Occasional Fail to Start when Warm - liberty / legacy

[Andy]

I posted most of this message some time ago. Since then I have
encountered an additional clue. The flashing of the Power light
indicator is new. Thanks to all who have had a go at this one,
particularly Carl 1 Lucky Texan.

My 1996 Liberty / Legacy 2.2 L Automatic has a problem starting once
the engine
has warmed. The problem was first noticed about 3 years ago. At that
time it only occurred about once every 2 to 3 months. As time passed
the rate of occurrence gradually increased. It is now likely to
happen more than once a week.

The symptoms of the problem are as follows:

" The car always, always starts from cold.

" There is never any indication of a problem when driving
(except for
occasional stall when idling).

" Occasionally the car fails to start once the engine is warm.
The engine cranks normally but there is no sign of firing.

" Once it has entered this failed state there is no spark at the
plugs.

" Once it is in this failed state it usually consistently fails
to start
for up to 20 minutes.

" Once it has eventually successfully started, it may then be
stopped and started immediately and consistently until the next
episode which may be one hour or up to several days away.

" If there is a stall from idling, then it often enters the
failed state (where it will not start for up to 20 minutes)

" The ECU has not indicated any diagnostic codes after a few
attempts to download.

" On three recent occasions, after a stall from idling, the car
has started with relatively little delay. On these occasions the
Power indicator light on the dash has then flashed on and off
approximately 10 times at a rate of approx 2 per second.



To date I have done the following:

Replaced both crankshaft and camshaft position sensors.
Replaced igniter with a Diamond unit.
Replaced HT leads.
Replaced main battery
Replaced ECU
Cleaned all visible earth points.
Coil primary and secondary resistance is always middle of acceptable
range when tested.
The main relay was opened today and its contacts look fine, and seems
to operate normally.

These actions have not had any effect on the problem.

3 Yrs ago there would have been about 70,000 Miles on the clock
Today it is 88,000 Miles

Any help in solving this riddle would be very much appreciated.

AndyH

 

[Wvd]

knock-sensor?

 

[Carl 1 Lucky Texan]

knock-sensor?

Seems unlikely to be the knock sensor, since, when running, the car
seems to run well. I suppose, filling with the highest octane fuel
available and clearing the ECU might be a reasonable test. If you
already run high octane, adding a few ounces of zylene or an
'off-the-shelf' octane booster and clearing the ECU might help with
diagnosis.

I don't know if your model is OBDII compliant. If it is, retrieving the
stored codes will be helpful. At present, we could guess IAC, TPS and
MAF. Also, having a good tech check the alternator for exessive ripple
might be a good idea. There real mystery to me is the missing spark
during failure. Something critical is missing or the ECU should start
the car in 'limp home' mode. I suppose it is POSSIBLE one of the new
items you installed is also bad. One of the position sensors or the coil
pack. I would try some 'freeze mist' or CO2 in a can, and chill each one
individually the next time the car fails to start and see if the problem
is cured significantly quicker than 'waiting'.

Have you posted this at www.ultimatesubaru.net ? There are a lot of
experienced folks there that are very good at keeping soobs on the road.
I mainly just parrot what I've read. Thosoe guys (some of which are
Kiwis and Aussies - so would have some idea of differences in your vs US
model) have a lot of experience.

Carl

 

[Carl 1 Lucky Texan]

Seems unlikely to be the knock sensor, since, when running, the car
seems to run well. I suppose, filling with the highest octane fuel
available and clearing the ECU might be a reasonable test. If you
already run high octane, adding a few ounces of zylene or an
'off-the-shelf' octane booster and clearing the ECU might help with
diagnosis.

I don't know if your model is OBDII compliant. If it is, retrieving the
stored codes will be helpful. At present, we could guess IAC, TPS and
MAF. Also, having a good tech check the alternator for exessive ripple
might be a good idea. There real mystery to me is the missing spark
during failure. Something critical is missing or the ECU should start
the car in 'limp home' mode. I suppose it is POSSIBLE one of the new
items you installed is also bad. One of the position sensors or the coil
pack. I would try some 'freeze mist' or CO2 in a can, and chill each one
individually the next time the car fails to start and see if the problem
is cured significantly quicker than 'waiting'.

Have you posted this at www.ultimatesubaru.net ? There are a lot of
experienced folks there that are very good at keeping soobs on the road.
I mainly just parrot what I've read. Thosoe guys (some of which are
Kiwis and Aussies - so would have some idea of differences in your vs US
model) have a lot of experience.

Carl

Oh yeah, the Power light flashing may be unrelated to the stall/restart
issue. But if you want to pursue that, it's a transmission problem -
MAYBE the speed sensor. check here;
http://www.surrealmirage.com/subaru/trans.html

Carl

 

[Andy]

Oh yeah, the Power light flashing may be unrelated to the stall/restart
issue. But if you want to pursue that, it's a transmission problem -
MAYBE the speed sensor. check here;
http://www.surrealmirage.com/subaru/trans.html

Carl

Thanks Carl. I'll look into that.
I've just read in my manual that a faulty neutral switch can cause
poor quality idling.
I'll have a look at that too.
Andyh

 

[Andy]

Oh yeah, the Power light flashing may be unrelated to the stall/restart
issue. But if you want to pursue that, it's a transmission problem -
MAYBE the speed sensor. check here;
http://www.surrealmirage.com/subaru/trans.html

Carl

I have now tried to check the O2 Sensor.
I am not certain that I am reading the correct lead-outs. (Black lead,
& white lead that is diagonally opposite on the connector)
The voltages I am getting seem to be quite different to the
recommended values.
When cold prior to starting I am getting variation between 35 mv and
300 mv. (should be a steady voltage between 100 and 200 mv)
When warm the voltage can be 0 mv for 5 seconds then jumping around up
to 400 mv for a few seconds, then falling again to 0 mv. It continues
to cycle like this. (it should be varying between 100 and 900 mv)
At one point the engine stalled and the voltage remained at 0 mv for
many seconds, it then jumped to around 300 mv then gradually fell to 1
mv over the next 2 minutes, and before I tried to start the engine.

If I am reading the correct sensor, this behaviour seems so far out of
normal range that I should definitely try a new O2 sensor next. My
manual is telling me I should have two O2 sensors, but I can only find
one. (Upstream of the catalytic converter)

Regards,
Andyh

 

[Carl 1 Lucky Texan]

I have now tried to check the O2 Sensor.
I am not certain that I am reading the correct lead-outs. (Black lead,
& white lead that is diagonally opposite on the connector)
The voltages I am getting seem to be quite different to the
recommended values.
When cold prior to starting I am getting variation between 35 mv and
300 mv. (should be a steady voltage between 100 and 200 mv)
When warm the voltage can be 0 mv for 5 seconds then jumping around up
to 400 mv for a few seconds, then falling again to 0 mv. It continues
to cycle like this. (it should be varying between 100 and 900 mv)
At one point the engine stalled and the voltage remained at 0 mv for
many seconds, it then jumped to around 300 mv then gradually fell to 1
mv over the next 2 minutes, and before I tried to start the engine.

If I am reading the correct sensor, this behaviour seems so far out of
normal range that I should definitely try a new O2 sensor next. My
manual is telling me I should have two O2 sensors, but I can only find
one. (Upstream of the catalytic converter)

Regards,
Andyh

O2 sensors do get 'lazy' in that they switch too slowly. And they can
cause problems without being 'bad'. How many miles on this sensor? Thay
can go bad as soon as 60K though many folks go 80-120K or more. If
you've ever had other emissions problems or a leaking headgasket, the
sensor can be damaged. The front one (the one you have examined) almost
always goes bad first as it is catching the most heat and the worst of
the un-catalyzed gasses. The rear one is supposed to assist in
diagnosing true converter problems. Bottom line, if it seems at all
suspicious, change it. NOTE! - many people have had problems with
aftermarket sensors. Some folks, no problem at all. Unless money is very
tight, get an OEM replacement.

Carl

 

[cobs]

Andy wrote:
[...]

I have now tried to check the O2 Sensor.
I am not certain that I am reading the correct lead-outs. (Black lead,
& white lead that is diagonally opposite on the connector)
The voltages I am getting seem to be quite different to the
recommended values.
When cold prior to starting I am getting variation between 35 mv and
300 mv. (should be a steady voltage between 100 and 200 mv)
When warm the voltage can be 0 mv for 5 seconds then jumping around up
to 400 mv for a few seconds, then falling again to 0 mv. It continues
to cycle like this. (it should be varying between 100 and 900 mv)
At one point the engine stalled and the voltage remained at 0 mv for
many seconds, it then jumped to around 300 mv then gradually fell to 1
mv over the next 2 minutes, and before I tried to start the engine.

If I am reading the correct sensor, this behaviour seems so far out of
normal range that I should definitely try a new O2 sensor next. My
manual is telling me I should have two O2 sensors, but I can only find
one. (Upstream of the catalytic converter)

Might be way off base, but have you checked the alternator? We had a
dodgy replacement on a 2k JDM rsk b4. Over time, symptoms were irregular
warning lights (bat / ! icons, no CEL) and hard/no starting when warm.
Inspection found a number of bulging components on the way out.


/C

 

[Andy]

Andy wrote:
[...]

I have now tried to check the O2 Sensor.
I am not certain that I am reading the correct lead-outs. (Black lead,
& white lead that is diagonally opposite on the connector)
The voltages I am getting seem to be quite different to the
recommended values.
When cold prior to starting I am getting variation between 35 mv and
300 mv. (should be a steady voltage between 100 and 200 mv)
When warm the voltage can be 0 mv for 5 seconds then jumping around up
to 400 mv for a few seconds, then falling again to 0 mv. It continues
to cycle like this. (it should be varying between 100 and 900 mv)
At one point the engine stalled and the voltage remained at 0 mv for
many seconds, it then jumped to around 300 mv then gradually fell to 1
mv over the next 2 minutes, and before I tried to start the engine.

If I am reading the correct sensor, this behaviour seems so far out of
normal range that I should definitely try a new O2 sensor next. My
manual is telling me I should have two O2 sensors, but I can only find
one. (Upstream of the catalytic converter)

Might be way off base, but have you checked the alternator? We had a
dodgy replacement on a 2k JDM rsk b4. Over time, symptoms were irregular
warning lights (bat / ! icons, no CEL) and hard/no starting when warm.
Inspection found a number of bulging components on the way out.


/C

I think the absence of any spark during the failure mode would tend to
rule out the alternator.

Andyh