I have the following impression/insights after playing with the Triple Tach that I have mentioned previously on this forum. The tachs for MG1 and MG2 are inductive loops, so they measure frequency instead of RPM and have a latency constant. I have "normalized" all three needles of the tach to read from 0-100%, where 100% is 6200 RPM for the ICE and 15,000 for MG2. MG1 is a special case and the RPM is offset by -12000 to permit a reading from -12000 to 12000 to equate to 0-100% (I am still fine tuning this.)

Things I have learned or impressions gained so far:

1. The HSD does “shift gears” – sort of. There is a small, but distinctive “notch” in the torque curve when the ICE is backed down after hard acceleration. When I watch my MG1 inductive tach, it shows that frequencies of between plus and minus 180 HZ or so are not used in any constant state, but are only transited when MG1 is reversing the direction of rotation. The ECU is apparently programmed to avoid the ratio between MG2 and the ICE that results in the MG1 RPM being at or near zero. The transition is not linear, but faster in near the zero crossover point and slower near the 150 HZ areas in each direction. From 150Hz outward the movement appears to become almost linear. I have not found a write-up on this, but I suspect it is to avoid the almost DC like currents that would be required to maintain the torque demanded with low frequency AC in the motor coils. My AC theory is not what it used to be, but I believe that single direction currents with high amplitudes and long half-cycle periods (relatively speaking) would create heat and other problems.

2. Starting the ICE is not solely a function of one MG or the other. The ECU apparently sets up a torque differential by varying the spped, and thus the torque on MG1 when an ICE start is desired. This results in MG1 actually making most of the required speed change during the ICE start, but MG2 does attempt to make a speed change also (albeit a very small one). It’s not manifested as an actual speed change, but only as a torque change. Probably the slight "jerk" you feel when the ICE starts.

3.When the drive selector is placed in “N”, the ICE is absolutely spun in fuel-cut once the speed exceeds the MG1 protection threshold (around 42-45 miles per hour). No fuel is flowing and the ignition is not turned on. Energy to do this is derived from the MG2 “coasting” output. The battery may be used as a current buffer, but I am not sure of this. Below the MG1 protection threshold, the MG1 frequency is run ratiometrically synchronous to MG2. This probably helps to keep the planetary carrier stationary so that the ICE does not spin. I do not know how the frequency difference is handled (I don’t have any form of current instrumentation yet, just the ability to look at the waveform frequency, not amplitude).

 

Whoa...Interesting read.

 

FastMover — A few comments and thoughts:

• Even at low MG rpm, the switching frequency of the Pulse-Width Modulated (PWM) signal to the MGs remains high. It's always either 1.25, 2.5, 5, or 10 kHz (as I recall). I assume that your tachs have low-pass filters to remove this switching frequency so as to get at the actual underlying drive frequency?
• I believe that Toyota use special "starting" and "stopping" algorithms for the ICE, so as to reduce "jerks." Their US patents describe such strategies. I think they torque MG2 by an amount to compensate for any jerk that would be transmitted to the wheels when cranking the ICE using MG1.
• You speak of a "ratiometrically synchronous" relationship between MG1's and MG2's frequencies when coasting below ~42 miles per hour with the ICE stationary. Of course, when the ICE isn't turning the planetary-gear set forces such a synchronous relationship mechanically. Since in this mode MG1 is freewheeling and not electrically loaded, the frequency you see is probably just the open-circuit voltage generated by the spinning rotor, and has nothing to do with the behavior of the drive electronics.

Stan

 

Stan,

1. Yes, a low pass filter was indeed necessary to get a meaningful signal across two legs of the ECU output to each MG.

2. Yes, there are frequency changes (small ones that translate more like phase shifts than a frequency change) transmitted to MG2 -- not just for ICE start but for all observable (read "larger ones") phase changes to MG1 (read torque demand changes) and most probably for all of them.

3. I don't know yet, but I think that in coast, the MG1 waveform is something more than an open loop. I can't prove it yet because I haven't discovered a way to measure amplitude and relate it to current, but the MG1 drive waveform has small "shifts" in phase constantly to MG2 while coasting. This could simply be due to exertion of torques that are changing in the mechanical linkage, but based on what I see, I strongly suspect that the software is actively attempting to keep the ring gear and sun gear RPM such that the torque on the planetary carrier is zero. Still a lot of speculaton here and not many facts, but I could see this as a plausable way to stop ICE mechanical wear caused by the crank moving forward and backward within the current compression cycle, or even to prevent "flipping" between compression cycles by the ICE. We know the ECU is active, We know it has a power source (MG2) and we know it can control MG1 by phase changes, so it is a reasonable and feasible thing to do.

To find out is not going to be easy. It will require someone else driving (coasting) wihile I fiddle with the 'scope, which must be 12v (or inverter), portable, and with a unbilical to the instrumentation termination points that I have installed behind the panel. (With all the stuff, this will be a back set job.)

The deeper I delve into the HSD, the more impressed I am with Toyota's level of technical sophistication. There is a lot more to it than first appears, even to a technical eye.