Startup Failure on 2005 Disco 3 (TDV6 Auto with Euro 2 HPFP)

[GrahamH]

Dear Abby,
I have this friend ....

A mate of mine with a 2005 Disco 3 is experiencing an issue with getting his TDV6 to start (the issue is that it flat out won’t start!). Sorry but this is going to be a long post - bear with me.

Background
Neil took the body off the D3 to replace the turbo and give it a good clean – the car is 19 years old after all but fastidiously maintained as he’s a retired automotive mechanic. He’s also had the body off my 2015 D4 to replace the 3.0Lt engine which seized at 66,000km (that’s a whole other saga). When we put the body back on mine, it took a while to bleed the fuel system but then it started fine and it now runs beautifully. So let me reassure you - Neil’s not a novice at this.

The Problem
Having carefully out it back together (paying particular attention to earth connections and to bleeding the low pressure fuel system thoroughly) it cranks well but refuses to fire. The only code coming up points to the High Pressure Fuel Pump (HPFP) – P0001 VCV Open. The Volume Control Valve solenoid measures 2.7 Ohms which is within the normal range. It's a Euro 2 pump and these pumps have the reputation of being very reliable so we were reluctant to replace it without being quite certain it was the culprit. So he had it tested (on a test bed) and the test report shows it to be good. This bears out our own belief that the pump itself is OK and is not the cause of the non-firing issue.

A Key Problem
Initially there was a problem with the 19-year old battery in the key having died, but I replaced the battery with a charged up new one (I’m an electronic tech and this is not rocket science believe me). The key was recognised by the RF charging circuit OK and the security light on the dash extinguished so the car could then crank. Initially the remote door locking wasn’t working but I held the key under the steering wheel console and operated the buttons and the key fully resynched and the remote locking now works fine so I’m confident that the immobiliser is not active – and if it was, I believe it would prevent cranking.

How it Should Work (briefly)
The P0001 code points to an open circuit around the VCV. It receives 12V from the ECM relay (R116) and 10A fuse 13E, and this is fed to one pin on the connector to the VCV. The other pin is fed to the Siemens ECM which has a switched output to ground (presumably via an open collector transistor) and the ECM modulates the valve to adjust the fuel supply from the supply (vane) pump inside the back of the HPFP (bypassing excess fuel back to the tank) to the 120 degree triplex high pressure pump which feeds the fuel fail at about 150 bar. There’s also a Pressure Control Valve on the output of the high pressure pump to compensate for pressure variations in the rail caused by the injectors opening. Both VCV and PCV are controlled by the ECM of course.
The Diagnostic Trail

All indications, but especially the P0001 code pointed to the likelihood that the VCV is not opening to admit fuel to the high pressure pump – one would expect that initially on cranking it should fully open and once the engine fires it should begin to be modulated by the ECM to regulate fuel flow.

The Diagnosis (so far)
First step was to try and establish that there wasn’t a problem with the VCV itself and then move to the wiring. So we have proceeded as follows:

• Check VCV solenoid coil resistance – OK at 2.7 Ohms

• 12V present at VCV from relay R116/fuse 13E – all correct

• Wiring continuity from ECM connector C0411C (the brown one) pin J4 to the VCV – all correct

• Continuity from C0411C (the brown one) pin J4 to the VCV through the solenoid and out towards fuse 13E – all correct

• Continuity from ECM power ground C0872L pins M1, M2, M3, M4 to battery ground connector – all correct

• 12V on the ‘top’ side of the VCV when the engine is cranking – correct

• Voltage developed across VCV solenoid when engine is cranking – not happening

• Current flowing through VCV solenoid when engine is cranking – not happening

Conclusion being that the ECM is not grounding the ‘other’ side of the VCV solenoid

The next step might seem a bit agricultural but we were desperate at this point:
• Disconnect ECM from VCV and apply 12V directly across the VCV
• Engine starts beautifully and idles smoothly, but there is no response to accelerator (not surprising I feel because the ECM knows it can’t control the VCV to regulate fuel supply in this condition

This last check has mostly allayed my fear that the Crank Position Sensor or the reluctor ring on the crankshaft may have failed (they’re original) and the 58 short pulses with 2 double-length pulses inverted may not be correct. I don’t have access to an oscilloscope right now (and if you want more info on checking this go to Christian and Vera’s LR Time YouTube channel) to check the pulse form but live data on the SnapOn code reader indicates 500RPM cranking and the smooth idle suggests all is well in this department but until I check with a scope I can’t guarantee this.
Conclusion

Conclusions (so far)
I have concluded now that the ECM is not opening the VCV and I can only think of two possible reasons for this:

• The driver transistor in the ECM for the VCV has failed

I think this is unlikely given the robustness of the Siemens ECM and the fact that there is nothing in the body-off procedure which is likely to have caused damage to the ECM. Also the ECM appears to be working perfectly in all other respects

• There is an input to the ECM which it needs before it will turn on the VCV which is not present

This is the more likely problem in my opinion, but I’m open to any suggestions at this point.

Questions

• First, does anyone have any useful idea of where I should take the diagnosis from here – have I missed something perhaps?

• Second, what is the startup process in the Disco3 as the engine is brought into operation?

I assume that it goes something like the following (but I'd be grateful for any corrections / amendments):

Key to ON
- Low pressure fuel pump activates and provides low pressure supply to the input of the HPFP (-0.3 to +0.5 bar required at the HPFP input)
- Glow plugs activate for appropriate interval then dash light extinguishes

Key to START
- Prerequisites: Transmission in Park or Neutral
- Starter motor engages (can be prevented by the immobiliser module) and engine turns over and the crank position sensor delivers pulses to the Engine Control Module (ECM)
- The Internal Transfer Pump (ITP) takes fuel from the filter and presents it to VCV thus on to the High Pressure (HP) pump
- ITP now delivers fuel to the HP Pump and the HP Pump delivers fuel to the rail

The HP Pump builds pressure in the fuel rail to at least 1500 bar (and the pressure sensor in the rail conveys this information to the ECM)
- The ECM now has the prerequisite information to commence activating the injectors
- The ECM opens the injectors and (hopefully) the cylinders fire and the RPM builds
- The ECM cuts the starter motor (or the driver releases the key) and the engine commences running on its own

The ECM uses the VCV to regulate fuel supply from the ITP to the HP pump and thus regulates supply to the rail. The Pressure Control Valve (PCV) controls pressure in the rail but its primary task is to even out fluctuations in rail pressure due to the injectors opening and due to the fluctuations caused by the pulses inherent in the operation triplex HP pump.
And SUCCESS – the engine is now running!

Is this an accurate sequence of events or have I missed something?

In order to cause the engine to run, the ECM must see:
- Immobiliser not active ie not preventing cranking
- Transmission in Park or Neutral
- Valid pulses from the Crankshaft Position Sensor (58 pulses then a double-length pulse followed by an inverted double length pulse (thankyou Christian of YouTube channel LR Time!)
- At least 150 bar in the HP fuel rail

If any of these is missing, the engine will not start.

Can the immobiliser allow cranking but not allow the ECM to open the VCV I wonder?

Thanks everyone for any information you can provide which might help – Neil and I will be very grateful

[RANDLOVER]

Can't you just ground the VCV coil's -ve side as the ECM seems to have lost that and hopefully it controls via the +ve side?

[shack]

I'd say you have a failing HPFP.

They can fault for some time before the die outright.

You will usually get injector control pressure errors for a while first.

The valves on the pump are what fail, they may fail mechanically not electrically...so hard to Diag

At least one of the valves is replaceable, but they are 30 - 40% of the value of a new pump post COVID.


But that's just my take.

As I side note I didn't think there were any EU2 tdv6s built, that's worse emissions spec than the last of the TD5s.

[discorevy]

Another vote for the HPFP, specifically the VCV and yep, I'd think EU3.
Is it possible the vehicle was starting to fault before and the Turbo was diagnosed?

[DiscoJeffster]

Can't you just ground the VCV coil's -ve side as the ECM seems to have lost that and hopefully it controls via the +ve side?

No electronics don’t work that way. One side is controlled, the other side is powered. The schematics will tell you that. As he pointed out, the positive is powered, the negative is driven, so you can’t ground the -ve and get control. Grounding the negative would drive it at maximum.

[DiscoJeffster]

I disagree with a failing pump. It’s not the symptom of one. You are not getting power to the positive side of the vcv which based on the diagram and how they work, means I believe you’re not satisfying a start prerequisite, which is why it’s not powering the relay to drive the vcv.

Why? You’re on the right track. Are you confident the code reader you have is seeing all the possible codes??

[Graeme]

Are you sure that the gearbox harness connection in the engine bay is sound?

[PerthDisco]

Are you sure that the gearbox harness connection in the engine bay is sound?

Yes it’s got the smell of something frustratingly simple and annoying.

[GrahamH]

No electronics don’t work that way. One side is controlled, the other side is powered. The schematics will tell you that. As he pointed out, the positive is powered, the negative is driven, so you can’t ground the -ve and get control. Grounding the negative would drive it at maximum.

Yes, absolutely correct. As you'll see in the diagrams and I explained in the text, one side of the VCV is powered from the battery +12 via a relay and a fuse - that's present and correct. The ECM provides a switched ground to the other side of the VCV - well it should but it's not. Grounding it permanently means the ECM has no control of the fuel supply to the injector rail, and although in that condition it allows the engine to idle, that's all it will do. The ECM has to control the VCV in response to accelerator position, engine speed and engine load for it to get beyond idle. There's no frigging it. I have to figure out why the ECM driver transistor is not switching, and as I said in the text, I can only think of 2 possible reasons:

1) the transistor is dead (seems highly unlikely) or

2) the ECM is missing a critical input to tell it that it's OK to allow the engine to operate at speed (much more likely I think).
But what input can it be missing? There are no codes to indicate any sensor which provides an input to the ECM is cactus. And I can't think of anything else it would need apart from the inputs and conditions I've listed. Well actually there are a couple when I think again ... it probably needs the 2 camshaft angle sensor inputs. Anyone else got any ideas?

And while I'm answering, you guys may be right - it probably is a Euro 3 engine. Not my car so I was relying on a comment from my mate, the owner about that.

Thanks for all your thoughts though - I really appreciate it.

[GrahamH]

I disagree with a failing pump. It’s not the symptom of one. You are not getting power to the positive side of the vcv which based on the diagram and how they work, means I believe you’re not satisfying a start prerequisite, which is why it’s not powering the relay to drive the vcv.

Why? You’re on the right track. Are you confident the code reader you have is seeing all the possible codes??

Thanks for the vote of confidence there! Yes, I am convinced that the pump is OK. The early ones have a reputation for not failing (unlike the later ones), and this one is an early one and has been tested and it performs to spec.

I'm absolutely sure the issue is around the switching of the VCV from the ECM (this based on my experience as a electronics tech with quite a lot of knowledge of car engines and automotive electronics - well high power broadcast transmitters, car electronics it's all basically the same stuff really just the power supply voltages are different! LoL)

[GrahamH]

Are you sure that the gearbox harness connection in the engine bay is sound?

Thanks Graeme, yes the gearbox harness is definitely sound and of course if the harness was disconnected, the engine ECM wouldn't know that it was in Neutral or Park and wouldn't allow cranking (I know this for sure from Christian's recent experience with Vera's D3 on LR Time).

However, the transfer box connector has had moisture in it. We do need to pull out the Transfer box ECU and check inside it to make sure that no water has got in past the connector. Apparently this is a common failing for owners who wash their engine bays. And Neil pleads guilty to doing that.

Suppose for a minute that there is an issue with the transfer box ECU or the connector. What would the engine ECM need to know from the transfer box ECU in order to start the engine by allowing fuel through the HPFP? If it needed something from the transfer box why wouldn't it inhibit cranking using the immobiliser ECU which is how the system prevents the engine from running given a reason that it's not safe to run for all the other reasons it prevents running? And if this is the issue, why is there no code to indicate that this possible input from the transfer case ECM is missing? (BTW, the Snap On code reader we're able to borrow is a professional instrument and costs a bomb, and the buss is only OBDII which isn't rocket science by any means so I seriously doubt it's not seeing codes which are actually present ... but never say never eh?)

However, thanks for the suggestions - I'll take this on board and investigate further. I am looking forward very much to being able to post that we've found the cause, corrected it and the car is now running normally. In the meantime there's more hair to be torn out I think. There will be a short hiatus as I head to Sydney to work for two weeks on Sunday. Keep watching this space!

[DiscoJeffster]

Every ECU will affect the engine as they’re all on the buses. You’d normally see cascading errors if there’s an upstream ECU issue though. I appreciate the snap on reader is expensive, I’m not convinced it can read all the ECUs in a Land Rover so I do wonder if there is an upstream issue it can’t see. That’s my only concern.

[Graeme]

The transfer case ecu isn't sealed - it is open across the bottom and the connectors aren't sealed. It's also a HS canbus gateway to lots of other modules on D3s, D4s and L320 RRS, although not on some L322s including my MY12.

[PerthDisco]

I’m sure you’ve carefully rechecked all the ECU engine harness plugs for any bent pins etc. and made sure they are firmly fitted.

[DiscoJeffster]

Josh Huber contacted me about this post. He suggested either trying another ECM and then switching back to yours or reflashing the current one. We have both had issues where the ECM won’t run the car after some time disconnected from the vehicle. My Indy had the same once working on mine too.

I don’t think your snap on tool will give you access to this, so you might need a dealer or an IID GAP tool to do a reflash.

Just another thing to try [emoji2373] and yes, it makes no sense, I know.

[GrahamH]

I’m sure you’ve carefully rechecked all the ECU engine harness plugs for any bent pins etc. and made sure they are firmly fitted.

Yep. They're all fine - straight and clean. But it was a possibility so thanks for suggesting that.

[GrahamH]

Josh Huber contacted me about this post. He suggested either trying another ECM and then switching back to yours or reflashing the current one. We have both had issues where the ECM won’t run the car after some time disconnected from the vehicle. My Indy had the same once working on mine too.

I don’t think your snap on tool will give you access to this, so you might need a dealer or an IID GAP tool to do a reflash.

Just another thing to try [emoji2373] and yes, it makes no sense, I know.

Thanks for the suggestion. It's certainly getting to the desperation levels now in trying to resolve this and this is kind of last resort stuff. Neither Neil nor I have a IID Gap tool (I've got a Nanaocom which I use for my D2 and my D4) but we're looking at purchasing one when funds allow. I'll certainly post an outcome if we do try this and let you know.

[GrahamH]

Well I'm delighted to report that Neil and I have finally tracked down the cause of this fault and, as you might expect, it was quite simple and very easy to fix.

You might recall that the ECM was reporting "VCV Open" and our detailed investigation confirmed that the Volume Control Valve on the high pressure fuel pump was not opening to admit fuel to the triplex high pressure pump from the vane type supply pump. We had also confirmed that the low pressure in-tank pump was supplying fuel up to the HPFP and the system had been thoroughly bled (several times). We had confirmed that there was +12V available at the appropriate terminal of the VCV when the ignition was live, but that there was no sign of a ground on the ECM side of the VCV, hence no current was flowing through the solenoid coil. We were able to confirm that the wire from the connection on the ground side of the VCV coil leading to pin J4 on the brown connector on the ECM, C0411C was intact. We also confirmed that the ECM's four ground wires (C0872L pins M1, M2, M3, M4) were indeed grounded and Neil had specifically cleaned the grounding point on the right front guard.

The way I saw it was that the ECM knew it was sending a signal to the driver transistor to turn on (probably an open collector arrangement as is common in such situations) and supply a ground to the VCV. But the ECM also knew that there was no current flowing through the driver transistor (I assume that there's an emitter resistor and the ECM monitors the voltage developed across it when current flows) thus the ECM reported that the VCV circuit was "open-circuit".

This all lead me to suspect that the driver transistor had failed for some inexplicable reason. Having run out of other options this afternoon I left Neil to remove the Siemens SID203 Engine ECM from the vehicle while I went home and emotionally prepared myself to perform surgery on it. This is not the sort of undertaking you approach lightly, even as an experienced electronics tech, and I knew that I could easily brick an otherwise perfectly good ECM if I got it wrong. We did however have a recovery path planned out which would have involved using a second hand ECU, swapping the EPROM with the VIN data, reading the old ECM and flashing it to the replacement ECM. But none of that was needed!

And the prize for correctly guessing the true cause of the fault goes to ...


**** PerthDisco! *****


Neil pulled out the ECM and then inspected the connectors in full daylight and square on. We had checked the pins carefully with a light in situ with it mounted on the firewall and we believed that we had seen everything satisfactorily. No so my friends!

On C4011C pin J4 was leaning from the vertical.

When Neil had first replaced the brown connector onto the ECM after putting the body back on the chassis, he had somehow caught pin J4 and instead of inserting itself into its socket, it had bent and ridden up the outside of the plastic moulding, fortunately into a void in the connector body. If there had been solid plastic there it would probably have folded the pin at 90deg and it would likely have broken when Neil straightened it. But luckily it was only slightly bent and he was able to straighten the pin successfully.

After replacing the three ECM connectors, and resetting the fault codes, the engine started and it now responds to the throttle as it should. At that point Neil broke out a beer to celebrate and phoned me.

So there you are. If you have occasion to remove the connectors on the Siemens ECM for any reason, you need to be very, very careful indeed about reinstalling them, by mating them squarely on top of the ECM. In fact, the safest way of doing it is to release the ECM from the firewall and leaning it forward so that you can see what you're doing because they are something of a PITA to get lined up. Then bolt the ECM back in position.

Neil now has his faith restored in his beloved "Hercules" (so named because of the load it carried when set up for major trips). However Hercules has been relieved somewhat of his burden over recent months as Neil has done a reappraisal of what he needs to carry, and the kitchen sink, along with a number of other accessories, will now be left in the kitchen.

Thankyou all for your suggestions, and encouragement. This has been a frustrating and prolonged investigation. But it has confirmed that if you look at the evidence, especially the fault codes, and proceed logically, you will find the resolution. Just don't make assumptions, and don't overthink it.

And now as I put this thread to bed, Neil and I wish you all good luck with resolving your own D3 and D4 issues.

[PerthDisco]

Just read this and have patted myself on the back. Well done and great news!!

I’ve just replaced my HPFP and just today gone through this exact process of reconnecting every ecu connection trying to be careful and straight as possible. Everything started and ran well.

They are lovely plugs with how they secure.