Toyota to Recall 1.9 Million Priuses to Update Software

During regen the back emf can complete a path through the diodes on Q1 and Q6 in the inverter. Then like you say Q7 on the Boost Converter will be needed to control the current flow to the battery when the back emf is higher that battery voltage.

 

I hate to say, "Me too!", but that's what I was saying before. Further, I'm betting that at 5-7 mph there's enough RPM in MG2 to create more than 200V; with Q7 on, that would be the beginning/end of the possible regen speed with forward biased diodes everywhere. Go faster than that and you'd control regen by mucking with Q7; with Q7 off, no current flows and no regen; with Q7 on, max regen is what you get, to the limits of the current being sunk into the battery.

Finally: Which pairs of diodes gets forward biased happens automatically as the rotor turns around, no software control of the motor IGBJTs required; the only bit the software has to do is control the PWM on Q7, a much simpler job. In something like this, simple is better.

KBeck

 

The MG2 is connected to drive shaft directly, assuming at 120 mph the MG2 is at 12,000 rpm and producing 650V back EMF, then at 200V, the MG2 rpm will be 12000 *200/650 = 3692 rpm or approx 37 mph.
This means under 37 mph will be no regen if the MG2 back EMF is not boosted up above 200V. The drive IGBT Q2 will act as the switcher of the boost converter.

 

All right. I think I'm confused about the numbers. Here's the graph from earlier in this thread:

And, no, I don't know where this graph came from, which MG (1 OR 2) it's associated with, and so on.
For the moment, however, let's suppose that you got your numbers straight: At 120 mph said MG is rotating at 12,000 rpm.
Then, the RMS voltage of one winding to neutral will be 450V from the graph above. However, we know from the schematics that the neutral doesn't seem to be connected anywhere, but we do have connections across two phases, separated by 120 degrees. Phase diagram is | 1/_0 degrees - 1/_120 degrees | = 1.732 /_30 degrees. Fine: so the rms voltage across two phases is 450*1.732 = 779 V. But, that's the RMS voltage: The peak voltage, which is what one is going to use when one wants the diodes to turn on, is 779 * sqrt(2) = 779 * 1.414 = 1102 V. That's a lot of volts. And that's open circuit, too, but that's not where we're going.

So, 120mph/1102V = X/200V, or X = 120mph * 200/1102 = 21 mph.

Hokay. My numbers are somewhat more aggressive than your numbers. I think that that 650V that you're quoting is the measured bus voltage when in regen or when the boost is on; but, in that case, we're drawing current out of the motor and that drops the voltage. (He scratches his head, tries to remember motor current equations, mumbles some more about resistances across the windings of transformers, and hides his head in shame from his EE 362 prof.)

In any case, that 200V peters out long before we get down to the 5 or 7 mph that we know is the threshold of regen, so, you and the other guys are right: there has to be a boost mechanism running with the windings of the motor as inductors. Or I've forgotten a factor of two (peak-to-peak voltages, anyone?) somewhere.

KBeck.

 

Hi there KB,

It comes directly from the Oak Ridge National Lab report, page 59. (I left that on deliberately, to make it easier to find.)

You raise a very good point about which MG it relates to. Surprisingly, from what I read in the report, it is not obviously stated which MG the graph pertains to. I imagine it's MG2, but even that is muddied by the fact that, for reasons of their own, they prefer to confuse matters further by referring to MG1 as "generator" and MG2 as "motor", and even referring to the "motor" as the Prius PMSM, when in fact both MG are exactly that, motors and generators, and both are PMSMs. However, I would also suspect that both MG1 & MG2 have similar BackEMF voltage/speed performance, even though MG1 has about half the power capability.

bedrock8x,

I'm not at all convinced that the figure of 120MPH at 12,000RPM for MG2 is that relevant (even if it's correct).

I'm fairly sure the ratio between the drive shaft and MG1 is quite a bit higher (around 3.6 from MG1 to ICE in the G1, not sure about G3 which has quite a different gear arrangement, but expect it's a similar ratio to maintain the ICE starting performance), meaning that, with the ICE stopped, MG1 will be turning faster than MG2, and will therefore have the higher (unmodified) BackEMF, and therefore is probably more usable at low road speeds with the ICE already stopped. That's not to suggest which generator is the better candidate for "conversion" to a rotating Boost Converter though. The effect of any slight load variations during regen (uneven braking) will be much more pronounced if they are happening in MG1, due to the higher gear ratio.

Actually, all the ground-connected IGBTs (4, 5 & 6?) will take turns at driving the Generator Boost Converter, depending on which winding is positive at the time. As the generator rotates, so must the transistor "roster". Hence (partly) the precision position feedback systems (Resolvers) fitted to these (Servo) motors.

So, IGBT4 will have a burst of PWM pulses, then they will switch to #5, then #6...

 

Not sure which Gen Prius this applies to (from eahart site), but it shows the speed ratio between MG1 & MG2 with ICE stopped.

 

That PSD is for the Gen 2. The Gen 3 max speed for MG2 is 13500 RPM but there is a Speed Reduction Unit with gear ratio of 2.6 to 1 between MG2 and the wheel.

 

Observation - while driving to work this morning I was trying to monitor BVB (Voltage after Boost) and HV Battery amps while braking. It is difficult to do in traffic and with the SGII between the steering wheel spokes.

When first touching the brake pedal the BVB was at either 500 or 650V depending on brake force requested. This voltage was maintained down to somewhere around the 20 mph mark when the voltage began to ramp down to the HV Battery level right around the 7 mph mark where the friction brakes kicked in. The regen amps remained high until about the 20 mph mark and then ramped down along with the voltage.

At one time when the BVB was at 500V during light braking I saw a point where the voltage jumped up to 650V just prior to ramping down to HV Battery voltage.

I am trying to get a handle on how the BVB responds now so that I can better compare after the software update.

 

I doubt you will see anything, that voltage will be stepped down to charge the battery ;-) I would expect power to be reduced with a hot battery with the change in software. Can you measure the power going into the battery?

 

I am not sure what you mean by this? I know what I am seeing now and can compare after update.

Yes I can read battery amps along with battery voltage.

 

Hi austingreen,

I believe he's measuring the DC Bus voltage, between the Inverters and the Primary Boost Converter. (Not the battery voltage.)

In Regen, from what people is saying, this voltage gets Boosted by the Generator Boost Converter, then down-regulated (by the Buck/Primary Boost Converter) into the battery. Although this seems to be an extra step, it makes some sense to me, as it's probably pretty difficult to accurately control the Generator Boost Output level (but easier to regulate the Buck Converter output).

 

Yes, I am monitoring HV Battery Voltage, Voltage After Boost (Output voltage of Boost Converter and DC input to MG inverters) and HV Battery Amps.

 

I also have the above parameters on my Engine Link with iPad. I set up plots ( 5 min real-time trends) and monitored these driving to work this morning. I want to gather as much "before" data as I can prior to the software upgrade.

 

Well, looks like the UK dealerships are up to speed now. My car is in today having the update done.

 

Last week two colleagues portuguese Prii were also updated.