alan_in_tempe and Droid13 — Let me try to clarify matters. The NiMH battery itself can only supply a relatively small amount of power to the MGs. But, as you say, the ICE can make up some of the difference between what MG2 can take (105 kW) and what the battery can supply. At medium speeds, MG1 generates electricity and sends (most of) it to MG2 to add torque at the wheels and allow the ICE to run more efficiently at lower rpm. Some of MG1's output may be sent to the battery if it needs charging. The battery may supplement what MG1 sends, so that the total power supplied to MG2 can be much larger than the capability of the battery alone. That's why MG2 needs a much larger power handling capacity than simply the battery's power capability. Whether it ever gets sent the maximum 105 kW I don't know, but if so, most of it will come from MG1 via the ICE. At even higher speeds electrical power flows from MG2 to MG1, driving the latter backwards (the so-called "heretical mode"), to give high wheel speed while keeping the ICE's rpm down for efficiency. All this is possible only because there are two MGs. The CVT is "synthetic" in the sense that it's the interplay between the ICE's and MG1's speeds that sets the effective drive ratio to the wheels, and hence the car's speed.

The Inverter has no means of dissipating power itself. If you are shuttling electrical power through the inverter, it has to go either to one of the MGs and/or the NiMH battery (or possibly to the air-conditioning compressor, which also runs off the high-voltage supply). At least, this is my understanding of things. Electrical power only flows if the voltage of the source differs from that of the load. By controlling the voltages, Toyota determine when and to where electrical power flows.

The PSD and Inverter share a single radiator, cooling fluid, and water pump. There are rubber pipes linking the two. The TCH does indeed have four radiators, but the fourth is the heater radiator for passenger comfort!

Stan

 

An update on fuel usage when FE reads 0 L/100 km (60 mpg):

I mentioned earlier that, when coasting at speeds above 65 km/h, the TCH's FE gauge sometimes reads PRECISELY 0 L/100 km (60 mpg in the US model), while my ScanGaugeII (SG) tells me that fuel IS still being consumed at a rate of ~1 L/h (and that the ICE is still spinning at ~1000 rpm, according to SG). The question is: Which is correct? Well, after some discussion with Ron DeLong of Linear-Logic (the makers of SG), and some perusing of the Prius Web sites, the answer appears to be as follows.

The vehicle can issue a FUEL-CUT command to the fuel injectors, to shut off all fuel consumption by the ICE at speeds above 64 km/h (~40 mph), even as it is forced to continue spinning by MG1 (presumably powered by electricity from MG2), so as to protect MG1 from over-revving (which could occur if the ICE was not kept rotating). I have found the attached document (ES-275to288.pdf) in the Repair Manual. See especially page 281 (top). This specifies clearly the conditions that must be met to induce fuel-cut mode. This is the high-speed equivalent of pure-EV mode at lower speeds. The TCH cannot run in pure-EV mode above 64 km/h (this agrees well with my own observations — I had deduced ~65 km/h). If one accelerates in pure-EV mode above this speed, the ICE starts up.

However, I had noticed that, when coasting (foot off the accelerator, and with the TCH fully warmed up) at speeds above 64 km/h, the FE gauge sometimes registered precisely on the tick mark at the boundary between the white ("fuel is being used") range and the blue pure-EV mode range. This is labelled 0 L/100 km (i.e., zero fuel usage) in the non-US version, and 60 mpg in the US version. This is what Toyota call the "fuel-cut" mode — the fuel injectors are given NO injection pulses. Because of the unequal compression and expansion strokes (remarked on earlier) in the Miller/Atkinson-cycle engine, thermodynamic work IS still being done by the ICE when spun without fuel in this way. This provides engine braking with no fuel-usage penalty, while protecting MG1. As the car's coasting speed drops below 64 km/h (~40 mph), the FE gauge moves above this bottom tick mark, showing that the ICE is again being supplied with fuel, even as the car continues to slow down. The ICE's speed stays at ~1000 rpm throughout.

SG is giving FAULTY fuel usage data during fuel-cut mode. Its ICE rpm numbers are correct, but its L/h (or galUS/h) and L/100 km (or mpg) readings are WRONG. This also means that SG's cumulative fuel-consumption and FE numbers will be slightly in error if fuel-cut occurred during a significant portion of the trip. Ron DeLong is aware of the problem (also with the Prius), and is investigating ways of working around it.

This is relevant for those planning to conduct further FE tests to compare normal driving with forced-EV-mode operation of the TCH.

Stan

 

Stan,... your observation of the Prius fuel-cut and SG operation are identical to my observations with the Ford Escape Hybrid.

During downhill coasting, ( can occur anytime, but especially in "L" or "B" gear selector position) the ICE is spun by the smaller MG, with all fuel cut, with electricity provided from the larger MG tied to the wheels. This provides electromagnetic drag, as well as engine compression drag to slow the vehicle.

The SG shows large fuel consumption during this event, while the built-in FE display in the Ford shows little or none. I think the SG looks at RPM and estimates fuel usage, but in a hybrid, the RPM's can go up in a fuel-cut mode.

-John

 

When in EV under 65kph with the ICE not rotating, does the ScanGauge correctly show 0 fuel usage in that mode?

 

Droid13 — Yes, it shows 0 L/h (actually, 0.1 L/h) fuel usage when the ICE isn't rotating. When the ICE is actually using NO fuel at all (either because of fuel-cut being in effect at speeds above 64 km/h; or because the ICE isn't rotating at lower speeds, and you're in EV mode), in metric units it displays varying, positive FE numbers in L/100 km, which increase as the car's speed drops, and reads 9999 L/100 km ("infinity") when the car comes to a standstill. [In US units, it displays decreasing FE numbers in mpg as the car's speed drops, and reads 0 mpg when the car comes to a standstill.] Ron DeLong is aware that one change that he needs to make is to "zero" the metric FE display when the ICE is consuming no fuel at all.

Stan

 

Yesterday I found myself back on the same stretch of road (between Wallenstein and Listowel, Ontario) that I had used way back (in Post #48) to test my theory regarding the FE penalty of forcing pure-EV mode operation. So, as discussed at the end of Post #48, I decided to take this stretch at ~60 km/h (~37 mph) this time, with the cruise control "on." The outside temperature was ~0 degrees Celsius (i.e., freezing), there was no wind, this time there were two people in the car, and the gas tank was "full." [Last time, there was only me, and the gas tank was half-full. So the car was significantly heavier this time.] Since this was part of a longer trip, this portion of highway was travelled only once, from Wallenstein to Listowel — the "downhill" direction. The 27.4-km trip took ~24 minutes to complete. The car kept switching in and out of pure-EV mode as the road's grades changed from slightly downhill to slightly uphill. The NiMH battery's SOC (State of Charge) stayed at ~3/4 the whole time, and the water temperature stayed constant at ~89 degrees Celsius.

The car used 1.3 L of gasoline (according to my ScanGauge), and this resulted in a FE of 4.8 L/100 km (~49 mpgUS) for this portion of my trip. I was interested to see how closely these numbers agreed with those that I obtained less heavily loaded on the same road in somewhat warmer weather last year.

One thing worth noting is that lowering my speed from 65 km/h to 60 km/h for the cruise-control driving modality, ALLOWED the vehicle to switch into pure-EV mode operation — something that I really didn't want to happen. So, this suggested modification to my previous procedure was NOT effective in equalizing the average speeds in the two modalities, while avoiding pure-EV-mode operation in the cruise-control modality.

Stan

 

Okay guys. Spring is here, even in Canada, and so it's time to begin testing the "forced EV-mode hypothesis" of this thread in earnest, before the hot weather sets in and the use of air conditioning is necessary!

Based on my experience, if one tries to run with cruise control at ~60 km/h (~37 mph) on relatively flat ground, a warmed-up TCH switches into pure-EV mode quite frequently. This means that driving at this speed doesn't lock out all EV-mode operation, whereas driving with cruise control at ~65 km/h (~40 mph) does basically eliminate EV-mode operation. [The downside is that the two driving modes are not undertaken at the same speed, but I don't know how we can eliminate this difference.]

We need to get more data comparing FE in "forced pure-EV mode" with FE in "normal mode" driving. My data provide a start, but yours are now needed. Over to you!

Stan

 

On November 13 last year I issued a serious challenge in this thread (post #36) for TCH owners to conduct properly controlled tests and report back with their data, either supporting or refuting my claims that "forcing" EV-mode operation will in general make your FE worse. Quite a bit of discussion ensued regarding how the tests should be conducted, but no-one reported any data (or did any tests?). This was disappointing, since I'm either right or wrong in my claims, and some posters had strong views one way or the other. Nobody, however, bothered to produce evidence to support either side of the argument. So, in the absence of data from anyone else, I set out on December 14 last year to obtain evidence from just such a test. This was reported in detail in my post #48 on December 16, 2006. On May 14 this year, I re-issued my challenge in post #97 to absolutely deafening silence! If no-one presents any data that contradicts mine by November 13, the anniversary of my challenge, I intend to claim victory and the correctness of my assertions. Please start doing your experiments now!

This said, I want to draw the following fly in the ointment to your attention. Have a look at the following Web site: http://privatenrg.com/
Towards the end of the page you will find State of Charge (SoC) measurements made on the Prius. These show that the 8-bar SoC display is highly nonlinear in the sense that each bar does not represent an approximately equal increment in the NiMH battery's usable charge. Indeed, their measurements show that bars 6 and 7 alone account for ~56% of the battery's usable charge range. I believe from my observations that the TCH's SoC display is similar. The implications of this on FE tests is that it makes it very difficult to ensure that you are recharging the NiMH battery back to its initial SoC at the end of each test! Without such a control, one cannot draw a reliable conclusion about trip FE from the fuel used. It is essential to return the battery to its initial SoC, or close to it! (The SoC reading error could be as large as 29% based on this Prius data!) How can we ensure this? Without being able to interrogate the battery ECU to read the SoC directly, the only feasible way is to charge the battery each time until the SoC display just illuminates bar #6 (or maybe bar #7) say. Then the SoC would be essentially the same at the end of each test run, provided only that one started the whole series of tests by charging the battery to this level in this same way too.

There is hope in the offing. Linear Logic is apparently about to release a firmware update for ScanGaugeII that would allow just such monitoring of internal hybrid parameters. The rub is that one needs to know the PIDs (Parameter IDs) for these items, and I'm not aware of such a list. I don't believe that Toyota makes one available. (Their Intelligent Tester obviously is programmed to interrogate the ECUs and display just such data, but the codes are hidden.) If we knew the PID for battery SoC, we could even use the Custom Commands option in the current ScanGauge to retrieve the hexadecimal data. Does anyone know of such PID codes for the TCH? They probably are very similar (or even identical) to those in the Prius.

To give you some idea of the parameters that we could potentially observe, I'm attaching some pages from the Repair Manual:
ME-36to38 show some Meter/Gauge System parameters that can be read and tested.
HV-42to49 show some Hybrid Vehicle Control System parameters that can be read and tested.
The latter list contains items: SoC, Battery SoC, and Delta SoC that relate directly to this question. It would be really nice to be able to read these parameters directly.

Stan

 

Well I'm not going to pretend to have hard test data, there's no practical way I can do that kind of thing because a) I don't have the NAV so my data access is reduced, and b) as someone pointed out elsewhere, to get a real set of data would take a lot of repeated driving and the only way I can measure my FE these days is by tank or instantaneous--my trip graph doesn't go high enough to extract meaningful info. So to get valid data I'd have to go tank by tank, and that means either filling up after minimal fuel use or eliminating all driving of errands and such to really track a consistent set of trips per tank.

But my non-scientific response is that my observed, consistent behavior so far this summer is that when I'm doing my commutes through the smaller towns and roads with fewer traffic lights and speed limits generally 30-40mph (one short stretch of interstate highway aside), I get a major FE improvement over driving the interstate at 65 miles per hour.

When I drove the highways almost the whole way each way, I was often getting 40 mpg, but regularly got less (during warm weather--much less in winter). This entire trip was using cruise control, with EV mode where I could for things like interchanges between highways. This is also about a 20 mile drive.

When a couple of months ago I went off-highway because traffic was very heavy, I noticed that even though my gas gauge showed me at about 2/3 to 3/4 of the way through my tank, each commute I made to/from work the tank FE kept climbing up, even though I was already over 40 mpg. So I decided to try doing that for an entire tank just to see what it would be like with no highway trips at all. My next tank I got mid-40s.

Right now my MFD shows 48+ mpg after... 350+ miles for the tank? That's starting out with an 85+ mile drive getting 56.7 mpg doing almost exclusively 30-50 miles per hour driving (50 miles per hour using cruise, anything less on pure foot control to maximize my EV mode driving while not running the battery down too far). Someone questioned driving conditions, but that was an elevation difference for that trip of a handful of feet, some hilly stretches, and obviously my commute is bidirectional so anything gained one direction there should be lost on the return trip.

I just know an experienced foot and insights about how to improve it seems to make a big difference at speeds when you can actually get prolonged EV mode distance with minimal infusions of gas. Highway I leave it up to cruise control because I can't begin to guess when coasting and when feeding gas will end up with a better result--I've tried second-guessing it but decided I just can't do it, and cruise is going to have finer-grain control over how much gas is needed in most circumstances.

That's just my 2c and my own personal observed experience. Far as I can tell the slower high-EV driving makes a big difference if you've got a good handle on it.

Sorry it's still not a scientific study.

 

It would be hard to get any scientific data without a closed circuit to run on. i ried to log the data for a bit, but the variability of the wind, state of charge, traffic, and the fact that 99% of my driving is between 55 and 65 MPH made it next to impossible to get any meaningful data.

To return the state of charge to the same point you would either have to force EV mode to bring it down or force a regen mode to bring it up. These forced modes could then skew the data you are trying to collect.

The best would be to try and set up a hybrid meet at the Kentucy Camry plant and see if we could use their test track for some data collection. Then have a couple cars equiped with secondary fuel tanks that are no more than a portable tank with a known volume of fuel in it. Both car travel at the same speeds with one forcing EV and the other not forcing EV. Then at the end of the test see which one used more fuel. The switch roles where the cars do the same circuit but the driver that was forcing EV is now driving normally and the driver that did not force EV forces it. Then average the two results to eliminat driver influence. This could even be done on a normal open road, but the key is to eliminate other external forces that could skew the results.