Planning gas-out experiment

We've been discussing 2010 running out gas and a recent driver found it would not start after running out of gas. They had tried to restart the car without having added gas and someone found a note that it may take ~2 gallons. Regardless, I'm thinking about running our NHW11 out of gas and seeing:

• how many attempts before it throws a code gives up
• what code might be raised
• try the 12 VDC reset

There is a risk I might have to get it towed but I'm good with that. Since I have a Graham miniscanner, I'm thinking about running a test:

• carry a spare can in the trunk
• run out of gas and park in a safe place
• read the codes
• count restart attempts until another code shows up
• add the extra gallon and verify it won't start
• try the 12 V battery disconnect for 5 minutes
• check for other codes
• attempt to restart the car

Suggestions?

Bob Wilson

 

I've done over three dozen gas outage tests but I never tried repeated attempts to start the car without gas. I always drove to a safe place, added the one gallon, and started the car. It always worked but I was aware of reports of repeated, failed starts leading to a 'code' that the dealer had to clear.

I want to replicate the problem with our NHW11, verify that it has the same logic, and document the code. Then I want to try the following fixes:

1. 12 VDC reset - disconnect the 12 VDC battery for say 2-3 minutes to clear the code
2. Graham scanner - clear the ECU codes
3. OBD clear (optional) - using an OBD extension cable, see if the Tc signal can clear the codes without having to disconnect the 12 V battery

If these work with our NHW11, I'll repeat the test with our 2010 Prius and document what works. Then comes the fun part.

Posting the proposed protocol has suggested another approach, splice in a switch on the fuel pump circuit. I also want to monitor the fuel pump current to get an idea of what happens when it runs out of fuel. So this may be a better protocol.

Bob Wilson

 

I like it but will continue the current protocol. I've got the spare can and everything needed in the car. Once I get the first test done, I can repeat it easily enough in the driveway and see if there are any other differences.

Bob Wilson

 

I've done it over three dozen times but my goal is to reach a safe place to park. So I drive at the slowest speed possible based upon the conditions and as quickly as possible, pull into a safe place to park and refuel.

The fastest I've been going is ~50 mph when the ICE ran out of gas. Rather than 'race the traffic,' I simply pulled over to the slow lane and let the car slow down. About a 1/4 mile down the road, I pulled onto a wide shoulder.

My longest EV mode driving is at about 20-25 mph cutting through the neighborhood returning home or the last mile into work. My EV goal is distance, not speed.

Bob Wilson

 

Executive Summary

After five attempts, there is no evidence of a 'poison' code to prevent further restart attempts. One simply has to turn the car "OFF" and then you get one 'restart' attempt per "OFF-to-ON" cycle.

Details

I was running just under 50 mph when the engine ran out of gas. Knowing there was a turnoff about a mile up the road, I shifted into "N" to extend the glide and turned on the emergency flashers. When the speed bled off from climbing a gentle grade, I put it back into "D" and crested the rise. I made a right turn onto the road leading to the airport and pulled onto the shoulder. Battery module lowest and highest voltages:

• 15.4V - 15.55V - at fuel exhaustion

The initial codes were:

• P3101 0000 0000 - Hybrid ECU
• P0000 0000 0000 - Battery ECU
• P3191 0000 0000 - Engine ECU

First attempt to restart had no effect but I had not turned the car off. So I turned the car OFF and back ON without trying to start and still had the same codes. So I tried start and the ICE spun up. The new codes:

• P3101 0000 0000 - Hybrid ECU
• P0000 0000 0000 - Battery ECU
• P3191 3191 0000 - Engine ECU

I repeated the "OFF - ON - START" cycle four more times and each time, there was no change in the codes. You got one attempt to start the ICE with each "OFF - ON" cycle. The traction battery highest and lowest module voltages:

• 14.58 V - 14.76 V - before 6th attempt to start after adding 1 gallon of gas

I then added 1 gallon of gas and repeated "OFF - ON - START" and it fired right up. No change in the codes.

When I stopped and later refueled at the station:

• 562.7 miles - trip meter to fuel exhaustion
• 9.3 miles @61 MPG - trip meter to gas station
• 11.219 gallons - to fill tank

Conclusion

Five attempts to restart the NHW11 after running out of fuel did not throw a "poison" code inhibiting further start attempts. The only tricky thing is turn the car OFF and then you get one restart attempt. Traction battery will decay but not to dangerously low levels.

Bob Wilson

 

Agreed and now we know:

One simply has to turn the car "OFF" and then you get one 'restart' attempt per "OFF-to-ON" cycle.
​

The only opportunity missed was recording the traction battery SOC loss per start attempt. But with a full fuel tank, it will take several more weeks to go back and run that test. In the meanwhile, I have the wife's 2010 Prius.

My past fuel exhaustion tests did not investigate restart attempts. But we've had one report of a Prius being towed to the dealer after running out of fuel. I'm interested in finding out why and how it can be avoided.

Bob Wilson

 

The best I'll be able to do, easily:

• traction battery dV
• traction battery SOC

I have a Graham scanner but it takes ~100 ms. per sample. So traction battery current and voltages will be at best a sampling for each attempt. Still, several attempts will provide enough sampling data to hopefully show what it takes.

Bob Wilson

 

The Prius does continue "when it runs out of gas."

Bob Wilson