the ECM will open the EGR vacuum regulator

solenoid very briefly. This allows manifold vacuum to pass through the

solenoid to the EGR valve. The vacuum should open the EGR valve. The

pressure in the exhaust tube that runs to the EGR valve is measured by

the PFE sensor. If the ECM has commanded the EGR solenoid to open yet

does not see the PFE sensor voltage indicate a pressure change the ECM

will assume the EGR is not working and will set the 332 fault code.

For testing purposes I like to use a length of vacuum hose and bypass

the solenoid, feeding the vacuum directly to the EGR valve with the

engine running at idle. If the engine nearly stalls then the valve is

passing plenty of exhaust flow.

I started to do that here but noticed that the hose that feeds

pressure from the EGR exhaust tube to the PFE sensor was broken. With

a broken hose the PFE sensor would not be able to monitor the exhaust

tube pressure change and would indeed set the 332 code. A replacement

hose and another KOER test and this time the repairs were done! A

complete system pass. No more light, no more high idle, no more codes!

posted by Deranger at 10:00 PM

1 Comments:

Blogger Mike said...

Hey Deranger!

Nice one, very nice!

8:23 PM EDT

Post a Comment

Links to this post:


No comments: