Scanned and no fault codes (of course!). Everything looked plausible except live data showed large deviation in fuel rail pressure from desired - both above and below. For me that and "OK at high revs/load" pointed away from injectors or the LP pump which I would be looking at if it were constantly low on desired FRP.
These engines have both a Fuel Pressure Control Valve (FPCV) on the rail and an inlet metering valve (IMV) on the pump @Iver posted a good summary of how they work together viewtopic.php?p=98778#p98778
I suspected one of these so I hooked up the Pico to the FRP sensor and to the control wire to the FPCV and amp clamped it's ground. I was going to do the IMV next.
The FPCV is PWM controlled so that raw data is dragged off the screen in this image - (though it looks perfectly fine of course) but the light green is a maths channel % on time, blue is FRP, red is amps flowing through FPCV
My interpretation of what is going on is that the pressure regulator is sticking closed so that the pressure builds until the "pintle bump" on the amps at 526ms and 774ms is the solenoid unsticking and going full open and the pressure drops to 0.5v which is 0 bar. The duty cycle then goes right up to fix it, tapers off when it starts to approach 1.3v, but the solenoid is still stuck closed so duty drops almost to 0 at which point it slams open again, pressure drops and rinse/repeat.
As the current anomaly occurs whilst the command duty does not change it would seem to be definitely a solenoid sticking issue. I do have the IMV signal on there as well but there seems to be a bug in PS7 getting it's maths channel to display.
Looking for an anomaly in the current draw by either the rail or pump pressure regulator should be a quick way to determine which unit is at fault. When combined with the duty cycle maths channel you can be certain that it is not an external influence - for example if the FRP sensor was dropping out and the regulator was reacting to that you would see the current as a reaction to a signal command.
This could also be extended to any PWM control solenoid/actuator prone to sticking - EGR valves I'm looking at you.
- 20220804-0001 T5 Pico Forum.psdata
- (10.35 MiB) Downloaded 96 times
Remembering that there had been some parts cannon of after market parts before I got there I suggested he replace the previously changed FRP sensor with the original, which he did and the engine then ran perfectly.
Thank you for sharing, I have enjoyed to read it.
thanks for the case study
does/should the rail pressure stay higher when the anomaly happens?
The marker line at 1.284v on the pressure sensor indicates more or less what pressure the ECU wants, as the pressure approaches this it starts to reduce the duty cycle (green) on the control valve to open it. However it is stuck closed so the pressure overshoots, the ECU then reduces the cycle even further but the pressure continues to rise until the valve lets go - indicated by the current spike. As the valve is fully open the pressure drops to 0 (0.5v), the ECU see's that so commands the valve shut again (green starts to go back up) where it sticks closed... and repeat.
These CR engines only have an inlet metering valve - IMV / flow metering valve on the HP pump. It is a normally open system meaning that if you disconnect it the pump goes full pressure (1800 Bar), so at idle it has about 0.7A to keep the fuel pressure down to 250-300 bar and it decreases that to raise the pressure.
Green is fuel pressure, blue IMV current and grey is the maths channel -ve duty cycle on the ECU side of the IMV. So I boot it at about 1.5s and as you can see the duty cycle drops, current decreases and pressure rises - not very smoothly.
Pressure peaks at 4200RPM and about 1500Bar, foot off and the DCy shoots up to close the valve which it does until it gets stuck at 6.5s, then the ECU kicks it with almost double the normal closed current to get it to fully close to idle pressure.
Here is what happens on a hard start up. The engine is cranking but not firing as the pressure is too low for the injectors to fire. The DCy and current is as low as can be waiting for the valve to unstick, which it does just after 1.127s, and then the DCy shoots up to catch it. There is then some rough running with the valve and the ECU fighting each other until 3.5s when the pressure smooths out and the car idles perfectly.