Mk6 1.4 TDCI Won't Start when warm

[wightstar]

Help,

Have a Mk6 1.4 TDCI which is slowly getting worse at starting when warm.  Happened a few times months apart - but now nearly every journey.  Ford say there is nothing wrong!

Starts beautifully when cold, however after a journey, it'll refuse to start.  Things all turn over as expected and it stops cranking after a while - which I understand is normal.  No codes showing and it doesn't look like immobiliser etc., which was once the thought.

We've replaced the air filter, the fuel filter, the camshaft sensor and are now looking at what next.

Don't think it's fuel pipes etc - but I wonder about replacing the primer pump and the pipes.  Reason being, can see some small air bubbles in the clear pipes - although also present when cold - and it'll start fine when it's cooled enough.  Not enough air for it to be a major leak, but searching on the internet this could be the cause.

Are there any relays or other sensors to check please?  Don't want this to get too expensive if it doesn't solve the problem.

Managed to bump start it last night when the problem occurred which surprised me as didn't think this would have worked.

Many thanks,

Wightstar

[wightstar]

By way of an update,

Have searched lots online and too many people have had this problem - it appears relatively common - (even with Peugeot HDI owners too), and then they've paid garages lots of money, only for it to still be a problem.  I've taken lots apart, cleaned madly and removed tonnes of carbon - which is from the EGR - what a bad idea that is!

Will put it all back together tomorrow, with a new crankshaft sensor - which is about all I really think it can be - since it will always start perfectly when cold, or after it's failed to start when hot, and then been allowed to cool down.

Is there anything to watch for when refitting the crankshaft sensor please - as the hole is fairly elongated? ;-)

Many thanks,

Wightstar

[Tdci-Peter]

2 hours ago, wightstar said:

refitting the crankshaft sensor

Crankshaft sensor is possible, they have been known to fail when hot. But for no codes, the failure must be partial or minor.

These are normally well located by a close fitting pin or hole. If the elongated hole only affects the gap, then I would aim for 0.5 to 1mm in absence of other information. If it affects the timing, then it is a mystery to me how it can be set.

If warm start still fails, I had that once in a old petrol car. It would start if stalled, but if more than a few minutes passed, then I had to wait for it to cool down. It turned out to be a compression problem due to a dodgy valve. May be a red herring, but diesels can be even more critical of compression problems than petrols.

 

[wightstar]

Okay,

By way of an update I thought I'd let you know what's going on.  New crankshaft sensor fitted - thanks Tdci-Peter - crankshaft sensor worked a treat, with very little movement - but it was difficult to fit.  Had to release the crankshaft pulley a bit and then slide the new one in behind it before retorquing.

Despite removing tonnes of gunk from the inlet manifold - which has had an astonishing affect on the performance!  I am very sad to say, the problem is still there.  We thought it was solved as various tests had shown it to be okay.  However, this afternoon - it failed to start when warm for no apparent reason.

So, back to the drawing board.  Is it a dodgy valve - and how would I test for that please?  Surely one valve would not stop the whole engine starting?

Any other suggestions as to why a warm engine will simply over, but refuse to start one minute, and then a couple of miles further on at home after a bump start, will start when hot having been stopped?

Many thanks,

Wightstar

[SpannerTheCat]

Been following this thread & it's baffling, @wightstar must be so frustrating? Here's a theory that @Tdci-Peter may be able to answer as he has much more experience than me. Could any of the fuel lines be in contact with any hot surfaces such as the exhaust manifold etc leading to the fuel boiling or being too hot? May be a dumb question but sometimes it's weird solutions to fix weird problems.

[Tdci-Peter]

15 minutes ago, SpannerTheCat said:

leading to the fuel boiling or being too hot?

Nice idea, and in this warm weather, underbonnet temperatures will be high. If it is anything like my 1.8TDCI, then any air or vapour entering the main fuel pump will prevent starting. But that is the down side to this theory, the vapour will have to get to the pump. The boiling point of diesel fuel is at least 180C, so the pipe from filter to pump, or the pump, would have to be pretty hot for vapour to reach there without condensing back to liquid. Vapour locks in the pipes to the filter might cause unexpected stalls, but not prevent starting.

My valve theory is floundering too. Thinking back, that was a 6 cylinder, petrol, carburetor engine. I think the valve was an inlet valve, and one bad inlet valve can stop the vacuum building up in the inlet manifold, so it would not pull in petrol / air mix. Diesel engines are different, no inlet manifold vacuum is generated.

On 16/08/2016 at 9:35 AM, wightstar said:

Managed to bump start it last night when the problem occurred

That is an interesting fact. There are two possible differences when bump starting: Faster cranking, and the battery does not have the starter motor load while starting.

Re-testing this bump start might help, or try a jump start from a car with its engine idling (14v at battery, instead of 12v).

If it is faster cranking, then engine compression is a 1st suspect, try getting a compression test done. Cranking compression will tend to be higher on a cold engine, as the oil gives a better seal.

If it is low battery voltage, then either the battery is poor (but still good enough to turn the engine over, say one cell gone poor), or the earth wire from engine to battery is high resistance, or some critical part that runs off battery voltage has become voltage sensitive. The fault would also have to be temperature dependent, which is possible, but perhaps improbable.

However, you seem to have check most of the probable causes, so as Ray says, weird or improbable causes have to be considered. An electrical problem like a bad earth could well be intermittent.

Got a voltmeter to check battery volts while cranking? It should stay above 9v really , on a warm start, though down to 7.2v is allowed for a cold start.

[wightstar]

Typical isn't it - no problems for a few days now.  Battery is about 12.2volts when going to start - dropping to 10.9 when starting from cold at the moment.  Will try and do some 'hot' measurements today to see what is happening.

Otherwise, it's looking for weird possibilities.

Not fully understanding the engine - is there something done to stop the engine when I tell it to.  I mean the fuel is not fed into the engine anymore - one possibility is that a solenoid? is still saying stop the fuel - anything I can check here please?  The other is that the throttle valve after the turbo on the air intake is still across - not letting air in - unlikely as it works very well under testing.  Is there anything else like this - which may get shaken lose on a bump start or spring back when it cools enough please?

Frustrated!

[Tdci-Peter]

2 hours ago, wightstar said:

one possibility is that a solenoid? is still saying stop the fuel

There is no separate stop solenoid, the injectors are solenoid valves, so just turning them off will stop the engine. There is a solenoid valve in the fuel pump that regulates the pressure, but I think this one opens to reduce the pressure.

The injectors will get quite hot when the engine stops, then the fuel will cool them once running. But it would be odd to affect all 4. I am not sure of an easy way to detect if fuel is getting in to the engine. With extended cranking (try short bursts of 5 - 10 sec, with 20 -40 sec gaps, and maybe with a battery charger connected), see if there is a smell of diesel at the exhaust.

A leak back test can be done at the injectors, there are YouTube vids of this, but it is a bit fiddly, and you don't get much leak back flow at cranking speed. But it might show if fuel is reaching the injectors and the injectors are operating.

Forscan can read the fuel rail pressure, to see if the pump is working ok, but I am not 100% sure it will operate while cranking.

I still suspect low compression when the oil is hot may be possible. Diesels only need 2 things to start, fuel (at the right time!) & compression. Glowplugs help in cold weather, but high pressure common rail diesels should start ok without them in warm weather, especially with a warm engine.

 

[wightstar]

Hello,

Sorry about the huge delay in replying SpannerTheCat and Tdci-Peter - any others who may be following this.  I've been monitoring things and measuring and waiting for it to go wrong - apologies this is a long post.

Good news is I now have FORScan, but the bad news is I still can't work out what is wrong.  At the risk of this being a very long post - I've also removed the first few lines which show the same sort of info.

This is the car starting:-

time(ms)	FRP(kPa)	RPM(rev/min)	ECT(°C)	IAT(°C)	MAF(g/sec)	LOAD_PCT(%)	MAP(kPa)	MIL_DIST(km)	OBDSUP	VSS(km/h)
10005	870	0	46	35	0.95	0	103	0	EOBD	0
10034	870	0	46	35	1.14	0	103	0	EOBD	0
10063	870	0	46	35	1.32	0	103	0	EOBD	0
10093	870	0	46	35	1.51	0	103	0	EOBD	0
10122	2120	0	46	35	1.7	0	102	0	EOBD	0
10152	3360	0	46	35	1.89	0	102	0	EOBD	0
10191	4610	0	46	35	2.08	0	102	0	EOBD	0
10229	5860	0	46	35	2.27	0	102	0	EOBD	0
10260	7100	0	46	35	2.23	0	102	0	EOBD	0
10288	8350	0	46	35	2.2	0	102	0	EOBD	0
10327	9600	0	46	35	2.16	0	102	0	EOBD	0
10356	10850	0	46	35	2.12	0	102	0	EOBD	0
10387	12090	0	46	35	2.08	0	102	0	EOBD	0
10414	13340	0	46	35	2.05	0	102	0	EOBD	0
10444	14750	0	46	35	2.01	0	102	0	EOBD	0
10474	16160	50	46	35	1.97	0	102	0	EOBD	0
10504	17560	100	46	35	1.94	0	102	0	EOBD	0
10543	18970	150	46	35	1.9	0	102	0	EOBD	0
10573	20380	200	46	35	2.37	0	102	0	EOBD	0
10611	21790	250	46	35	2.84	7.84	102	0	EOBD	0
10651	23200	300	46	35	3.32	15.69	102	0	EOBD	0
10681	24600	349	46	35	3.79	23.92	102	0	EOBD	0
10709	26010	399	46	35	4.26	31.76	102	0	EOBD	0
10738	27420	449	46	35	4.73	39.61	102	0	EOBD	0
10775	27820	499	46	35	5.2	47.45	101	0	EOBD	0
10804	28230	597	46	35	5.68	55.29	101	0	EOBD	0
10833	28630	694	46	35	6.15	63.53	101	0	EOBD	0
10871	29040	792	46	35	6.62	71.37	101	0	EOBD	0
10902	29440	890	46	35	8	79.22	101	0	EOBD	0
10940	29840	988	46	35	9.38	74.12	101	0	EOBD	0
10981	30250	1086	46	35	10.77	68.63	101	0	EOBD	0
11009	30650	1184	46	35	12.15	63.53	101	0	EOBD	0
11049	31060	1282	46	35	13.53	58.04	101	0	EOBD	0
11082	31460	1379	46	35	14.91	52.94	101	0	EOBD	0
11119	30520	1477	46	35	16.29	47.84	101	0	EOBD	0
11146	29570	1466	46	35	17.68	42.35	101	0	EOBD	0
11186	28630	1456	46	35	19.06	37.25	101	0	EOBD	0
11215	27690	1445	46	35	20.44	31.76	101	0	EOBD	0
11244	26740	1434	46	35	20.02	26.67	101	0	EOBD	0
11284	25800	1424	46	35	19.61	23.92	101	0	EOBD	0
11313	24860	1413	46	35	19.19	21.18	101	0	EOBD	0
11343	23920	1402	46	36	18.77	18.82	101	0	EOBD	0
11381	22970	1392	46	36	18.35	16.08	102	0	EOBD	0

 

And this is it not starting:-

time(ms)	FRP(kPa)	RPM(rev/min)	ECT(°C)	IAT(°C)	MAF(g/sec)	LOAD_PCT(%)	MAP(kPa)	MIL_DIST(km)	OBDSUP	VSS(km/h)
2779	650	0	77	54	1.2	0	103	0	EOBD	0
2805	650	0	77	54	1.36	0	103	0	EOBD	0
2845	1020	0	77	54	1.52	0	103	0	EOBD	0
2872	1400	21	77	54	1.68	0	103	0	EOBD	0
2899	1770	42	77	54	1.84	0	103	0	EOBD	0
2930	2140	64	77	54	2	0	103	0	EOBD	0
2968	2510	85	77	54	2.02	0	103	0	EOBD	0
2995	2890	106	77	54	2.03	0	103	0	EOBD	0
3025	3260	127	77	54	2.05	0	103	0	EOBD	0
3065	3630	148	77	54	2.06	0	103	0	EOBD	0
3094	4010	170	77	54	2.08	0	103	0	EOBD	0
3120	4380	191	77	54	2.09	0	103	0	EOBD	0
3154	4720	212	77	54	2.11	0	103	0	EOBD	0
3199	5060	215	77	54	2.12	0	103	0	EOBD	0
3226	5400	218	77	54	2.14	0	103	0	EOBD	0
3253	5740	220	77	54	2.15	0	103	0	EOBD	0
3282	6080	223	77	54	2.18	0	103	0	EOBD	0
3309	6430	226	77	54	2.22	0	103	0	EOBD	0
3343	6770	229	77	54	2.25	0	103	0	EOBD	0
3381	7110	232	77	54	2.28	0	103	0	EOBD	0
3409	7450	234	77	54	2.31	0	103	0	EOBD	0
3440	7790	237	77	54	2.35	0	103	0	EOBD	0
3475	7960	240	77	54	2.38	0	103	0	EOBD	0
3513	8120	242	77	54	2.41	0	103	0	EOBD	0
3544	8290	244	77	54	2.45	0	103	0	EOBD	0
3572	8460	246	77	54	2.48	0	103	0	EOBD	0
3610	8630	249	77	54	2.45	0	103	0	EOBD	0
3646	8790	251	77	54	2.42	0	103	0	EOBD	0
3672	8960	253	77	54	2.39	0	103	0	EOBD	0
3700	9130	256	77	54	2.36	0	103	0	EOBD	0
3728	9290	258	77	54	2.33	0	103	0	EOBD	0
3755	9460	260	77	54	2.31	0	103	0	EOBD	0
3783	9540	262	77	54	2.28	0	103	0	EOBD	0
3822	9610	262	77	54	2.25	0	103	0	EOBD	0
3853	9690	261	77	54	2.22	0	103	0	EOBD	0
3880	9770	260	77	54	2.19	0	103	0	EOBD	0
3910	9840	260	77	54	2.23	0	103	0	EOBD	0
3937	9920	259	77	54	2.27	0	103	0	EOBD	0
3967	10000	258	77	54	2.31	0	103	0	EOBD	0
3995	10080	258	77	53	2.35	0	103	0	EOBD	0
4023	10150	257	77	53	2.39	0	103	0	EOBD	0
4050	10230	256	77	53	2.43	0	103	0	EOBD	0
4089	10300	256	77	53	2.47	0	103	0	EOBD	0
4117	10360	255	77	53	2.51	0	103	0	EOBD	0
4144	10430	254	77	53	2.55	0	103	0	EOBD	0
4172	10490	252	77	53	2.59	0	103	0	EOBD	0
4202	10560	251	77	53	2.56	0	103	0	EOBD	0
4230	10620	250	77	53	2.53	0	103	0	EOBD	0
4258	10690	249	77	53	2.49	0	103	0	EOBD	0
4297	10750	248	77	53	2.46	0	103	0	EOBD	0
4324	10820	246	77	53	2.43	0	103	0	EOBD	0
4352	10880	245	77	53	2.4	0	103	0	EOBD	0
4382	10900	244	77	53	2.37	0	103	0	EOBD	0
4419	10910	246	77	53	2.33	0	103	0	EOBD	0
4447	10930	247	77	53	2.3	0	103	0	EOBD	0
4473	10950	248	77	53	2.27	0	103	0	EOBD	0
4512	10970	250	77	53	2.28	0	103	0	EOBD	0
4539	10980	252	77	53	2.28	0	103	0	EOBD	0
4569	11000	253	77	53	2.29	0	103	0	EOBD	0

The only thing I can really see is that the engine is warm - even after the air filter was replaced. The warmth has been a common theme, however, this may not be the cause.

Again, I still have no errors:-

===PCM DTC ===
Successful DTC reading, no error codes found

Module: Powertrain Control Module
===END PCM DTC===

===OBDII DTC ===
Successful DTC reading, no error codes found

Module: On Board Diagnostic II
===END OBDII DTC===

===IC DTC ===
Successful DTC reading, no error codes found

Module: Instrument Cluster
===END IC DTC===

===RCM DTC ===
Successful DTC reading, no error codes found

Module: Restraint Control Module
===END RCM DTC===

===ABS DTC ===
Successful DTC reading, no error codes found

Module: Anti-Lock Brake / Traction Control Module
===END ABS DTC===

A few questions please:-

Can anyone else spot anything else from the figures please?

Is the engine designed not have have the injectors working if the engine is above a certain temperature - seeing as the pressure is there?

I could smell no diesel at the exhaust when trying to start it and it failing. We tried to do bursts and gaps as Peter suggested, but still no luck - until an hour later when we tried again, and it just started.  This should not be viewed as conclusive, but perhaps indicates the injectors are not injecting?

Just run it up in the garden and the thermostat opens just over 80 degrees C and then the radiator gets warm.  I have seen the fan working, but not sure when it does this and can't say for certain if it happens when the engine is turned off each time.  

Next step would appear to be compression testing and perhaps resetting the valves - just in case.  Loath to do either as it starts after a couple of turns when cold.  Surely it would be unlikely for this to be the case on all four cylinders at the same time (84,000 miles), which is why I am currently leaning to electronic "magic" stopping it from starting when warm - what else should I look at here please?

Very frustrating problem and trying to set up FORScan every time, I've missed a couple of captures - but the above two are fairly representative.

Many thanks for your help - anything would be useful.

Wightstar

[Tdci-Peter]

7 hours ago, wightstar said:

Is the engine designed not have have the injectors working if the engine is above a certain temperature - seeing as the pressure is there?

It should start at any reasonable temperature. Like out in the sun in southern Europe after a long drive.

Brilliant to get those results from Forscan, It can take a bit of getting used to!

What does really strike me is that the fuel rail pressure at 200rpm was 20,400 when cool and starting. But it was only 4,600 when hot and failing. And about 200 looks to be the point it fired up.

There are quite a few potential causes for low FRP:

Air in line to pump. Due to leaks in fuel filter, or pipes. But the cranking pressure looks fairly consistent on the fail to start run, I would think air bubbles would make it vary.

Filter blockage. But actual flow is very low when cranking, seems unlikely.

Excessive leakage on injectors. I have not heard of this being temperature dependent, and it would have to be quite bad. Surely the power / performance when running (hot) would suffer.

Poor pump. Leakage in the high pressure pistons or valves. Again, this would usually have a pro-rate greater effect at high power, when the pump is being asked to deliver up to 150,000 kPa.

Bad pressure regulator valve in the pump. This could be highly rpm dependent, so ok when running, but bad at cranking rpm? Just a guess, needs verification really.

My suspect list would be: Air ingress (filter leaks), Pressure regulator valve, Others

I have not done any tests while cranking, I was not 100% sure if it would work. The rpm readings look a bit suspect. I think I would cut the monitored channels down to just rpm & FRP, for a few tries. May get better data. Remember that the car is using the same (HS-CAN) bus as Forscan & the ELM, and should have priority over them. So there may be some timing variance.

Logging a run while driving is quite easy. Start recording, ensure cable & computer is in a safe state. Drive for 10 mins to 30mins, with a range of speed & power. Stop, stop logging and save the data. (Never look at the computer while moving!). This will check the pump over its full range. As comparison, I recently got the results below. The flat bit at APP=98% was accelerating hard up a steep hill in 3rd. FRP is about 150,000, with a little overshoot when I closed the throttle to change up to 4th. (The FRP scale is 27,330 to 170,580.)

 

 

 

 

 

[wightstar]

Thanks for the reply Tdci-Peter.  The car is now getting annoying and very time consuming.

Understanding that the car should start at any temperature - it's not :-( So, by was of a recap, it has a new fuel filter, a new air filter, new camshaft sensor, new crankshaft sensor, lots of cleaning of the inlet manifold and still doesn't start when it's hot.

The fuel rail pressure does seem to be key to this - somehow.  When warm, it manages to get to around 10,000-12,000 and still fails.  When cold, it seems to start around here, and just get faster than the 200 rpm and starts.

I don't think there is any air in the fuel pump - when it's failed I've manually primed it lots and it still doesn't work.

Likewise for a fuel blockage - and I think the numbers say otherwise, I can't see how there is one.

Power seems fine when it's hot and it simply refuses to start, so I think injector leakage is unlikely - it could possibly be leaking on one injector, but all four?  Nothing shows even one leaking though. The pump is quite happy delivering the pressure when required:-

time(ms)	FRP(kPa)	RPM(rev/min)	ECT(°C)	IAT(°C)	MAF(g/sec)	LOAD_PCT(%)	MAP(kPa)	MIL_DIST(km)	OBDSUP	VSS(km/h)
720596	112580	2964	81	32	52.58	71.37	167	0	EOBD	25
720623	114910	3001	81	32	54.22	72.55	167	0	EOBD	25
720650	117240	3038	81	32	55.87	73.33	171	0	EOBD	25
720690	119570	3075	81	32	57.51	74.51	175	0	EOBD	26
720717	121910	3112	81	32	59.16	75.29	178	0	EOBD	26
720755	124240	3149	81	32	60.81	76.47	182	0	EOBD	26
720780	124740	3186	81	32	62.45	77.65	186	0	EOBD	27
720820	125240	3215	81	32	64.1	78.43	190	0	EOBD	27
720847	125740	3244	81	32	65.74	79.61	194	0	EOBD	27
720875	126240	3273	81	32	67.39	80.39	197	0	EOBD	28
720913	126740	3302	81	32	68.01	81.57	201	0	EOBD	28
720941	127240	3332	81	32	68.63	81.57	205	0	EOBD	28
720970	127740	3360	81	32	69.26	81.18	205	0	EOBD	28
721000	128240	3390	81	32	69.88	81.18	205	0	EOBD	29
721027	128740	3419	81	32	70.5	80.78	205	0	EOBD	29
721054	129240	3448	81	32	71.12	80.78	205	0	EOBD	29
721086	130140	3477	81	32	71.74	80.39	204	0	EOBD	30
721131	131040	3511	81	32	72.37	80.39	204	0	EOBD	30
721163	131950	3544	81	32	72.99	80	204	0	EOBD	30
721197	132850	3578	81	32	73.61	80	204	0	EOBD	31
721240	133750	3612	81	32	74.15	79.61	204	0	EOBD	31
721266	134650	3646	81	32	74.69	79.22	204	0	EOBD	31
721294	135550	3679	81	32	75.22	79.22	204	0	EOBD	31
721324	136460	3713	81	32	75.76	78.82	203	0	EOBD	32
721353	137360	3746	81	32	76.3	78.43	203	0	EOBD	32
721382	138260	3780	81	32	76.84	78.43	202	0	EOBD	32

Or this one, from slightly earlier:-

time(ms)	FRP(kPa)	IAT(°C)	MAF(g/sec)	ECT(°C)	RPM(rev/min)
433286	151000	27	74.42	84	3730
433325	151150	27	74.8	84	3736
433358	151310	27	75.18	84	3742

 

21 hours ago, Tdci-Peter said:

My suspect list would be: Air ingress (filter leaks), Pressure regulator valve, Others

I like the suspect list, but rule out the first - hand priming seems to remove this.  Pressure regulator valve - unlikely as it goes like a train when required and other numbers discount this.

I have quite a few sets of readings and can see nothing obviously wrong - other than it won't start when warm!  However, I take the point about cutting the number of channels being monitored.  FORScan is really brilliant - not sure I like my passenger knowing I've put my foot on the accelerator though ;-) .  Also, perhaps your car is newer, giving the Accelerator Pedal Position, or maybe the IOS version does not show this.  This is a 2006 model.

What do you think of another idea where the Catalytic converter is so bunged up with tar, deposits and rubbish that the engine cannot get high enough rpm's and therefore the fuel pump can't generate enough pressure?  Also supporting this is the fact that we bump started it once when it wouldn't start.

Engine is sort of in bits at the moment - looking for a solution to the problem.  Wish this was simpler and jumping out at me!

Many thanks for any ideas please.

[Tdci-Peter]

1 hour ago, wightstar said:

Engine is sort of in bits at the moment - looking for a solution to the problem.  Wish this was simpler and jumping out at me!

These things are never easy. You can go to a garage with an open cheque, if you are lucky they get the right bit first, but more often they just change expensive parts until they get it right. Or you have to spend a lot of time digging & testing, to narrow it down.

On the APP PID, I have 2006 Focus, but also use the full Forscan on a laptop, so it could be either.

On the Cat, or anything in the airflow, this would surely impact the full power end, more than starting. 200 rpm sounds a sensible cranking speed to me, it takes a lot of effort to turn a diesel engine over. I have tried starting single pot engines with a handle. It needs a de-compressor, the compression is impossible by hand. Just googled it, and 200 comes up lots of times!

I have been trying to find out how the pressure control stage works for some time. Valves & pressure are part of my job, I find it (sadly!) interesting. I have just come across a nice description, I am not 100% sure it is right for these Ford pumps, but it all seems to match.

These pumps are all two stage units. A low pressure vane type pump pushes fuel into the main piston pump. The pistons are pushed out by a cam to create the high pressure, but are only pushed back in by the pressure from the first stage pump. That is why air at the inlet, which prevents the rather poorly sealed vane pump from working, then completely disables the high pressure stage.

To control the outlet pressure, a throttle valve after the vane pump reduces the flow into the main pump, so the pistons do not travel back so far, making it a variable displacement pump. So reducing the flow out of the main pump. It also reduces the load on the engine, as the cam is not doing so much work, making the pump efficient at low flowrates.

This throttle valve is controlled by the regulator solenoid, which is driven by a PWM signal from the pcm to control the pump flow, and hence to control the rail pressure. When the pcm energises the solenoid, it closes, and the intermediate pressure then has to push the throttle valve across, against a spring.

I can see a few leakage or sticking points that could result in no flow at 200rpm, and plenty of flow at 800rpm or more, in the intermediate pressure regulation stage. The viscosity of oil is temperature dependant, reducing by a factor of 2 or more between 20C & 80C, which will increase leakage when hot.

When cold, the pressure shot up to 20,000kPa very quickly, and by 200rpm. When hot, it grizzled up to about 11,000kPa. There is a spring to overcome in the injector, and 11,000kPa may not be enough to open it.

 Like you say, it all seems fine once it is running at 800rpm or more. (Though there is a slight unexplained fall, from 151,000kPa to 136,000kPa between the two runs. Same air flow & rpm, so this suggests a regulation error?)

I am guessing that this is not really the answer you wanted! But it is the best I can come up with, and seems to fit the results so far.

The pressure regulator valve is available for about £60, there seem to be quite a few sellers, so it may be a common fault. But in this case I think it is 60% likely the problem is in the pump itself, in the other parts of the regulation system. A good diesel specialist could test and possibly repair the pump. The diagram below indicates why I think this. The solenoid valve only seems to be one part of the system.

Spraying the fuel pump & fuel line with freezer spray when the engine is hot, might test whether it is the pump?

Example link to valve: http://www.ebay.co.uk/itm/FORD-FIESTA-FUSION-1-4-TDCi-FUEL-PRESSURE-REGULATOR-INLET-METERING-CONTROL-VALVE-/281213562578

 

[Michael9]

I know this thread's 2 yrs old but did you or he ever find what it was

[wightstar]

No - still have the car, just don't stop it in hot weather if we want to use it again soon.  Any ideas gratefully accepted though please.

[Ticktock]

Just read through this and I have an identical 1.4 tdci. I would be looking at the battery and the fuel pump.

battery - Is there a drain on the battery when driving, could be not enough power to turn it over quick enough.

Fuel pump - Might not be enough fuel getting pumped when it's hot.

[wightstar]

Thanks for taking the time to read this ticktock.  Reading it back there is a lot here and I'd still like to find out what it is - an older car would just work - modern ones have too many electronic variables!

I've just done a bit more experimentation and since we fitted a new battery earlier in the year, I was hopeful.  Drove it around yesterday and got things nice and warm, went to the top of a big hill and turned it off.  It did restart immediately, but then about 10 minutes later there was nothing - it just turned over and refused to start - presumably it was too hot.  The same old problem.  Had to bump start it and drive it home.

So it looks as though the only thing to try now is the fuel pump.  Point is these things are a few hundred pounds for new ones (assuming we're talking the high pressure pump?), and I have to ask if the gamble is worth it?  Sadly it's getting to the point where a replacement car could be cheaper and a better gamble.  Even if that one had a few things wrong with it!  This car is just unreliable in warm weather, unless the car is allowed to cool down which is rather sad.  I don't like being beaten, but the work involved in tracking this problem down has been rather large.

[TomsFocus]

That's some dedication to keep a car for a further 2 years that won't start hot! 

I know it's not the same engine, but the VAG 2.0 TDIs have a common hot start problem as well, they usually still start, but take several seconds of cranking when hot.  The hot starting ECU map needs to see a higher cranking speed than the cold one, but because of the resistance of old wiring and starter coil when hot (they soak up engine bay heat), the starter can't 'quite' spin the engine fast enough, despite sounding fine!  So some have the map tweaked to stop it being an issue, some change the starter, some change the battery.

I haven't read the whole thread, but have if you haven't tried changing the starter, I'd try a good used one where the seller can confirm no hot start issues.  Easier and cheaper than a HP pump!

[Tdci-Peter]

On 9/12/2016 at 10:12 PM, wightstar said:

The fuel rail pressure does seem to be key to this - somehow.  When warm, it manages to get to around 10,000-12,000 and still fails.  When cold, it seems to start around here, and just get faster than the 200 rpm and starts.

 

3 hours ago, TomsFocus said:

The hot starting ECU map needs to see a higher cranking speed than the cold one,

Interesting idea, I had assumed the starting threshold rpm would remain the same, but these cars all seem different, sometimes defying normal logic. It makes it dangerous to assume anything.

But the Forscan data from 11-09-16 above (Tom admits he may not have gone that far back!), suggests to me it is the fuel pump. When hot it did get up to 260rpm, but the fuel rail pressure was really struggling to rise. When cold, the pressure just shot up.

However, It is a lot easier to experiment with starters than with IMVs, high pressure pumps or injectors!

My guess is the first stage of the HP pump is a bit worn, and can not deliver enough fuel to the second stage when the diesel is warm and low viscosity. But it is just a guess! Leaking IMV (Fuel metering control valve) or leaking injectors are also possible.

I think the car has the luxury of a priming system. Trying to use this to boost the pressure into the 1st stage while actually cranking (if safe to do so!) might be worth trying. The pressure from the 1st stage is not very high, so even a bit of extra pressure might help. If it did, I think it would remove suspicion from the injectors.

 

[isetta]

this is not very likely the answer but I will mention it in case.

when the engine is turning on the stater motor does it sound like it is spinning fast enough?

About 18 years ago I have a Mk3 Escort 1.3 (CVH engine). when hot it was difficult to start and I felt that the starter motor was not turning as fast as when cold. 

I took the starter motor apart and could see scrape marks where the inner spinning part had been scraping on the outer static magnets. I think the clearence was too small and the heat (conducted from the hot engine) causing expansion was closing up the gap. (I know you might say the outer part would also have expanded, not just the inner part but maybe the different types of metal in there expanded at different rates. 

So I replaced the starter motor and it was fine.

Bit of a long shot that your car is suffering from this but you could have something very uncommon and that no one would guess is the problem.

[plpawzaw]

Hello!

First of all - i dont speak english fluently - so sorry for bad english.

I have Fiesta 1.4 tdci 2006. 

I have had same problem with fiesta - everything works fine until i have to start on hot engine. 

I also read this post to find out solution :)

You will call me crazy but i found solution that works for me and it was the cheapest one in my life :)

Try it, maybe it will help you.

I just replace F15 fuse (fuel pomp fuse) under my glovebox. Fuse wasnt burn or something, it have some discolorations.

I just replace it to new one and now i can start engine every time.

Greetings from Poland :)

 

[Abbadon]

This issue stinks of a leak off related issue - either worn injectors leading to excessive injector leak off or worn pump leading to excessive pump leak off.

Leak off is always present by the operating physics of the designs, but should only be small.

If pressure does not build far enough soon enough, the injectors will not open, and no fuel is delivered to the cylinders.

Easy way to demonstrate that the rest of engine is working under this condition is spray a SMALL amount of easy start /ether in to the inlet - the engine should fire up and then die again.

What may happen when you do that, is that it may start, as the engine will run slightly faster and pressure will exceed that critical level long enough for injectors to fire....

 

 

[chebhou]

Hi all, i have the mk7 with 1.4tdci and i'm facing the same problem but i think i have narrow it down to the injectors, seems one of them is faulty and when it heats  it gets stuck open and drops the pressure then the car dies, I can start the car normally when it's cold but once it heats up  I can't start it again unless i use a starting fluid ( i use insects spray ) or let it cool down for hours; but when i was experimenting with it i tried cooling down the suspected parts with water then trying to start the engine again ( before that i used a scope to inspect the signals and all was fine ) so i tried the crank,cam sensors, fuel temp, the high pressure pump, the fuel rail ( in case it had a crack ) and no difference, then i tried the injectors and it worked , so if the car is hot and it doesn't start i pour some water on the injector for a while and it starts immediately (although other parts are still hot ) .
I have checked the waveforms and voltage levels when the car is starting ( once it's cold and once it's hot ) and didn't see any difference except for the fuel pressure which is 4000PSI and can go up to 6000PSI  when the engine is cold, while it stays under 1200PSI  ( since the startup time is short when the engine is cold I removed the injectors' plugs to prevent the car from starting while i log the pressure readings )

I could clearly see the effect of cooling the injectors as the pressure start to rise up  ( from 1200PSI to 1500PSI then it starts ) but I'm still not sure

i rolled out :

-the starter : because of the cranking RPM was same hot or cold.

-the crankshaft and cam sensors :  because the wave forms are the same ( and the car starts with a starting fluid even when hot and runs fine which means they do work when hot ).

-the fuel  pressure, flow regulators + the pump : also the wave forms were the same and i tried cooling them down but got no results ( but they are still suspects).

-the ecu blocking the injector waiting for higher pressure : because at first fooled the ecu by rising/lowering the voltage on the pressure sensor line with no difference, then probed the injector lines and there was a firing signal ( it above 100V since they are piezo ) even at low pressures while the car is hot.

- the fuel evaporating in the pipes and making bubbles :  also because it runs fine once started with starting fluid and keeps running ( where did the bubbles go ? )

to confirm my suspicions I'll take the injectors to a repair-shop and see how it goes from there ( I had taken them before only to get worse results ) 

so if anyone have any advice or notes you are welcome to point it out , and I hope this helps you @wightstar  .

should note that i have poured the water on the engine while the "car cooling water reservoir"  was open from the start ( when it was cold ) first to avoid any effect of the  vapors condensing quickly and secondly  to not burn my self opening it when the engine is hot ( done that once and burned my arm ) and i think someone will complain that it's not a good idea to pour cold water on hot metal but if you have this problem for a while you'll be ready to push this car off a cliff into the sea.

 

 

 

[isetta]

I am impressed with what you have done and the way you have clearly laid it out to help others in the future.

This is the kind of thing we need more of so people with future problems can find stuff on here to look at straight away.

Assume you will now change the injector, be interesting to hear further from you if there are any more findings when you do this eg. some visual defect with the injector (but I doubt there will be).

I don't have anything to point out other than

"i rolled out "  should be " I ruled out "

[Tdci-Peter]

21 hours ago, chebhou said:

to confirm my suspicions I'll take the injectors to a repair-shop and see how it goes from there

+1 for being impressed with the testing and diagnostics. Tests like that are often not as easy as they sound on a car, with poor or awkward access.

The ECU will not energise the injectors until a certain rpm & fuel rail pressure is reached, on my car is seems to be about 200rpm & 200Bar (3000PSI), but it looks like the rail pressure is lower on that 1.4, if it starts at 1500PSI. Or just possibly there is a calibration error, which could be temperature dependant, in the fuel rail sensor. But that would not explain why cooling the injector allowed it to get up to pressure and start.

A good diagnostic system like Forscan can record fuel rail pressure, RPM and other stuff during a start, even during a normal rapid start. I have posted pictures of start-up logs from my car. It can also check rail pressure & boost pressure while driving, to get readings which can not be obtained on the driveway.

Measuring the leak back flow from the injectors might help. This is relatively easy. While an injector is open there is a continuous flow into the leak off outlet, from the pilot stage of the injector. If the pilot stage is stuck open or leaking, then there will be a leak-off flow even with the injector not plugged in to the loom, while cranking. If the main stage is leaking into the cylinder, then I would expect it to affect performance, raising idle, making the engine uneven at low power, making fuel consumption worse, and making smoke.

A lot of places that offer to test and repair this complex stuff are not very good. It needs the right equipment, experience, care and knowledge to work on these critical parts. Sometimes a used injector from a known good car may be a better bet. If the 1.4 is like the 1.8 (I think it shares similar Siemens equipment), then it is not too fussy about replacing injectors, the ECU will self adjust.

 

[chebhou]

@isetta thank you for the kind words and happy to be here and of help,  as for the injectors I'll try to fix them and if i can't then I'll have to buy another ones . it's the weekend so I'll take them for repair on Sunday . (it seems that I can't edit my old post )
@Tdci-Peter Well what can I say once I read your replay I run to the car and performed the leak-off test ( i have just finished ) and it revealed the faulty injector, so Thank you for the suggestion and the explanation.

For the startup pressure measurement it was not very accurate because i just used an ELM237 clone and the refresh rate was low, but it was enough for the long periods.

as for the repair-shop  you are right, the first one that i took them to claimed that he have the tools for it but apparently he didn't ( cause he messed up my friend's Focus with 1.8 engine it came running to the shop and left towed after the service ) and if i recall correctly the injectors on the 1.8 are different and have no other compatibles while the 1.4  ones are more common ( because of  PSA).

Now back to the leak-off test, I did it with the elimination method since i didn't have enough pipes and plugs ( each injector has a Y junction on the leak outlet ) so i collected the leaked fuel from all the injectors once it was cold and once it was hot for two minutes each time, and the when-hot-quantity was three time that of the cold one, then i started eliminating the first injector by connecting it back to the car leak pipe while the other three are under test ( their leak is collected in a bottle ), then 1 & 2 are eliminated and 3 & 4 are under the test and so on. at each time observe if the leaking flow changed too much from the last test.

but the results were too obvious and showed that the second injector was leaking too much when hot. and i didn't bother with the rest of them . probably I ll just take this one for repair cause the last repair costed more than the price of two injectors.

all I needed for the test was two pipes, a plug and a bottle , a pipe from the tested injectors to the bottle and the other pipe with the plug just to close the other outlet from leaking.

thank you all for the help .