Idle Fluctuation

Hello,

I've got a '91 Civic Si with the stock 1.6L MPFI setup and a manual
tranny. I put an AEM cold air intake on in August and I don't suspect
any problems from it directly. It has 136k miles and is well maintained
as of late (IMHO).

Over the past few days the car has developed an idle problem. I have no
check engine light indication. The ambient air temp. has been around
30-40 deg. F lately. The fast idle seems normal when I start the car,
but after I have driven the car up to operating temp. (either cruising
on hwy. or around town) the idle bounces repeatedly from 750 - 1250 rpm
in neutral with the clutch out. I also seem to be getting lower gas
mileage (low 30's vs high 30's).

I changed my coolant this summer, and never bled the system. I just
squished my upper rad. hose a bunch after filling it up and running it.
I know that the dashpot diaphragm that keeps the throttle plate from
slamming shut sticks sometimes, but I don't think it would be the source
of my problems here.

I don't have my Helm manual currently and am going to work on my car
this holiday weekend. So far here is what I am going to try:

bleed cooling system
check all vaccum connections
check for leak around intake manifold and throttle body gaskets
check EACV

Does anyone have any other ideas?

Thanks for your time,

Mark

 

I think I learned the somewhat-hard way that bleeding must be taken very
seriously, or air pockets and so inadequate cooling results around vital
components like the Idle Air Control valve (a.k.a. EACV).

I'd just add that it will likely take at least 40 minutes to get the
radiator fan to come on the first time (which the manual says to do when
bleeding) and so do a thorough, proper purge.

I did a bleed this summer, and the ambient temperature was 79 degrees F.
Forty minutes is about how long it took for the fan to kick on. One is
supposed to let it come on twice.

Remember that both www.autozone.com and the UK site
http://www.honda.co.uk/owner/workshopmanuals2.html
have detailed guides for your Honda. At the UK site for my 91 Civic (1.5
liter), I used the Concerto manual a lot. It's engine is the same or
identical to my Civic's.

I took off and disassembled my Civic's EACV this summer while
troubleshooting idle problems, but could have spared myself this trouble
had I first simply purged the cooling system of air.

 

After the engine heat up, the ECU goes into feedback mode during idling.
That's when the ECU uses the O2. I replaced the O2 when that happened on my
91 Lx.

 

Thanks for the advice Caroline. It seems kind of strange to me that it
took your car 40 min. for the fan to come on. I think mine comes on once
the car is nice and warm and it is sitting idling.

What do you think of the other poster's response about my 02 sensor? Is
it reasonable for it to fail after 14 years and 136k miles? AFAIK it is
the original and I have heard different sides about replacing 02
sensors. I've heard that it is the sort of thing that is either working
or it isnt, and I've also heard it can slowly degrade in sensing
quality. I could see a bad 02 sensor causing my problems, but I have
also have heard the many stories about air in the cooling system causing
similair problems. I guess I hope there's not something wrong with my
car (running too rich?) that caused the sensor to go bad if it is indeed
bad.

Could my loss in mileage be caused by having some air in my cooling
system? I may be really lucky, but I have had to drain and refill my
radiator a couple times with my current car and a couple with my old '91
Civic DX sedan, and I haven't really had idling/mileage problems
previously. I did my timing belt (and water pump) earlier this summer
and haven't had idle issues until now.

I did forget to mention that my idle is a little off with regards to my
power brakes. Stepping on the pedal causes my idle to dip maybe 100 RPM
, and then it adusts up to maybe 50-75 RPM below my normal (no lights no
brakes, etc.) idle. Maybe this is normal, but I don't remember it on my
old Civic. I have been figuring it is something with the booster, but
maybe it is related to my recent idle fluctuation problem. I have done
the check for vaccum at the check valve on the booster vaccum line and
it was okay.

It looks like I can get a Bosch OE style sensor for around $30 on the
net. If I can get one cheap locally I might just buy one. I just fear
the bolts on the manifold heat shield. I've never tried to unfasten them.

Mark

 

it could be the sensor, but it's likely you'd get a code in that case.
insufficient coolant however is a common problem. most people are in
the habit of just looking at the expansion bottle. if there's a small
leak, any contraction with the cooling that takes place on switch off
sucks air back through the leak, not fluid from the expansion bottle.
so the coolant level can easily drop without being obvious. and
unfortunately, the fans come on ok if there's low coolant because
they're switched from a seperate circuit from the engine temp circuit,
so again, it looks like everything's ok. the trouble is, that if the
/engine's/ temp sensor is not fully bathed in coolant, the ecu goes rich
assuming cold start mode, and gas consumption goes way up. it cooks
cats too, so you need to fix it asap.

 

I agree it seems strange. Yet several months ago someone posted here that
his Honda's fan had taken a very long time to come on when he was doing
some job on it in the garage. So I decided that, by gum, I was going to
wait at least an hour next time I did a purge.

After replacing the thermostat a few months ago my car was having idle
problems. I discovered the empty EACV cooling line. I went through the
cooling system purge steps and waited, as I mentioned, some 40 minutes. I
got the purge done exactly per the manual's steps. The fast/irregular idle
problem went away. The car seems runs great. It's never had a cooling
problem. It gets 40+ mpg nine months of the year.

I've heard otherwise: They do degrade in performance over time, and
degraded performance won't necessarily cause a check engine light (CEL) nor
be catastrophic. I think discussion of this is on the 'net, so you might
want to double check me on this.

But if the idle is messed up the way you say it is, and it were due to the
O2 sensor, I think I'd expect a CEL, like Jim said and from some commercial
shop experience I've had with other cars.

As it happens, I decided to be pre-emptive about this. Since I found an OEM
oxygen sensor for only about $45 total online, I decided to replace my
Civic's this summer at 154k miles. https://www.automedicsupply.com/. (This
was a big summer for my car. Sort of a mid-life tuneup.) Autozone loans O2
sensor wrenches at no charge. Some might say I wasted money, but I figured
I'd be somewhat up a creek without a paddle if the O2 sensor failed on a
long trip somewhere, and that this would probably be the only other O2
sensor the car would need for the rest of its life. So I went for it.

I do recall reading of people here who have the original O2 sensor with
over 200k miles and ten years on their car, and all's well.

On a 91 Civic, I don't think it's likely a cooling system with air in it
would mess up the computer so much that the exhaust would also be so messed
up it would damage the O2 sensor. The O2 sensor sits in the exhaust
manifold and I think it would take a very unusually foul exhaust to mess it
up. I'm really doubtful it could be messed up by any exhaust, from my
reading of its operation. But I haven't googled on this point. I do know
that if one gets anti-seize on the sensor while installing, that's another
story. There are warnings about this.

I see Jim's post and theory about how air in the system messes up the
engine's temperature sensor. This certainly seems like a possibility, too.
He's worked on way more Hondas and other cars than I have, too.

I like Jim's theory and also the EAC (electronic air control or idle air
control) valve theory. It seems reasonable to conclude that air in the
cooling system thwarts proper cooling of not just the engine as a whole but
sensors that determine how to run the engine.

There are several regular posters who advocate a proper purge of the
cooling system as a first step to troubleshooting this. I didn't understand
it well until I went through it with my own car this past summer. Now it
makes complete sense.

I just got introduced to vacuum power brakes recently. Your theory and
checks with the vacuum booster line seem to me to be barking up the right
tree, yet the problem persists. But I don't really have any hands on
experience to share on this one. Hopefully one of the regulars or some
other person does.

I can say I don't see this with my 91 Civic's brakes.

Yes, they can get a little tight. I replaced mine this past summer when
doing the O2 sensor job. I trust you know about PB Blaster, the miracle
penetrating oil. Fortunately the three bolts (on my 91 Civic) are fairly
accessible.

The web site I give above says $32 for an OEM Walker O2 sensor and $39 for
a Denso sensor, not counting shipping and handling.

My old O2 sensor had the (formerly Nippon) Denso "ND" logo on it, so I went
with the Denso sensor for my Civic.

Hey, ya know doing a purge is not a big deal. You can do it in a parking
lot with minimal tools and the appropriate compatible spare coolant. I'd
just go for it a.s.a.p. You know where the procedures are online, right?

 

Thanks Jim and Caroline!

Bleeding coolant did the trick. By the way, does the bleed valve leak
air around the threads when you are bleeding? I could always get air to
come out after a bubbleless stream left the valve.

Mark

 

I'm not sure I'm exactly following you. My impression is that the step where one
adds coolant with the engine cold and not running, such that a steady stream of
coolant comes out the bleed valve, does *not* ensure all air is out of the
system. Just a good deal of it. For one thing, liquids hold entrained gases
better when cold. For another, without the engine and so water pump running, air
bubbles will tend to stay lodged in high spots.

It's the subsequent heating of the coolant while running the engine and with the
radiator cap loose that gets the last bit of air out. At least, that's my take.

When I last did a bleed, I certainly saw air bubbles at the filler neck of the
radiator when I ran the engine. Made sense to me. Sort of the final purge of the
last bit of air.

BTW, how long did it take for your radiator fan to come on, and what was the
ambient temperature as you did this?

Glad to have the update. Good for the archives!

 

Indeed!

My neighbour across the road was a w/shop foreman and I wanted
to put an oversize washer on one of the heatshield bolts so I asked
him what the chances were of getting the bolt out with its head still
connected.

He advised to find a socket that was an absolute tight fit on the head
of the bolt even if it needed a firm tap to get it on fully. I found an
imperial one I had did just that. He said once the socket had been
'attached' to the bolt-head that one should apply torque progressively
increasing and to (attempt to) turn it in both directions. Do this
sensitively and the bolt will break its 'sticktion' and come out clean.

I did all of that and it did, indeed, come out clean!

HTH

 

Oxygen sensors do eventually get poisoned no matter how well the car is
cared for, that's why they do eventually need to be replaced. Some sources
even recommend replacement every 60K miles, which to me is overkill.

A LAZY sensor--one that is still operating but does not fluctuate its
voltage as rapidly as it should--will not necessarily set an error code,
but will cause you to fail emissions. It will not necessarily even affect
your fuel mileage or drivability either.

Google for:
lazy oxygen sensor

 

Those bolts never seize. The constant heat/cool cycle keeps them free.

The heads do rust though, so can appear undersize for the 12mm sockets that
they were originally made for. If so, a 7/16" will fit with a bit of
persuasion.