I know it has been done, and I know it is in use on some of the commercial PHEV conversions. But I would like to know if anyone here knows how? I know they do not cut the fuel supply as it is instant, and I know it throws codes so it is a global error. I have read that some on here (bwilson mostly) do lots of tests including out of gas tests. I am wondering if there is any data gathered that would allow a normally functioning and fueled vehicle to "think" it was out of gas. Then on reset be ok (assuming the codes were cleared). My thought is that there has to be a fuel system error message of some sort transmitted over the CANbus or there is an easily altered sensor line. In normal driving conditions, what trips an "out of gas" fault? There must be a sensor close to the engine heirarchy that signals this. So we/I should be able to signal this sensor line artificially and cause it to think it is out of gas. Before I go sit at my local gas station for hours running it out of gas and back again trying to capture any sensor data, I am hoping the answer already exists. If you have any ideas on the cycle I would glad to hear it.
I assume it is a viable state, so if I can stimulate the environment in a way to engage such state that would be great. So stimulate the environment to get the car to simulate being out of gas...
Well, in my 2008, I flip a switch to "simulate" an out of fuel condition. The switch cuts power to the fuel pump, the engine sputters for a few seconds, and BAM! red triangle, and now I can drive up to 52mph on battery. Don't know if that is the same thing you are looking for, but it seems unlikely that the ECU would rely on the gas guage to decide when to declare "Out of Gas". Anyone else know? BTW, I don't regularly kill the pump since the engine does its little coughing fit and you do have to reset the code, which can be a PITA.
But the next stage of the question is, during it's sputtering fit what tells the car that you are out of gas and lights up the dash like a Christmas tree? There must be a sensor in the engine area that monitors this, or maybe even a message broadcast over the network about the event... And to confirm, once you flip the switch the remaining fuel in the line is used. Then the car sputters, then the ICE dies. The dash lights up, and you get 52mph stealth mode. This mode is maintained until power reset. Correct? Now... upon reset I am guessing the error codes persist? But the ICE should spin and everything OK, just you have a shining dash board right? Or does the car refuse to restart the ICE until the codes are cleared? Do you clear the codes the normal way of sending commands over the CAN bus? Is there an easily accessible wire in the cabin? Are you using a physical contact to break connection between the fuel pump power line and a power source? This might be perfect for me to start experimenting with the fault conditions. My understanding is that the HV ECU is the "master brain". Almost all decisions regarding making the car go and stop is at one point or another processed through the HV ECU. I believe the HV ECU also controls power to some of the lesser ECUs, and allows a graceful predetermined startup sequence within the network. Is this anybody else's conclusion?
Most likely the crank-shaft sensor which would be detecting one misfire after another, until the engine quits turning. I'm not sure the vehicle knows that it is out of gas, but it knows the engine will not run.
If it is the crankshaft sensor, then with the right amount of work we should be able to override the normal signals and have it produce erroneous reports of misfires (without actually misfiring the engine). This should inturn tell the car to fault and shut off the ICE with no actual or possible damage to the mechanics. Anyone know where this sensor is, and how to access it? It could be other sensors too, or a combination of sensors that must all report faults, who knows. Although I would think in a situation like this a single sensor error should OR together with other sensors and create a single fault flag. I would hope this is one of the sensors that makes its way back to the HV ECU in raw form...
Actually it appears the appropriately named Engine ECU controls the engine It is hard to trace the wires back, but I see no direct connection between any sensor that could detect an exhaustion of the fuel supply or a rough/erroring engine back to the HV-ECU. Since the HV-ECU must be at a minimum notified of this event it stands to reason that the CAN bus is used to transport a fault message to all systems. (Pure speculation at this point btw)
Sorry, I can't really help with the NHW20 as we only have the NHW11 and ZVW30 and both are substantially different from the NHW20. I've never tried to 'spoof' an out of fuel condition . . . I just go out and run it out of gas with a spare can in the back. The usual excuse is I want to burn off the last of the current tank so I can switch brands or grades of gas. There have been speculations ranging from shutting down the igniters, injectors and fuel pump. But most folks with the NHW20 just hack the EV switch and don't worry about it. Bob Wilson
Well thankyou anyways Bob. The ZVW30 does not have a "limp" mode after fuel exhaustion correct? Not sure I understand that design decision... As an aside, can anyone confirm that a NHW20 can lubricate parts sufficiently with the ICE spinning but not burning fuel? i.e. using MG to spin the engine while denying fuel into the injectors. Obviously not efficient, but it seems like it would work to periodically lube up the parts if not driven with the ICE for a long-ish period of time...
I've been doing a lot of work on the CAN bus lately, and I've got lots of cool tools to analyze things like this. When I get done with the software in my replacement battery ECU, I'll take a sniff. I'm reasonably sure the ECM (engine control module) must determine that there is a problem (out of fuel) and set a code. Either the HV ECU is informed of this code, or by some other parameter on the bus. I'll put my CAN Isolator on the ECM and log a few minutes of normal engine run, then I'll cut the fuel by removing the power to the fuel injectors. This should result in a faster "die" without all the stumbling. Then I'll see what happens on the CAN bus and see if I can isolate the frame that actually triggers the HV ECU to go into 52mph EV. Shouldn't that hard. The trick would be some way to reset that condition without stopping the car.
Well I think there is some two way comm going on. Engine ECU: Warning Flag, I'm dead HV ECU: Oh Crap, battery power. Sends message out on bus for everything signaling this event ??? Perhaps now the HV ECU sends a special subset of signals to the Engine ECU to control only electric power or it sends a single message to the Engine ECU and it manages the electric only... Either way, if we capture this message or set of messages, then theoretically, we can isolate the Engine ECU from the CAN bus. Send specific messages back to the bus, other trap. Similarly, we can confiscate all HV ECU messages and alter them to be what they would be had it died. This way the HV ECU thinks everything is kosher, but the Engine ECU is throwing a fit silently. That should not require a restart to bring back to life. Should work just by allowing the messages back through from the HV ECU. Thoughts?
The HV ECU doesn't have to tell the ECM it's going all electric, It simply does it. The HV ECU controls MG1, MG2, the A/C compressor, and contains the Hybrid decision-making logic. It's also where the EV parameters are decided, and it actually decides when to start the ICE and provides the starting torque. Once the ECM complains that it cannot run anymore, the Hybrid ECU "locks it out" and continues to run in all EV mode. Apparently, it will not reset unless powered down.
