fuel pump primes 17 times then stops, engine won't run

[Toilet Duck]

Evening knowledgeable people,

My Ariel Atom (supercharged Honda K20 engine) has developed a weird issue. I was at a track day yesterday and all was well. Parked up in a pit garage and put some fuel in and then switched ignition on just to get a reading off the dash fuel guage. The fuel pump normally primes and then cuts out after 3 seconds or so, but for some reason it kept priming. Having investigated a bit further today, it primes 17 times and then stops. It does it 17 times every time you put the ignition on and repeat. If you try and start the car whilst it is doing this fuel pump priming/cycling it will start but idle very rough. After the 17 cycles of priming the engine dies. The engine will not start at all after this (the fuel pump is not running)

Here is a video I took today whilst investigating:

https://www.youtube.com/watch?v=5vwAjCGWtNc

There is a relay at the front of the car that clicks/energises when the fuel pump activates; I swapped that out today and it makes no difference. At the back of the car near the ECU there is this thing which also clicks every time the fuel pump runs/cycles:





Googling the part number at trackside yesterday reveals its a "fuel pump module relay" and is from (amongst other vehicles) a Freelander. I found a localish Freelander breakers/specialist and they took one off a vehicle they had there. I got them to connect up the module from my Atom. Although I didn't see them start the Freelander, they said it ran but didn't sound "right". I was hoping it would make the Freelander cycle the fuel pump like my Atom and confirm this was the faulty part but no joy. However, they kindly let me borrow the unit they took off their Freelander so I took it back to the track and fitted it, it didn't make any difference. Today I also resoldered all the pads on the printed circuit board inside my original fuel pump module relay and again no difference. I'm fairly confident it has eliminated this part from being the cause.

The Atom has a factory fit immobiliser which blocks (among other things) the fuel pump when activated. If I turn the ignition on without deactivating the immobiliser, both the relay at the front of the car and the fuel pump module relay click 17 times but the fuel pump does not run. I don't think the immobiliser is causing the issue.

There is a quick release connection on the fuel line feeding the fuel rail. I disconnected this and aimed it into a bucket. When the ignition is on and the fuel pump starts cycling/priming fuel pisses out.

I've googled to see if the Honda K20 engine has a fuel pressure sensor incase that is the problem but it doesn't appear to have one.

I don't know if there is a fuel pressure regulator (maybe that red thing visible in the video?) or if a faulty one could even cause this issue?

I am at a complete loss now. I hate electrickery faults on cars. Can anyone help or point me in the right direction?

I would really appreciate any help as this has me stumped!

Many thanks

[GreenV8S]

That is wierd and I haven't encountered anything like that myself. I don't have an Atom or a Freelander and don't know how the fuel system works on those, but the vehicles I'm familiar with behave the way you describe - the pump runs for a few seconds when you switch the ignition on, then runs continuously when you start cranking the engine and as long as the engine remains running. If the engine stops, the pump will stop after a few seconds. Fuel pressure is regulated via a mechanical restrictor and this has no influence on the pump.

I agree the symptoms do not suggest a faulty immobiliser.

Based on this I suspect the ECU may be being reset at the end of the priming pulse. The fuel pump relay is normally controlled directly by the ECU and should include a fly-back diode to prevent a back EMF spike to the ECU when the relay is switched off. It's a WAG, but the relay or its diode might have failed in a way which is spiking the ECU and causing it to reset.

You might be able to confirm that by removing the fuel pump relay and connecting a small test lamp across the relay control terminals. You should see it power on for the length of the normal priming pulse and then switch off. If it still does the cycling thing, it disproves the relay theory and suggests you need to look elsewhere.

Do you normally see any other symptoms when the ECU switches on, such as sweeping dials or cycling warning lights? If so, that could also confirm/deny that the ECU is resetting.

[Toilet Duck]

GreenV8S said:

That is wierd and I haven't encountered anything like that myself. I don't have an Atom or a Freelander and don't know how the fuel system works on those, but the vehicles I'm familiar with behave the way you describe - the pump runs for a few seconds when you switch the ignition on, then runs continuously when you start cranking the engine and as long as the engine remains running. If the engine stops, the pump will stop after a few seconds. Fuel pressure is regulated via a mechanical restrictor and this has no influence on the pump.

I agree the symptoms do not suggest a faulty immobiliser.

Based on this I suspect the ECU may be being reset at the end of the priming pulse. The fuel pump relay is normally controlled directly by the ECU and should include a fly-back diode to prevent a back EMF spike to the ECU when the relay is switched off. It's a WAG, but the relay or its diode might have failed in a way which is spiking the ECU and causing it to reset.

You might be able to confirm that by removing the fuel pump relay and connecting a small test lamp across the relay control terminals. You should see it power on for the length of the normal priming pulse and then switch off. If it still does the cycling thing, it disproves the relay theory and suggests you need to look elsewhere.

Do you normally see any other symptoms when the ECU switches on, such as sweeping dials or cycling warning lights? If so, that could also confirm/deny that the ECU is resetting.

Cheers for the reply GreenV8S

The dials sweep when you first switch the ignition on. This is functioning correctly i.e. they sweep once whilst the fuel pump does it's cycling/priming dance

Sorry but what does "WAG" mean?

I'll give the fuel pump relay lamp trick a go and report back, probably be after work tomorrow.

I might also see if I can borrow an ECU off another Atom owner and see if I can do a quick swap, but that will take some time to organise

Someone on the Atom forum kindly posted up some wiring diagrams which I think may help:









I have a multimeter etc but on the face of it this issue is far more complex than anything I've worked on (electrics wise)

Many thanks again

Edited by Toilet Duck on Wednesday 19th September 19:33

[Toilet Duck]

This gets weirder....

The relay at the front of the car that clicks 17 times in sync with the fuel pump priming isn't the relay for the fuel pump. It's the throttle pedal relay (it's drive by wire). Could the throttle position/pedal be somehow causing this?

[GreenV8S]

What does a "throttle pedal relay" do?

[Toilet Duck]

GreenV8S said:

What does a "throttle pedal relay" do?

I have no idea, google has not helped. I assume it's because it has a drive by wire throttle so "electric" rather than cable and the relay feeds that, but even if that's correct surely a DBW throttle wouldn't pull enough current to require a relay?

[Toilet Duck]

I just did a test. With the ignition on (and the fuel pump dancing on/off), when I operate the throttle pedal you can hear the butterfly valve in the throttle body opening closing. I'm fairly confident that's normal i.e. if the fuel pump was behaving and just priming once, it's "normal" for the throttle pedal to actuate the throttle butterfly with the ignition on but not the engine running.

What I did notice is that once the fuel pump has cycled 17 times, the throttle pedal no longer activates the butterfly, it goes dead. So Is that pointing more towards an ECU fault i.e. its making the fuel pump dance and then killing the DBW?

With the ignition on and the fuel pump dancing, if I remove the "throttle pedal relay" the throttle pedal no longer actuates the butterfly

[Steve_D]

Have you tried pulling that relay or disconnecting the throttle pedal.

Steve

[Toilet Duck]

Steve_D said:

Have you tried pulling that relay or disconnecting the throttle pedal.

Steve

Yep tried both, neither effects the fuel pump dancing 17 times. It is bizarre!

[Toilet Duck]

Toilet Duck said:

Steve_D said:

Have you tried pulling that relay or disconnecting the throttle pedal.

Steve

Yep tried both, neither effects the fuel pump dancing 17 times. It is bizarre!

Just tried again for piece of mind and the above is correct.

Just to confirm though, whilst I can actuate the butterfly valve with the throttle pedal, each time the fuel pump completes one of its 17 cycles and both the fuel pump module relay and the throttle pedal relay click, the butterfly valve shuts as the relay closes then opens up as it energises. So effectively, both the fuel pump and drive by wire throttle cycle 17 times in perfect sync before both going dead

[GreenV8S]

I'm not familiar with that type of DBW system. I would have thought the actuator would be driven electronically without needing a relay, but maybe there's some clever scheme for powering the whole thing down while the engine is off. I suspect the fact it's clicking is a consequence of whatever is cycling the fuel pump relay, rather than the cause.

It seems to me that you're already quite a long way into understanding how the fuel pump control system works and have already logically ruled out the obvious component failures. I think the next step is to research the wiring diagram and understand how the fuel pump is controlled (presumably eventually coming back to an ECU) and see what is happening within that ECU. I suspect it will involve interrogating the ECU with a suitable scanner.

Unless you're inclined to take on quite a challenging DIY problem, this seems a sensible point to get a professional on the case.

[Krikkit]

On each occasion when the pump primes, then stops, does that take longer or about the same as it usually would to prime and get up to rail pressure?

I'm wondering if it's the pump failing to pressurise sufficiently, then the ECU resetting for another try.

[Kccv23highliftcam]

Krikkit said:

On each occasion when the pump primes, then stops, does that take longer or about the same as it usually would to prime and get up to rail pressure?

I'm wondering if it's the pump failing to pressurise sufficiently, then the ECU resetting for another try.

Is there such a thing as a fuel rail pressure or pump pressure sensor ? Is the Honda K20 engine supercharged "by Honda" [sorry I don't know just throwing thoughts in]

Is the inertia switch environmentally sealed?
What should the output pressure of the fuel pump actually be?


But why the 17 times???

Edited by Kccv23highliftcam on Thursday 20th September 13:45

Edited by Kccv23highliftcam on Thursday 20th September 13:53

[Toilet Duck]

Krikkit said:

On each occasion when the pump primes, then stops, does that take longer or about the same as it usually would to prime and get up to rail pressure?

I'm wondering if it's the pump failing to pressurise sufficiently, then the ECU resetting for another try.

I think it takes about the same time as a "normal" single prime when the car would ordinarily start and run fine

Kccv23highliftcam said:

Is there such a thing as a fuel rail pressure or pump pressure sensor ? Is the Honda K20 engine supercharged "by Honda" [sorry I don't know just throwing thoughts in]

Is the inertia switch environmentally sealed?
What should the output pressure of the fuel pump actually be?

I can't find any reference to a fuel or pump pressure sensor when googling K20 Honda engine etc.
The supercharger (Jackson Racing) is aftermarket and fitted by the Ariel factory.
I don't know if there is an inertia switch or what the actual pump pressure should be unfortunately.

Kccv23highliftcam said:

But why the 17 times???

I only wish I knew, it is bizarre! I think I have exhausted my DIY skills. I suspect it's ECU related so I'm going to contact a place familiar with Hondata ECU's and get this looked at by professionals.

Thanks for everyones help, if anyone has any ideas in the meantime that are DIYable I'm all ears. I changed the fuel filter earlier but it didn't make any difference.

Cheers

[jeremyc]

Toilet Duck said:

Kccv23highliftcam said:

But why the 17 times???

I only wish I knew, it is bizarre! I think I have exhausted my DIY skills. I suspect it's ECU related so I'm going to contact a place familiar with Hondata ECU's and get this looked at by professionals.

Thanks for everyones help, if anyone has any ideas in the meantime that are DIYable I'm all ears. I changed the fuel filter earlier but it didn't make any difference.

An uneducated guess, but the ECU could be using a 4 bit "retry" counter that starts at 0000 (binary) and increments by one after each cycle.

By the time it's got to 1111 (binary) it will have tried 17 times.

It doesn't help identify the problem though, which presumably is some kind of failure with the ECU deciding it can't (or shouldn't) start, so tries the cycle again.

[eliot]

if the fuel system is "returnless" it will almost certainly have a fuel pressure sensor in order regulate the pressure. Older returned systems have a pipe running back to the tank after the mechanical fuel pressure regulator - I would be surprised if that was the case on a modern engine.

As suggested above, it might be cycling trying to get a valid fuel pressure and giving up - would also explain rough running. Is there a shrader valve on the fuel rail where you can check the pressure (you may need someone with the same engine to tell you what normal looks like, i.e. static pressure, pressure during prime pulse, pressure on crank and running)

[stevieturbo]

You're wasting your time fking about.

Interrogate the ecu, check for codes and ensure all sensor feedback appears normal and that the ecu is actually alive and happy ( or not )

[Toilet Duck]

eliot said:

if the fuel system is "returnless" it will almost certainly have a fuel pressure sensor in order regulate the pressure. Older returned systems have a pipe running back to the tank after the mechanical fuel pressure regulator - I would be surprised if that was the case on a modern engine.

As suggested above, it might be cycling trying to get a valid fuel pressure and giving up - would also explain rough running. Is there a shrader valve on the fuel rail where you can check the pressure (you may need someone with the same engine to tell you what normal looks like, i.e. static pressure, pressure during prime pulse, pressure on crank and running)

No shrader valve that I can see. I can't see any pressure sensors etc in the fuel line/rail either





I'm wondering if this red thing plumbed into/near the fuel filter is a pressure regulator, and if so could this cause the issue if it's somehow gone tits up?:

stevieturbo said:

You're wasting your time fking about.

Interrogate the ecu, check for codes and ensure all sensor feedback appears normal and that the ecu is actually alive and happy ( or not )

I know, I hoped I could fix it but it's pretty obvious it's beyond my skill level. I've already emailed a Hondata ECU specialist who are due to do some work on my car in a few months anyway, I'm hoping that they can have a look soon.

[stevieturbo]

it has the potential to be a FPR....although who on earth is using crappy worm drive clips on those fuel lines. Totally the wrong type of clip.

But as an OEM Honda ecu to start with, you'd like to hope any decent fault reader/scanner should be able to view at least a good amount of data, although ecu specific would be better.

[Toilet Duck]

stevieturbo said:

although who on earth is using crappy worm drive clips on those fuel lines. Totally the wrong type of clip.

Ariel Motors finest