Technical Starter issue in freezing conditions

[nigelvan]

A weird issue has come up on my 2007 3.0: it won't engage the starter when it had a cold night...

Symptoms: 2 weeks ago the first sub-zero temperatures at night, in the morning I tried to start the van (no issues before, glow plugs replaced) and I only hear one solid click from the solenoid, the starter doesn't crank, nothing. No issue from the battery. I put a heat gun on the starter motor (stupid, won't do a thing) and try again. still nothing. 5 minutes later it starts like nothing happened, not a weak start, just a normal solid start. After 10 min I shut it off and it won't start again. I try a few more times and leave with another vehicle...

Next day I take the connections off the starter motor and clean everything. Lots of corrosion taken away with a gentle brushing of the dremel. I put everything together and add a bit of grease on top as a layer of protection. So I thought I sorted it. I had many more starts between then and now, no issues. So, case closed, I thought.

This morning it had frozen again and the van won't start again. Same symptom: solenoid clicks but nothing more. (I recall from the first time I hooked up a DMM and saw no voltage drop when the key is in the cranking position, so the starter is simply not drawing current). I jiggle the connection a bit and yes, it fires up right away. I shut the engine off and again, won't start. I jiggle the starter's cables from the top and hurray it fires up again. I drove 30 km on the highway and I decided to test it again when I arrived at my destination. Engine good and hot, lot's of bumps in the roads here... And again I have to shake the cables a bit before the starter cranks the engine.

Now, I know this all sounds definitly like a bad connection somewhere and not an issue with the starter, unless the shaking with the starter connection is a coincidence. But since I cleaned everything thoroughly I wonder if I'm missing some kind of weak spot in this system? I've seen there's a short lead between solenoid (where the starter cable connections are) and starter that is clamped on there and has no insulations and looks quite corroded that could be an issue. But shaking the starter cables/connections doesnt affect this lead...

 

[jackwhoo]

The short lead between solenoid and starter motor carries the high current from the internal switch in solenoid

 

[jackwhoo]

Does the small wire that triggers solenoid have a clean tight connection to the solenoid?

 

[nigelvan]

Does the small wire that triggers solenoid have a clean tight connection to the solenoid?

Jep, cleaned and brushed as well. The solenoid always engages. but it also draws less current than the starter itself I guess...

 

[jackwhoo]

Jep, cleaned and brushed as well. The solenoid always engages. but it also draws less current than the starter itself I guess...

Yes a solenoid draws between 15 and 30 amps roughly depending on type.

Your first post is a great description of symptoms.

Could you fit a temporary small wire to solenoid and touch it to the jump start point to see if symptoms reliably away over a number of days? If they do the fault is in the trigger circuit, if not then either starter motor assembly or main current wire.

 

[nigelvan]

Yes a solenoid draws between 15 and 30 amps roughly depending on type.

Your first post is a great description of symptoms.

Could you fit a temporary small wire to solenoid and touch it to the jump start point to see if symptoms reliably away over a number of days? If they do the fault is in the trigger circuit, if not then either starter motor assembly or main current wire.

Thanks for the suggestion @jackwhoo . Actually while re-reading the symptoms I realized, since the solenoid always engages, 12V must reach at least the solenoid. I think you mean by trigger circuit a switch in the solenoid that engages the starter once the pinion is engages to the flywheel.
I forgot to mention that I did tap the starter with a hammer as well the first time.
Another possibilty is a voltage drop somewhere that's not bad enough for the solenoid to fail, but is bad enough for the starter to fail. I guess the starter draws a few 100 amps?
Tommorow I'll be under there again trying to fit the new A/C belt, I snapped one of the long bolts that keep the compressor in place... So I'll hook up a DMM to the solenoid if the issue is present again. (There was no issue as long as temperatures stayed over 0, which is a weird variable in this situation)

 

[jackwhoo]

By trigger circuit I meant the circuit that includes the small wire that triggers the operation of starter.

 

[nigelvan]

By trigger circuit I meant the circuit that includes the small wire that triggers the operation of starter.

That small wire is the signal from the ignition switch I think, it seems that one is working. I found the schematic online here. I know it from my alternator issues.

 

[Communicator]

A weird issue has come up on my 2007 3.0: it won't engage the starter when it had a cold night...

Symptoms: 2 weeks ago the first sub-zero temperatures at night, in the morning I tried to start the van (no issues before, glow plugs replaced) and I only hear one solid click from the solenoid, the starter doesn't crank, nothing. No issue from the battery. I put a heat gun on the starter motor (stupid, won't do a thing) and try again. still nothing. 5 minutes later it starts like nothing happened, not a weak start, just a normal solid start. After 10 min I shut it off and it won't start again. I try a few more times and leave with another vehicle...

Next day I take the connections off the starter motor and clean everything. Lots of corrosion taken away with a gentle brushing of the dremel. I put everything together and add a bit of grease on top as a layer of protection. So I thought I sorted it. I had many more starts between then and now, no issues. So, case closed, I thought.

This morning it had frozen again and the van won't start again. Same symptom: solenoid clicks but nothing more. (I recall from the first time I hooked up a DMM and saw no voltage drop when the key is in the cranking position, so the starter is simply not drawing current). I jiggle the connection a bit and yes, it fires up right away. I shut the engine off and again, won't start. I jiggle the starter's cables from the top and hurray it fires up again. I drove 30 km on the highway and I decided to test it again when I arrived at my destination. Engine good and hot, lot's of bumps in the roads here... And again I have to shake the cables a bit before the starter cranks the engine.

Now, I know this all sounds definitly like a bad connection somewhere and not an issue with the starter, unless the shaking with the starter connection is a coincidence. But since I cleaned everything thoroughly I wonder if I'm missing some kind of weak spot in this system? I've seen there's a short lead between solenoid (where the starter cable connections are) and starter that is clamped on there and has no insulations and looks quite corroded that could be an issue. But shaking the starter cables/connections doesnt affect this lead...

I have only had time available to very quickly scan this thread. Years ago I had a similar but reversed problem, in that starter would not engage when the engine was hot. Electrical resistance increases with temperature, and my problem was due to the starter getting hot from the adjacent turbo, despite bing fitted with a heat shield.
I am wondering whether in this case the starter solenoid operation is marginal. Batteries do not like low temperatures, and if extreme climates battery heaters are used.
I am wondering whether the increased internal resistance of the battery at sub zero temperatures , is sufficient to drop the terminal voltage below the critical point for pinion engagement, and solenoid contact closure.
Are ther any aftermarket alarm systems connected into the solenoid operating circuit. Also consider poor connections at ignition switch, and elsewhere.

 

[nigelvan]

I have only had time available to very quickly scan this thread. Years ago I had a similar but reversed problem, in that starter would not engage when the engine was hot. Electrical resistance increases with temperature, and my problem was due to the starter getting hot from the adjacent turbo, despite bing fitted with a heat shield.
I am wondering whether in this case the starter solenoid operation is marginal. Batteries do not like low temperatures, and if extreme climates battery heaters are used.
I am wondering whether the increased internal resistance of the battery at sub zero temperatures , is sufficient to drop the terminal voltage below the critical point for pinion engagement, and solenoid contact closure.
Are ther any aftermarket alarm systems connected into the solenoid operating circuit. Also consider poor connections at ignition switch, and elsewhere.

Thanks @Communicator, I see you quickly scanned this one because I think you missed a few details, but no worries ;-). Short story:
- Battery is fine, either no crank, or cranking with full power (I had a bad earth before so I know how poor cranking sounds) - voltage stays ok during cranking
- Solenoid always engages (very solid 'clunk') - starter does not always crank
- Problems arise at sub-zero temps - weird since this shouldn't affect the starter
- Temporary solution is fiddle a bit with the connection, it seems, although connections have been checked and cleaned

I'll get back here after tomorrow when I have the chance to check some things. Chances are, if it isn't so cold there won't be any issues.

 

[nigelvan]

Problem solved! Put the van on the bridge today at my mate's place for easier access. We needed to get the A/C belt sorted. Turns out they shipped the wrong size... Anyway a nice trick is to have one guy put the belt halfway on the A/C compressor and then put the car in 6th gear and turn the wheel whilst blocking the other one. So you can turn the crank without a special tool that fits on the crank pulley. No special tools needed.

Anyway the van wouldn't start again so on we went with the multimeter. Took off the starter but everything seemed to be working fine on the bench. Turned out the cable between junction box and starter was corroded at the starter. We measured 0.4 ohm over that cable so we took off the L shaped cable lug, cut a bit off the corroded cable and put it back on. Seems to be starting fine now. Makes sense that jiggling the cable before did solve it temporarily.

So the solenoid was still working through the voltage drop, but it was bad enough so the starter itself wouldn't even do anything...

 

[jackwhoo]

Was it the non insulated wire between the solenoid and motor or the large red insulated wire?

 

[Communicator]

Problem solved! Put the van on the bridge today at my mate's place for easier access. We needed to get the A/C belt sorted. Turns out they shipped the wrong size... Anyway a nice trick is to have one guy put the belt halfway on the A/C compressor and then put the car in 6th gear and turn the wheel whilst blocking the other one. So you can turn the crank without a special tool that fits on the crank pulley. No special tools needed.

Anyway the van wouldn't start again so on we went with the multimeter. Took off the starter but everything seemed to be working fine on the bench. Turned out the cable between junction box and starter was corroded at the starter. We measured 0.4 ohm over that cable so we took off the L shaped cable lug, cut a bit off the corroded cable and put it back on. Seems to be starting fine now. Makes sense that jiggling the cable before did solve it temporarily.

So the solenoid was still working through the voltage drop, but it was bad enough so the starter itself wouldn't even do anything...

If I have understood correctly, the main cable connecting to the CAL4 starter motoer fuse at the battery was the faulty one. This would not affect the operation of the solenoid, as the solenoid current is routed separately via the engine bay fusebox, and ignition switch.

 

[nigelvan]

For clarification, it was the cable between junction box at the left hand headlight, D004 on the schematics, it connects to the positive jump start point, and the connection on the starter solenoid, so NOT at the battery box. At the point where that cable and the alternator cable are connected together with big cable lugs (L-shaped) to the starter solenoid.
The non-insulated wire between solenoid and motor was fine.

the cable is about 35mm2 and when we took the cable lug off we noticed that the strands of that wire had so much dust and stuff on them that probably only the outside strands touching the lug were conducting current. we cut off about 2cm of wire to get to some fresh copper (the wire is long enough) and put the same lug back on after cleaning it using a bench vice.

 

[Communicator]

For clarification, it was the cable between junction box at the left hand headlight, D004 on the schematics, it connects to the positive jump start point, and the connection on the starter solenoid, so NOT at the battery box. At the point where that cable and the alternator cable are connected together with big cable lugs (L-shaped) to the starter solenoid.
The non-insulated wire between solenoid and motor was fine.

the cable is about 35mm2 and when we took the cable lug off we noticed that the strands of that wire had so much dust and stuff on them that probably only the outside strands touching the lug were conducting current. we cut off about 2cm of wire to get to some fresh copper (the wire is long enough) and put the same lug back on after cleaning it using a bench vice.

Thanks for the clarification, but perhaps for D004, I should be reading A005.