Toyota to Recall 1.9 Million Priuses to Update Software

Sorry about all the snippage.

Actually, I think I'm going to argue, probably incorrectly, that the Prius doesn't really resemble an induction motor. One second...

Did a quick search on Google, which led to a much older thread on Prius Chat. The motor/generators on a Prius are not induction motors: They're permanent magnet in the rotor with no coils in there (no coils on the rotor means its not an induction motor), with all the coils and such in the stator, to which our transistors are connected.

This kinda does mean that the Prius and my cooling fans are a bit more alike than not. I'll agree that it looks like a three-phase wiring job in there, but I'm not really sure, without digging, just how many windings, poles, and what-have-you are on my cooling fans. I just apply the voltage: The manufacturer says how fast they go .

I do remember generators like what you describe: Back in the day when I banged on naval aircraft radars for a living, the same shop I was in had the test fixture for airborne power generators. You know, the ones hooked up to attack/fighter jet engines. The fixture was a box 7 feet high, 10 feet long, and 8 feet across, powered by some serious electricity from the aircraft carrier's power systems, and noisy as all get out when it was running. The aircraft generators (400 Hz, with a pretty serious transmission) had a small permanent magnet generator; this would drive into a stator winding, which was then used as a source of current into a Much Bigger induction-style set of windings that would generate the Megawatt or so of three-phase power when the thing was running. I tended to stay on my side of the shop with its 20 kV radar power pulses, where it was safe, rather than near this monster, which, frankly, scared me.

They did have me checking for shorted windings one time when they needed an extra pair of hands and things were critical. Amazing how a milliohmmeter can measure 7 mOhm (good!) or 0.1 mOhm (bad!). Of somewhat more interest is how a jet aircraft handles a failure like that.. But I never asked.

KBeck

 

Hi KBeck,

No need, I prefer it.

Yeah, I explained that. It's NOT an induction motor! But there are some similarities.

Exactly as I said. Like an induction motor stator, with a PM rotor. That turns it into a PMSM. Permanent Magnet, Synchronous Motor.


If you're interested, there's a video on YouTube of a burned out Prius stator being replaced. The car still ran, amazing enough, but as they said, it shuddered a lot - as you'd expect!

Cheers,
Al

 

Toyota has said regen ends at 7 mph not 20 mph. Think of it as a black box as you are not seeing the full electronic. When anti-lock sensors don't detect a problem, regen will generate up to Y watts at X volts, where Y watts is the lower of a software limit or the request coming from the brake pedal. If it is above the software limit, abs has a problem with a wheel, or its bellow 7 mph, then mechanical brakes will be used as well as regen in the first case, or without regen in the second and third (anti-lock problems or bellow a speed limit). Electronic hardware is present to step up or down the voltage to charge the battery regardless of speed.

I believe the changed software reduces Y bassed on temperature of the battery, state of charge, and load on electronics. This reduction would put less of a strain on the electronics, reducing the chance that that they fail.

 

At 15 kmh braking on a step hill into underground garage I get steady -20 AMP battery power measured by Torque. I'm going to have firmware changed within an hour but maybe it's less related to boost voltage/power but different cooling pattern? Inverter has a variable pump speed maybe it needs to speed up faster when system is under high load, not wait as it will be too late?

 

Hi KiwiAl,

Here is another hint for you. The model of the motor winding is as the diagram.
Ue is the back EMF, during regen, Ue is the voltage in the winding and the inductor is the inductance of the winding. You now can figure out how the drive transistor can be use as the switch for boost converter.

 

Hi again bedrock8x,

Thanks for the tip...

Haven't had time today to think much about this, unfortunately. I have a Transaxle Oil Change to do this weekend, and some interesting test results to publish...

I'm not sure how the substitution diagram helps, as yet. I know it's a 3-phase motor. Although motors can be, and often are, connected either way, I'm pretty sure that the windings in a Generator cannot be connected in Delta, as the idle winding will then be shorting the two active windings. This would mean the motors have to be connected in Star (or 'Y' or Wye).

The (unmodified, unloaded) three phase Generator / Back EMF waveforms are as per the graph below. I have put the green box around the zero-crossing point of the black phase, and the grey lines represent the sum of the other two (the active) windings.



I know we want to boost the output voltage from say 100V to over 200V, and I'm sure this is done by alternately dropping the Star point voltage across the idle (black) winding, first one way then the other, but haven't figured it out yet, and haven't Googled any of it. Let me persevere.

 

Had mine updated today- took less than an hour and the dealership was very well aware of the campaign. In fact most cars in today were Priuses getting updates. So far I did not observe any difference in the car's behavior.
The paperwork indicated they were supposed to put a label somewhere , indicating update has been made. Any idea where that is supposed to be? I checked all the usual spots under the hood, on the bottom of the hood and door sills but am yet to find one. Just wondering

 

Glove box? It also maybe in an out of way place, like under dash or behind the glove box, with the 'computer' label. That's where the master key code is for Fords with the keypad doors.

 

Hi again bedrock8x

I conclude after a few minutes more on this that my initial approach (and graph) above, was incorrect. I now think that what they must actually be doing is PWM "shorting" to ground via the relevant IGBT, each winding in sequence during its +ve voltage cycle peak (e.g Black phase at around 90 degrees), to cause the winding currents (all three of them) to increase (as in the regular Boost converter), so they can then harvest the increased voltage in the same way.

I will now go away and do some more research to see if this is correct, but if it is, then I guess there are plenty of good opportunities to overcurrent the IGBTs - for example, leaving them on too long, or accidentally overlapping the ON times so that two of them end up shorting the rail capacitor... I can imagine that in the "Battery Getting Too Hot" routine, they could have tried to compensate for the loss of Regen Braking by trying to soak up energy in the windings instead as per the substitution diagram (but not a great idea?)... Who really knows?


Thanks for your tips,
Al

 

I have no idea what toyota does. Two proper techinques would be to shunt regen braking to a resistive load if the battery can't take the power, or to reduce regen and apply mechanical brakes, or a combination of both. It should not overcurrent IGBTs on purpose. It might do that with a bug. I am satisfied that toyota has identified software to prevent electrical problems. I just wish I could install it myself instead of going to a dealer.

I expect toyota will be tight lipped abou the real problem and solution.

 

KiwiAI,

I think I see it. Not as complicated as one might think.

First: I'm going to claim that when the motor(s) are acting as generators, all the transistor connected to the motor windings are simply turned off. When the voltage coming off a particular winding is more positive than the positive rail, the body diode in the IGBJT will be forward biased; when the voltage coming off a particular winding is more negative than the negative rail, then the body diode in the transistor connecting that winding to the negative rail becomes forward biased. No transistor switching is required.

Next: Now, that fancy graph you showed with EMF generated from the motor was winding to neutral. And it was RMS, to boot. If you look at the schematics you'll notice a distinct lack of a neutral connection anywhere. So, the implication is that what we should be looking at is the difference between two of the windings, not the difference between a winding and neutral! So, on our way to figuring out how much rpm we need to start charging the batteries, let's start off with 1.0. Then, the peak winding-to-winding voltage would be 1.0 x 2*sin(60) = 1.732. (This is how, in the U.S., when one has 120 V three-phase, one can get 208 V between any two of the phases; in some areas this is called 220, which is a different kettle of fish).
Next, the values in that graph were listed as RMS, root-mean-square. But if we want to see what the peak voltage is, that's RMS*sqrt(2); which gives us 1.732*1.414 = 2.45.

(For the lay people: Back in the day, Edison's power generators put out DC voltage, so, say, 110V DC at 1A would give a power consumption of 110V * 1A = 110 W. However, with a bread toaster, to get the same heating effect with an AC voltage which regularly passes through zero volts, the peak voltage needs to be 110*sqrt(2) = 155.6 V peak. Then, if one hooks up the toaster to this 155.6*cos(2*pi*60*t) waveform, one will get the same heating effect and one's toast will be done on time. AC voltages are still specified that way, so if one has 110Vrms * 1 Arms = 110W. A 120V circuit in a house (common in the U.S.) has a peak voltage of 120*sqrt(2) = 169.7V, measured to the return in the wall socket. The peak-to-peak voltage is twice that.. But that's enough math for one day.)

In any case: Going back to the graph, that means that if one wants 200V out of the diodes in the transistors, then the rms voltage of one field winding to neutral has to be 200V/2.45 = 81.6V. And that's around 1700 rpm. Which is a lower RPM value than I think you've been using.

Just so we're clear: Only two windings at a time will be forward biased, one negative, one positive, while the other winding voltage is "inbetween"; as the motor rotates around and voltages drop, probably both diodes will turn off (total voltage less than the rails); then another pair of diodes will forward bias. My brain is thinking that the current pulses will be at three times the generator rotation rate, but my brain may be out in left field somewhere.

If the motor/generators are turning fast enough so too much current is being dumped into the battery (limits on charge current) then we start turning the upper "boost" transistor on and off; in fact, my guess is that it would be some kind of control loop where the maximum current into the battery is some kind of set point, below which there is no control, with the additional caveat that if the battery voltage gets too high we simply turn off the upper transistor and leave it that way.

Where did I go wrong with this one?

KBeck

 

I had the upgrade and the transmission oil changed. After 30k service miles that is some knarly looking oil. Monday the sample is off for analysis.

Bob Wilson

 

Hi there austingreen,

FWIW, I agree with you on every point. I'm sure they would not deliberately overcurrent the IGBTs, but they may have done so inadvertently. And I really doubt they would have attempted to dump power into the electric motor windings. That WOULD be stupid!

Yes, agreed 100%. That's why I think there's spin when they refer to "the Boost Converter", because really, there are two of them.


BTW, there is another way to dump excess energy, which I'm sure they also use - spin the ICE. If they wanted to waste more energy, they could change the valve timing, but that's counter to their true objectives.

 

Hi again KB,

Indeed! bedrock8x's tip was actually quite relevant, when I got past my incorrect notion.

However, please do forgive me for wanting to straighten out some of your evaluation.

As I see it, you're not quite right on a few points. Like me, you tend to over-complicate things...

Yes, most of the time. I.e. when the MG speed is high enough to generate enough voltage to charge the battery at the (ECU) determined rate. Remember that the battery voltage is nominally 200V. However, to charge it, you have to push current through it backwards, and in order to do that, you have to raise the terminal voltage somewhat. Don't know by how much, but maybe to 220 or 230VDC. The more current you want to feed in, the higher the voltage, of course (but not exceeding the voltage at which the battery will overheat, or start gassing...)

However, as I mentioned earlier, the MG output voltage (aka Back EMF) is directly proportional to RPM. I didn't want to get into RMS and phase angle and so on because most people here don't understand that, and you don't need to, to get the basic concepts. The point is that for the sake of understanding, (let's say) the Back EMF is 600V at maximum RPM. At 1/3 of maximum RPM, the output voltage will therefore be (say) 200V. Don't worry about Phase to Neutral voltages or Phase Angles, RMS voltages and all that. The basic concept is all that really matters, since we (ok, I) don't know the exact gearing is, and what the motor speeds to vehicle speed ratios, etc, are anyway.

Get the basic concept first, then worry about the details. If (say) maximum MG RPM occurs at 60MPH, then at 20MPH, the MG(s) will be running at 1/3 of full speed, and producing only 200V output. As above, this is not enough to charge the battery, as over 200V is required.

The question I was puzzling over is: How then do you regenerate enough voltage to charge the battery BELOW 20MPH - which is quite clearly happening in my G2, and must, in order to achieve that economy also be happening in the G3? The identified, modular Boost Converter doesn't work backwards. It's only a Buck converter in reverse, so how?

I could be wrong, but I don't think that's part of the equation. IGBTs are 4-layer devices. In one analysis I have seen, they have a parasitic SCR, which, as they said, you definitely don't want to turn On. Once you turn that On, it behaves exactly like a normal SCR - the only way to turn it Off is by voltage / current reversal. So, if you trigger the parasitic SCR at the peak of the waveform, it's going to remain on for the next 90 degrees of rotation - probably more than enough time to "deform" the IGBT! It did occur to me that triggering the parasitic SCR might be the problem they are having, but I doubt that.

As I understand it, the diodes in the Inverter circuit are normal, separate, deliberately created devices, which have one major purpose in the normal 3-phase motor control circuit. That is to give the current that has just been established in the motor winding, a safe path through which to flow once the relevant IGBT turns off. By creating an easy current path for the inductor current to continue flowing in the same direction as it was before, no potentially damaging inductive spike voltage is created. They are called "free-wheeling" diodes for that reason - they allow the current that is already flowing to carry on flowing until it runs of of energy - okay - until the resulting magnetic field decays, if you wish to be precise.

Anyway, in the Motor/Generator scenario, the Freewheeling diodes have two functions, (1) as above, and (2) to create the 3-phase, full wave bridge rectifier needed to convert the AC voltages coming out of the generator into DC to charge the battery. I surmise that this is the reason that Toyota decided to make their own power semiconductors, because I think the diodes in the standard IGBT motor drive packages ain't big enough to use for Regen Power Rectification as required in the Prius /HSD system.

Yes, I mentioned that in the original post, but for conceptual purposes, it doesn't really matter. What actually matters is the peak voltage, not the RMS value. If you don't have enough peak voltage, there won't be any battery charging.

The diagram I posted makes no mention of RMS, peak or any other voltage. All it has on the Y axis is +1, 0 and -1, for conceptual purposes.

Yeah... In fact, I have yet to see any diagram of the Prius Inverter / MG circuitry which shows any connections inside the motor, though as previously, I believe it has to be Star / "Y" / Wye connected in order to work as a generator. (Second thoughts on that say even Delta wired motors produce Back EMF, so maybe that is not valid, but never mind.) The reason I mentioned that it is Phase to Neutral is because that is what my source, the ORNL Report, from whence the BackEMF graph came from, said.

The fact that they state the readings were the average Phase to Neutral voltage doesn't prove that they measured it that way, but again, it's only for conceptual purposes anyway.

[snip]

It depends on the instant you are looking at... Don't worry about this, but if you look at the 3-phase AC waveforms, you'll see that at different times, one winding is positive, 2 negative, then 1 +ve, 1 -ve, 1 0, then 2 +ve, 1 -ve.... They all change around as the machine rotates.

Pairs applies to the Single phase Bridge. In a 3-phase bridge, you need to think in triplets. They transition from one to the next - one starts to decrease while the next starts to increase, and so on. It's like the difference between a 2-cylinder engine and a 6-cylinder engine. In a 6, the power strokes overlap, so you get much smoother power, (and a completely different, continuous exhaust sound, FWIW!)

Six times! If you look at the "fancy" graph, you'll see it relates to one revolution. There are three pulses in the top half and 3 in the bottom. The Full wave Rectification effectively turns the bottom three up the other way. You'll see they fit between the existing three, to give six in total. Much smoother DC than single phase, full wave.

Just think PWM. That's what I had to do to get my head around the idea of shorting the generator winding to ground. It's all happening extremely fast - well, maybe 10kHz, which is rather slow in electronics terms these days. So, the Buck converter is running, in PWM, and drops the voltage to the battery when the Generators are in the upper end of their range. (There is never "No Control" in the Prius system - unless it's off or failed.) At high regen voltages, the Buck converter is continually (down)regulating the battery charge current (by PWM) to the desired level.

The tricky bit is when the MGenerators are running too slowly to produce the required more than 200V output. The "more than 200V" is created by a very cunning technique. (I believe, and it makes perfect sense to me now, and the clues thrown by bedrock8x also fit): They actually convert the entire Motor/Generator and Inverter into a rotating Boost Converter. If you forget about the irrelevant green box and extra lines on the 3-phase waveform diagram, and look at the Black phase: It gets to peak +ve voltage at 90 degrees. Let's say the peak voltage is 100V. If you now switch ON the IGBT connecting the Black phase to Ground, and momentarily try to short out the entire voltage on the Black phase (to ground), what will happen? Well, what normally happens when you try to short the output of a generator is that you get massive currents flowing, and something blows!

But in this case, you only want to "short" it out for say, 1/20,oooth of a second. In that time, being that the current is flowing in a "collective" inductor (collective because all three windings will conduct some part of the current) , the current will start to increase at a rate determined by the output voltage and the effective inductance of the generator windings. It won't increase instantaneously to maximum, but will build slowly, storing extra energy in the motor windings in the form of a magnetic field.

You are better at these calculations than I am...

Then , you switch Off the IGBT... What happens? The stored energy causes the voltages in the inductor to increase until it finds somewhere to go. In this case, through the bridge rectifier, into the DC Rail capacitor, which it charges to a higher voltage than 200V... And which the existing Buck converter can then control, into the battery.

Thus, a Motor/Generator/Inverter becomes a Generator / Boost Converter, and you are then able to keep regenning and charging the battery right down to about 7MPH. (Looks like about 7km/h in the slow revving G2 MGs.)

That's my take on it anyway...

 

Hi Bob,

I'm guessing you mean your G3? Since this is a G3 group/thread.

Changed my G2 Transaxle fluid yesterday and read your info about the Transmission Breather - Yeah, I know OT. (Actually, we're both OT...) Will post separately, have some interesting Oil Analysis photos/story, but I'm curious if you kept / developed the Breather system? I'd be very nervous about condensation - though the low pressure inside the housing is almost an invitation to water ingress as well.

 

Maybe this has been forgotten, but the animation at the eahart website. You'll need to search as I can't link yet, still not enough posts! Anyway, it shows that at 7MPH, which I believe is the agreed cutoff for regen, the ICE is 0rpm, MG2 is 390RPM but MG1, which I believe is what generates the recharge current, is actually spinning at -1014RPM. So 1000RPM surely must be enough to generate a useable current for recharging?
Absolutely loving the talk from our resident EE's. I'm understanding most, but am also learning so much more!

 

7mph is the software limit that I remember. I believe the explanation was that it it is exteremly difficult to give good regen pedal feel at low speeds, which makes it more of a limit for usability (better brake feel) than because you can not generate any power. The power though would be very low though, kineticic energy recovered is proportional to velocity squared, so there is only 1/15 of the energy available between 7 mph and 0 mph as between 28 mph and 7 mph.

IIRC either mg1 or mg2 or both mg1 and mg2 can provide regen braking.

 

After testing virgin oil in a sample jar that had the transmission sealant around the bottom, I concluded the extra Si and a trace of the Fe came from the sealant. The Fe from the pigmentation used to make the sealant 'pink.' So I'm no longer worried about contamination. But I did put in the breather tube.

It is very difficult to reach the breather tube as the first step is to disconnect the inverter enough to move it aside. Then you have to reach into a 'blind spot' between the transaxle and firewall. Let's just say it is a royal pain and long job. Since subsequent changes have show no ill-effects, I'm leaving it in. I do check the clear hose when I change the engine air filter but other than a light coating that has taken years to form, I don't see a problem.

Bob Wilson

 

Hello Spacejelly,

Thanks, never knew about it...

Toyota Prius - Power Split Device (Obviously already linked here (in PC) in other places...)

Check out the interactive Planetary Gearset further down the page!

Yes, 1000 rpm would be easily enough to generate useful regen power.

Thanks! You're welcome. Glad to hear we're not confusing you too much. Yes, the Prius is more fun to learn about than any other car I know! Not to mention drive. The worst part about it is the other drivers, especially the ones who can't see as far as the Red light 50 yards in front...

 

Here is another hint.

Take one phase of the winding, the back emf cannot complete a path through Q1 and Q6 . If no IGBT is turn on there is NO re-gen. If Q7 is on regen is possible if back EMF is higher than battery voltage, but when back EMF is close to battery voltage, regen is very small.
Q2 can be PWM switching to generation higher voltage to create more regen. Of cause regen creates braking power, if the driver brake hard, the Q2 can go to higher PWM period to create higher voltage, then brake force increases.


Sorry I made a mistake, Q7 should be in the boost converter not the drive IGBTs.