P610 transaxle

In both the Gen-1 and Gen-3, I've seen the MG1 torque change sign to start the engine but immediately go to either negative or zero torque. Even using "B" descending a hill, the engine runs against MG1 and MG2 torque reverses, in effect a mild "R", with the engine generating the reverse power. I've never found a mode beyond the initial cold start and much shorter, hot starts, where the engine is turned by MG1. Even the cold-start, is just a couple of seconds to get the oil pressure up.

The transaxle, engine driven pump has been constant through all four generations. Both the transaxle and engine oil pumps are the last two that beg being electrically driven.

Engine drive makes sense because running an oil pump in reverse would pull the lubrication oil the wrong way. The problem is a mechanically driven pump has to be sized for the lowest rpm. At higher shaft speeds a pressure relief valve opens, wasting mechanical energy. Speculation on my part, an electrically driven, engine oil pump might provide another 1-2% gain in engine thermodynamic efficiency, especially at higher power ranges.

Speculation, tweaking the engine water pump control laws could have reduced the mechanical overhead and added to the 40% thermodynamic efficiency. However, that awaits another day.

Bob Wilson

 

I've not found a mechanism for 'valves shut' in the engine. I see the variable valve angle adjuster for the intake valve but nothing that would keep the valve cam shafts from operating the valve. Would you have an engine diagram or sketch that shows where it is?

Now I know GM is fond of shutting down cylinders with some sort of valve disabling mechanism:


All I've found in our Prius engines is a hydraulic, back-lash adjuster.

Bob Wilson

 

I remember reading something about the Honda engines in the original Insight. But I lost interest when I saw the cone-belt CVT. @chuck is the Honda subject master.

Bob Wilson

 

Source: Test Car List Data Files | Cars and Light Trucks | US EPA

I trimmed the attached ".txt" file, (CSV format), to the header row and four Prius models. From the description, you may be interested in:

• AXLE-RATIO - From VI
• N-V-RATIO - Engine rpm divided by vehicle speed in miles -- from VI

I'll try to get the New Car Features description and add it too.

Bob Wilson

 

Corrected:

If I understand correctly (not sure about this, I admit), you are stating those figures apply to both torque and power - that is, 27.8% of the power would take the electrical path and the rest would go through the mechanical path. I'm afraid this is incorrect though - statically or dinamically. This is only true for the torque due to the PSD geometry, as you said. However, the power being the torque multiplied by the rotational speed of the element the torque is applied on and the part providing the power, the engine shaft rpm [rjw], there is only one special situation in which the above figures also hold true for the power, that is when all the PSD parts - sun/MG1, carrier/ICE, and ring - spin at the very same speed. In the general case though, the power split shares depend on the relative speeds of the components and are therefore vastly different from those figures most of the time. "laevus" above has written the related formula.​

Newton's Third Law:

When one body exerts a force on a second body, the second body simultaneously exerts a force equal in magnitude and opposite in direction on the first body.​

In this life, we obey the law.

Bob Wilson

 

I suspect they meant to post "MG1" and include a "]" on the initial quote tag.

Bob Wilson

 

It took me a couple of years before I realized the 3d law made some of this moot:

One of the tricks we pay in math (and clever engineering) is to translate the problem into another 'coordinate system.' There are problems found in signals that can only be solved by using "i" the square root of "-1". That is what happened when I realized the Power Split Device (PSD) is splitting the engine power. Suddenly everything made sense.

Now what folks are trying to figure out is "how many volts and amps" are in the MG1-to-MG2 and MG2-to-MG1 transfers. That way they can apply the 92% rule for the MG1 <=> MG2 transfer. Then your excellent example shows why it does not work:

Creep to a big wall, resting the front bumper on it, and leave the car in "D" - The engine will remain in idle; MG1 will generate power and send it to MG2, and; MG2 will generate torque that added with the power from MG1 will generate a motive force on the wheels trying to push the bumper through the wall. A sufficiently strong wall will leave the car motionless so no power will flow through the final gears and transmission to the wheels.​

In reality, the power that takes the electrical path will always be -28% of the engine shaft power. The power that takes the mechanical path can vary from 0 kW to a big number but that path is load dependent. Once you see the problem that way (and it is hard!) understanding the Prius transmission becomes trivial.

Many years ago, I bought an SAE paper by a Georgia Tech professor that claimed at high-speeds the two-mode transmission was more efficient than the Toyota system. As I read the paper, it slowly dawned that he was trying to model the electrical and mechanical paths over different loads and that was nuts. His conclusion was the Toyota transmission was too inefficient at high speeds yet he never did the dumb thing . . . put the car on a dyno (or in my case, the road.) The had transformed the problem into 'the briar patch' and the whole paper was almost a waste of my $45. What I did learn was his approach was flawed.

Bob Wilson

 

I have MiniOBD and an earth backed, wall. So I can actually do the first experiment.

The second is to put the front on jack stands, start the car, and let it run:

1. Start car and leave in "P"
2. When engine stops, shift into "D"
3. Record what happens

In the second test, there is no drive wheel load beyond the drive friction. So what happens to the power in creep mode on jack stands with no load?

UPDATE: It will need some way to couple the left and right wheels. If one turns faster than the other, it will trigger the traction control and bugger up the results. It may be easier to rent dynamometer time and a heck of a lot safer.

Bob Wilson

Let me know if this addresses your questions.

Bob Wilson