Idle Speed Dips Quite Low Before Warmup

In my ongoing quest to ensure my 1991 Civic lasts forever:

Recently, before the car warms up fully, the engine RPM has begun to dip quite
low, near stalling I'd say, when I come to a stop sign or stop light. It comes
right back up again, but I am not accustomed to this dipping. I am not turning
on or off electrical accessories when this dipping occurs.

I suspect the Electronic Air Control Valve (EACV) is acting up, per
http://www.honda.co.uk/owner/ConcertoManual/62sk301/6-120.pdf and other reading.
Can people here who've had EACV problems share their experiences and thoughts
on my preliminary diagnosis?

Other info about my Civic:

--154k miles on it, manual transmission, no air conditioning, 1.5 liter, fuel
injected, 4-door sedan, driven in salty nasty northern U.S. winters for half its
life

--No indicator lights at the dash, suggesting anything is wrong, come on.

--Plugs, wires, distributor cap and rotor, ignitor, coil and PCV valve are all
about a year old or up to date. All OEM. I've been checking the PCV valve every
other gas fillup (since about March, and due to previous concerns about this),
and it's staying clean.

--I replaced the O2 sensor a few weeks ago. This seems to have stopped
oscillations that caused high RPM. I see only the dipping low RPM now. The new
O2 sensor is a Denso, just as the old, OEM one was.

--I added a bottle of Chevron Techron injector cleaner to the gas tank in May.
Changed oil two weeks later.

--Mileage is great: It continues over 40 mpg in the warm months. I've been
recording it after each fillup since early May.

--I took off the EAC valve (=EACV or a.k.a. fast idle valve?) last week to
inspect it. I did not fully disassemble it to inspect the cooling passages. The
EACV looked okay, though I figure the cooling passages (for the radiator coolant
that passes through it) could be fouled and might be behind the car's idle
problems. I do note that one EACV cooling line was chock full of coolant. What I
think is the return line was empty... hm. There is a return line, isn't there?
The one I think is the return line is at a higher elevation than the other line,
so I was thinking the coolant might drain out of it when the car is stopped?
Dumb notion?

--After the car is warmed up, the idle is better and appears in spec. (about 770
RPM, according to the dash tachometer, with the radio on)


I'm going to replace the throttle body gasket and O-ring and clean the throttle
body, with carburetor etc. cleaner in a week or so. I'm going to replace the
EACV gasket and possibly the little filter screen in it, too. Or, to go to
extremes, at my car's age, might it be worthwhile to replace the whole EACV?

I'm thinking a small leak in, say, a (vacuum?) hose that affects engine controls
may be the culprit. At my car's age, should I consider replacing all the
gas-containing (non-radiator coolant) hoses in and around the throttle body?
Who's done this?

How about the tandem valve? I am working from
Ihttp://www.honda.co.uk/owner/ConcertoManual/62sk301/6-162.pdf to troubleshoot
it, but if it's probably prudent to just replace it, I will.

IIRC, I read somewhere that I can use liquid dishwashing soap to identify vacuum
hose leaks. True? Also, I now have my Mityvac vacuum pump which I suppose I
might apply judiciously to test for leaks. Any suggestions beside reading the
manual that Mityvac provides?

I am going to place an order with Majestic online Honda parts soon for the seals
I mention above but also using input from this thread.

 

Have you checked the ECU for codes or gone through the diagnosis at
http://www.honda.co.uk/owner/ConcertoManual/62sk301/6-122.pdf

You should be able to easily verify coolant flow through the EACV. Can you
get clean water to flow through it when it's off the car?

I'm not sure why you've decided to do this. Have you determined that the
gasket was leaking? You should be able to check it by spraying some carb
clean on it while the engine is running. If there's a boost in the idle
speed then it's leaking. Note that you should perform this test with all of
the air cleaner plumbing attached in order to avoid a false positive.

You can clean the screen by carefully using a vacuum such as a shop vac. Be
careful not to apply too much suction (hold the vacuum hose at a distance
and slowly move it closer without actually contacting the valve). Or, to be
more cautious, remove any built up carbon with a pick while holding the
valve such that the carbon falls away from the valve. As you've probably
already determined, replacing the EACV is a little costly. It would be good
to test it first.

I've never seen or heard of anyone using soapy water. However, the carb
cleaner test that I mentioned above is standard practice and works quite
well.

 

Have you checked the ECU for codes or gone through the diagnosis at
http://www.honda.co.uk/owner/ConcertoManual/62sk301/6-122.pdf

You should be able to easily verify coolant flow through the EACV. Can you
get clean water to flow through it when it's off the car?

I'm not sure why you've decided to do this. Have you determined that the
gasket was leaking? You should be able to check it by spraying some carb
clean on it while the engine is running. If there's a boost in the idle
speed then it's leaking. Note that you should perform this test with all of
the air cleaner plumbing attached in order to avoid a false positive.

You can clean the screen by carefully using a vacuum such as a shop vac. Be
careful not to apply too much suction (hold the vacuum hose at a distance
and slowly move it closer without actually contacting the valve). Or, to be
more cautious, remove any built up carbon with a pick while holding the
valve such that the carbon falls away from the valve. As you've probably
already determined, replacing the EACV is a little costly. It would be good
to test it first.

I've never seen or heard of anyone using soapy water. However, the carb
cleaner test that I mentioned above is standard practice and works quite
well.

 

Per the above site I did go ahead and remove the Hazard fuse, wait ten seconds,
then start the car. Still no check engine light. I didn't see the little light
on the ECU (under the passenger side floormat) blink, either.

So to me, that eliminates any serious, "direct" problems with the electronics,
if I'm understanding the "drivability" fault detection system correctly. It
won't tell me much about inadequate coolant so as to cause the EACV maybe not to
function at optimum. It will tell me when a sensor's electronic "guts" have
failed completely.

Yes. I took the EACV's little cooling compartment cover off yesterday. I saw
some posts by you about rust and debris buildup in the cooling passages, so I
wanted to be thorough.

The compartment looked very clean. The two hose lines were clear. All I took off
the surfaces, and it took a Q-tip to see, was a light, grimy, grey film.

No. I just figured a throttle body cleaning might be a good idea on a car this
old.

But if you say that, like my brake job, 'this is likely overkill. The throttle
body will likely last the life of the car without a cleaning,' then I'll believe
you and skip it.

Yes, strike this baloney about the dishwashing soap. I was confusing this with a
test for other leaks in other machinery.

Using the carburetor etc. cleaner for leak tests is what I meant.

The two lines I talk about above are (doh! of course) for coolant supply and
return. I began to think more about why the one line was empty. I finally
decided that had to be wrong. I also remembered I'd replaced the thermostat in
March, and despite some air purging of the radiator system that I did then,
maybe I hadn't got it all.

The EACV sits rather high, the one coolant line sits even higher, so air might
very well accumulate in a pocket there.

In this hope, yesterday, I primed (filled) both EACV coolant lines and the EACV
coolant passage with coolant as best I could, re-connected all, and then did the
most thorough purge of air on the coolant system I'd ever done. I know it was
more thorough than ever, because I actually got the cooling fan to come on
twice! That took about 35 minutes of the car idling (to get the fan to come on
the first time) in an ambient temperature of about 77 degrees F.

In all my coolant changes for this car, I'd never waited this long. So I never
saw the fan come on before.

Someone posted here a few months ago that it took at least 20 minutes with his
car, thus my new patience with the effort.

Test drives yesterday and today seem to indicate way more stable idling. I'll
watch it for the next week or so but I think the air purge might have fixed the
problem.

 

Per the above site I did go ahead and remove the Hazard fuse, wait ten seconds,
then start the car. Still no check engine light. I didn't see the little light
on the ECU (under the passenger side floormat) blink, either.

So to me, that eliminates any serious, "direct" problems with the electronics,
if I'm understanding the "drivability" fault detection system correctly. It
won't tell me much about inadequate coolant so as to cause the EACV maybe not to
function at optimum. It will tell me when a sensor's electronic "guts" have
failed completely.

Yes. I took the EACV's little cooling compartment cover off yesterday. I saw
some posts by you about rust and debris buildup in the cooling passages, so I
wanted to be thorough.

The compartment looked very clean. The two hose lines were clear. All I took off
the surfaces, and it took a Q-tip to see, was a light, grimy, grey film.

No. I just figured a throttle body cleaning might be a good idea on a car this
old.

But if you say that, like my brake job, 'this is likely overkill. The throttle
body will likely last the life of the car without a cleaning,' then I'll believe
you and skip it.

Yes, strike this baloney about the dishwashing soap. I was confusing this with a
test for other leaks in other machinery.

Using the carburetor etc. cleaner for leak tests is what I meant.

The two lines I talk about above are (doh! of course) for coolant supply and
return. I began to think more about why the one line was empty. I finally
decided that had to be wrong. I also remembered I'd replaced the thermostat in
March, and despite some air purging of the radiator system that I did then,
maybe I hadn't got it all.

The EACV sits rather high, the one coolant line sits even higher, so air might
very well accumulate in a pocket there.

In this hope, yesterday, I primed (filled) both EACV coolant lines and the EACV
coolant passage with coolant as best I could, re-connected all, and then did the
most thorough purge of air on the coolant system I'd ever done. I know it was
more thorough than ever, because I actually got the cooling fan to come on
twice! That took about 35 minutes of the car idling (to get the fan to come on
the first time) in an ambient temperature of about 77 degrees F.

In all my coolant changes for this car, I'd never waited this long. So I never
saw the fan come on before.

Someone posted here a few months ago that it took at least 20 minutes with his
car, thus my new patience with the effort.

Test drives yesterday and today seem to indicate way more stable idling. I'll
watch it for the next week or so but I think the air purge might have fixed the
problem.