Good deal! I was hoping to work on this tomorrow, but we have some pretty good winds in the forecast. I'd prefer to eliminate that as a confounding variable, so I'll look at next weekend.

Meanwhile I went out to the test course today for a "test" run (25-43) to get comfortable with the protocol, especially the manipulation of the shift lever and cruise control in quick succession. I screwed up halfway through the first attempt and at 25 MPH pushed the CC lever down instead of up. So I had to abort the attempt. I can see it's going to take some concentration.

Today's run was a little longer than first planned; the total distance was 5.5 miles. I timed it for curiosity's sake: 9 minutes, 38 seconds, for an average speed of 34.2 MPH. Glad that puts me in the ballpark with your findings.

 

Preliminary data using 25-43 P_G and 34 mph CC:
Direction,Temp,Style,MFD MPG,Distance
S,73,CC,73.9,2.6
N,73,CC,84.8,2.6
S,73,P_G,77.8,2.1
N,73,P_G,94.1,2.2
S,73,P_G,84.9,2.1
N,77,P_G,93.0,2.6
S,77,CC,76.5,2.6
N,77,CC,90.2,2.6
S,77,CC,70.0,2.6
N,77,P_G,90.3,2.7

Averaging all five samples:
88.0 - P&G
79.1 - CC

Averaging the middle three samples:
87.0 - P&G
77.1 - CC

I've recorded the data with a Graham scanner and will do a detailed energy analysis later. However, I'm not very happy with the variations:

Of the five samples, I'll start with the best and the worst of P&G and CC. This will hopefully give insights to improving the protocol and understanding the results but that will take a little time. Still, a cursory look at the data suggests the north bound tests did better than the south bound tests.

Bob Wilson

 

Good stuff! I notice your distances vary in some of the P&G segments. Did you, by chance, allow your pulses and/or glides to complete instead of stopping them in midstream to comform to an exact distance?

Any theories as to the source(s) of the variation?

 

I let all P&G segments complete and in some cases, it required running further than expected. But this does bring up one problem with P&G, fitting it into a route.

I've got to look at the data and that will take some time. But there are some known aspects.

For cruise control, the ICE cycles on-off to charge the battery and then use the battery to sustain velocity. There is an energy loss charging the battery and later discharging it into the motor and if we're looking at 95% each way then we're looking at a 10% loss for the battery powered sections.

In contrast, P&G put the ICE energy into vehicle kinetic energy and drag slows it down. What surprised me was the near linear speed decay, which suggest a closer look at rolling drag.

Bob Wilson

 

Hi,

Here are the first five samples:

Don't worry about small errors. I make a few fubups but in the big picture, they don't matter. Just make a note and move on.

Here are the last five samples showing:
One unexpected result was how little time the ICE was off and the vehicle sustained speed on just battery power. Although the transaxle operated in "energy recirculate" mode, there was not enough battery energy to "stealth" mode on battery power. The accumulated ICE run time led to the lower CC MPG. This is a good strategy for battery life, avoiding charge/discharge, but not so good for low fuel efficiency.

There are more details to look at but this gives a big picture of what is going on at 34 mph CC and 25-43 PnG.

Bob Wilson

 

I made more practice runs last night after work, and hoped to begin the real thing if traffic allowed. Traffic wasn't too bad. I had to abort one or two runs for traffic, but if last night is any indication, evenings look pretty good. My fiancee will appreciate me not being out just driving back and forth through an industrial park all weekend.

It took a few tries to figure the best way to manipulate my various measurement methods. I chose to use the car's trip odometer, the ScanGauge's reset-able "current trip" MPG reading, and the stopwatch on my wristwatch. The toughest part was executing several tasks nearly simultaneously at the beginning of each run: suspending CC, starting the stopwatch, resetting the ScanGauge, and resetting the trip odometer. The best sequence seems to be: 1) reset the trip odometer (since it requires holding the button for about a second) with my right hand, timing the reset with the start of the course; 2) reset the ScanGauge with my right hand while simultaneously suspending CC with my left (while holding the wheel); and 3) starting the stopwatch. End of run measurements were simpler: quickly glancing at all readings and recording them on the tape recorder.

There was one mistake after another: I pushed the CC lever the wrong way, failed to bush a button here or there, forget to put it in N at 40 MPH, whatever. I finally managed some good runs, but by then it was 9:00 and 55 degrees. That's considerably cooler than what we expect later in the week, and though I know temperatures won't be identical for all tests, I'd like them to be as close as possible. So last night was regarded as strictly practice.

The ScanGauge also records maximum per-trip RPM and MPH readings. I will report those for additional documentation of test variables.

If only I had real-time data logging ....

 

More frustration. I made several runs the last two evenings. A few were good, many were suboptimal, and some were aborted outright. I want to adhere to the protocol as strictly as possible and make my measurements as precise as possible, but I missed on both counts several times.

I think I need to find a data logging device. Then I can give most of my attention to driving the car. Bob (or others), are you familiar with any of the inexpensive OBDII to PC scan tools on the market? I searched the web today for some, and found a couple below $100. One that caught my eye is ProScan. Anyone familiar with it?

 

All I can suggest is a calm state of mind and don't worry if a step needs repeating, say getting the ICE stopped.

I can't help much since I've got a Graham miniscanner and it is fantastic little box.

Bob Wilson

 

For me, "calm state of mind" is almost an oxymoron at times -- like last night. And I'm **** retentive.

I'll keep looking for a data logging tool. I'll probably hold off on more testing until I either find one or conclude there's not an inexpensive one that suits the need. It's too bad there's not a Graham miniscanner for the NHW20.

 

There is one candidate called "Prius CAN Monitor"...
http://priuschat.com/index.php?showtopic=25759&hl=

Ken@Japan