Car will not crank sometimes

JERRY SCHAUSS

MEMBER
2002 BUICK CENTURY
3.1L
6 CYL
FWD
AUTOMATIC
110,000 MILES

My car will not crank sometimes. When this happens you can try to start it a number of times and it will still not crank. Come back in a few minutes and it will start with no problem. You may go a week without any problems and then problem returns. Engine can be hot or cold from over night and this makes no difference to when problem occurs. My neighbor also has a Buick with the same problem. He replaced the starter (new starter not rebuilt) and still has the problem.

Battery has been replaced and contacts cleaned.

Commutator interrupter switch has been replaced (part of ignition switch, not a easy job to replace).

Security light is working correctly, security key okay.

Dash lights, radio, A/C – heat fan, head lights, etc, work correctly.

Starter has been checked by local auto parts store a few times.

Getting twelve plus volts to starter solenoid.

Connected a Mallory Sonalert to the twelve plus volt starter solenoid wire so that you can hear when car will not crank. So I know I am getting the voltage to the solenoid.

5 Replies

CARADIODOC

MECHANIC
33,916 POSTS

When you do your testing, use the engine block for the ground point for any test equipment. Measure the voltage on the large cable on the starter solenoid. You should find twelve plus volts there all the time, but it will get drawn down to around ten volts during normal cranking. If that drops to zero volts when trying to crank the engine, that cable has a bad connection, but the clue is you will hear the single, rather loud clunk of the starter solenoid. If that voltage drops to zero volts and you do not hear the solenoid, suspect the ground cable has a bad connection. Typically the strands of wire corrode away under the insulation where it is not easy to spot.

SPONSORED LINKS

Was this

answer

helpful?

Yes

No

+1

Wednesday, August 9th, 2017 AT 4:15 AM

JERRY SCHAUSS

MEMBER
3 POSTS

After reading your reply and more investigation into the problem, I think the problem has been resolved. I always have the voltage at the starter solenoid when I try to start the car and the solenoid does not engage. Checked the ground cable with an ohm meter and moved the cable around to check for high resistance, looks ok. Disconnected the battery and then removed the starter solenoid assembly to bench test assembly. Was able to get the loud clunk of the starter solenoid as it engaged the bendix and the starter worked correctly. Suspecting a problem with the solenoid I took a resistance check of the solenoid coil. Resistance looked ok but then I noticed a cold solder connection on the coil wire. See photos of the problem found.

Cold solder connection to right of large bolt to starter.

Cold solder removed to show solenoid coil wire not soldered to terminal. The small coil wire is just bent over the terminal. I’m sure this is the problem as this would cause an open coil wire from time to time. Soldered the connection (forgot to take picture) and checked operation on the bench and then installed the starter assembly in the car. Seeing how the problem was intermittent, it will take a few weeks to prove this was the only problem. This is a problem you would not expect to find, but you never know what you will find with an intermittent electrical problem that happens once or twice a month.

Images (Click to make bigger)

Was this

answer

helpful?

Yes

No

Friday, August 11th, 2017 AT 6:10 AM

CARADIODOC

MECHANIC
33,916 POSTS

Dandy. I've done similar things with Ford's fender-mounted solenoids to create "bugs" for my students to diagnose. You found what they could find after four weeks of classes.

As a side note, resistance tests are only good to verify you found a problem or to check for a completely broken connection. In high-current circuits like this one, resistance readings aren't that useful. A starter can draw up to 300 amps to get started spinning, and then it will drop to 150 amps or less. The entire system will be rendered ineffective if there is as little as.04 ohms total in the ground cable and starter cable. That is WAY too small to measure, but we can measure the results of that resistance with voltage drop tests. Your meter's leads will have two to five ohms of resistance which adds to the confusion.

Very happy to hear you solved this.

Was this

answer

helpful?

Yes

No

+1

Saturday, August 12th, 2017 AT 11:31 PM

JERRY SCHAUSS

MEMBER
3 POSTS

Thanks for the side note information. I will do a voltage drop test on the cables to make sure there is not another problem brewing. Thanks for the great help on this problem.

Was this

answer

helpful?

Yes

No

Monday, August 14th, 2017 AT 6:15 AM

CARADIODOC

MECHANIC
33,916 POSTS

Happy to help. For the voltage drop tests, it's the mechanical connections we're interested in. Those are the connection between the battery's positive post and the cable clamp, and the terminal on the other end and the stud on the starter solenoid. Same battery post and cable on the negative side, and between the terminal and engine block on the other end. We don't count where a cable is pressed into the lead cable clamp or the terminal, but those get included in the tests.

You can start with just two measurements. Put one meter probe on the battery's negative post, and the other probe on a paint-free point on the engine block. Set the meter to the lowest DC volts scale, then read it while a helper cranks the engine. The most you're allowed is 0.4 volts in that entire circuit, and no more than 0.2 volts at any one connection. Even if you find 0.39 volts or less, you should test each connection separately to be sure neither is over 0.2 volts.

The resistance in these connections is way too small to measure, but these voltage drop tests allow us to see the RESULTS of that resistance. On the positive side, put one probe on the battery's post, (not the clamp), and the other probe on the stud on the starter solenoid, (not the terminal). Same 0.4 volts is the maximum allowable. If you find 0.2 volts or less, there's no need to test each individual connection. Since they can only be less. It's when you find more than 0.4 volts that you need to check each connection in that circuit to see which one has too much resistance.