Q&A of Prius Brakes

oh and a horn to warn you, your going to be there as well as them. At least that's what most do here. Although some don't bother with the horn just use the one finger salute as the bust thru.

 

<div class='quotetop'>QUOTE(Frank Hudon\";p=\"67223)</div>

Tell me about it.

Around here, the other driver glares at you and gives the "salute" as if it's *your* fault he went through a red light.

 

<div class='quotetop'>QUOTE(jayman\";p=\"67318)</div>

Around here too. Bad driving is one thing, but my blood really boils when someone not only drives like an idiot, but gets mad at me for not being an idiot as well. Here in Vancouver I've had someone honk at me for not driving through a crosswalk while a pedestrian was not just crossing, but directly in front of my car! (my Corolla - not a tall car - easy to see over).

I learned to drive in England, and drove in Europe before driving in North America. Since then I've driven in lots of other countries too, and although bad driving isn't unique to one area, I think we are a little more rude about it here in North America - just my opinion.

DanMan,
I've been following your thread on PO too... you have good reason to be frustrated. It's tough when people get entrenched in their argument rather than open their minds to reason.

mdacmeis,
Great post. Exactly the kind of stuff I come here for.

 

A few additional comments to minimize confusion.

To understand the brake pressure apply function of the brake system, rather than using the green portion on the fail safe diagram Dan provided, look at the "brake map.jpg" diagram posted first. In normal operation, accumulator pressure feeds the SLAXX (A= apply) solenoids which are pulsed to provide pressure addition control. The SLRXX (R= release) solenoids are used to release brake pressure during normal braking and ABS, VSC, etc. This brake pressure is "vented" or returned to the brake fluid reservoir. By turning on and off the various solenoids, one can apply, hold, or release pressure to each individual wheel.

Under failed conditions, with all the solenoids closed except for SMC1 & 2, brake pressure is reduced by reducing brake pedal travel, hence master cylinder piston travel, and the resulting action reduces pressure as displacement decreases. Thereafter one reaches a point when the master cylinder chambers are exposed to atmospheric pressure via the fluid source hoses from the brake fluid reservoir. Also under these conditions, the only source for rear brake pressure is via solenoids SLARR and SLARL which are closed thus no pressure. Further, without accumulator pressure present, opening the solenoids provides no pressure, thus no rear brakes.

If anyone is curious about the relief valve, this is provided so that if the pump were to run continuously, such as an electrical short or failed pressure sensor, the hydraulic pressure cannot build to a point where the pump stalls and then burns out or worse, pressure exceeds design limits and the hydraulic system bursts. Under normal conditions, this valve is always closed. If it opens, excess pressure is vented or returned to the brake fluid reservoir.

 

<div class='quotetop'>QUOTE(DanMan32\";p=\"67395)</div>

LOL!

 

Dan it dosen't lock because of the rear piston inlet check valve being off the seat till the piston gets depressed far enough to have the front piston compress the stroke simulator.

 

Dan look at the diagram of the cylinder and you'll see the master cyl. piston is larger than the stroke sim. and the fluid just returns to the reservoir. The cc displacement of both pistons is probably the same till the stroke sim bottoms and the inlet check goes on seat.

 

Rick:

CANbus used in the process control environment - which is where I'm familiar with it - can handle 16 bit resolution.

I think the limiting factor with CANbus or *any* shared bus like Foundation FieldBus, ProfiBus, or DeviceNet, is the number of devices sharing the bus.

The "stickiness" you refer to is also experienced in "critical" process environments. The bus polling tends to create this problem unless you dramatically reduce the device count.

Of course, even by minimizing bus device count and working out critical bus masters, you reach a point of diminishing returns on performance. Not to mention dramatically increased costs.

In the aviation world of fly-by-wire, you can create control feel in something like a Boeing 777 that "feels" almost like a "real" airplane. That is exotically expensive. That's why Airbus chose the sidestick control with no tactile feedback whatsoever, though most pilots detest it, due to the cost issues.

I think more cars will start using this approach, and we will soon have a new generation of drivers who get used to dead feeling controls or decidedly non-linear responses.

 

part of the problem might just be the rubber cups in the master cylinder as they have a stickiness to them in that they grip the bore then release and move ahead to compress more brake fluid and that jerkiness might account for part of the feeling you got. Same when stopped and release them the piston seals in the brake caliper tend to resist movement. Just move the rod in a hydraulic cylinder and you'll see what I'm trying to describe.

 

<div class='quotetop'>QUOTE(Frank Hudon\";p=\"73495)</div>

Frank:

Good point. Like how in the industrial process control field they went away from modulating valves to variable speed drives on pumps to control flow.

The modulating valves tended to stick slightly in movement, especially in a slurry environment. This created slight pressure pulses both upstream and downstream, really raising h*** with the line pressure sensors.

If you tried to trim this effect out by fiddling with deadband in the PID algorithm, your control loop would inevitably run haywire. So you put up with that "stiction."

Thank God for Variable Frequency Drives!

 

Rick:

Unfortunately no. The Sticky Friction or "stiction" is highly unpredictable and very non-linear, that's why modulating valves are rarely used now. That's why it was a giant PITA for the process control industry as pressure sensors on piping would go absolutely bonkers due to stiction.

OTOH the steady and *very* predictable pulsation from a centrifugal pump being operated by a Variable Frequency Drive is easy to compensate for with math in the PID control loop.

 

<div class='quotetop'>QUOTE(htmlspinnr\";p=\"73485)</div>

Much of this sensation is a result of the limited use of pure hydraulic braking. The regenerative braking is resistive in nature, by that I mean there is a constantly variable resistive force being applied via the motor, unlike the constantly variable frictional force generated by the brake linings. Not sure I am making sense here, but friction is never constant and has a lot of variation to it. Consider that friction is really the high frequency process of "sticking and releasing," kind of like a high frequency vibration. As a result, when you experience electrically generated resistance, which is more constant when held to a constant value, one achieves a more pure resistive force than that achieved with a frictional material. Again, I am not sure I am putting this in words very well. The linings used on the Prius definitely have that "sticky" feel at slow vehicle speeds.

Now consider the Prius brake system versus a conventional one. The Prius system has very little compliance. Most of the compliance one feels in the pedal is generated by the brake stroke simulator. Flow through the solenoids used to meter brake fluid to adjust hydraulic pressure to each wheel is a function of temperature (fluid viscosity), pressure applied to the solenoid (accumulator pressure), and pressure resisting flow through the solenoid (pressure already in the brake line and the orifice size in the solenoid). We have a large number of potential flow rates with limited control over them (solenoid response time). Thus, at slow apply rates (low pedal travel), solenoid on times are very short and thus feel stepped as you observe while the feedback system (pressure sensors in the modulator which are shown at each wheel in the mechs) waits for the impulse pressure to stabilize somewhat and digital filters try to normalize the pressure such that the control system can generate the target pressure. So the system basically adds an initial pressure to get most of the way to the target, then when valid feedback information is available additional small pressure adjustments are made to achieve the target pressure. At higher pedal travels, the solenoids stay open longer and the target pressures are closer to the accumulator pressure, thus the steps may feel smaller as less fluid flow is needed to increase pressure to the target when additional adjustments are made. Conversely, if there is significant "ringing" when pressure additions are made, the higher pedal travel adjustments could feel worse than those at lower pressures.

Most of this phenomena is not noticed by most drivers because we are modulating pedal travel (desired deceleration) when the vehicle is moving, thus much of this "hunting" for the correct pressure is done while other forces are in place and mask the effects, primarily regenerative braking. Thus, by the time full hydraulic braking is in effect, much of the desired pressure has been ramped in as vehicle speed, and regenerative braking force, has decreased. If the vehicle is stationary, large changes in brake line pressures may be more noticeable since the apply process is controlled by solenoids metering accumulator pressure, rather than by the displacement of master cylinder pistons which provide fluid displacement based strictly on pedal travel. Then again, you could simply be feeling the affects of the brake stroke simulator.

 

IN a digital environment, nothing can be "true" linear or analog. Even with 16 bit samples (65,536 individual values) you will, at most, "approximate" something like a pure sine wave. There is a corresponding exponential increase in system complexity and cost when you go from 8 bit (256) to 16 bit to 32 bit (4,294,967,296) sampling.

Depending on the resolution of the oscilloscope you use to view the A/D conversion, the sampling errors become obvious. Human beings are remarkably sensitive to these obvious errors: when CD's first came out, a lot of people complained they sounded "cold" even though a DDD format CD is technically "pure."

In a lot of digital applications, you just don't "need" so many samples to get good results. In many chemical process applications using ORP for pH samples, all you have to know is a value of say 4-12, with a resolution of 1. So 8 bit sampling works fine, is very fast in an embedded RTOS application, and very robust. I shudder to think of 32 bit sampling in that application!

Rick I don't think this is approaching a "dangerous" lack of resolution, although it could if other cumulative factors become apparent. I'm thinking of the tracking problems that some Prius owners such as myself have reported: the system may not be causing the problem, but it tends to amplify the problem.

Then again, if you don't tune the algorithm to prevent oscillations, a high-resolution sample system can cause even more problems if an underlying issue like wander should happen. In either a high resolution or a low resolution system, something like stiction will create non-linear responses.

The speed of the system isn't very predictable once you deal with a networked environment like CANBus or ProfiBus or Foundation FieldBus. It goes without saying the most critical device is the Bus Master, but low-level devices can really slow overall response. The system response can be made "acceptable" but there is always room for improvement.

Just like sample resolution, sample speed can also cause instability too. In a PID control loop, you invest a lot of effort to fine-tune the integrals. Fortunately, that is now done with software so you don't have to scratch your head while scribbling integrals on a piece of paper.

 

all this talk of the by wire brakes and its necessary mechanical redundancy incase of power failure.

steering is also by wire, does it have a mechanical redundancy or when i get my unpowered car stopped, will i be in the middle of the highway?