[Joule Thief]

I've been doing some experimentation with MM Gaming the past 2wks to collect data on if I can truly improve my MPG and/or EV range (AER). Here are my results, I was able to improve both MPG and AER by an average of 25% over my baseline of using no MM. However, the standard deviation on the MPG is fairly high while the EV range improvement is rather predictable.

For each test, I attempted to minimize variables by maintaining consistent driving habits and use of climate controls. For climate I used Eco mode set to 75 and seat heater on 1 bar with occasional use of front/rear defogger as needed. For driving, my flat highway speeds were 65-70mph, and down to 50-55mph through the windy mountain pass. I drive in Normal L mode.

To understand my data, let me first explain my commute. I drive from south of Santa Cruz "over the hill" into San Jose/Santa Clara on SR 17 (one of the most dangerous highways in the state according to Wikipedia). Along this commute I traverse the Santa Cruz Mountains, climbing 1800ft to the summit before descending back down into the valley. The total distance is about 44.5 miles from home to work which breaks down as follows:
1) 0-1.5m Leaving my house is about 1.5 miles of cruising through the neighborhood at 25-30mph
2) 1.5-16.5m Then I'm on straight freeway (Hwy 1 and Hwy 17) at 65-70mph until I get to about 16.5 miles into the trip where the curves begin
3) 16.5-20.9m During the climb I'm at speeds of 50-55mph through the tight curves, at 20.9 miles I'm at the highest point of the summit and begin descending
4) 20.9-25.9m During the descent I get heavy regen braking, still travelling generally at 50-55mph, at 25.9m the sharp turns are over and I'm back to straight roads and higher speeds (traffic permitting)
5) 25.9-44.5 For the remainder of the trip I'm on flat freeway back to 65-70mph traffic permitting

Next I'll explain the 3 methods or strategy's I tested.
1) No_MM: First, my baseline or control is not using MM at all and drive in Normal L for the entire trip. I have 2 data points for this method.
2) MM_Incline: My second method was to turn on MM at or above 15 EV miles which tends to engage the ICE at about 16.5 miles into the trip, which just so happens to align perfectly with the start of the mountain climb! Then I disengage MM and return to Normal mode at the summit, 20.9 miles in. I regen all the way down and remain in Normal L for the duration of the trip. 3 data points for this method.
3) MM_Decline: My third method was to wait until I reach the summit OR was down to 4-5 miles of EV range (whichever came first) and then turn on MM and engage the ICE through the decent. Then turn off MM at about 26 miles into the trip at the base and back to Normal L for the remainder of the trip. 3 data points for this method.

UPDATE: Here's the elevation profile showing where the ICE turns on and off in MM_Incline and MM_Decline:

You can feel free to scrub my data for yourself, but here are my conclusions. First, both the 2nd and 3rd strategy showed significant improvements over the baseline of not using any MM. The 2nd method had the best overall performance with a 25% average improvement in both MPG and EV range, although the MPG stdev is rather high at 24.1. I'm planning to get a few more data points for methods 2 & 3 next week and see if the stdev tightens up. But for now I'll be sticking with method 2 as my default.

Summary data:
Method 1 Baseline No_MM (n=2)
135.8mpg +/- 11.7
31.0 AER +/- 1.5
70% electric

Method 2 MM_Incline (n=3)
169.7mpg +/- 24.1 (25% improvement)
38.6 AER +/- 1.5 (25% improvement)
87% electric

Method 3 MM_Decline (n=3)
150.9mpg +/- 46.8 (11% improvement)
36.3 AER +/- 2.4 (17% improvement)
82% electric

You can copy the data table below into Excel and perform a space delimited text to columns if you want to see the details or do your own analysis:

Strategy No_MM No_MM MM_Incline MM_Incline MM_Incline MM_Decline MM_Decline MM_Decline No_MM MM_Incline MM_Decline No_MM MM_Incline MM_Decline No_MM MM_Incline MM_Decline
Date 10-Jan 12-Jan 2-Jan 3-Jan 11-Jan 5-Jan 6-Jan 9-Jan
Start_Time 7:17 7:08 8:18 7:05 7:01 7:35 7:05 7:05
Stop_Time 8:44 8:15 9:05 8:55 8:05 8:33 7:56 8:22 Avg Avg Avg %Delta %Delta %Delta Stdev Stdev Stdev
Garage_Temp 51 51 54 51 54 54 50 46 51.0 53.0 50.0 0% 4% -2% 0.0 1.7 4.0
Low_Temp 41 41 49 41 37 42 38 38 41.0 42.3 39.3 0% 3% -4% 0.0 6.1 2.3
SOC 34 35 43 41 34 43 39 37 34.5 39.3 39.7 0% 14% 15% 0.7 4.7 3.1
MM_EV_Buffer_Miles N/A N/A 14 14 14 N/A N/A N/A 0.0
Peak_EV_Miles 2 3 8 10 9 5 4 4 2.5 9.0 4.3 0% 260% 73% 0.7 1.0 0.6
Base_EV_Miles 2 3 13 15 15 9 8 10 2.5 14.3 9.0 0% 473% 260% 0.7 1.2 1.0
MM_Start_(ICE_On) N/A N/A 16.8 16.7 16.2 20.9 20.4 20.3 16.6 20.5 0.3 0.3
MM_Stop_(ICE_Off) N/A N/A 20.9 20.9 20.9 26 26.2 25.6 20.9 25.9 0.0 0.3
CD->CS 29.9 32 42.1 44.4 43.3 44.3 40.1 41.1 31.0 43.3 41.8 0% 40% 35% 1.5 1.2 2.2
Trip_Distance 44.4 44.5 44.5 44.5 44.5 44.5 44.4 44.5 44.5 44.5 44.5 0% 0% 0% 0.1 0.0 0.1
Trip_Gallons_Used 0.3 0.3 0.2 0.2 0.2 0.2 0.3 0.3 0.3 0.2 0.3 0% -33% -11% 0.0 0.0 0.1
Leaf_Gallons_Used 0.36 0.33 0.26 0.23 0.26 0.23 0.36 0.36 0.3 0.25 0.32 0% -28% -8% 0.0 0.0 0.1
Trip_MPG 127.5 144 161.8 196.7 150.5 204.8 127 120.9 135.8 169.7 150.9 0% 25% 11% 11.7 24.1 46.8
Leaf_MPG 127 143 161 196 150 205 127 121 135.0 169.0 151.0 0% 25% 12% 11.3 24.0 46.9
EV_Miles 29.9 32 37.1 40.1 38.6 38.9 34.2 35.8 31.0 38.6 36.3 0% 25% 17% 1.5 1.5 2.4
ICE_Miles 14.5 12.5 7.4 4.4 5.9 5.6 10.2 8.7 13.5 5.9 8.2 0% -56% -40% 1.4 1.5 2.3
kWh 10.2 10.3 9.9 9.8 9.9 10.2 9.9 9.8 10.3 9.9 10.0 0% -4% -3% 0.1 0.1 0.2

 

[bonaire]

MM_incline seems to add .1 gallon to the fuel used. Which may be part of the reason for higher AER which helps mpg.

BTW - I'm jealous of your commute and home/work location...

Do you charge at work for a return trip? That would help the whole case of a Volt for that trip.
Nevermind - I see that your other thread shows your nice new awesome at-work stations. It's just too bad you can do all-electric for the whole drive. Full 16kWh available would let you do that. Did you ever consider a Leaf for that commute? I think it's about the right kind of range a Leaf owner would find a good match.

 

[Rooster]

Good start on the data--that is interesting. However, if you want it to be truly meaningful, you need more data points. At a minimum, I'd recommend you target a 80% confidence level in the mean/sigma results. In English, that means you need 42 data points in each of the 3 scenarios you tested. That should keep you busy for 2012. ;-)

 

[sparks]

Wow, JT, thanks for the superb data gathering and analysis!

I have also found that gaming MM can buy a significant increment in MPG, especially in cold weather. Getting heat from the ICE at the beginning of the drive is a major benefit ("heat from the ICE"-- hee hee). But even in summer, running the ICE in MM when at highway speeds tends to capture more of the gasoline energy burned by the ICE, I think because it loads the ICE more heavily (more HP going to drive the wheels AND charging the battery simultaneously) where it is more efficient.

 

[Joule Thief]

bonaire, my average gallons used in MM_Incline was 0.25 which was the lowest overall, did you mean MM_Decline added 0.1g? My baseline average was 0.30g, MM_Incline was 0.25g, and MM_Decline was 0.32. Since the overall trip distance is constant, theoritically it wouldn't matter if you looked at gallons used or MPG. However, the resolution on gallons used is in incriments of 0.03g on the leaf menu whereas the MPG "appears" to have much better resolution because it updates continuously even as the gallons used remains unchanged (hopefully that makes sense). So I tend to rely more on the MPG number than gallons used for fuel efficiency indication.

BTW, something I think is funny is that while the leaf menu shows fuel used in 0.03g incriments, the trip menu only shows 0.1g incriments, so the leaf indicator is better. BUT, the leaf mpg is in 1m incriments whereas the trip mpg is 0.1m incriments??? I don't get the logic here by GM, I think they should have stuck to having the better resolution data on the leaf menu and lower res on the trip or at least be consistent. Also, I reset trip A at the start of every trip, but the numbers don't always align with the leaf menu, it's like there's slight differences in the algorithms?

And no I never considered a Leaf, I'm not too fond of the way they look. I did consider a Prius, but since I can charge at work the Volt was a better choice, and all else being equal I think the Volt is a much nicer vehicle overall than the other 2.

 

[Joule Thief]

Good start on the data--that is interesting. However, if you want it to be truly meaningful, you need more data points. At a minimum, I'd recommend you target a 80% confidence level in the mean/sigma results. In English, that means you need 42 data points in each of the 3 scenarios you tested. That should keep you busy for 2012. ;-)

Haha, yeah I don't think so I will gather a few more data points but I don't think I'll need 42 to convince myself there's a definite improvement. I've actually tested methods 1 & 2 many many times, I just never recorded the data. The results have always been comparable accounting for temperature fluctuations primarily. So I'm very confident that method 2 absolutely improves my AER and MPG over baseline, I think the only thing more data points will change is exactly how much of an improvement is gained, but not whether or not there is an improvement.

I've only tried the 3rd method a few times based on some of tboult's posts here. That's the one I want to gather more data on because the first data point was very good but also an outlier compared to the subsequent 2 data points that were lower but more consistent with one another. With such a high stdev on method 3 it's entirely possible that it could be as good or better than method 2, although I doubt it. I expect future tests will be more like the last 2 data points. But hey we'll see! I'll report back.

 

[Drathen]

I'm curious how much difference you'd see between Normal L and Normal D. I've driven the 17 between SC and SJ many times and that is one crazy commute you have! How does the Volt handle through the twisties?

 

[Joule Thief]

I'm curious how much difference you'd see between Normal L and Normal D. I've driven the 17 between SC and SJ many times and that is one crazy commute you have! How does the Volt handle through the twisties?

Yeah, it can be very scary some days! I think driving in D would have a big negative impact, maybe I'll add that to the test plan. But you can see in my data I gain about 5 miles of EV range coming down from the summit to the base in L mode. I doubt I'd gain as much in D. And the Volt handles the twisties amazingly well! That's the other benefit of L mode, it makes the decline very easy to navigate with the self-slowing effect. Most of the time I'm crusing down with my foot completely off the accelerator or brake.

 

[PAredvolt]

I have a 60 mile round trip for my daily commute. On days when morning temps are <30, I put the car in MM for the drive to work thinking that I can use the rest of the battery charge more efficiently for the drive home when temps are typically in the upper 30s or higher. I think it helps. I wish there was a way to just run the ICE for the morning drive (when I typically use the most energy for heat and headlights), and them use the battery for the evening drive home.
Nice job on data collection. I don't have the patience to do that.

 

[tboult]

Thanks for the data.. interesting stuff. Your MM Incline being more efficient is interesting. You incline data is at speads well below what I've done (at 55mph I'd be run over on the highway). So it quite plausable your incline has lower RPMs and hence still keeps it in the sweet spot with the generator running.

When/how did you measure SOC?

I noticed your MM decline was both best and worst. but the two poor mm decline days were also colder.. With a fixed climate goal that could be an real impact. Also knowning when you needed to defog as that is both heater and AC and really sucks the juice.

what is MM_EV_BUFFER_MILES

Any noticable difference in the need to use the brake on the downhill?

Your commute seems to be a good testing ground for MM games. Mine is not (only airport runs). Would be interesting to log RPM and other data while you commute. I could lend you an OBD2 -- do you have an android phone? Or maybe a laptop you could be running on the commute? Then the logging could be very easy.

 

[Joule Thief]

Tboult thought this might catch your eye . I'll answer in order. Driving above 55 is not possible most days due to traffic. But even with no traffic driving much over 60 is a death wish due to the tight corners and narrow lanes. Some of my terms are probably not the best descriptors. SOC refers to my initial estimated EV range when starting the trip. It's what the car reads when I first start it up. The EV buffer miles is just how many miles are deducted from my EV range when switching to MM in the 2nd method, which is always 14 miles. No noticeable difference in downhill braking between any of the methods. I do have an OBD2 and laptop. I'm not sure my OBD2 has logging ability, I'll check. I have an iPhone.

 

[tboult]

Thanks.. Was not suggesting you drive faster, just why your reports would be different than mine. When I use MM to recharge uphill I'd have to be doing at least 65 andthe engine is then reving high (more than 3500rpm). And when I talked about MM slightly downhill, I'm still using charge to drive, no regen. So you are more extreme at both incline and decline.

SOC is state-of-charge for the battery (which it does not report, though can be estimated with bars). I use EER estimated Electric Range for what you called SOC.

Few OBD2 devices can directly log, but almost any can if connected to a laptop. What model ObD2? is it ELM327 based? I could write a custom logging program you could run on the laptop. what do you run: Win/Mac/Linux?
Don't think the iphone allows bluetooth so only wifi OBD2 devices could be used there.

These trips are short so gas used is not very accurate. I estimated actual gas using TRIP_LEN/TRIP_MPG Giving
Actual gas 0.348235294 0.309027778 0.275030902 0.226232842 0.295681063 0.217285156 0.349606299 0.368072787 0.327687776 0.26222746 0.294897283

I generally try to estimate MPG_CS, which for your trips I estimated assuming 12.5 of needed ICE (based on the best non MM trip).
That give a line of data:

MPG_CS 35.89527027 40.4494382 45.4494382 55.25280899 42.2752809 57.52808989 35.7545045 33.96067416 38.17235424 47.65917603 41.3539057

Where your MPG_cs for MM include averages 47mpg_cs which is around what I've gotten in various MMgames. Pretty darn good!


I was wondering about the braking since the regen needs to put charge into the battery and with MMrecharge going on I wondered if maybe you exceeded the rate at which it can regen and store. But it could also be that regen is less efficient if the battery is already getting MMrecharge, either because of the configuration of the engine/generator or because of the battery limit to accept hich C charging.

 

[Joule Thief]

Oh yeah I know you weren't suggesting I drive faster, that was more tongue in cheek . For SOC I think someone else here had it that way in their sig so that's probably where I got that. I know true SOC is a percentage which should always be the max allowed (80/85%?) at the start of every trip. I wasn't thinking much about it when making the table as I was more interested in the numbers.

I'm in AZ this weekend so I'll check my OBD2 model when I get back tomorrow. I'm running Win7. Yeah I think back calculating gas used is probably more accurate and your method for mpg_cs makes sense. When I do more testing next week I'll pay more attention to the braking in case there's a subtle affect I didn't notice before.

 

[MrEnergyCzar]

Why wouldn't it be better to only put it into MM when only on the flats or going downhill on the highway? I thought it was more efficient to use battery going uphill on the highway?

MrEnergyCzar

 

[saghost]

Why wouldn't it be better to only put it into MM when only on the flats or going downhill on the highway? I thought it was more efficient to use battery going uphill on the highway?

MrEnergyCzar

I don't know. I don't think we know. We're still figuring it all out with experiments and whatever data we can find online or pry out of the car. The bsfc map from ecomodder suggests that the best speed to run the engine is ~2200, or at least in the 1700-2700 rpm range. Switching into MM from dead at freeway speeds will push the engine harder than that except on downhills.

The problem is, there's other sets of losses to be considered and avoided where possible. Generator/inverter losses, battery charging losses. The Volt's strategy in steady state driving changes substantially in the 55-65 speed range, presumably based on which losses cost the most.

It's entirely possible that the OP's switching into mountain before hitting the hill and starting with a "full" mountain mode battery as he runs up the hill is letting the engine run in that range due to road load (not as hard as it would run if shifted later with that slope, since it's maintaining battery charge, not rebuilding it.)

 

[Joule Thief]

Ok tboult/all, my OBD2 scan tool is an Actron CP9580. It does have the ability to record frames of data which can then be downloaded to my laptop. The problem is it only records for about 150sec before stopping (presumably due to memory limitations). There are 61 PIDs I can record, the more variables I record, the fewer data points I get per variable. For example, if I log all 61 PIDs, I only get 25 samples of each. If I only log 5 PIDs, I get 275 samples of each.

So as it stands, I can only get about 2.5min of data per trip unless there is a way to continuously log that we can come up with (tboult?).

Oh, BTW something a little scary I noticed in the Actron manual - both the warranty and scanning suite software are provided by SPX!!! Yikes...

Here are the PIDs available for recording with descriptions:

Tool Description
1 MIL STATUS Off
2 MIL STATUS Off
3 MIL STATUS Off
4 MIL STATUS Off
5 MIL STATUS Off
6 ABSLT TPS(%) 34.9 Absolute Throttle Position
7 ENG SPEED(RPM) 1400 Engine Speed
8 BARO PRS("HG) 29.8 Barometric Pressure
9 CALC LOAD(%) 69.4 Calculated Engine Load
10 MAF(LB/M) 1.26 Mass Air Flow Sensor
11 MAP("HG) 22.7 Manifold Absolute Pressure
12 OUT TEMP(°F) 55 Outside Air Temperature
13 OUT TEMP(°F) 55 Outside Air Temperature
14 COOLANT(°F) 133 Engine Coolant
15 IAT(°F) 84 Intake Air Temperature
16 IGN ADV(°) 22.5 Ignition Advance
17 ST FTRM1(%) -1.6 Short Term Fuel Trim Bank 1
18 LT FTRM1(%) 0.8 Bank 1 Long Term Air to Fuel Ratio Correction Factor
19 VEH SPEED(MPH) 0 Vehicle Speed
20 ABS LOAD(%) 48.6 Absolute Load Value
21 FUEL PRES(PSI) 44 Fuel System Pressure
22 FUEL PRES(PSI) 43 Fuel System Pressure
23 FUEL SYS 1 CLSD Fuel System 1 Loop Status
24 FUEL SYS 2 CLSD Fuel System 2 Loop Status
25 REL TPS(%) 21.2 Relative or Learned Throttle Position
26 THROT CMD(%) 30.6 Throttle Actuator Commanded
27 ABS TPS B(%) 34.1 Absolute Throttle Position B
28 ACC POS D(%) 19.6 Accelerator Position D
29 ACC POS E(%) 9.8 Accelerator Position E
30 CAT TEMP11(°F) 1294 Catalytic Converter Temperature Bank 1, Sensor 1
31 EVAP PURGE(%) 0.0 Evaporative Emmision System Vapor Pure Solenoid
32 EVAP VP(H2O) 1.141 Evaporative Emmision System Vapor Pressure
33 CMD EQ RAT 0.999 Commanded Equivalence Ratio
34 O2S11(V) 0.155 O2 Sensor Output 11
35 ST FTRM11(%) 8.6 Short Term Fuel Trim Bank 11
36 O2S12(V) 0.750 O2 Sensor Output 12
37 CLR DIST(mi) 5909 Distance Since Codes Cleared
38 CLR DIST(mi) 5909 Distance Since Codes Cleared
39 CLR DIST(mi) 5840 Distance Since Codes Cleared
40 CLR DIST(mi) 5887 Distance Since Codes Cleared
41 CLR DIST(mi) 5907 Distance Since Codes Cleared
42 CLR TRPS 153 Number of Warm-ups Since DTCs Cleared
43 CLR TRPS 153 Number of Warm-ups Since DTCs Cleared
44 CLR TRPS 153 Number of Warm-ups Since DTCs Cleared
45 CLR TRPS 63 Number of Warm-ups Since DTCs Cleared
46 CLR TRPS 153 Number of Warm-ups Since DTCs Cleared
47 MIL DIST(mi) 0 Distance Since Malfunction Indicator Lamp Came On
48 MIL DIST(mi) 0 Distance Since Malfunction Indicator Lamp Came On
49 MIL DIST(mi) 0 Distance Since Malfunction Indicator Lamp Came On
50 MIL DIST(mi) 0 Distance Since Malfunction Indicator Lamp Came On
51 MIL DIST(mi) 148 Distance Since Malfunction Indicator Lamp Came On
52 FUEL LEVEL(%) 46.3 Fuel Level
53 VPWR(V) 13.696 Vehicle Power
54 VPWR(V) 13.806 Vehicle Power
55 VPWR(V) 13.781 Vehicle Power
56 VPWR(V) 13.791 Vehicle Power
57 VPWR(V) 13.582 Vehicle Power
58 VPWR(V) 13.757 Vehicle Power
59 VPWR(V) 13.440 Vehicle Power
60 ENG RUN(m:s) 01:02 Engine Run Time
61 OBD2 STAT JA/CA On Board Diagnostics 2 System Type

 

[tboult]

Ok tboult/all, my OBD2 scan tool is an Actron CP9580. It does have the ability to record frames of data which can then be downloaded to my laptop. The problem is it only records for about 150sec before stopping (presumably due to memory limitations). There are 61 PIDs I can record, the more variables I record, the fewer data points I get per variable. For example, if I log all 61 PIDs, I only get 25 samples of each. If I only log 5 PIDs, I get 275 samples of each.

So as it stands, I can only get about 2.5min of data per trip unless there is a way to continuously log that we can come up with (tboult?).

Oh, BTW something a little scary I noticed in the Actron manual - both the warranty and scanning suite software are provided by SPX!!! Yikes...

Well custom software for that model is probably not worth the effort. Could use a batch job that run their software on a timer (maybe using a UI testing tool), but even that is a lot of work. Maybe I'll just get an extra OBD2 with elm327 and do something for windows with that. At least that way I could have stuff for sharing with windows users too. ( My code is currently unix/android, but there is lots of basic windows stuff I can get going to grab the data, then log it for processing later.)

 

[Joule Thief]

More data!

Ok data junkies, this week I captured 3 new data runs, 1 for each method. I was off on Mon and had a miss last night (forgot to plug in, doh!) so that's why I don't have 5 new data points. But the big news is this time I used my Actron OBD2 scanner to data log the first 2.5 miles of ICE operation which represents about the first 50% of my MM runs.

To start, my results still show a slight edge to using MM on the incline vs. decline. The temps here dropped significantly on Tue and Wed when I gathered the new data, by about 8d F, so that had a big overall drop in both scenarios. Because of this change in temperature, you can't really average the new data in with the old data. But just comparing the individual runs you have:

01/17/12 MM_Incline
128.1 MPG
37.5 AER
37d F +/- 1 avg temp
56mph +/- 3 avg speed

01/18/12 MM_Decline
114.0 MPG
35.6 AER
40d F +/- 1 avg temp
56mph +/- 4 avg speed

Now here's a couple interesting observations. MM_Incline performed better despite having a few disadvantages to MM_Decline. First, the average temperature was 3d colder. Second, the fuel tank was almost twice as full at 41.5% in MM_Incline vs. 23.6% in MM_Decline. Also, in MM_Decline the ICE is on for about 1 extra mile (providing more time to recharge the battery), and yet after returning back to Normal/CD mode still results in less total EV range than MM_Incline.

As it stands, MM_Incline appears to be the best strategy. To better understand why, here is the OBD2 data for the first 2min30sec (or 2.5 miles) of ICE operation during both MM_Incline and MM_Decline modes. One of the first questions to answer is engine speed:

In MM_Incline, the average speed is about 2500 RPM +/- 1000
In MM_Decline, the average speed is about 2230 RPM +/- 850

It would seem the Volt is more efficient when it only needs to maintain the battery SOC rather than replenish it. Attached is the entire data set and a couple graphs I put together of some of the PIDs for comparison. Thoughts?

Update: Apparently the zip file is not working so I've replaced it with 3 txt files comma delimited (.cvs) for easy import into Excel:

View attachment Run Data.txt View attachment MM_Decline_OBD2.txt View attachment MM_Incline_OBD2.txt

 

[DCFusor]

I think the trick here is that when you run the engine, you want it loaded to the hilt to get best BSFC. It's a little more complex if some of that load is going into batteries due to round-trip losses, but these seem amazingly good to this old fart used to the lead acid ones in my home PV system. Maybe one of our good GM engineers here can comment on this, since surely they have these numbers.

I'll try to get some good data myself, but what I've been doing that seems to work based on some pretty sloppy observations is this sort of things. Gotta get me a DashDaq or something like it.

Drive down to no battery - when the ICE kicks in, switch to MM to use it to both charge the battery and push the car. When you get back to the 45% or so - switch back to electric. This works great on flats and constant speed. Sadly, the display doesn't tell you when you're there - you have to pay attention to the animation or the engine kicking off to know when to do the switch. I consider this a bug - the battery fullness ought to update in real time, it's not like it changes that fast.

I adapt this for much of my actual driving. If I'm going out on the interstate, I'll try to be in MM with less than 45% when I get there, and run the engine for those high speeds (around here, interstate means 70-80 mph or you'll get run over). Then back to pure electric for in town driving.

One thing to take very special note of here is that the console numbers are kind of "not quite right" for estimating all this.

One issue - if you return with more battery than you left with - you used some gas to charge and that will show up as reduced MPG you haven't "realized" yet - you could still drive that extra charge off getting more mile for zero gallons.

Another - if you got that charge in MM, but then switched back to electric - your gasoline MPG looks terrible as only some of that gas was burned to move the car - some was going into the batteries - so you will show more miles "pure electric" than is correct - some of that electricity came from gasoline use.

Seems very important to always be full electric for in-town driving at slow speeds with waits at lights. If the engine is running then in MM it's not very much loaded and doesn't get as good a BFSC. So I'll kick in MM on the way into town if there otherwise wouldn't be enough battery to handle the short errand loops in the town. Also, in town, especially at a light, it's kind of embarrassing to have the engine running where all the other drivers notice...having it racing at a stoplight is nasty.

I see above that Joule says otherwise. Guess we need more data! But do take the other considerations into account here, they really matter (and have good theoretical basis).

 

[Joule Thief]

DC, I agree. If you arrive at your destination with some battery left it would really mess up the numbers. In my tests, I always run out of battery and switch to CS for the last few miles so I know I'm comparing apples to apples. Another thing to note is you can absolutely artificially inflate your AER, but not your MPG. For example, if you drove on a perfectly flat road for 100 miles, when your EV range got down to 4-5 miles you could turn on MM, then turn it off after the engine replenished the battery and drive in CD mode again. You could keep repeating this until you arrived at your destination and would have a better EV/ICE ratio, but the same or worse MPG. You are correct that MPG will suffer dramatically as you run the ICE in MM, but when you switch back to CD the MPG number will start climbing back up, so no matter what you do the car will always show MPG as miles/gallons, you can't cheat it but at the same time you do get credit back anytime you're in CD.

I think when climbing and descending my data shows you can improve both AER and MPG. On flat roads I think you can only improve AER. So the only reason I could ever see using MM on flats would be to boost AER and/or hold EV for city driving as you mentioned, but not to gain any true efficiency improvements.

There is certainly an argument to be made that EV miles from MM replenishment shouldn't count towards AER (I'm impartial on it). But if you take that stance, then that would be another reason to use the MM_Incline rather than MM_Decline since the former is just holding your wall charge until you want it and the latter is using gas to recharge.

BTW, what did I say otherwise to?