Your EV maytake you hostage and try to kill you !

Last edited:
it could be that a software (or hardware) failure might just totally prevent the “human” activation of the brakes ?😱😳😳
Precisely. A software failure in the ABS system can disable brakes completely. And pretty much every car on the road today has ABS. Has that ever happened out of hundreds of millions of vehicles on the road? I haven't heard of it.

The only "software" issue in cars that I recall is the few unintended acceleration incidents in Toyota cars. And all evidence points to it not being software at all but mass hysteria and possibly floor mats.
 
Precisely. A software failure in the ABS system can disable brakes completely.
Check 2:35 in the Fortnine ABS video I linked. In motorcycles and older cars, zero software is required for an effective ABS. Just a speed sensor (which doesn't require much, if any, logic) and a solenoid. No reason to get software involved at all.

I don't know how ABS is controlled in modern cars, but it can be done without software, or at least with a minimal, low-level, separate logic circuit. Which is exactly what the post is about.

Furthermore:
Since the ABS can interviene to prevent the “human” input locking the wheels, it could be that a software (or hardware) failure might just totally prevent the “human” activation of the brakes ?
A software failure in the ABS system can disable brakes completely
In the video, you can see that if for some reason, ABS fails, or stops working, you still have brake control. All that fails is the system that electronically assists your foot in preventing brake lock. The hydraulics is still a closed system, it's not like ABS failure opens up your hydraulic line and kills brake pressure. That would be lawsuit-inducing stupid design. So it's very unlikely to exist.

Has that ever happened out of hundreds of millions of vehicles on the road? I haven't heard of it.
Are you asking, have cars ever had ABS failure? I've had it happen to two cars that I've owned in the past. You've never seen that warning light come on? It can and does happen, and you can choose to get it repaired, or not, and just make do with brakes that don't have ABS. Are you asking, has there ever been an ABS failure that results in complete loss of brake control? No, that would be silly, for reasons that I mentioned above. Are you asking, has there ever been a software failure that leads to inability for the driver to engage brakes? Yes, that can happen, if, as you say:
The gas pedal, the brake pedal and the steering wheel have been physically decoupled for a long time now
and one example was linked by the video that started this thread, a runaway tesla. And what several people on this forum are saying is, having software control critical systems is not always a good idea. Because as rare as incidents might be, the consequences of said incidences are much higher when humans are unaware of how much control computers have over our safety. They're also much more unpredictable, whereas human error is pretty much guaranteed, and if the majority of people operate under the assumption that fallible, dumb humans are in charge of cars, planes and power plants, then the majority of people can then make more risk-based decisions with their lives, their driving, and their behavior around others. As opposed to going around with an assumption like: "computers are in charge of everything, and they're much safer than people, so I don't have to be as careful anymore."

Also pretty silly of you to say "I've never heard of software disabling brakes", when the first video in the thread is an account of a Tesla's brake software not working.
 
in the interests of accuracy….just sayin’..
Mazda 2 Neo ..2299 lb
pre 67 VW combi bus …2515 lbs
This source says 2095 lbs for '63 Kombi:

Discrepancy might be due to the wide variation in interior equipment for Type 2 VWs. That was one of the few stinking death boxes ever made whose payload was similar to its empty weight.
 
Last edited:
You seem to be trying to argue that no software can be 100% reliable. Discounting the fact that there are billions of things in this world that are controlled by software, gas furnaces, ovens, elevators, literally everything modern, running software code 24x7. And they are functioning just fine. I find your argument asinine. So l'm going to leave it at that. :mrgreen:
 
Yeah, that's why a piddly little Mazda 2 weighs more than a pre-'67 VW bus.
The road arms race hurts all of us.

I absolutely hate it. But we can't do anything about it, it's a matter of picking your poison.

I watched the crash test results for my car before i bought it. It will take a big hit from all angles and the cabin seems to only deform by a few %, meaning, it's very likely you aren't getting severely injured.

Here's my car after having been slammed with 4200lbs ( IIHS side test )

1697041654997.png

Frontal crash:

1697042148845.png


I would definitely prefer a 2000lbs vehicle, however, this is how they fare in the same situation:

1697042355300.png

I think the extra 400lbs is totally worth it!
 
Precisely. A software failure in the ABS system can disable brakes completely. And pretty much every car on the road today has ABS. Has that ever happened out of hundreds of millions of vehicles on the road? I haven't heard of it.

A majority of cars have been designed as such that ABS doesn't control control braking but intevenes during braking.. from what i understand, if ABS fails then the symptom is that the brake pressure is not rapidly modulated during traction loss.. IE.. the brakes still work except without their function.

The only "software" issue in cars that I recall is the few unintended acceleration incidents in Toyota cars. And all evidence points to it not being software at all but mass hysteria and possibly floor mats.
...yeah as a result of this my 2018 car's accelerator pedal ( totally electronic ) has 7 wires coming out of it last i checked. Only 2-3 are required.. i would guess the extra wires provide a second opinion on what the throttle % should be to add safety.
 
You seem to be trying to argue that no software can be 100% reliable. Discounting the fact that there are billions of things in this world that are controlled by software, gas furnaces, ovens, elevators, literally everything modern, running software code 24x7. And they are functioning just fine. I find your argument asinine. So l'm going to leave it at that. :mrgreen:

Ah you just mentioned software applications that are so miniscule in complexity that previously, small and fairly dumb electronic circuits used to handle those applications.

A car that can drive itself using inputs from dozens of sensors and also a cellular network is what keeps me up at night. It has a large attack vector, and no real battle testing on that yet. It is closed source, but even if it wasn't, the software is sufficiently complex that 1 human couldn't understand it all.

I'm going to be an extreme, or maybe never-adopter of a car like that, as long as it is still legal to operate a car that isn't basically a CPU on wheels.
 
A majority of cars have been designed as such that ABS doesn't control control braking but intevenes during braking.. from what i understand, if ABS fails then the symptom is that the brake pressure is not rapidly modulated during traction loss.. IE.. the brakes still work except without their function.
From what I understand the ABS module has a pressure bypass valve for each caliper, 4 wheels 4 calipers/drums, so if the ABS module software decided to bypass pressure to all brakes, it would effectively prevent you from not crashing into something.

Ah you just mentioned software applications that are so miniscule in complexity that previously, small and fairly dumb electronic circuits used to handle those applications.
My point was that software has long been relied on for life critical systems, where analog circuits used to do it before. It's a natural progression and there nothing inherently wrong with software. It's not all buggy, insecure and unreliable. And there is no evidence whatsoever that auto manufacturers in tandem with all the regulators that exist are putting anything even remotely dangerous on the market. Yes, cars are getting much more complex software, but life critical parts are still compartmentalized and are designed with many fail-safes.
 
From what I understand the ABS module has a pressure bypass valve for each caliper, 4 wheels 4 calipers/drums, so if the ABS module software decided to bypass pressure to all brakes, it would effectively prevent you from not crashing into something.
This is incorrect. If ABS fails, you still have brakes. The bypass valves you are describing, which briefly release and then reapply pressure to prevent wheel lock, are normally closed valves. If they fail, they fail closed, maintaining break pressure, albiet without electronically assisted antilock.
 
This is incorrect. If ABS fails, you still have brakes. The bypass valves you are describing, which briefly release and then reapply pressure to prevent wheel lock, are normally closed valves.
What if the ABS software keeps the normally closed valves open for as long as the software wants?
 
What if the ABS software keeps the normally closed valves open for as long as the software wants?
I don't understand, I thought you were convinced that software in charge of critical systems never fails and we should rely on it with no qualms? I dont feel this way: the simplest motorcycle ABS systems, as I have mentioned, can use speed sensors to fire solenoids, and not software, and that's what i would prefer if I had a choice in the matter. Which most customers do not, and it seems likely that we will have even less choice in the future.

Your hypothetical would depend on the specifics of the particular ABS in play. In the FortNine video above, the closed brake system would still be a closed brake system regardless of what valves were firing, and braking pressure would still be available even if valves were open. In lots of modern cars, ABS engaging would involve both the opening of specific valves, as well as an engaging of the brake booster pump, but most would still also be some kind of closed system. This is because ABS can and does fail, so manufactures will make sure that ABS failure, of many types, won't cripple basic braking. But unfortunately, while I do have a good understanding of the mechanics of ABS, I do not know all the specifics of all systems on the road, so I can't say that with 100% certainty. Your hypothetical critical failure could happen, for all I know. In which case, systems should be built assuming that software can fail catastrophically. Do you think so as well?
 
ABS is extremely unlikely to fail in a 'brakes open' state; it has multiple checks against faulty sensors and will fail to operate if the checks at startup fail. It's a pretty good system that saves alot more lives than it removes. It's also got some 40 years of refinement in how it's designed.

I drive a car with a handbrake and manual transmission. In an emergency braking situation, i could downshift the motor and yank on the brake and get some pretty good deceleration. The ability to do that makes me feel a lot more confident in the vehicle.

I have had a car with an automatic transmission have the transmission computer fail on a busy highway. The result was rapid and unexpected braking. The car could not be shifted into neutral and we could have been plowed into. Luckily there was a shoulder on the highway.. but that car design could have been the end of Neptronix as you know him.

So yeah, i'm a little averse to a car that has a computer capable of operating the entire car itself. Because we are using external sensors and data from the internet to help drive the car, we have an additional thing that could happen: we could lose total control over ALL the vehicle's functions in the case of a software error or hack.

It's not a risk i'm interested in taking, i'd prefer as little software in my car posible, and zero connectivity to anything resembling an internet or inter-car-net.
 
the simplest motorcycle ABS systems, as I have mentioned, can use speed sensors to fire solenoids, and not software,
Sorry, but if you think a sensor and a solenoid can function in any useful way without some “logic” circuitry (firmware ?) , then i will be amazed.
Software , firmware, even those “sensors” and solenoids, can all fail, no matter how well tested and prooven they may be.
A simple rock , kicked up and rupturing a brake line can render the best engineered brake system usless.
5hit happens as they say !
 
I never thought that, of all places, a DIY electric bike forum, there would be so many Luddites. :ROFLMAO:
 
Well, now you know us better :es:(y)
 
You seem to be trying to argue that no software can be 100% reliable. Discounting the fact that there are billions of things in this world that are controlled by software, gas furnaces, ovens, elevators, literally everything modern, running software code 24x7. And they are functioning just fine.
Nothing is 100% reliable. The question is - is it reliable ENOUGH?

Well designed software based systems for mission critical tasks (flight surface control for aircraft, steering assist for cars, ABS etc) are generally far more reliable than the systems they replace. Sure, you can screw them up and make them failure-prone - but that's true for any system, from cable operated brakes to a fly by wire system for an airliner.
 
Sorry, but if you think a sensor and a solenoid can function in any useful way without some “logic” circuitry (firmware ?) , then i will be amazed.
Then be amazed!

A 1960s era house furnace has no firmware. But it does have a sensor (a thermostat) a solenoid (fuel or gas valve) and even some electronics (igniter.) And many devices still work that way. I have a toaster, for example, that does just that. Just a bimetallic strip, a few switches and an actuator.
 
Then be amazed!

A 1960s era house furnace has no firmware. But it does have a sensor (a thermostat) a solenoid (fuel or gas valve) and even some electronics (igniter.) And many devices still work that way. I have a toaster, for example, that does just that. Just a bimetallic strip, a few switches and an actuator.
My house that I bought a few years ago, it had a newly installed "smart" thermostat, programmable and whatnot. Failed in a year, stopped being able to turn on the heat. Replaced it with what I thought was a more reliable electronically controlled thermostat, no smarts, just programming on and off. Also failed within a year. Scoured facebook marketplace for a decades-old Honeywell mercury thermostat, no issues for years, and it was already a couple decades old when I got it.

Despite my comments on this thread, I don't actually have any inherent problem with software running things, when it works. I'm just someone who is distrustful of assumptions that software is always working and always better, and I don't like when analog solutions are portrayed as automatically worse, inefficient or inherently bad.
 
Last edited:
Nothing is 100% reliable. The question is - is it reliable ENOUGH?

Well designed software based systems for mission critical tasks (flight surface control for aircraft, steering assist for cars, ABS etc)
Firmware in a modern toaster is just as reliable as mission critical stuff. There are 0 bugs in that firmware. It fails much more often than mission critical stuff because of hardware issues. Like overheating the logic chip beyond specs every time the toaster turns on.
 
My house that I bought a few years ago, it had a newly installed "smart" thermostat, programmable and whatnot. Failed in a year, stopped being able to turn on the heat. Replaced it with what I thought was a more reliable electronically controlled thermostat, no smarts, just programming on and off. Also failed within a year. Scoured facebook marketplace for a decades-old Honeywell mercury thermostat, no issues for years, and it was already a couple decades old when I got it.

I had two identical WiFi enabled smart thermostats for about 15 years. Just removed them a few months ago when the "cloud" part of the product went out of business. Nothing wrong with them. Still work great, just no more remote access.
 
Firmware in a modern toaster is just as reliable as mission critical stuff.
I've worked on mission critical code. And the reliability framework, processes, tracking and testing that it goes through are NOTHING like the code in a toaster.
 
Back
Top