Fuel pump not priming - car slow to start - why?

[Kaldek]

I can only think of the big O ring gasket (that I reused - and not to be confused with the O Ring for the Walbro aftermarket pump discussed in this thread) not sealing 100% and on certain days (depending on temp inside the car) this affecting the built in line pressure.

I fitted a fuel pressure gauge to test exactly that. Pressure held but the car still stumbled.

 

[NuffNuff]

back to the warmer weather, stumbling again. occasionally needs a 2nd attempt to start. once its started once that day, its perfect for the rest of the day, even if you stumble it once in the early morning then in the late evening move the car it starts perfectly first time.

 

[Kaldek]

Well here is something interesting folks. The car started to stumble again right after I re-flashed it and was doing some power pulls to check for air-fuel ratio and knock. One particular session I had EvoScan logging right when the car stumbled - check out this log data!

It might be a little hard to see what's going on in the attached image, but for about quarter of a second right after cranking the ECU reported knock counts of 38, coolant temp of -24C, set the octane rating to 90% (due to knock), pulled engine timing and reported an airflow/load of 80. The knock counts and coolant weirdness dropped off immediately but it continued to stumble and log airflow of around 80 for the duration of the stumble. You can see as the airflow reading drops right around 6 seconds that the engine starts to pick up speed and the stumble goes away.

 

[eddyvr4]

That is very interesting.... i assume this was a one off and it didnt recur.

 

[Kaldek]

I'd need to keep EvoScan running every day to confirm if the stumbling is consistently related to an incorrect airflow measurement.

And I may even do that!

 

[jungle]

Mine has the odd random load reading. Like cruising @ 100km/h and suddenly there's a single spike to 200%+

Happens during load pulls but rarely. Difference is it doesn't affect the engine ie there no noticeable change, only when reviewing to log do I notice it.

 

[Kaldek]

OK here we go. Another stumble start with log data.

What's consistent with the last log? High airflow/load readings of 80. I am ignoring the whacky data when I'm cranking the engine as I think that might be a red herring.

The airflow/load however surely isn't a red herring. It is consistent and matches the timing of when the stumble ends and engine rpm comes up.

So, now that we know what is going on, what do we think could be causing the high airflow readings? They are *very* high, much more than an air leak would give us. @TME_Steve if you have a few minutes to look at the graph and let me know if that updates your previous belief it is an air leak that will help my next steps.

 

[Kaldek]

Wow actually that data is even MORE interesting than I thought. The engine load is at 80 before the car has even been cranked!!

 

[Kaldek]

Right, I did another test start. The MAF signal when you turn the ignition to ON is always around 80. When the car is working fine, immediately after cranking the MAF signal drops right down to around 20. It's only when that signal hangs at 80 that the car stumbles.

So, why does it hang I wonder?

 

[Mulvi]

Following, my pfl has just started doing this

 

[Kaldek]

It's worth asking @BCX if he has anything to say about this as he's intimately familiar with the code having reverse engineered a chunk of it. Bill, anything to say if you know how pre and post start MAF signals are treated?

 

[Kaldek]

Here's a log of a clean start with no stumble. Note the airflow reading drop rapidly after the crank event at 19.7 seconds (blue line).

 

[Blissfully]

Here's a log of a clean start with no stumble. Note the airflow reading drop rapidly after the crank event at 19.7 seconds (blue line).

View attachment 42366

Could it be as simple as a dodgy MAP sender? Have you tried switching it out to another unit? My fl has this same stumble on cold starts, however I have attempted to diagnose as of yet.

 

[Kaldek]

Could it be as simple as a dodgy MAP sender? Have you tried switching it out to another unit? My fl has this same stumble on cold starts, however I have attempted to diagnose as of yet.

A dodgy MAF you mean? Tempting to think so but it only happens once a day at start and otherwise the car runs mint.

 

[Blissfully]

A dodgy MAF you mean? Tempting to think so but it only happens once a day at start and otherwise the car runs mint.

Yeah soz, meant MAF. Could be worth just changing it to another unit and seeing whether the readings are the same

 

[Kaldek]

I think Steve is proving me wrong. Apparently the value of 80 is the "atmo, no air moving over the sensor" value. Since the engine rpms are so low when the problem occurs, it's not moving enough air to change the value.

I know what I need to do next - I need to measure the MAF raw data rather than the calculated load. That will confirm the above and stop me from looking at MAF/airflow data.

 

[NuffNuff]

Anyone with this problem using a paper airfilter? I used much less oil on my filter and seems to start a lot better. Could purely be coincidence of some other elements though
I did over oil it once long ago and perhaps dirtied the maf

 

[Blissfully]

Anyone with this problem using a paper airfilter? I used much less oil on my filter and seems to start a lot better. Could purely be coincidence of some other elements though
I did over oil it once long ago and perhaps dirtied the maf

Actually that might make sense, my air filter is filthy which would restrict proper airflow until it starts sucking enough in to get the engine running properly. I'll try replacing it and see if that fixes the problem

 

[NuffNuff]

I have new spark plugs and fuel pump and cleaning the air filter and using much less oil in it made the biggest difference

If I clean my maf I'll be using a paper filter from then on

 

[Kaldek]

Hmmm I'm reading that MAF (analogue ones anyway) would report low signal with key on engine off, and would only increase signal with airflow.

Definitely need to capture the MAF hz in EvoScan next.