Occasional Issue when braking over bumps

Well first of all, if we are comparing the symptom to ALL ABS vehicles, there are precious few non-hybrid vehicles that have regenerative braking.

I think this symptom is unique to Prius and Hybrids.

I've driven many "standard" ICE vehicles with ABS and traction control systems and never had a similar symptom. The only time I've ever felt anything like this symptom was when driving my Prius.

I speculate that it has something to do with the Prius, regenerative braking, the traction control system, and it's safeguards for protecting the electric motors.

And I'm not an engineer and could be wrong.
But I'm still saying the symptom as described by the OP....is pretty unique to Prius.

 

My Lexus GX has both traction control and ABS, I never feel this.
I drove Nissan Leaf for 2 years, I remember I got similar feel. So maybe it's from the regenerative braking part.

 

I only ever experienced it in my 2005 Prius. Not in any other car on the same roads. I wonder is regen braking cars with AWD experience it.

 

I have been getting 3.6l/100km in my Prius by monitoring lots of data in real-time (while watching the road obviously), but learning the correct amount of braking to maximise regen (and then, to half the current, to quarter the energy loss (to some degree)) and being able to keep the ICE at 1280RPM and about 11-13HP, really helps save a lot of fuel - especially with my car being very worn out.

During this time, I have also seen what happens when the 'brake loss' thing happens. Basically, the regen stops completely; then it drops to about 5 horsepower worth of regen for some time after a slip (the max is maybe 32-35HP regen).

I believe it is a safeguard for protecting MG1; it can activate in wet weather (without me noticing, except that regen gets limited to 5HP). I was thinking about it; while overspeeding MG1 is one thing to worry about (traction control reduces the risk of this), MG1 is highly dependent upon the speed of the roadwheels, and spins much faster than MG2. Although MG1 is 'connected' to the ICE, it is all connected together.

What I believe is happening, is that the car is protecting MG1. In particular, I think that in addition to 'overspeed', a bigger risk to MG1 is acceleration of the motor; if the car is in regen and suddenly the wheels lock, MG1 can instantaneously change speed; the sudden acceleration would put huge stress on the motor, as it has a decent amount of 'intertia', and could tear it apart. This probably explains the conservative traction control. I don't know how this would impact on the actual ABS; it may be that when the actual ABS is activated, the car 'ignores' the risk to MG1, given it is an emergency situation, whereas small slips during regen / acceleration occur much more often than true emergency stops, so the car does everything it can to avoid these stresses to MG1.

Just a theory, but it could explain some things (including why MG1 is 'protected' at a lower speed on the Gen2, to allow headroom if there is wheelspin etc.

 

The car, however, continues to slow at roughly the original rate, except for a brief "whoopee" feeling during that transition. For the remainder of the stop, if regen is only accounting for 5 HP of the braking, something else has picked up the rest. The description usually offered on PriusChat is that the hydraulic system is what has picked up the rest.

Are you suggesting that you think that is not what is happening? If so, what are the hydraulic braking parameters that you have simultaneously monitored to establish that, and what are they doing?

There is also a fairly long-accepted explanation for why the car has chosen to switch at that moment from regen to hydraulic braking, and one attractive feature of the explanation is that it has to do with braking (namely, braking with four wheels offers more control and lower traction demand than braking with two), rather than simply waving it all off onto an effort to protect an MG.

Are you suggesting that the data you've gathered call that understanding into question? If so, can you say more about the data leading you to that conclusion?

 

I could capture a lot of data and may do soon. I have been observing the data, but am using a live data display, rather than a data capture device. The hydraulic brakes have picked up the rest; I can say that I have observed that regen drops to about 5HP, after the smallest of slips (even ones that don't activate the warning light on the dash), and it stays at 5HP for a while (which can be e.g. a minute) - it seems to revert to 'full regen' mode after you have gone some time without doing any braking. This happens after either a big slip (as discussed earlier in this thread), but also an imperceptible one. The reason I paid attention is because a tiny slip (almost imperceptible) can result in 5HP rather than 30HP regen down a rather large hill.

I won't call the understanding into question. I still find it unusual that, even after an almost imperceptible slip, with no skid warning lights, caused by e.g. a few leaves on the road, or a slightly uneven road surface, the regen can stay at 5HP for a long time down the rest of the hill, and this remains the case with the same level of gentle downhill braking, and despite the fact that the brakes can be applied much more heavily at times without activating the ABS.

If it is due to braking, rather than protecting the MG1, then it is, at least, something to be aware of if trying to improve fuel economy in areas that do not have flat terrain. If it is due to braking, then it is 'bad coding' in the control system somewhere, because it almost completely disables regen for a long time. Thinking about it more, it actually completely disables regen, as the residual braking is actually the same as when driving with no brakes (just the simulated 'engine braking'). It would be very simple to code 'disable regen fully until no brakes are used for x seconds after even a very slight, nearly imperceptible slip'.

I was probably overthinking it, trying to think of a reason for the aggressive loss of regen. However, all that said, I know the speed limits of MG1 are about 6000 and 10,000 rpm. If I was to take an MG1 spinning at 10,000rpm, and change the speed to 0rpm almost in an instant, the force would probably be a lot higher than the 'centrifugal' force, so sudden changes in speed might be something the designers protected against? They put a clutch into the Prius presumably for similar reasons to protect the transaxle against excessive forces (I presume).

 

I expect mg2 would do all the Regen. The speed of mg1 only depends on the speed of the wheels and the speed of the engine.

Also, this happens in EV mode on the prime when I don't think mg1 or the engine are turning at all.

It's interesting that controlling the brakes with the pedal it can seamlessly transition from Regen to friction braking. But if the wheels slip even a little there will be the feeling of the car slipping forward while it transitions, if the friction brakes aren't already engaged.

 

Another possible overthink going on here is conflating "activating the ABS" with simply "switching to friction braking."

In a conventional car, your only braking is friction braking from the get-go, and the only "special" transition that can happen is when the car goes from plain friction braking to "activating the ABS" (that is, putting the valves under rapid computer control to hold braking right near the tires' traction limits, only needed when things get pretty wild).

A Prius has a braking mode where the friction brakes are out of play, so it has two possible transitions: 1) from regen to plain friction, and 2) from plain friction to "activating the ABS".

Again, the ABS gets activated only in fairly extreme circumstances, where the sustained brake force needed is near the tires' grip limits. Quite unusual to see during light or moderate braking.

Just switching from regen to plain friction is easy to trigger. Any fleeting instant of wheel-speed input suggesting that there even might be a traction issue will trigger it. This transition has its own benefit, even without going on to activate ABS: braking on four contact patches instead of two immediately brings the traction demand to comfortably within the grip of four patches, if it was at all iffy with two. It also has to happen before any activation of ABS can happen, so can be thought of as a preparatory step in case the situation further turns out to demand ABS. But nearly always in light/moderate braking, the mere switch to four-patch braking will totally solve whatever iffy traction might have been seen for two patches, and the car will see no need to go any further and activate ABS.

You can also observe the effect (in a Gen 3 anyway) by simply pressing the brake pedal while cruise is set, without canceling cruise first. The cruise control interprets that as a sign of at least some urgency in braking, so goes directly to friction with minimal regen. If you have cruise set and watch the regen amps when braking lightly, they'll be low, but if you just cancel cruise first and then press the pedal, you'll see them high. Very rarely, of course, will ABS have anything to do with it.

I have not observed what you describe, with the system getting "stuck" in direct-to-friction mode for any particular length of time. My observation has been that completely releasing the brake and braking again will show the normally-expected behavior. That's why there's another way of getting regen braking while cruise is set: instead of canceling cruise first, just give the brake pedal a tap-release-apply instead of just apply. The tap will be seen as urgent because the cruise is set, but also cancels the cruise; releasing the pedal and applying again will get regen braking brought in.

 

I must wonder -- is Regen always exclusively from MG2? Or does this depend on particular speed and load conditions, with MG1 sometimes generating too?

Without engine braking, MG2 obviously does all the regen. No torque is passed along to MG1 to put MG1 into generator mode. In engine-braking-only mode without battery recharging, MG1 is clearly motoring, not generating.

But in B mode with simultaneous engine braking and Regen into the battery, I not so sure. Without going back to find the RPM nomographs and doing various computations, it intuitively seems that under some light regen situations with low ICE RPM, MG1 could simply be 'dragging' and thus need to be in generator mode. But with heavy B-mode ICE braking at high RPM, MG1 would need to be in motoring mode to drive that high ICE RPM. The split would happen at the car speed / ICE RPM ratio where MG1 can be stationary.

But I need to think more about this before making any assertions.

 

There are definitely scenarios where mg1 is braking. Probably a lot of the time when the engine is on, actually. But most of the time if you're coming to a stop, the engine is off (at least in the Prime), so all the braking would be from mg2.

 

Wouldn't that be generating, not braking?

 

But a generator is a brake ...

 

True, but when M/G1 is braking the engine, it isn't doing so to brake the car.

 

Yes, I meant breaking with respect to the motor only. I think of that and generating interchangeably. I can see how that could be confused with braking the car.

When engine braking, MG1 must be doing something other than free spinning. Probably generating/braking. So in a way you could say it is involved in braking sometimes.

 

Sure, and the speed of mg2 only depends on the speed of the wheels, period. But both of them can be either "generating" or "motoring" at any time they are spinning, depending on how the HV ECU wants the power balance to work out.

You can pick one or the other to not be turning at all. If the car is moving at all, they can't both be stopped. If the engine is stopped, MG1 is turning. You can have moments when the engine is turning and MG1 isn't, but only when the engine and MG2 speed are in one exact relationship to each other, so it's usually just a moment on the way from one MG1 spin direction to the other.

All the right ideas. It really goes to show something about the usual kind of PriusChat post that tries to say too much about which of MG1 or MG2 is "doing" what to what in which driving condition. More of those kinds of posts get made by newer owners still working through how the stuff works, and fewer are made later on, after realizing how that way of thinking didn't really help very much with the learning process. Really, all these parts are all connected all the time, just as sure as gears are cut from steel, and always have their coordinated roles to play.

It ends up being easier to just remember the algebraic relationships between them that are always satisfied by what they are all doing, and not try too hard to wring out a folksy explanation like "the relationships are satisfied because right now A is doing B to C." The relationships are satisfied because they can't not be satisfied, and that's enough to let you solve for what everybody is doing any time you know a sufficient subset of the conditions.

 

I think the Prime has a one-way clutch that allows both the engine and mg1 to be stopped, but I don't know the details so I didn't mention it.

The algebraic relationship is true for the speed of each motor/engine, but what is more interesting that no one talks about as much is, where is the power going.

 

Prime's one-way clutch prevents the engine being spun backwards when MG1 is motoring hard in EV mode. It doesn't revise the nomograph. (Previous generations also avoided spinning the engine backwards; they just did it by not letting MG1 motor that hard in that direction.)

Gets talked about plenty, including in a rather long, quite recent thread. The New Car Features manual has a series of nomographs paired with schematic diagrams and arrows showing where the power is going in various different driving and braking modes.

Conservation gives you some more relationships (on power flows) that will be satisfied (to first approximation, as if without losses, heat, hysteresis, etc. ... figuring in the losses accurately too would be strictly engineer stuff).

Thinking of 'motoring' and 'generating' as separate things can make the picture untidy; letting numbers be negative can make it tidy again.

Remembering that the MGs are not really DC, but three-phase AC, and the voltages/currents aren't so much 'positive' and 'negative' as phase-shifted can make the picture untidy again; letting numbers be complex can make it tidy again.

 

All very useful info; I think that Toyota have improved the software design with each generation (which makes sense); mine is a Gen 2 so is more primitive; some more testing today (or observations during winter storms) revealed that the regen is disabled with any wheel slip (even a tiny twig on the ground under one wheel) and stays disabled until the brake is released; when letting go of the brake, even two seconds doesn’t ‘reset’ the regen ability, but does increase the allowable regen, so it had a max of 5HP regen, then 7,8 etc - this is a winding, worn road which gets debris whenever it rains - hence my experience. The Prius C copes with it much better!