Mechanics

CAR WON'T START WHEN ITS EXTREMELY COLD OU

1998 Honda Civic

Engine Mechanical problem
1998 Honda Civic 4 cyl Front Wheel Drive Automatic 114000 miles

At temperatures under 20 degrees for an extended period of time the car won't start-more specifically the engine won't fire. It sputters and attempts to start but will not fire completely up without assistance and the longer it sits without running in extremely cold weather the higher the probability that it won't start at all. I have been able to get around the issue by placing a small space heater under the hood for 30 minutes or so prior to attempting to start and it usually does the trick, although the initial try after doing this is met with resistance, it will eventually start but have a rough time getting there. The other way I found to combat the issue is to start it and let it run for a little while at regular intervals (if I can) usually about every 8 hours but that is shortened depending on the severity of the cold temperature. Otherwise any other weather conditions and the car starts normally and without issues. I'm leaning towards a problem with either a sensor or a component in the distributor. I know from past experience that there is both a coil and and an ignitor in the distributor but I don't know which of those components this issue seems more indicative of being the culprit so any suggestions there would be helpful. I'm hoping to avoid the ritual of replacing everything to solve the problem if I can focus on one area or component that is the most problematic so I welcome the forum's thoughts in helping me get it fixed.
Avatar
KW77
January 3, 2010.




Hi KW77,

Thank you for the donation.

Difficult cold starting can be due to various causes. As you mentioned heating up the engine compartment works, the more likely cause would be the Engine Temperature Sensor.

Get it check to see if it is within specs. I could not find any data to manually test the ECT so you would need appropriate diagnostic equipment to test it.

Does the Check Engine Lamp indicates when the engine is running?


http://www.2carpros.com/forum/automotive_pictures/192750_EngComp98Civic16LFig01_1.jpg

AD

KHLow2008
Jan 4, 2010.
In response to your follow up question, no, the Check Engine light has never come on during this issue nor for anything other than the normal sequence when the key is put to the on position.

Doesn't the ECT sensor just control the cycling of the cooling fan as not to allow the engine to over heat? Would this sensor interfere with the ability of the engine to receive spark or fuel? Also, I tried to elude to this in my original posting but during the process of getting it started with this issue it sputters and fires intermittently but won't completely start up. Would a defect in that sensor cause it to behave in that fashion?

Tiny
KW77
Jan 4, 2010.
The fan switch is for turning the fans on when the coolant reaches a pre-fixed temperature.

The ECT is a separate sensor for informing the PCM the engine temperature to provide the correct amount of fuel according to engine temperature. Cold engine requires a richer fule mixture whereas a warm engine requires a leaner mixture,

If the problem is fuel repalted, it could be the ECT that is not sending the correct signal for the PCM to control the injectors to spray the correct amount of fuel for ignition.

You can try pumping a little starting fluid into the throttle body and see if it is a fuel issue. Get the fuel pressure tested as well.

If it is an ignition spark problem, it could be that the sparks are too weak whe cold and has difficulty igniting the fuel. A test of the ignition coil would be recommended.

KHLow2008
Jan 4, 2010.
That makes sense as its consistent with the starting behavior I've described. I had also tried the starting fluid method as well which has worked somewhat in the past instances of this problem with mixed results but lately heating the engine compartment has seen more of a consistent way to overcome the problem. I will focus my efforts in the areas you've highlighted as the ECT sensor is an inexpensive part and the coil can be isolated with testing and its the least expensive of the distributor components.

If I have any further questions I'll continue the thread otherwise I will followup if I'm able to rectify the issue so you have more knowledge in future cases. Thanks.

Tiny
KW77
Jan 4, 2010.
If you have follow up, just post reply here and I would get back to you once I receive notification.

Have a great day.

KHLow2008
Jan 4, 2010.
I wanted to followup on the post and get your thoughts about some of my findings thus far.

*Note: the weather hasn't been cold enough to recreate the issue since the last post.

Following your suggestion I used the service manual to find the parameters for testing the ignition coil since I was leaning away from it being a fuel mixture problem. I got the following readings from the coil tests: Resistance between A & B terminals = 1.3 ohms (spec should be between.68-.77 ohms which is out of spec)

Resistance between A & Secondary terminal = 15.9k ohms (spec should be between 12-19 ohms which is in spec)

However, at the time of the testing I got a temperature reading of 90 degrees from the metal portion of the distributor housing and the manual suggests that the optimum temperature for testing the coil itself should be at an ambient 68 degrees. Do you think the slight increase in temperature would cause the one reading to be out but the other remain in? Do you think the out of spec reading is accurate therefore rule that the coil (likely being under-powered) is the issue?

Also, while reading through the service manual I found a section dedicated to the ECT. It says that this sensor on this model is the on/off type in that once the coolant temperature reaches 100 degrees it closes the circuit and suggests a warm engine mixture. That would lead me to believe a way to rule out a defect in this component would be to disconnect it when the issue occurs forcing the ECM to see a cold reading. Would you agree this would be a way to eliminate this component from the problem?

Your thoughts.

Tiny
KW77
Jan 14, 2010.
I would agree on your findings that the ECT is most probably not the cause.

As to the ignition coil test, the recommended testing temperature would give you the most accurate results.

Any difference in temperature affects the readings cannot be used as an accurate gauge. I would suggest testing it again at the recommended temperature.

If the reading is out at the recomended temp, then the coil is weak and could be the cause of the problem.

KHLow2008
Jan 15, 2010.

AD