Wiring/power issue to fuel pump connector - HELP (lots of detail)

[Neeek]

OK, this may be a bit long...

I changed out my fuel pump on Saturday - went from stock unit to a Walbro 255 (because M30 turbo). Everything went OK, got everything back in, test started the car and it chugged into life, I was happy. An hour or two later, I went to head out somewhere and started her up (all fine), was checking a message on my phone before reversing out and the car spluttered and died. Tried to start it again, couldn't hear the pump prime before cranking. Hauled the whole lot out again.

Things I have tested, and are known to be good and functioning correctly:

1. Main relay terminals
2. Fuel pump relay terminals
3. Fuses
(I did replace all the relevant relays and fuses with new items, just in case something was bad - main relay, FP realy, fuse 23 and fuse 17)
4. Power to the connector at the fuel tank (to sender and pump assembly in boot/trunk, pins 4 and 5 on the connector)
5. Pump verified as working using an alternate 12v power source
6. Crank position sensor gets 545 Ohms
7. Fuel sender, when plugged in, working 100% correctly - you can turn it upside down and the gauge on the dash moves from empty to full

Here's where it gets interesting. As soon as I hook ANYTHING directly up to the connector in the boot/trunk, voltage disappears. With a test light, key in position 2, voltage at the connector drops to 8v from 12.6v. Plug in the pump by stuffing wires into the connector directly? 0.04v.

I have also checked the voltage on the two brown wires that connect to the pump through the sender unit when the connector is attached. Nothing. Again, remove the plug, and 4 and 5 have 12.6v with the key in position 2, as it should do.

I'm running out of options, here. My only thought at this stage is to run additional wiring from the fuel pump relay (because it clicks on and off with the key in position 2, as expected to prime the pump, so it's functioning correctly) and run that directly down to the fuel pump with another relay and fuse under the back seat, bypassing the actual loom and the plug. But I can't for the life of me figure out what's going on with the power supply to that connector, and why voltage just disappears when you add anything other than a multimeter to the circuit...

Hope you guys have a couple of ideas for me to try.

[jehu]

Did you examine the wire bundle where it goes through the backseat/trunk wall near the battery? I had that power wire break and a wire was run through a nearby hole directly to good section of the power lead. Problem solved.

[moroza]

What ^ he said - likely damaged wiring where it exits the body near the RR wheel. Some strands get broken resulting in non-linear high resistance; when current is low (such as checking with a voltmeter or ohmmeter), resistance is low so there is no voltage drop, but when current is higher (such as when the pump draws its full power), resistance becomes higher and significantly drops voltage.

How much power does your aftermarket pump draw? Perhaps it damaged the wiring by overheating it, which would explain the timeline of failure better than random wire damage.

[Neeek]

Good suggestions, both - will check, thank you.

As for the new pump drawing more power, that may be the case. But if drew more than 7.5A, the fuse would've blown first, right?

EDIT - according to this page, it' might've been more than 7.5A... http://www.roadraceengineering.com/f...pflowrates.htm. Mine is the 342 model...

[moroza]

Not necessarily. Funny thing I've observed about how wiring (mostly on this car but also on others) appears to be designed: there seems to be an assumption that under normal conditions, current will be well below what the wire and the fuse can handle, and that any overload is going to be severe, like a short-circuit. However, countless slightly melted fuse holders and occasional damaged wiring - with intact OE-specified fuses - suggest that OE wiring designs often don't account for slight overload. I suspect the reason is that wiring is sized to be the smallest possible that will deliver the required current, with primary focus on acceptable voltage drop, instead of on how little heat the wire can generate or how much it can absorb.

For illustrative purposes, pulling numbers out of where the sun don't shine...
0.75mm wire over 3 meters can draw 5 amps before the insulation starts slow-cooking (hardening), 7 amps before the 13.0V at the source becomes an unacceptable 11.0V at the consumer, 15A before insulation starts fast-cooking (melting), and 50A before the conductor itself melts. When it's feeding a 3 amp fuel pump, everything's hunky-dory. If it develops a short, its 10A fuse blows and everything else is saved. If it's asked to feed a 6 amp pump, this wiring can overheat and become damaged, especially anywhere it has a spot of higher resistance - connectors, tight bends, pigtail splices. Insulation that gets baked hard is prone to cracking, which then leads to the conductor fatiguing from lack of support (even if it doesn't flex in use, the vibrations in a car can fatigue it; that's why cars don't use solid conductor wire), which leads to a partial and eventually a full break. Meanwhile there may be areas with melted grommets, plastic connectors, or the wire slightly stuck to adjacent wires in a harness, and the 7.5 or 10A fuse may slightly melt its housing but remain unblown.

[Neeek]

You're probably right, and to be honest I'll never know now... we just bypassed the factory loom for the FP. Used the 87 output on the FP relay to trigger another relay under the back seat, fed by the battery right next to it. Dry tested the FP in the boot/trunk, works exactly as it should. Now to get the pump back in the tank with the right wires running to it, and away we go again! Only this time, my M30 turbo will have enough fuel to not lean out horribly under very high loads...