If you look at the database you will see that the two cars in the title are tied dead even at 48 mpg right? Closer inspection reveals (to my surprize) that the MT HCH is at 48.4 and the Prius is at 47.5, almost a full mpg difference in favor of the MT HCH! They have been almost tied in the past being only a few tenths apart. Anybody want to guess why the HCH is ahead by almost a whole mpg?

I'll guess: it's because it has a manual transmission vs automatic.
Compare the MT HCH to the CVT HCH - two cars that are otherwise identical, and the effect of the MT becomes much more obvious.

The sample size of the MT is pretty small, so the more hypermilers in that group, the bigger the skew is, as compared to a group with a larger sample size it makes less of a difference because there's more "normal" drivers pulling the numbers down. Also MTs are harder (imho) to waste gas with, such as if you punch it in 5th gear, you don't get revs, the car just slowly accelerates, whereas in a CVT or auto you drop gears (or change ratio) and boom, there goes your fuel economy.

2001 Prius Photos       2005 Prius Photos
'97 Prius Simulation Prius inner workings

Prius MPG takes a pretty solid hit in cold weather. My recommendation is to address this question again in say 6 months . . . August or September.

Bob Wilson

Operation Iraqi Oil Freedom:

Automatic, stock, project car.

My other 1500 cc car:

Automatic, stock, backup car.
Free speech, dialog and knowledge thrives without the poison of SPAM.

OK, deal!

This thread will still be here in 6 months so someone please "bump" it at the end of summer.

Bob the HCH takes a hit in the winter too. I wish I could get Jason to make the old database graph available. Voltswagon showed a perfect example of how mileage goes down in the winter. Voltswagon had twin peaks in early and late summer went it is warm but no AC is required. Mid summer was a little lower due to AC use (I assume). Voltswagon is buzz70's MT HCH BTW.

Works for me. BTW, I've found it possible to parse the detailed records from the GH database. I did this with my Prius data to load it into Excel and do some correlation studies. One of my 'get around to it' tasks is to use a Perl script to do web data harvesting: (1) Ebay for 'fair market value' and (2) GH for other correlation studies.

The Ebay script will automate looking for 'new listings' as well as provide 'fair market value' data for the price of hybrids. From what I'm learning about our Prius, there appears to be some 'price-performance' sweet spots. Tracking may lead to a second hybrid electric for our family which would be a Prius.

The correlation studies I'm interested in are 'trip duration' and 'temperature'. I have noticed in my data extracted from GH database a step function between 45(F) and 50(F) for 20 minute duration commutes. I suspect the warm-up fuel penalty has two componets, initial temperature and trip duration. Simular warm-up functions should exist for each make and model of hybrid-electric and the parameters are embedded in the GH database.

One item I forgot to mention is a curious 'aging' effect the Prius fleet data shows in the first year. Within the first year, the Prius fleet data shows the initial high MPG seems to decrease by 5-10% and stays constant thereafter. My Prius cold weather experiments suggest this may be due to Prius transaxle oil fatigue from over-stress and an unfiltered breather tube.

I found the viscosity of my Prius transaxle fluid was significantly lower than the virgin Type T oil. Although a lower viscosity oil would reduce internal resistance and energy loss, if it also came with a lower 'film strength' (unmeasured), the higher friction would eat up any viscosity savings. Oil can be stressed by mechanical and/or heat and this is something I'm looking into.

The second problem, possibly unique to my Prius that came from Ft. Worth Texas, is silicon contamination. I'm still working on this problem and plan to install a filtered air breather tube. However, it is likly to be a problem other Prius owners are unaware of. Certainly those still under the spell of 'warrantee' are not likly to even ask much less investigate.

Now if we compare the size and architecture of an ordinary manual transmission versus the Prius CVT, several differences stand out:

- Prius transaxle has both an 18 kW and 40 kW motor generators (Prius I)
- Prius transaxle has two distinct gear assemblies and chain transfer

The higher powered Prius motor generators means there is a significantly larger heat load than in an ordinary manual transmission. This provides more opportunity for heat stressing the transaxle oil.

The larger Prius transaxle galleries for the motor generators means that thermal cycles can ingest more dirt laden air than the smaller galleries of an ordinary manual transmission. The Prius planetary and normal gear assemblies are linked via a chain which again increases the internal air gallery space that thermal cycles can fill with dirt laden air. Contaminated transaxle oil would significantly increase energy loss.

Although possible, I don't think the Prius helical gears are likly to mechanically stress the oil. Helical gears that make a 90 degree power transfer are more likly to cause mechanical stress than those that don't make the bend. But I would note the Prius differential is embedded in the transaxle. Ordinary differentials have a heavier grade of lubricant than typical transaxle oil.

This is probably more information than you were interested in but we are comparing two dissemular transmission technologies. A simple, manual transmission and a constant velocity transmission. My wife has forbidden us from ever having a manual transmission again.

IMHO, I suspect changing the Prius transaxle oil annually would make a significant improvement in Prius fleet performance. But it will take a Toyota technical change notice to make that happen for warrantee vehicles.

Bob Wilson

Operation Iraqi Oil Freedom:

Automatic, stock, project car.

My other 1500 cc car:

Automatic, stock, backup car.
Free speech, dialog and knowledge thrives without the poison of SPAM.

bump

Hi,

As follow-up to the bump:

I still need to 'get around to it' but that effort is taking a backseat to collecting oil samples. Although not completed, I've got a first cut at a web form to capture hybrid vehicle oil samples:

http://hiwaay.net/~bzwilson/hybrid_fleet/oil.html

I've got to complete the e-mail confirmation page and backend database. Then I'll add all of the known samples. At that point, we can start taking new reports. The goal is to provide simular oil analysis like this for all of our hybrids:

http://hiwaay.net/~bzwilson/prius/pri_toil.html

This can lead to smarter, more efficienct oil changing strategies based upon oil testing, not marketing hype.

Doug's Ebay tracking will have to do for now but I've not forgotten the need. This is especially important since "CNW Marketing" is describing the eventual softening of the 'white hot' demand as nobody likes hybrids any more.
This summer, I'm adding an engine block heater and transaxle pre-heater with thermostat control. This will allow controlled experiments when the cold weather returns to quantify the effects. The database records still exist but I prefer analysis of more controlled tests.
This should be visible in the existing mileage database after factoring out weather
effects. However, it remains a lower priority after getting the oil testing program working.
I was especially impressed by the difference between the Type T, NHW11 Prius oil, and the Type W, NHW20 Prius oil. The virgin Type W oil has simular viscosity to the worn-out Type T. I'm thinking about going with Type W in a controlled test in my NHW11 this fall after I complete my current vent and Amsoil ATF oil testing. Toyota had mentioned a change in transaxle oil along with adding a roller bearing for a bushing.
This experiment is at 6,000 miles, about half-way:

http://hiwaay.net/~bzwilson/prius/pri_vent.html

So far, I can not find evidence of oil overheating:

http://hiwaay.net/~bzwilson/prius/pri_tt.html

The hottest I've seen is ~70C, about 10C cooler than the ICE coolant temperature. There may be local hot-spots and some have reported MG temperatures above 100C.
As pointed out earlier, this is something that has my interest. Curiously, the used NHW20 sample had half of the particulate matter of my NHW11 and better viscosity wear patterns. It looks like NHW20, Prius II, transaxle oil is good for a 60,000 mi. change interval. The wear materials appeared to be higher but the contaminates were half of the NHW11.
I'm begining to suspect that the higher viscosity of the Type T oil leads to mechanical stress wear. A thinner oil would not suffer the shear stress and would be less likly to lose viscosity. This is consistent with the different results between the NHW11-TypeT and NHW20-TypeW results that we have today.
These results beg the question of what happens if Type W is used instead of Type T in an NHW11 transaxle. I expected to perform that experiment this winter.

Bob Wilson

Operation Iraqi Oil Freedom:

Automatic, stock, project car.

My other 1500 cc car:

Automatic, stock, backup car.
Free speech, dialog and knowledge thrives without the poison of SPAM.

Mid summer update on July 31st. Median/Mean

Gen 1 MT Civic 48.2/48.3
Gen 2 CV Civic 46.8/46.6
Gen 2 CV Prius 47.6/47.6

So far the Prius has picked up a tenth (which make sense) and the MT Civic as LOST 2 tenths (which does not make sense).

End of Summer results:

Gen1 MT Civic 48.2/48.3
Gen2 CV Civic 46.9/46.6
Gen2 CV Prius 47.9/47.9

Gen1 MT still RULZ (if only by a hair)!