What the Prius Could Do with a Mechanical CVT

  #11  
Old 11-17-2006, 07:21 PM
SoopahMan's Avatar
Pretty Darn Active Enthusiast
Thread Starter
Join Date: Sep 2006
Location: North Hollywood, CA
Posts: 374
Default Re: What the Prius Could Do with a Mechanical CVT

Wow, a ton of great info in this thread!

Towing I meant any situation where the front wheels would make contact with the pavement. With a mechanical CVT between the wheels and engine, if that CVT engaged during a tow despite the car being off, it could protect the rest of the drivetrain at speeds over 42mph, which pairs up with the 10000rpm limit of MG1.

Atkinson cycle optimal range Bob I just know you have a graph of this somewhere - can you link me? I'd love to see the optimal range; I thought the Prius engine followed the typical "curve" where there's a sharp hump at lower RPMs peaking somewhere between 2000 and 4000, then a gradual dropoff from 5000. I thought wrong.

Originally Posted by bwilson4web
With the exception of going down a hill, you can drop the transmission into "N" at any speed. Driving down the interstate at high speeds, dropping into "N" is perfectly safe.
Safe but not 100% fuel-efficient; the ICE ignites and idles to protect the electric motor. In fact, at 85mph, a common speed on highways here, the ICE revs up even higher (~2500rpm) to protect MG1. I'd like to Glide at that speed - coast with 0 fuel usage. I'm frequently caught in traffic moving faster than 42mph (often 55-85), but moving consistently at that speed. Gliding would make a huge impact on my fuel efficiency possibilities.

That's interesting that you can over-rev only by putting it in Neutral then exceeding 42mph - there's another thread here where some confusion was created around that.

Electric Motor Efficiency Based on what I've read, electric motors are most efficient at low speeds and high loads. This PDF presentation has several graphs of various electric motors and their performance under different speed and load conditions. This means using the electric motor to aid in accelleration at high speeds may cause the Prius to run this motor inefficiently. I'm not sure what sort of load/speed conditions would occur if the power were delivered via MG1, the sun gear - but any power delivered via MG2 at speeds above a few MPH are going to be far less efficient than if a gear ratio were in place to help.

Originally Posted by bwilson4web
I've gone up Brindley Mountain, ~160m rise, at 80 mph operating the ICE and MG1 at maximum power.
So, flooring it? Did it worry you that you were flooring it? I went up the mountain off of the 73 here on my way to San Diego and was too worried about flooring it to do so and found myself losing 30mph of speed attempting to climb it. I'm unsure of the grade but it seemed to be about 30 degrees - possibly a little steeper.

brick, very good point about dependability, I didn't mention that in my initial post. Another thread here has Lifetime rental car data showing Civic Hybrids as losing their transmission frequently - however this may be from abuse.

I'm definitely not suggesting there be a clutch on the Prius; the Reverse approach is brilliant on Toyota's part, and the clutch is basically a thing of the past.

I also wonder how they got the 25% efficiency they recently announced; I wonder if they further increased the RPM safe range on MG2, or something more. Probably something more.

Originally Posted by alan_in_tempe
In case you are interested, neutral is achieved by letting MG1 spin freely (open circuit). Similarly, by "shorting" MG1, it will be locked up
Hi Alan, what do you mean by shorting/locked up? I always wondered how this worked, and that sounds pretty fascinating - what exactly are they doing to the motor to lock it in place?

And I think I see where you guys are going with the "slippage" part - in addition to slippage being annoying when driving, it's inherently inefficient - any slipping is motive energy lost to friction/heat. I wonder what efficiency loss the Civic Hybrid typically experiences in its belt-CVT?

The HSD continues to be one brilliant, brilliant design. Planetary gears are cool, but the HSD takes it to ultracool. At least out here in Geekland ;o)
 

Last edited by SoopahMan; 11-17-2006 at 10:28 PM.
  #12  
Old 11-17-2006, 08:17 PM
bwilson4web's Avatar
Engineering first
Join Date: Oct 2005
Location: Huntsville, AL
Posts: 5,613
Talking Re: What the Prius Could Do with a Mechanical CVT

Originally Posted by SoopahMan
Wow, a ton of great info in this thread!

. . . Bob I just know you have a graph of this somewhere - can you link me? I'd love to see the optimal range; I thought the Prius engine followed the typical "curve" where there's a sharp hump at lower RPMs peaking somewhere between 2000 and 4000, then a gradual dropoff from 5000. I thought wrong. . . .
I think you are ready for the next level:

http://autos.groups.yahoo.com/group/...us-sat1/files/

Check out the folder, SAE.

This one is pretty good too:

http://autos.groups.yahoo.com/group/...l_Stuff/files/

On my local data:

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

You also need a copy of SAE 2004-01-0064 -- Google it up.

Also "Comparative Assessment of Hybrid Vehicle Power Split Transmissions"

Another good source: "ORNL/TM-2004/137"

I've got some excellent links under:

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

I highly recommend ORNL/TM-2004/247

I've pointed to a lot of technical stuff and it'll take a while to go through it. But you'll finally have exposure to the real data. You don't have to understand everything on first read but get a feel for the data and you'll eventually find it 'jells.'

Bob Wilson
 
  #13  
Old 11-18-2006, 10:08 AM
alan_in_tempe's Avatar
Veracitorian Muser
Join Date: Jun 2006
Location: Tempe, Arizona
Posts: 335
Default Re: What the Prius Could Do with a Mechanical CVT

Originally Posted by SoopahMan
Electric Motor Efficiency Based on what I've read, electric motors are most efficient at low speeds and high loads.
The linked graphs are not saying exactly what you think. Electric motors tend to increase efficiency with increasing speed. More importantly, electric motor efficiency is fairly constant, especially compared to an ICE, which varies widely over its operating range. For practical purposes, the MG1 and MG2 efficiencies are nearly constant. My guess is that they stay between 80% and 90% throughout most of their operating range (from just above zero to nearly peak RPM, across most loads). By comparison, the ICE goes from 0% at idle to a peak well under 40%. A primary advantage of the HSD is the ability to use high efficiency electric motors to control the operating load and speed on the ICE in order to keep it operating near its peak efficiency most of the time that it runs. No problem sacrificing a 5% loss in the electric motor for a 20% gain in the ICE, especially with that 5% loss is 5% of a 50 HP motor, and the 20% gain is 20% of a a 150 HP engine.

I also wonder how they got the 25% efficiency they recently announced; I wonder if they further increased the RPM safe range on MG2, or something more. Probably something more.
Not sure what 25% efficiency you are talking about.

Hi Alan, what do you mean by shorting/locked up? I always wondered how this worked, and that sounds pretty fascinating - what exactly are they doing to the motor to lock it in place?
If you have a small DC motor to play with, it is pretty easy to demonstrate. With nothing connected to the motor's electrical terminals (an "open" condition), give the motor a spin with your fingers. It should spin pretty easily, and gently slow down. Now connect a wire between the two terminals (a "short"), and spin it again. With the motor shorted, it will be difficult to turn, and will bring itself to a stop nearly instantly. With a high quality 3 phase induction motor such as MG1 and MG2, the effect is dramatically increased. When shorted, it is nearly impossible to turn the motor beyond a very slow turn. When open, the motor will spin very freely, and continue spinning like a top, save for the friction in the bearings. Also, the max RPM of these motors are well below 10,000 RPM, and the reason for the max RPM is to avoid throwing the permanent magnets, for the most part.

Some added tidbits: The PSD efficiency should be fairly close to that of a manual transmission, above 95%. However, there is probably quite a bit of efficiency variance in the electric power control, a rough guess may put it in the 80-99% efficiency across the operating range. I know a lot about the NiMH battery efficiencies when used in very high demand applications (all-electric model airplanes is where I learned most about the motors/batteries/speed controller technologies, in addition to EE degrees), but I am guessing that Toyota and Honda and Ford have sacrificed charge/discharge efficiency potential of these cells to greatly extend service life. On the other hand, treating these cells so gently may keep them at an even higher efficiency, perhaps over 90%.

-- Alan
 
  #14  
Old 11-18-2006, 02:03 PM
SoopahMan's Avatar
Pretty Darn Active Enthusiast
Thread Starter
Join Date: Sep 2006
Location: North Hollywood, CA
Posts: 374
Default Re: What the Prius Could Do with a Mechanical CVT

Thanks Bob, I'll have a look at those numbers. It's definitely going to take me a while to absorb though.

Alan, the numbers I'm finding run the range of 75%-95% efficiency depending on load for an electric motor - a much narrower range than the ICE certainly, but at the same time, an opportunity for some MPG savings.

However, the point about the friction of the band-CVT in the Civic has made me reconsider whether a CVT - or at least, that kind of CVT - would actually improve efficiency. If the electric motor's efficiency can only be improved by about 20%, if the friction losses are 5% you're playing a losing game in some driving situations. I'm not sure what the losses are though, I can't find data on that.

What originally got me excited about the HCH ironically was a new transmission Honda was supposed to use but ultimately didn't use, that had near-0 friction loss despite being a CVT. It used a new transmission fluid that becomes a solid when placed under pressure and heat, so the "clutch plate" was created and eliminated instantly between the bearings, causing 2 jagged metal surfaces to transfer the force from the engine to the drive shaft. I don't know why they went with the belt ultimately instead, but I was very disappointed to find out.

Point being - perhaps that kind of mechanical CVT would actually improve Prius performance though.

Originally Posted by alan_in_tempe
Electric motors tend to increase efficiency with increasing speed.
The graphs show a dramatic loss of HP when speed is increased even a little for those motors - but those aren't necessarily directly comparable to those in the Prius. I'm trying to find something that more directly compares efficiency and speed in a motor with a wider range, and Google just isn't pulling it up for me.

Originally Posted by alan_in_tempe
Not sure what 25% efficiency you are talking about.
The PriusChat homepage had about 2 weeks ago a news item linking to a video where a lead engineer for Toyota went over their progress with the Prius update. He showed a diagram of larger batteries in the trunk, plug-in capability, and one slide showed "25% increase in MPG" along with a few small diagrams. I can't figure out how to get to old news items on PriusChat's home page.

Originally Posted by alan_in_tempe
If you have a small DC motor to play with
I'm not an engineer, just a very curious programmer. However I have several mechanical engineer friends back in Boston I'll be seeing for Thanksgiving - perhaps I'll ask them for a motor to mess with.

Originally Posted by alan_in_tempe
Now connect a wire between the two terminals (a "short"), and spin it again. With the motor shorted, it will be difficult to turn, and will bring itself to a stop nearly instantly.
Very cool. So, in the open condition, spinning it is creating a current that isn't going anywhere. Now that you've connected the 2 terminals, trying to turn it causes the polarity of the magnet to grow, which in turn resists your attempt to turn it - is that correct? This sounds like it's bad for the magnets. It also sounds like there's some risk of the shorted motor occasionally turning anyway, which would cause some transmission slipping. Perhaps the Prius electric motor is too powerful for that to be realistic though.

Originally Posted by alan_in_tempe
The PSD efficiency should be fairly close to that of a manual transmission, above 95%.
That's interesting that there's still some loss even in metal gears tightly coupled like that. If you had to guess, what would be the efficiency of a belt system like the Civic's? 90%? 80? Lower?
 
  #15  
Old 11-18-2006, 02:20 PM
SoopahMan's Avatar
Pretty Darn Active Enthusiast
Thread Starter
Join Date: Sep 2006
Location: North Hollywood, CA
Posts: 374
Default Re: What the Prius Could Do with a Mechanical CVT

...Found the presentation where they mention 25% increase in efficiency!
http://www.evworld.com/view.cfm?sect...e&storyid=1127

The first half is about climate change and what their focuses are at Toyota. The second half is about the Prius. He vents a little bit about people who buy SUVs the second gas prices fall 10 cents a gallon - pretty entertaining. He also stresses some surprising concerns about Biodiesel that I had never heard before. He mentions their margin for putting the HSD into a given car model is about $2000/car. He says car designers at Toyota are allotted a "carbon budget," which determines the total amount of carbon they're allowed to generate in manufacturing the car, fuelling the car for 20 years, and destroying the car.

The major slides go from about 3/4 of the way through to the end. He mentions the upcoming 5-year refresh of the Prius achieving about 25% increase in efficiency. He says this refresh for the Prius is "2008," though he doesn't say if that's the 2008 model, or the one introduced in 2008 (the 2009 model). He briefly mentions the Prius' switch to LiIon and their work with plug-in capability, but suggests the plug-in part may have to hold off until consumers give a crap.
 

Last edited by SoopahMan; 11-18-2006 at 02:29 PM.
  #16  
Old 11-18-2006, 06:02 PM
alan_in_tempe's Avatar
Veracitorian Muser
Join Date: Jun 2006
Location: Tempe, Arizona
Posts: 335
Default Re: What the Prius Could Do with a Mechanical CVT

Originally Posted by SoopahMan
Alan, the numbers I'm finding run the range of 75%-95% efficiency depending on load for an electric motor - a much narrower range than the ICE certainly, but at the same time, an opportunity for some MPG savings.
It is not hard to keep a high quality motor at pretty high efficiencies. When the efficiency is low, so is the power.

Consider a very contrived example: If you are asking for 40 HP and you have 90% efficiency, then you need to supply 44 HP equivalent energy in electric power to the motor, and are losing 4 HP. When the motor is needed to supply 5 HP and is getting 75% efficiency, then 7.5 HP input is needed, and there is LESS loss at the LOWER efficiency then when the motor is operating at a higher efficiency (2.5 HP lost at 75% efficiency vs. 4 HP lost at 90%!).

My point is, the HSD can keep the motors around 90% efficiency when it matters, whether the speed is low or high, but keeping the ICE most efficient when it matters the very hard to do.

The bad part of this is that most of the electrical energy that is lost in motor inefficiency comes less from regeneration (deceleration) then it does from the ICE, which means the losses in the ICE is further reduced by the losses in the motors (generators). The HSD program does try to maximize the use of otherwise wasted ICE power (e.g., when idling) for charging the batteries, which is nearly "free" power as is decelerating regeneration.

The graphs show a dramatic loss of HP when speed is increased even a little for those motors
I think you are confusing power with efficiency. Ironically, for most of the operating range of an electric motor, HP increases pretty linearly with speed, from near zero, to typically 80% of peak RPM, and tapers off, but continually increasing above those speeds until nearly peak RPM (for a well designed motor). Note that the graphs you are looking at are only showing a very narrow speed of the motors. They are focusing on that narrow speed range because that study is for "constant" speed applications where the motor speed is held to a very narrow speed range. The effect you see is greatly magnified by the specific application of those motors.

The PriusChat homepage had about 2 weeks ago a news item linking to a video where a lead engineer for Toyota went over their progress with the Prius update. He showed a diagram of larger batteries in the trunk, plug-in capability, and one slide showed "25% increase in MPG" along with a few small diagrams.
Thanks. That 25% represents the end to end efficiency from the gas in to the distance traveled out. There are many ways to get good improvements in end-to-end efficiencies: larger batteries to enable more regeneration capture; smaller ICE; higher planetary gearing ratios; less total vehicle weight; etc. The problem is most of these methods also reduce the overall desirability of the vehicle to the consumer (less power; higher cost; poorer handling; etc.). I am confident that Toyota is shooting for a 25% improvement with no reduction in desirability. The far higher (double) capacity of the lithium cells at the same weight of the NiMH cells can almost do that by themselves (if their cost can be kept down, and safety maintained).

So, in the open condition, spinning it is creating a current that isn't going anywhere. Now that you've connected the 2 terminals, trying to turn it causes the polarity of the magnet to grow, which in turn resists your attempt to turn it - is that correct?
Not exactly. A motor has a voltage constant called Kv, which represents the voltage per RPM. A 2000 Kv motor will turn 2000 RPM if one volt is applied and there is no load on it. Give it 3 volts, and it wants to turn 6000 RPM. If you turn it 4000 RPM, then it will measure 2 volts on its terminals (at zero current; as when it is open). When the voltage is higher than the no-load RPM, then the motor will either be generating torque, or will turn faster. Also, when the voltage is lower than this RPM, the motor will generate current, or will slow down. The power controller in the HSD regulates the voltage on the motors to cause them to generate torque, or to generate current. When the motor is open (zero current), the voltage can be anything and the motor can spin freely. However, when the motor is shorted, the voltage is forced to zero, and the natural speed for any motor with zero volts on the terminals is zero RPM. When the voltage is held zero, but the motor is turned, then a current is generated. Since the resistance in the coils is very low, the current generated will be very high, and that will create a strong magnetic field driving the magnets in the opposite direction that they are being turned. Driving a shorted motor is much harder on the coils than on the magnets. The coils get hot when lots of current is going through them, and too much heat can damage the wire. Heat can also damage the magnets, but these are very high temperature magnets used in these cars.

That's interesting that there's still some loss even in metal gears tightly coupled like that. If you had to guess, what would be the efficiency of a belt system like the Civic's? 90%? 80? Lower?
I have very little idea what the belt CVT efficiency is. 85%-90% sound plausible to me, but that is only a guess.

-- Alan
 
  #17  
Old 11-20-2006, 08:10 PM
ken1784's Avatar
Pretty Darn Active Enthusiast
Join Date: Mar 2004
Location: Yokohama, JAPAN
Posts: 499
Default Re: What the Prius Could Do with a Mechanical CVT

Originally Posted by SoopahMan
One of the main differences between the Honda Hybrid and Toyota's is how the "CVT" part is implemented: Honda's is a true CVT, allowing both the electric and gas engine to have their gear ratio varied for optimal performance from both. Accellerating in the Honda, you can hear the gas engine rev up briefly, then sit at its optimal RPM continuously through 85mph.

Toyota's is an "eCVT," and I know I'll stir up trouble when I suggest this, but it basically allows the gas engine to run near-optimal RPMs, by making up for it with the electric engine. This basically means the electric engine falls outside of its optimal range easily. In practice it also means the gas engine isn't running at a constant optimal speed like it does in the Honda.
Toyota tested all types of transmissions during the Prius design stage in 1990's.
If you had more technical knowledge than hundreds of Toyota engineers, you will win.
So, what if the Toyota model was a real CVT - if it had a mechanical CVT transmission tied in?

1) You could tow it in Neutral without concern - the mechanical CVT could modify the gear ratio to protect both motors.
Not true.
You can't tow the mechanical CVT vehicle too. You have to raise the front wheel when towing.
2) The gas engine really could run at its optimal RPM at all times - in the eCVT design you can hear the gas engine rev low at 15mph and high at 70mph.
Not true.
The Prius engine runs at most efficient rpm and torque combination meeting the driver's power requirement.
You have to understand how it works referring to following article...
http://www.engin.umd.umich.edu/vi/w4...Miller_W04.pdf
3) You could Glide at any speed, even 85mph if you like, because the mechanical CVT could protect the electric motor from revving over 10000rpm despite wheel speed.
Not true.
The mechanical CVT also have mechanical loss at high speed.
Gliding at high speed is tough against the air resistance.
Forget about gliding at high speed.
4) I'm not sure of this, but you could probably get more optimal power out of the electric engine - because you could place a more optimal load on it at all times.
I can't understand this. What is the point against m-CVT vs eCVT?

Ken@Japan
 
  #18  
Old 11-24-2006, 09:17 PM
msantos's Avatar
Eco Accelerometrist
Join Date: Jun 2006
Location: Winnipeg, MB
Posts: 1,191
Default Re: What the Prius Could Do with a Mechanical CVT

Originally Posted by SoopahMan
That's interesting that there's still some loss even in metal gears tightly coupled like that. If you had to guess, what would be the efficiency of a belt system like the Civic's? 90%? 80? Lower?
The loss of efficiency on Civic depends greatly on the condition of the CVT fluid. Gen 1 Civics had a greater tendency to suffer from slipage particularly when the fluid began to break down. To add to that, Honda (through their own admission) was unreasonably optimistic about the fluid replacement frequency. As they later found out, the fluid really needed to be replaced at shorter intervals.

The current HCH-2's are less susceptible to similar losses and the nominal efficiency with fresh fluid makes the HCH2 CVT (according to Honda's technical literature) the most efficient CVT of its kind. In the latest redesign of the HCH CVT they've accounted for better cooling, a more robust ATF pump, a heavier duty start clutch, and a vastly improved logic & electronic clutch control. According to them, the latest CVT reduced the losses to a little less than 10% mostly because of the more acurate ratio selection and improved input to final shaft matching. (hummm... too optimistic if you ask me).

...anyway...

All in all, it is a good engineering effort that so far is delivering very good and consistent results on HCH-2.

Putting a mechanical CVT on a Prius could work but I doubt it would improve things without significant re-engineneering. It would perhaps amount to nothing more than an academic exercise without real benefits.

Cheers;

MSantos
 

Last edited by msantos; 11-24-2006 at 10:21 PM.
  #19  
Old 11-25-2006, 05:32 AM
bwilson4web's Avatar
Engineering first
Join Date: Oct 2005
Location: Huntsville, AL
Posts: 5,613
Default Re: What the Prius Could Do with a Mechanical CVT

Originally Posted by msantos
. . .
The current HCH-2's are less susceptible to similar losses and the nominal efficiency with fresh fluid makes the HCH2 CVT (according to Honda's technical literature) the most efficient CVT of its kind. In the latest redesign of the HCH CVT they've accounted for better cooling, a more robust ATF pump, a heavier duty start clutch, and a vastly improved logic & electronic clutch control. According to them, the latest CVT reduced the losses to a little less than 10% mostly because of the more acurate ratio selection and improved input to final shaft matching. (hummm... too optimistic if you ask me). . . .
What sort of cooling system is used? ICE coolant from the radiator or a separate loop?

The Prius has one coolant loop shared with the inverter. However, the Dept. of Energy report indicates most of the cooling happens from air passing around the housing.

Thanks,
Bob Wilson
 
  #20  
Old 11-25-2006, 07:26 AM
msantos's Avatar
Eco Accelerometrist
Join Date: Jun 2006
Location: Winnipeg, MB
Posts: 1,191
Default Re: What the Prius Could Do with a Mechanical CVT

Yes, from an active cooling standpoint we could say the the HCH-2 has separate fluid loop. Essentially, several relief valves along with the pump, use the ATF pressure to regulate the temperature inside the CVT assembly.
But in reality, the main dissipation of the excess heat is done via the CVT assembly. This time the engineers added a significantly larger surface area and allowed for a proper cooling stream of air to pass over and around it.

<<EDIT>>
Oh my. I guess I overlooked something important on my last description of the CVT cooling. On the HCH-II, the ATF fluid actually has its own radiator which is actually called quite simply "The ATF Cooler". This actually is the main pathway for heat dissipation and completes the active cooling loop.
<</EDIT>>

I believe ICE coolant from the radiator was always a less attractive option because it would still be too hot with average temps well above what is healthy for long CVT life.


I guess I have to order the tech documentation for my (wife's) Prius too. I just hope it is an easy process.

Cheers;

MSantos
 

Last edited by msantos; 11-25-2006 at 11:14 AM.

Thread Tools
Search this Thread
Quick Reply: What the Prius Could Do with a Mechanical CVT


Contact Us -

  • Your Privacy Choices
  • Manage Preferences
  • Archive - Advertising - Cookie Policy - Privacy Statement - Terms of Service -

    When you click on links to various merchants on this site and make a purchase, this can result in this site earning a commission. Affiliate programs and affiliations include, but are not limited to, the eBay Partner Network.

    © 2024 MH Sub I, LLC dba Internet Brands


    All times are GMT -7. The time now is 04:55 PM.