99 Accord Has Me Stumped

For almost a year I have been struggling with a recurring problem that
appears to be triggered by sustained driving on the highway. It has come up
three times now and I thought I had it fixed after the second episode so I
don't know what to try next.

It's a 99 Accord LX 4 door 4 cyl 5-speed, 184,500 miles (lots of highway
driving).

The problem is after the engine hits operating temperature, it cannot idle
at the bottom end -- when you slow down it just drops off to zero RPMs and
dies unless you feed it gas with the clutch down. It first happened during
an all-day drive last summer. I had just changed the oil, filter and air
filter for the trip. To try to fix it, I replaced the spark plugs and the
PCV, ran some injector cleaner through, and replaced the coolant and the
ignition coil. That seemed to help a little and the car stopped dying so
much but still ran uncertainly at low RPMs. Then I took another long drive
(this time a warm day in December) and the same problem came back. This
time I switched out the ignition wires, distributor cap and rotor. Lo and
behold the car seemed almost back to normal. I have driven it up to 50
miles at a time without incident. But this weekend my son drove over an hour
on the highway (warm abient temps, too) and sure enough, it started dying
again.

There are no trouble codes being activated. I'm looking for a good idea on
what to try next and this group is a great help. Thanks in advance for your
suggestions.

 

I am no expert, but I suggest you look at ignition and fuel (filter, timing,
plugs, module). When was the last time you tuned it up?

 

check for the great ignition switch recall. applies to your year.

[low revs cause vibration which can make the switch contact "noisy".
always return the key to the "run" position by hand, don't rely on the
string to return it fully. and make sure you have no other heavy
objects on the key ring.]

 

Jim, You're right. I had the recall service done at the time it was
announced. Thanks.

 

Thanks, Spam. The fact that it always waits to start dying until heating up
suggests to me that something electronic is happening that causes the
computer instructions on where to set low idle get interrupted somehow. The
ignition, timing and idle all work perfectly fine while the car is warming
up. Also, they have been working fine on short trips for the last few
months. The problem only re-occured after a longer drive at high speeds.

 

Maybe you should monitor the reference voltage going to the Throttle
Position Sensor(TPS)and the TPS output voltage.Perhaps the ref V is
dropping(it's source is the ECU),or the TPS output is changing (bad TPS)and
telling the ECU to lower the idle....to a point where it will not run
properly.
Take a "cold" reading on the reference V,then another when the engine
starts to stumble. You'll have to backpin the connector.

Or maybe the igniter (in the distributor)is the problem;perhaps altering
the firing of the coil once it warms up.

 

Thanks, Jim. Your suggestions sound promising.

Would high or low voltage out of the TPS trigger a trouble code? (I haven't
had one with this problem.)

The igniter also sounds like a possible answer, especially given it's
closeness to the hottest part of the engine. But replacing it will be a
$130 gamble. Is there any way to test the igniter?

 

yes. it's written up at tegger.com. but if the vehicle runs at all,
it's unlikely to be the igniter.

bottom line, if working on the distributor scratched the itch, and it
seems to have, keep poking about with that. suggest removal, cleaning,
and thorough inspection. read all of tegger's write-ups on ignition
problems.

also, make sure you use denso or ngk plugs.

 

the OUTPUT is supposed to vary with throttle position,0-~4.5V.If the
Reference input V went high(more than 5V),I suppose that could "set a
code",but I doubt it.
A code is probably generated if there's NO output V from the TPS.
Like if the ref V input failed or the TPS went open.

If anyone knew the value of the variable resistor in the TPS,one could sub
a pot for it and see if that fixed(steadied) the idle.
Maybe a 10K or 50K linear pot.

Scope it's output,with an oscilloscope. See if the pulse train is too jumpy
or erratic.
Since most people don't have those instruments,the only other way is to
swap it out with a good one.

 

there's an IC and a switching transistor in there that could be affected by
heat or temperature sensitive.IIRC,there's white heat sink grease used to
help heat transfer from the ceramic substrate of the igniter to the metal
of the distributor.maybe that grease is missing?

Someone else mentioned using only OEM Honda distributor cap/coil and
rotor,that aftermarket parts caused him a lot of wacky trouble.

 

if it is, it's unlikely to be the igniter. the igniter's just an
electronic power switch. like most other semicons, they usually either
work, or they don't.

kinda related to this, i'm a big fan of premium coiled core spark plug
leads. they suppress a lot electrical noise under the hood that the
rubber core leads don't. this means less noise for the sensor inputs to
filter and i think it makes the vehicle run much more smoothly all ways
around.

 

Since the igniter contains an IC to adjust the "dwell" time of the
transistor that is the actual power switch,it's possible for that IC to get
flaky at a high temp,or even at a narrow temp range.With electronics,
"either works or it doesn't" is not always true.

also,doesn't the igniter feed back data to the ECU,thru the tach output?

 

the added complexity compounds the problem, not mitigates it. ever had
computer componentry flake on you? one minute it's there, next it's
gone. pow.

no, the ecu relies on the sensors in the distributor for rpm input. if
you think about it, the tach line is just a phased version of the ecu's
own output - not much good for ecu feedback purposes.

 

Jim Yanik said the following on 6/13/2007 12:01 PM:

I had a similar idle problem with a Dodge Caravan (3.0 liter Mitsubishi
Engine) that turned out to be a gummed up throttle body. There were no
engine failure lights, the idle speed would just drop to zero.