Gen 3 PSD Question

The sun gear is connected to the ring gear thru' pinion gear of the carrier. When ring gear rotates, the pinion gear also rotates so does the sun gear. Although the pinion gear rotates, the carrier remains still cos' no torque is acted on it. The sun gear is rotating but just idling (no torque). If the torque come from the sun gear, MG1 must be either applying torque (motor) or generating electricity (generator).

Give yourself sometime to think over this. It took me a while to figure this out.

Vincent

 

Now I see why I missed this passage. It isn't in that document at all. These are your words, not Toyota's.

Like these nomographs, your gear descriptions seem to address only the static case, where everything that is moving is at a constant speed, or very nearly so. In this case I agree that MG1's torque is essentially zero.

But dynamics are different than statics. Another document shows the Gen2 MG2 rotor weighing in at 22.5 pounds. MG1's weight (mass) is not displayed. While it is lighter, and lighter still in Gen3, it is most definitely not zero. Combined with its higher RPM, it should make a fairly respectable flywheel.

Flywheels do not spin up by themselves. The torque, energy, and angular momentum aren't created out of thin air, but must by applied externally.

When the car makes a fast standing start in EV mode (ICE unpowered), as MG2 is spun up to move the car, MG1 needs to spin up too -- much faster, due to the gear ratios. But its flywheel-like rotational inertia will resist, so a significant torque must be applied. This torque comes either from the sun gear on its shaft (idle case), or from the stator field (active drive case).

Any torque on the sun gear shaft requires a corresponding counter-torque on the planet-gear/ICE shaft. That will cause the ICE to turn too, unless it has lots of static friction (denied earlier in this thread) or is otherwise locked (also denied). In the steady state case, the ICE will have a lot more friction than MG1, and a bunch of compression drag, so within a few seconds it should slow to a stop, leaving MG1 spinning at full speed. But that can't happen instantly, so I ought to see the ICE RPM rising above 0 for a few moments, and the (scaled) MG1 RPM lagging behind MG2's RPM. But I haven't seen either happening. Is it too small or brief to see, or is the ICE really stationary?

The other option is for the required torque / energy / angular momentum to be applied by MG1's stator field, i.e. by active motor drive. If it is properly phase-locked to MG2 (scaled by gear ratios), it will provide only the torque to overcome its own inertia, leaving no net output torque. And it will hold the ICE stationary. This is what I believe is happening.

But as long as we are speaking different languages -- you use statics (steady state), I use dynamics (transient conditions) -- we are not going to be able to agree.

 

FWIW, I sometimes work on dynamic, controlled systems from time to time. Mainly phase locked loops (digital, mainly), but other systems where one is playing control.

To my mind, the equations of state of the Prius HSD and associated components sound very much like something that would be amenable to analysis, ending up with a second or third-order polynomial in the transfer function. This would include the masses of the various gears, shafts, motors, and so on, and their impulse response to electrical stimuli (motors) and gas (ICE). Given all the fixed gear ratios and such, keeping the ICE from turning when one wants the car to move forward on electricity only seems like child's play; further, starting and/or stopping the ICE without jerking the car (or only causing a slight, minimal shudder) seems fairly straightforward as well.

The point is, the engineers mucking with the algorithms for this thing have full control of the currents going into and out of the motors/generators. It doesn't, and very likely isn't, step voltages/or currents being applied; instead, I fully expect that cool-looking double exponential-style currents that have the masses of MG1/MG2 and the various gears already figured into the fun are what starts, and, for that matter, stops the ICE.

Think of one of those funky fighter jets, whose basic flying characteristics are to nose-dive or flip into a spin; they're inherently unstable, and it's only by the grace of control theory that they stay up in the air at all. The inherent instability allows them to change attitude and turn/dive/climb much faster than an inherently stable aircraft. (A commercial airliner, by comparison, is built for stability all over.) If engineers can do that with a fighter aircraft, I suspect that a good controls system engineer could have a heyday with the Prius HSD.

For what it's worth, back when I was an undergraduate the controls system class I was in had a homework assignment that allegedly was identical to the problem a summer hire was given. It went like this:

You've got this big sheet metal mill. As part of that mill, you've got this big pair of electrically powered rollers squishing a bar of metal, making it slightly thinner as it passes through. There's a sensor on the output side that tells you the thickness of what comes out; you've got control over the motors, and we were given the control parameters of that, including the mass of the roller/motor combination: Now, come up with an optimal control circuit that guarantees stable operation (no ripple in the output metal). Fun. Again, the HSD sounds easy by comparison.

KBeck

 

^^ I presume the problem statement also included a minimum distance (or bar travel time) from the roller squish point to the thickness sensor? Of course the control system designer wants this to be as short as possible.

 

Yep. And the inductance of the motors that controlled the rollers, lots of fun. I'd have to go back and look at the controls book and my homework answer to give the fine details; it's only been 32 years!

And having said that: Delay is handled by an e^-T term, or something like that, and can be made part of the whole control function matrix, right down to the point of handling varying metal bar speeds. So, in principle, or at least as a first-order issue, it's not that important to have the sensor close to the squish point. Second-order issues, I'll give you.

KBeck

 

Q: Are the actual 'gear-ratios' the same or different between the Gen-II and Gen-III PSD units?

 

Hi 70AARCUDA,

The PSD is basically the same for all generations of Prius, all have the following gear ratios:

Power Split Sun Gear: 30
Power Split Pinion Gear: 23
Power Split Ring Gear: 78

MG2 = (30/78) * ((18/5) * Engine - MG1)

However, starting from Gen-III, Toyota added the SRU (Speed Reduction Unit) so that MG2 can rotate at a higher speed:

Motor Speed Reduction Sun Gear: 22
Motor Speed Reduction Pinion Gear: 18
Motor Speed Reduction Ring Gear: 58

MG2 = (145/143) * ((18/5) * Engine - MG1)

For those who love to calculate, 2010 gear teeth counts | PriusChat

Gen III MG2 RPM vs MPH | PriusChat

Vincent

 

Thanks! Do the above ratio numbers also include the different final drive-ratio gear ratio(s) that are different in Prius/V and PiP and -C models?

For newbies (me) the problem is (1) knowing WHAT to look for, followed by (2) knowing WHERE to look for it. Again, thanks.

 

Nope, it is just for PSD only.

All the Prius generations use different final drive gear ratio:

Code:

Gen 1 - 3.905
 
Gen 2 - 4.113
 
Gen 3 - 3.267
 
PiP  - 3.267
 
v   - 3.703
 
c   - 3.190

Vincent

 

Thanks, again!

So, the "transmission" ratio-numbers are set/established/determined by the SRU & PSD numbers, which THEN drive the final-drive (axle) ratio by belt (Gen 1&2) or gear (Gen III).

The *reason* I'm interested in those MG1-ICE-MG2 ratios, is that I would like to have my ScangaugeII™ unit (displaying engine RPM & MPH values) calculate and display -- in realtime -- what the actual effective gear ratio (GR) is, using X-Gauge programming and this equation:

MPH = [ 60/( GR×AR) ]×[ RPM/rpm ]​

where:

MPH = vehicle speed, miles-per-hour​

RPM = engine (ICE) speed, revolutions-per-minute​

rpm = tyre speed, revolutions-per-mile​

60 = proportionality constant, minutes-per-hour​

GR = effective GEAR ratio​

AR = AXLE (final drive) ratio.​

...if possible. At least, that's the plan.

Has someone already done this? I searched on "Scanguage / SGII / SG" but didn't find anything specific.

 

That is incorrect. There is actually only 1 fixed ratio.

​

Unlike conventional car, the engine speed in HSD is not proportional to vehicle speed. You can have 0 RPM when the car speed is for e.g. 40 mph or 1000 RPM when the car is stopped. What you want to do can be achieved by monitoring MG2 RPM since MG2 is proportional to vehicle speed. jdcollins5 already did that calculation in the same thread that I posted earlier.

Vincent

 

Oh well...as you can tell, I (obviously) am totally NEW to the Prius and its' eCVT assembly (PSD, SRU, and ??).

But, I do have both a curious mind and a willingness to study & learn...so that's what I'm fumbling around trying to do.

Again, thanks for the insightful (not Honda-like) knowledge sharing...I appreciate it, a LOT!

 

With Torque app, I can see some MG1 torque in action in EV mode, especially during sudden speed change (i.e: braking or PHV acceleration).

 

The only thing that's spinning right now is my head. Oy...

 

You guys stayed at a Holiday Inn Express last night, didn't you?

 

...is that the one where they say they'll "...leave the (red) light on..." for you?

 

Hi Vincent
I've seen this equation for MG1 rotational speed in different sites :
MG1 rpm= 3.63*ICE rpm - 2.63*MG2 rpm
which is correct?
...............i'm confused..........
if it's possible please tell a reliable source to find the correct equation that could be referred!
Thanks a lot!

 

That is the equation for the Gen-II Prius which vincent1449p quotes: MG2 = (30/78) * ((18/5) * Engine - MG1) which equates to MG1 = (18/5) * ICE - (78/30) * MG2. (18/5 = 3.6 and 78/30 = 2.6)

Gen-III Prius is MG2 = (145/143) * ((18/5) * Engine - MG1) which equates to MG1 = (18/5) * ICE - (143/145) * MG2

 

------------------------------------
I appreciate you Maarten28.
I understand the equation, I need some article or book that have use'd this equation.
I want to use that as source in my project.
a book or article is more valid in comparison to a website.
Thanks a lot

 

If you can't find something in print, simply credit it something like this:

"Extracted from Post ####, PC-forum, XX-April-2015; original source: Toyota Service Manual/Bulletin."