Dual speed fuel pump relay wiring

[Kaldek]

For anyone that cares, I dug up a wiring diagram for an Evo that matches the wiring of the Legnum dual-speed fuel pump relay.

This is from an Evo 5, but it's identical to the Legnum/Galant. The "Fuel Pump Relay No. 2" is mounted to the Firewall of the car and has a rubber boot over it. It is a 5-pin relay that switches between two different circuit paths rather than just open and close a circuit. The ECU either controls this relay to force power to the fuel pump via the path with a resistor in it, or via the direct path that has no resistance. The resistor reduces the volts which reach the motor which in turn lowers the speed of the pump. As this is a high amperage circuit, you can tell that the resistor is the big square with the heatsink on it located right next to the fuel pump.

Just ignore the fact that mine is painted a nifty shade of cyan.

 

[Kaldek]

Note that right next to the fuel pump relay and resistor there is supposed to be another bracket with a single-pin connecor that goes to a blanking plug. This is the "fuel pump check connector" and supposedly allows you to read the voltage that is being delivered to the fuel pump.

I checked this plug and was only seeing 1 volt though, so I'm not sure what's going on yet.

 

[AKKO]

Also previously covered in some posts here for reference:
http://www.ozvr4.com/threads/how-to-hotwire-your-fuel-pump-to-battery-voltage.740/ not advocating the practice suggested by the thread title however

 

[Kaldek]

Thanks Akko, I didn't know of that thread. It's very handy and explains this all from different point of view.

I think it's fair to say that cranking up the flow all the time is overkill and may cause over fueling with the stock FPR (hence the two speed design in the first place).

 

[naughtika]

thanks for this thread, I gotta admit.. I have been lazy with looking into my own issue..

if we were to replace that relay should a 30 or 40 amp 5 pin change-over relay be enough?

diagram below "should" work..

I think connection should be..

1 - 85
2 - 87a
3 - 86
4 - 87
5 - 30

 

[Kaldek]

Actually mate I think the best test for your issue is a relay bypass test. I would start by hotwiring the fuel pump directly off the battery by running a fused link from the battery positive terminal to the fuel pump power feed. This is the middle pin of the three pins that come off the relay.

Next I would test that this relay is energised by the first fuel pump relay (apparently it's behind the stereo!).

Also see next post for some extra very interesting data I just gathered.

 

[Kaldek]

OK folks I have just confirmed some very interesting facts (video to come later).

• When the car is cranking (key in the "Start" position to crank the motor over), the ECU provides the full 12 volts to the fuel pump
• When the car catches and an rpm signal is detected, the ECU immediately switches the fuel pump relay over to the resistor path (~10 volts)
• With the resistor in-path, the idle speed fuel pressure is 39psi
• If I bypass the resistor using a jumper wire, the idle speed fuel pressure is 44psi
  • If you disconnect the resistor circuit completely, the fuel pump stops and the car will stall
  • The car can be started with the resistor circuit disconnected, but stalls as soon as the ECU switches over to the low speed circuit (duh)
• The resistor has a resistance of about 0.5-0.7 ohms.

Mitsubishi was correct to reduce the fuel pump speed at low load; the fuel pressure regulator cannot resist the flow and as a result the fuel pressure is too high.

Could they have dealt with this through software in the ECU instead of a dual speed relay? Possibly, but I suspect there are too many "analogue" variables that would require the ECU measures the pressure in the rail to overcome.

The downside to this approach is that IF your relay craps out you may end up in a situation where the car runs but then goes super lean under load because the 12 volt path is not switching to speed up the pump.

 

[Kaldek]

There is some very interesting relations to my own lean-running problems when I had the car tuned:

• The previous owner replaced the fuel pump with a Walbro, but did he do it for performance or to fix a problem?
• The previous owner tried to grease up the fuel pump relay with dielectric grease (and probably didn't understand what it does), so was he also trying to fix something?
• The previous owner had the injectors cleaned, but did he do it for performance or to fix a problem?
• The car ran lean under load

A 255lp/h walbro pump gets its higher flow from a higher kV rating (which basically means motor rpm per volt, at the expense of more amps drawn). It's quite possible that the stock fuel pump in the car was seriously underperforming due to electrical issues, and when he installed a Walbro pump which has a higher flow per volt applied it was almost keeping up with demand, but not quite.

It's almost like the previous owner discovered it was running lean, tried a bunch of stuff to fix it and almost got it dialled in but not quite. Possibly due to a dirty relay? Certainly worth testing.

 

[Kaldek]

This morning I hotwired the fuel pump directly off the battery. To do this, you must disconnect the fuel pump relay and run a 10-amp capable wire between the battery positive terminal and pin #2 of the relay plug. This is the middle pin of the three pins in a row on that connector. It will spark when you first connect it so be ready for that, coupled with the sudden noise of the pump engaging.

Hotwiring the pump allowed me to measure the current being drawn by the pump at battery voltage. The Walbro 342 255lp/h pump was pulling roughly 6 amps during the test.

 

[naughtika]

I thoroughly cleaned my relay's pins and connector housing today and the car had started up just fine.. but then again the engine was a little warm already. Might try it again some other time.. like next week ?!

 

[Kaldek]

Updating this thread to point out that my car no longer runs lean and the reasons for it running lean were purely bad sealing of the fuel pump in the tank. Fitting the right O-rings and spacers behind the O-ring solved the issue. This was diagnosed fully using a Wideband O2 meter and comparing it to the desired/target AFR in the fuel map using EvoScan.