Hacker News Reader

Comment by sbradford26 | Hacker News Reader

So a big reason why this probably wouldn't fly is that it has a common mode failure which is loss of power. Most EVs don't use permanent magnetic motors so it needs power to be able to use the motors for regen. So lets say the relay fails from the main battery pack suddenly you have no power, and no way to energize the coils in the motors to use them to regen. This could be mitigated by using permanent magnet motors but you still have the issue if control power is lost the vehicle cannot stop. The beauty of hydraulic brakes is that they work really well with power, but still function without it.

View on HN

Interestingly enough, I've had hydraulic brakes fail on me for two different cars (thankfully it was 20 years apart, but still I would have preferred that the first time was the last time).

iirc both times it was some sort of brake fluid leak.

So, while I totally agree that having motorized breaks is really scary, hydraulic brakes can fail as well.

I don't deny that there can be physical failures to the systems. And I also drive a hybrid currently that most of the time only uses the electric motor for slowing down, I don't find using the motor for braking scary. I personally see having both regen and hydraulic brakes on a vehicle a great step forward in redundancy and safety.

Electrical slowdown in hybrids does not provide redundancy for emergency braking. The battery has to absorb like 500kWt during 1s, this is well outside what a small battery can do.

Redundancy doesn't mean it will bring the car to a stop as quickly as hydraulic brakes, it is just another means to bring the car to a stop safely. From personal experience I could safely get my car to a stop in many situations without using the friction brakes at all.

I've also had two cars lose brake pressure. I'm very thankful that the cable-operated parking/emergency brake was functional on both cars; it saved one of them from going through the back wall of the garage.

The cable-operated E-brake definitely doesn't stop the car anywhere near as well as the power-assisted hydraulic brakes, but having two different braking methods can be a literal lifesaver when one system fails. (And all systems fail! I'd previously had two other cars where the E-brake didn't work.)

I've never had my breaks fail on me. I also don't use the parking break that often. I would be so proud of myself if I realized what was going on, remembered that the e break exists, and used it before hitting the back wall of my garage.

Good on you.

When the hydraulics failed (brake line exploded) on my old vehicle I discovered that the cable is also attached to the brake pedal, so I didn't have to play games with the parking brake mechanism. There is a world of difference in braking force though, it's like pushing on the side of a brick building to gently encourage the vehicle to stop. I was lucky that I was driving a manual so I could also downshift to kill the speed.

I had the engine fail (timing belt broke) and that also killed the hydraulics in that car, very different behaviour of the brakes compared to normal and you need a hell of a lot of pressure to stop but it still works. Downshifting was not an option since the engine didn't work anymore.

> that also killed the hydraulics in that ca

You are probably referring to the brake assist mechanism in all cars after the 90's. It uses vacuum pressure from the engine to help you push the pedal down. Without the engine running, the brakes will work ~3 times, and then they'll get super hard to press - but they will still function if you have leg muscles strong enough to work them.

> I've never had my breaks fail on me.

Just a friendly reminder that your car has brakes. When they break, then they will fail.

I had _a_ brake caliper jam on my tundra. Completely frozen sliders on the caliper. The good news about hydraulic brakes is that the other circuits on the brakes still worked fine! I actually didn't realize the caliper was dead for (likely) a few months. With modern cars, each wheel is independent, so leaks and failures are more failsafe.

“It still kind of works without power” was an idea that kept hydraulic actuators in planes for a long time. But now most new planes are fly-by-wire (fully electric). Reliably stopping an electric car in an emergency should be a solvable problem.

So the fly by wire just means it is digital control to the hydraulic actuator instead of a physical cable running from the yoke or something to the actuator. They have not gotten rid of hydraulic actuators yet, they are working do that in the future but that is still a long ways out. Aircraft power systems also have significantly more redundancy so loss of control power is not a common mode failure due to have multiple dissimilar power sources.

It could be solved but it might not be worth it.

Planes have a whole different set of expectations on their reliability and upkeep.

For example I don’t think people would be excited about an electric car with an aux power unit.

They already have an auxiliary power unit of sorts: the 12V battery.

The air brakes used on heavy vehicles (semi trucks, etc) have a very simple fail-safe. You need power to not brake. I.E. Air pressure is used to disengage the brakes. If you lose power/air pressure, the brakes engage and are stuck engaged.

That is easy to accommodate as you can design the electronics such that when there is no power they cause maximum braking force and dump energy into the braking resistors. This can be done completely with power taken from the motor itself and if it’s not turning to generate power, then I would expect the parking brake system to engage, eg: a spring defaults it to “On” and power is used to keep it disengaged.

This can only be done with permanent magnetic motors which is not the norm for EVs for many reasons. In a non permanent magnetic motor on loss of power there is no energy to dump into braking resistors since the motor is not causing any resistance because the coils have been de-energized.

Most EVs have permanent magnet motors of some flavor right now (some have both a PM motor and a pure induction). I would assume that if you were having the motor be the primary braking system, and it was induction, there would be a tiny cheap magnet installed in it to generate bootstrap power to prevent exactly this scenario.

An even simpler design is to assume that the car can only start moving when there is power. If there is a loss of power while moving, then you can use the energy in the HV capacitors in the inverters to make a bootstrap field. That can be maintained till the car stops (and the user still has control of when/how much to brake). When the car comes to a halt, the parking brake is applied, and it won't move again till repaired.

All of that can be done with software-only - no hardware changes needed.

Parking brakes are pretty useless to slow down a car from any decent speed.

This is entirely dependent on how your vehicle is designed. Pulling the electronic parking brake in my car (Audi S4) at a decent speed applies maximum braking effort to all 4 wheels simultaneously. It'll practically pop your eyeballs out.

Interesting. Although if that brake is electronic you can't really use it as backup in case something goes wrong with the car's computer systems.

Define "decent." I lost all hydraulic brakes on my 87 Toyota pickup due to a hole in a brake line and I had to drive it to the repair shop using only the cable-operated handbrake. Granted, I didn't go on the interstate, but at some point I was still doing 35-40mph.

Having a stick shift definitely helped, though.

Letting the driver decide whether to coast or brake is far safer than making the decision for them.

Slamming on the breaks is not a solution. Part of what makes hydraulic breaks safe without power is you can still control the breaking force.

That would be the default of this system if the brakes are still able to communicate with and provide power for the brake pedal. Basically if the brake system lost external power, it would get the power needed to run via scavenging it from the motors. Once the vehicle came to a stop, it would then loose power and the parking brake would then engage.

The slam on brakes case would only trigger if there was gross damage to the braking system, where you might have bigger things to worry about.

I think you've lost sight of the initial concern that was raised. There's still a common-mode failure of loss of control power.

Control power is different from actuation power, the latter of which is what you're addressing. And it's not "gross damage" because it's a common-mode failure, so you can't just ignore it like that.

To lose local regen braking derived power requires a failure of the cars systems similar in nature to one that would cause loss of hydraulic fluid or air in a traditional braking system.

Edit: made more concise.

I see what you're saying now

It is trivial to derive control power from actuation power - at least enough for a few can-busses that run from each wheel to up-to-12 independent sensors on the brake pedal.

It's trivial if you're doing an opportunistic design, but it's challenging to match the safety and reliability of a hydraulic system while beating the cost.

So rather than coasting to the shoulder when you run out of gas your tires immediately slam full on coming to a screeching, uncontrolled stop in the middle of the highway?

hah. I didn't see this before I wrote my comment. Really makes you wonder about this common SWE fantasy of mew ways to turn cars into big electronic touch screen devices, whether the people dreaming these ideas up have ever actually driven a car.

Nope. If your car lost power for whatever reason, the system would scavenge power from the inertia of the car and the brakes would still operate as they do when the car is operating nominally. The 1-10w needed to run the braking system is a rounding error in comparison to the 4-10kW of wind/rolling resistance at freeway speeds.

1-10W will run the entire braking system, from foot pedal through computer systems to activating the breaking resistance?

It is possible. Some embedded systems can work in the milliwatts to microwatts region.

>That is easy to accommodate as you can design the electronics such that when there is no power they cause maximum braking force and dump energy into the braking resistors

This is still uncontrollable failure. What if the car is in the third lane doing enough speed and there's some cars behind? I assume no ABS, because ABS can't work with locked up brakes. This means no steering, just skidding whatever way.

Having no power would be hard as the brake system can easily generate power to operate itself if the vehicle is moving (and if its not moving that is what the parking brake is for). This is more of a scenario that occurs when there is a catastrophic failure in the braking system, such as the motor control electronics getting damaged.

So fallback in the case of power loss is you don't come coasting to a stop, your car just suddenly and unexpectedly slams on the brakes? What's it do for recovery, accelerate into the car in front of you?

Maybe dedicated capacitors it’s enough power for one stop? Or a mechanical fail safe? Just a big pad that drops down on top of your tires.

Hold up capacitors definitely could extend the length of time the brakes would work but it would depend on what the power failure was and how long you need to brake. A mechanical fail safe just seems like a worse version of hydraulic brakes.

Maybe the capacitors could store enough energy to start braking, and then the braking energy could recharge the capacitors.

Interesting. Is there no way to design the circuitry so that when given no input, it defaults to a fully brakes-engaged mode? I’ve only worked with PM motors, so I know how that sort of h-bridge could be designed, but my knowledge of non-PM motors is fuzzy at best.

If you're driving 75 mph and one or more motors suddenly lose power and it fully brakes, that's a certain death in my opinion.

Well, it is certainly a very shaky proposition, and the car will probably be a total loss.

But there are ways you could walk away from this.

You see some pretty crazy crash survivals in race driving, for example.

And I remember stories from a few years ago, when Ford had a recall on F150 tires because they tended to blow out at highway speeds, with an effect similar to full wheel lockup-- yes, people died, but others didn't.

life == hope

are people under the impression that when you're driving down the highway at 75mph, there isn't already a hundred ways you could die?

I mean, now we have 101 ways, and not all those ways are equally likely...

I'd imagine this was introduce adverse failure modes like - slamming to a stop without warning within seconds on the highway if power is lost, rather than coasting to a halt... and then being untowwable as the car can't be put into neutral.

Tesla's that lose battery are already a challenge to tow as the parking brake cannot be disengaged without power.

So looking into it a bit there are what is called Eddy current brakes. In the section of disk eddy current brakes it talks about the power tool implementation where there is just a metal disk and the magnetics are pulled away from the disk when the device is powered on and passively return to the disk when power is removed. Something like that could work, but I am unsure of the size of magnets or disks necessary to adequately stop a vehicle. Most likely you could also add a normally closed relay hooking the motor to a resistor load. So when power is applied the resistor load is disconnected, and in a loss of power event the motor is directly tied to the resistor load.

https://en.wikipedia.org/wiki/Eddy_current_brake

Eddy currents develop in proportion to the speed of the moving magnetic field. So if you used eddy currents for braking, it would be most effective at high speed and gradually have no stopping power at low speed.

That is mostly true about regenerative braking as well. Once you hit a certain speed the vehicle has to use some energy to get the motors to actually stop the vehicle. Usually EVs and such will just rely on the friction brakes for the last bit but some cars like the Bolt will automatically use some energy to slow the car to a stop so you can use do one pedal driving.

Good point, and a good reason to have friction brakes also installed.

Hmmm. So you're travelling along the interstate at night, you lose power and your car screeches to a stop in the middle of high-speed traffic and your lights are off.

I would love to be present at the DFMEA (Design Failure Modes Effects Analysis) meeting when this came up!

I see a lot of good ideas for how to get around this failure mode, I'll throw one more in the ring: "giant spring loaded spike to jab into the road"

I actually don't think that's a bad idea... Specifically, most cars can only brake at about 1 G, and if that's insufficient, they use crumple zones and airbags which brake at far more than 1 G.

If emergency braking could use an explosive-spike-in-ground deceleration at say 5 G, then far more crashes would be survivable.

I think the reason it hasn't been done are human reasons - users don't like the idea that jabbing the brake pedal hard will mean they have to pay tens of thousands of dollars to repave the road and replace their spike.

Good point. Feels a little like airplanes where at some point it was decided that mechanical backup is not needed for fly by wire.

So a fly by wire aircraft might not have physical cable moving surfaces, but most of them have a backup mode that relies basically entirely off of physical switches and direct communication with the surfaces. Those modes are only used in extreme emergencies but they are still designed into the system. Also what works in airplanes doesn't always translate to cars. Aircraft have rigorous maintenance requirements, while cars leave the manufacturer and might not have any sort of inspection or maintenance done for 20 years.

Reply on news.ycombinator.com