Figuring out why sometimes the engine won't stop

[weave]

I do a lot of long distance travel, hence in HV mode. It has baffled me that sometimes the engine will stop when I coast or slow down, and other times it keeps running and seems to never shut off unless I come to a complete stop.

Some observations after a lot of miles:

• Speeds above about 70 tend to drain the battery more as the car tries to use battery power to supplement the direct drive -- if it's on. Especially on a slight incline. I've watched my EV miles drop quite a bit when going over 70. Also direct drive engages a lot less on speeds over 70 from my observations.
• HV mode seems to want to keep the battery charged up to where it was engaged.
  
• If I keep the battery deficit from dropping below about 2 miles of EV range, the system will cycle the engine on and off while driving, like if I coast or have very light acceleration.
• Once the HV mode drops EV range by about 2 miles, it goes into a mode where it won't stop the engine unless you come to a stop. It wants to charge the battery back up.
• However, the recharge aggressiveness seems to slow down once it gets within about a mile of original level when HV mode started, which means likelihood of engine shutting off at highway speed diminishes greatly once you hit that ~2 mile deficit threshold.
  
• If you have a lot of downhill and recharging and the battery level goes above what it was when HV mode was engaged, it gets far more aggressive at using the battery, shutting the engine down and pushing that blue bar fairly high so you an even keep up some speed on a flat road with the engine off.

Due to this behavior, it explains why if you switch to EV mode and then back to HV mode, it starts to cycle the engine more often. You've basically reset the level when the HV mode started.

So really it goes into some short periods of HV charge mode when battery gets too low below that set point. Hence why I think HV charge mode being used on purpose and then use EV mode isn't THAT inefficient. You're just basically taking a bunch of little charge/EV times and making it into a longer charge then EV time. It all washes out in the end as far as gas mileage goes.

I'm going to keep trying to look for patterns the more I drive. One thing that bugs me is the 2 mile deficit measurement. I would think if it is using a value to determine when to change behavior, it'd be measured in kilometers, not miles.

Basically the car is controlled by a computer, so it doesn't have a mind of its own. It's operating on rules. So when it's behavior changes, then something trigged it. That's what I'm trying to figure out. So any other observations you have are appreciated.

Also one more observation. If I preheat the cabin from a L2 charger then drive almost right away in HV mode, it cycles the engine on and off almost right away instead of waiting the ~10 minutes to warm it up. Hence I think preheating the cabin is also warming up the ICE ahead of time, meaning less stress to engine starting when cold.

So far all theories but still testing to see how often and reliably I can reproduce them all.

 

[insightman]

Basically the car is controlled by a computer, so it doesn't have a mind of its own. It's operating on rules.

I like to compare the Clarity PHEV's behavior to a mainframe-based chess program--even the programmers are sometimes surprised by what the program is doing. The program in both cases is goal-oriented and there may be multiple ways to achieve those goals.

 

[Viking79]

I do a lot of long distance travel, hence in HV mode. It has baffled me that sometimes the engine will stop when I coast or slow down, and other times it keeps running and seems to never shut off unless I come to a complete stop.

Some observations after a lot of miles:

• Speeds above about 70 tend to drain the battery more as the car tries to use battery power to supplement the direct drive -- if it's on. Especially on a slight incline. I've watched my EV miles drop quite a bit when going over 70. Also direct drive engages a lot less on speeds over 70 from my observations.
• HV mode seems to want to keep the battery charged up to where it was engaged.
  
• If I keep the battery deficit from dropping below about 2 miles of EV range, the system will cycle the engine on and off while driving, like if I coast or have very light acceleration.
• Once the HV mode drops EV range by about 2 miles, it goes into a mode where it won't stop the engine unless you come to a stop. It wants to charge the battery back up.
• However, the recharge aggressiveness seems to slow down once it gets within about a mile of original level when HV mode started, which means likelihood of engine shutting off at highway speed diminishes greatly once you hit that ~2 mile deficit threshold.
  
• If you have a lot of downhill and recharging and the battery level goes above what it was when HV mode was engaged, it gets far more aggressive at using the battery, shutting the engine down and pushing that blue bar fairly high so you an even keep up some speed on a flat road with the engine off.

Due to this behavior, it explains why if you switch to EV mode and then back to HV mode, it starts to cycle the engine more often. You've basically reset the level when the HV mode started.

So really it goes into some short periods of HV charge mode when battery gets too low below that set point. Hence why I think HV charge mode being used on purpose and then use EV mode isn't THAT inefficient. You're just basically taking a bunch of little charge/EV times and making it into a longer charge then EV time. It all washes out in the end as far as gas mileage goes.

I'm going to keep trying to look for patterns the more I drive. One thing that bugs me is the 2 mile deficit measurement. I would think if it is using a value to determine when to change behavior, it'd be measured in kilometers, not miles.

Basically the car is controlled by a computer, so it doesn't have a mind of its own. It's operating on rules. So when it's behavior changes, then something trigged it. That's what I'm trying to figure out. So any other observations you have are appreciated.

Also one more observation. If I preheat the cabin from a L2 charger then drive almost right away in HV mode, it cycles the engine on and off almost right away instead of waiting the ~10 minutes to warm it up. Hence I think preheating the cabin is also warming up the ICE ahead of time, meaning less stress to engine starting when cold.

So far all theories but still testing to see how often and reliably I can reproduce them all.

Interesting results, and good to track. It sounds like hold mode in the BMW i3 REx where you know the engine will never be able to keep up so it runs hard continuously trying to do everything it can to maintain charge. This could be bad for gas mileage as it might go into a less fuel efficient more powerful mode. I will try to watch for this now too.

 

[leop]

I took a longer trip on an interstate highway in HV mode this past week. I had an ELM327 in the OBD2 port and the Torque app on my tablet. In particular, I looked at and logged the engine rpm, the coolant temperature, the vehicle speed, and the throttle position. Along with this logging, I kept the energy flow diagram on the center (or driver's display when I needed to view the navigation maps). I was interested in observing when the Clarity went in direct drive mode.

As Insight man says, the computer algorithms have different means of achieving the same goal. Much of the time that I thought the Clarity should be in direct drive mode, the gear icon in the energy flow diagram was visible and the engine rpm matched the road speed (I have calculated out the engine rpm v. speed when the ICE is in direct drive mode). But, the ICE and power train was not always in the direct drive mode when I thought it should be. This was a concern of mine since the recent software updates. However, the data showed a surprising but logical result. Most of the time when I thought that the ICE should be in direct drive mode but the gear icon did not show, the ICE rpm approximately matched the road speed and the throttle position was at the same value as when the ICE was in direct mode. If the gear icon display is correct (that is, the sensor and sensor reading for the clutch to be engaged is working correctly), the ICE was just supplying all the energy to the battery/motor and the total ICE power output to the battery/motor matched the combined power output to the wheels and battery when in direct drive mode. The difference is that the traction motor is providing all the vehicle propulsion.

I had not taken this data before the software updates so I cannot compare before and after. I can say that the engine rpm and load when I think he direct drive mode should be engaged is the same as when the direct drive mode is engaged (with the caveat that the clutch sensor is working and the gear icon matches the clutch sensor output). I do not know why sometimes the direct drive is engaged and why it is not at times, but the physical effect on the ICE output is about the same. All in all, this is a clever system.

LeoP

 

[David Towle]

However, the data showed a surprising but logical result. Most of the time when I thought that the ICE should be in direct drive mode but the gear icon did not show, the ICE rpm approximately matched the road speed and the throttle position was at the same value as when the ICE was in direct mode. If the gear icon display is correct (that is, the sensor and sensor reading for the clutch to be engaged is working correctly), the ICE was just supplying all the energy to the battery/motor and the total ICE power output to the battery/motor matched the combined power output to the wheels and battery when in direct drive mode. The difference is that the traction motor is providing all the vehicle propulsion.
LeoP

I am not doubting you've seen this, but it seems physically impossible for everything to be the same in both gear drive and EV. There has to be a heat energy loss when the engine hp gets converted to electricity, and another loss when the electric motor turns it back into wheel hp. The losses could be small, but they have to be something. And when you are in gear drive neither of these losses exist.

I thought previously in electric drive the engine might be turning at a more efficient speed, which could make up for these losses, but now that you've seen the same rpms I don't get it.

 

[leop]

As I said, the caveat is being sure that the clutch sensor signal is communicated all the time to the head unit that displays the gear icon. But, I think that the clutch is disengaged at those times as the ICE rpm has more fluctuation that when the gear icon is on. Also, remember that there are gear friction losses when direct drive is engaged just as there are gear friction losses and electrical losses for the traction motor. You are correct to look at the efficiency of the generator to traction motor energy path but I do not think it is that large given publicly known experience with modern such systems. However, this is a bit more complicated with the Clarity. The Clarity generator power output can be varied even when the generator is driven at a constant rpm. This is also true of the traction motor output. I should have said that I observed that the throttle position value was "about" the same in both modes. I do not have enough data to be sure but I think that the throttle position may have been small bit more open in the generator-traction motor mode than in the direct drive mode. What I need is access to the clutch engaged sendor output. This sensor output is available to whatever drives the center and dash displays so it should be available on the OBD2 output on some bus.

Also recall that the Clarity often is generating power to the battery when in direct drive mode and this power generation is subject to the same gear friction and electrical losses. I also would like to have access to the generator power flow to the battery to fully understand the Clarity operating modes.

LeoP

 

[insightman]

The Clarity has more data than we do--it knows when it is more efficient to travel in Engine Drive mode. How can I be so confident? This is Honda's 2nd-generation i-MMD plug-in hybrid system. As random as the clutch activation seems to be, the extensive engineering, space-allocation for the mechanism, and materials expense were all judged to be warranted by the increased fuel economy Engine Drive mode provides. The engineers certainly wouldn't then keep the clutch engaged at times when Engine Drive mode doesn't provide increased efficiency.

Wow--except for the Clarity's fabulous fuel economy, the entire basis for my argument is nothing but confidence in Honda's engineers and the engineers' ability to convince Honda's bean-counters. I'm sure there was data to back up their design decisions.

 

[weave]

The engine programming can't know what you are doing, like I'm going to drive all day or I'm 10 miles away from a charger. The HV mode clearly seems programmed to want to keep the battery at a level near where the car started or HV mode was engaged. I can clearly change the behavior of the car by either keeping it in HV or switching it to EV and then back to HV.

I believe it'd be more efficient in HV mode to slowly let the battery deplete and not worry about charging it back up from the engine in cases where you'll end the drive and charge before it depletes to zero. Or maybe use that same ~60% level that you can only HV charge to as a lower level for this slow drain from full. But that'd probably generate a lot of user confusion, so they chose to keep it about the same.

We already have people complaining to NHTSB about the angry bees, so a lot of people clearly don't understand how this car works unfortunately. Just saying I'm sure there were some design decisions trying to minimize some of the differences in end user experiences.

 

[David Towle]

We already have people complaining to NHTSB about the angry bees, so a lot of people clearly don't understand how this car works unfortunately. Just saying I'm sure there were some design decisions trying to minimize some of the differences in end user experiences.

I don't think anyone is complaining to NHTSB about the angry bees nor would that be a legitimate complaint. They are complaining about the big loss of power and inability to accelerate that sometimes accompanies the angry bees.