Strange starting problem

[rattlenbang]

Hey all, Ive got a 2002 with 360k km that started giving starting problems a few months ago, getting progressively worse. It doesn't seem to matter of its warm or cold, maybe more often cold, but there's times when it starts right up, other times when it won't or I have to crank and crank to get it to go. It's not glowplugs because those check ok and no engine code comes up. They're also getting voltage, as is the cutoff solenoid. I've checked and fuel is getting to the injectors when cranking, and the air valve in the intake is wide open.
The engine never dies once started, just hard to start sometimes.
It has a new battery and seems to be spinning the engine over quick enough.
I wondered about compression, but it runs great, no smoke, and most times it lights up immediately even when stone cold. You can then start it, run it for 20 min, shut it off, then it won't start again.
I've dealt with bad glowplugs and bad cutoff solenoids before, but this is strange in that it's only starting causing the problem but it's not those components. A weak battery can cause difficulty starting but that's not it. If it was a worn engine it should always be hard to start. If there was something in the injection pump or fuel supply it should affect running performance. Weak starter should be obvious as slow cranking. I'm getting stumped with this one. My son opened up the injection pump to replace a seal awhile back which might have done something, but hard to see what.
Any ideas?

 

[Mongler98]

Check to fuel going though the clear line when its struggling.
Seriously sounds like a leaking fuel line. Aka braking prime just enough to drain some of the fuel out of the filter.
There are check valves at the tank and your filter could be leaking air..

That's where I would start.
109 relay and battery test are also a area to quickly rule out.

 

[Mpaw]

I would have thought something like fuel or timing, assuming glow plugs etc. are ok. Obvious is the injection pump. If you've got a vcds, have a look at the timing given that it's been getting progressively worse.

 

[Mozambiquer]

I second checking timing with vcds. If it's off, it'll be hard to start.

 

[rattlenbang]

Ive got a a VCDS, but how is it used to check timing? Isn't that a mechanical measurement with a dial gauge off the ass of the IP?

 

[Mozambiquer]

Ive got a a VCDS, but how is it used to check timing? Isn't that a mechanical measurement with a dial gauge off the ass of the IP?

TDIGraph - TDI Timing Checker

 

[Vince Waldon]

Ive got a a VCDS, but how is it used to check timing? Isn't that a mechanical measurement with a dial gauge off the ass of the IP?

Pre-TDI pumps did it that way because there was no computer to ask. For TDI pumps we can just talk to the brains of the operation.

Also on the TDI fuel temperature is an important element of setting the timing, hence the handy graphing tool rather than a straight measuring block.

 

[rattlenbang]

Pre-TDI pumps did it that way because there was no computer to ask. For TDI pumps we can just talk to the brains of the operation.

Also on the TDI fuel temperature is an important element of setting the timing, hence the handy graphing tool rather than a straight measuring block.

That's cool, I didn't know that. But here's the thing: if it were an adjustment issue like this, wouldn't it always be hard to start? While the computer is supposed to monitor things and an electrical fault should show up as a CEL, the way it fires fine one time and the next time very hard to start, sounds a lot like an electrical fault to me.
Is there a cold start valve on these IPs? Can there be an electrical fault causing mistiming, that the computer might not be able to detect? I'm quite familiar with the various Bosch rotary pump versions, but all mechanical, not the later electronic controlled.

 

[Vince Waldon]

In general, the timing stays pretty fixed over the life of a timing belt, and so you'd not expect it to be a sudden factor.

That said, cold weather tends to bring out issues that were previously masked, and in particular retarded timing tends to show itself most when the temperatures drop.

Like the rest of western Canada I believe Victoria has been experiencing record lows? (Record lows for Victoria...balmy for Edmonton)

In this case, it's a quick thing to check and rule out... so you can move on. Or dig in, if it's not correct, and see if that improves the issue.

The computer back in 2002 was mostly designed to run the engine and monitor systems related to emissions-contol systems... there's a ton of other faults it will miss. Grossly-retarded or advanced timing, for example... which I confirmed the other day while doing an ALH timing belt. After replacing the timing belt the timing was initially so far retarded it was well off the screen... but normal engine startup in a 20C garage. Had I left it there there would have been a lot of white smoke the next morning out in the yard.

The pump itself is a very close cousin to the straight rotary versions you're familiar with... very close. Electric throttle actuator but pretty much everything else is mechanical like the previous version and hidden from the computer for the most part. Static timing is adjusted mechanically by loosing three bolts in slotted holes on the IP sprocket, confirmed using VCDS.

 

[jmodge]

asv sticking shut is another thing to check

 

[rattlenbang]

asv sticking shut is another thing to check

That was checked. opened and closed it manually no diff. We'll run the VCDS asap and see if that's it.