There seem to be two problems, one that there is a steady load on the battery in the off condition. Modern but conventional  cars with radio keys etc have this problem to an extent, but it is all designed to 'sleep' except for a few milliseconds every second and then wake up if activity is detected. It is not clear to me why the 12V battery cannot be charged from the traction batteries but there seems no appetite to do this. And as there is no electric cranking to be done, the 12V battery is a relatively undersized.

The second part of the problem is the amount of 'stuff' in the EV in way of settings that are held only in RAM. It could be made in a way that saves this system-critical data to something non-volatile as a last act when the battery dips below some lower limit like 11.9V then disconnecting totally. Such low voltage battery saving cut -off is in military kit that spends a lot of time parked up, but never in normal cars or EVs as far as I can see.

Rather like tow charging for EVs, which would be easy enough to design in - the braking is regenerative after all, so charging the battery by pulling it along ought to be a thing, it seems to have been overlooked in favour of an attitude of "design only for success" when " design to cover every possible failure" is a better maxim, in my book at least.

Mike.

I have previously posted on immobilized EVs,

https://engx.theiet.org/f/discussions/28694/evs-immobilised-by-flat-batteries-or-breakdowns

The situation doesn’t seem to get any better. Is the problem that EVs are seen more as software than mechanical systems and too much is in the hands of software engineers, who don’t generally have reputation for producing robust software. It is possible, but difficult and expensive, to produce reliable software. The railway and aviation industries have been doing it for a long time (although Boeing seems to have taken it’s eyes off the ball. Someone on here, possibly Andy M, said that for a reliable system it must be either simple enough that every failure mode can be checked or complicated enough that every failure can be detected.

As an aside I recently went to pick up a relatively new Suzuki S Cross as a hire car. It was still wet from the cleaning both and I could open the doors with the remote key. We loaded our luggage and I went to start the engine, everything went dark. The hire company said ‘battery failure’, gave us another car and managed to get the tailgate on the first one open so we could get our luggage out. By the time we had transferred to the new car and were ready to go a mechanic had already changed the failed battery in the first car and got it started. No complicated reprogramming required.

I'm surprised that they still use 12V batteries as you no longer need to run a starter motor. Some IC's used higher voltage batteries, but nothing above 48V (for obvious reasons).

The actual electronics runs on much lower power rails (often 3.3V, 1.5V, sometimes a negative rail also). The exceptions will be input/output actuators - electric windows, speakers, etc. This means that every module has a 12V to multiple lower rails conversion and usually a supervisory device providing specific power rails for the System-on-Chip. There are always going to be some level of loss in such a device even with a high efficiency stepping power converter.

They won't want to have 400V (or 800V for some vehicles) running to every module in the vehicle as that means you have to consider the regulations around such voltages. But I agree that you could have a single power conversion module providing a vehicle wide lower rail to power the non-traction modules. That could then include charge and conditioning for the back up battery (the 12V).

I doubt that an EV has that much more being held in NVRAM then a complex IC vehicle. But there might be fewer modules holding more data. I assume you that modern IC vehicles do have strategies for handing 12V drop outs, given its a common occurrence on crank in cold temperature and then you have stop/start strategies. I'm not going to detail any of the actual strategies that were used but I'll point out that if you try write NVM when your battery is dropping like a stone - you risk corruption.

As you mentioned, there are also strategies to minimize the power drawn in extended sleep modes to as near to zero as you can get (which is pretty low). Even on an IC you need to deal with someone parking the car up for a while and going on holiday. You can't have it fail to start when they return. My experience is that the commercial vehicles have been based around the architecture for a passenger vehicle, so it still had all the same design aspects considered.

EVs are slightly different. Some manufacturers are developing entirely new architectures, some are entirely new entrants into the industry. Others are adapting existing IC architectures with some changes to suit EV use (usually minimizing quiescent power drain and efficiency). But I haven't had a significant involvement in an EV project - I left automotive a while back.

I think there is just so much car stuff already 'out there' that runs off a nominal 12V that to not have 12V and to have to redesign,  either up or down to a new voltage,  makes little sense for companies racing to get their latest models  out of the door as fast as they can. That pressure means as much re-use as possible from existing catalogues /parts bins (window motors, wipers, locks cameras etc all would have to be re-done.. )
 I think a large part is as you say, the new entrants and a 'software defined' vehicle approach - which is arguably a paradigm error.

Professionally I am very aware of equipment that both remembers some things and very deliberately forgets others once the external battery is removed - NV writing in that case is just about having enough capacitance on board to complete that, and a method of ensuring  power to  all other peripherals is removed, and all other computer processes are dropped, so the writing  happens.

Arguably something that has to be reprogrammed to restart it it could be seen as an incomplete design. (That said my  normal car loses the radio key when the battery is changed, and if I was not an electronics person maybe I'd be going to the dealer..cheers Vauxhall !! )

Even trains sometimes don't manage it..(....GTR Fleet Control diagnosed the problem rapidly and a global GSM-R broadcast instructing drivers to perform a battery reset saw traction power restored to 27 of the affected units. This left 31 units that could not be reset by the driver and had to wait to be rebooted by a technician using a laptop....)

Here's hoping that once the technology stabilises things are better organized.

Mike

"difficult and expensive, to produce reliable software"

I would argue that it is potentially more expensive not to. The cost of stopping a manufacturing line or a vehicle recall can easily outweigh profit. Let alone reputation damage. How many people are avoiding flying Boeing now?

It is the same as the justification for doing proper systems engineering, for a complex project (and a modern car surely is), you will save yourself lifetime costs against not doing it. However, given we don't have parallel universes to check the alternative in, its not always easy to sell.

Hello Mark:

"Difficult and expensive" also applies to the maintenance and repair of cars.

System complexity is becoming a major problem for the smaller car repair shops here in the US.

Even when handheld portable code readers are available, they sometimes do not provide adequate information to help repair the car.

To a certain extent this also applies to the brand name car dealer's maintenance shops, who's prices have gone sky high.

Based on what i have read a large number of car mechanics working for the smaller car shops have simply given up.

Peter Brooks

Palm Bay 

I would expect that pushing users to only use official dealers is an intent. Something I would expect to see more of, especially given the concept of monetarising the features which seems to becoming more common.

In a competitive environment, you make less profit on the product itself and you need to recoup more of that profit by selling services to your existing customers.

I would expect that pushing users to only use official dealers is an intent. Something I would expect to see more of, especially given the concept of monetarising the features which seems to becoming more common.

In a competitive environment, you make less profit on the product itself and you need to recoup more of that profit by selling services to your existing customers.

Except that in some circles in the UK at least, lock-in to the dealer has  been seen as a bad thing, and a mix of pattern parts and hobbyist car bashers keep cars going long beyond the point where the dealer will be happy to touch it. - I've been driving since the 1980s,  and have never felt the urge to buy a new vehicle, although I now tend to buy at the 8-10 years point rather than older as I would have done years ago.

Now that may not be typical , and I am aware I keep odd company, but I know a lot of folk who are quite happy to invest time to avoid paying over the odds for repairs etc. I'm not sure what maintaining a 10 to 20 year old EV will be like, but I guess in another 5-10 years we will find out. (as right now I'm driving a machine from 2010 ;-) )

Mike

Ah, one of my former colleagues has a collection of older cars and bikes - and by that I mean at least 20 years old and going back to post-WW2. These are all self-serviced. The only thing he can't do is issue his own MOT certificate.

As an aside, its interesting to be  driven in a car that doesn't have seatbelts, or indicators, or a roof but will still happily belt along at 40/50 mph. But at least has a spade on the side if you get into trouble.

But the real issue will be that modern vehicle electronics is based upon modern mobile phone technology. Shrinking the lithography also shrinks the lifetime of the silicon. It isn't cost-effective to keep alive the older fabs that support the older processes just for spare parts.

So I suspect you are not going to see the equivalent of a 40 year old Volvo 240 still on the road in the future.

Mike:

Have you priced cars recently?

 As of 6 months ago, second hand cars ( 3 years old)  here in the US where only about $2,000 less than new cars. With new cars going for a little under $30,000.

The car market is now going south due to the high interest rates and the companies that sell second hand are taking a blood bath. Companies like Ford are in trouble with their F-150 trucks sales as they were overpriced and their dealer inventory has sky rocketed.

The trouble with older cars is the degrading of the plastic parts like door handles . The dealer wanted $180 for one, while parts supply chains were selling good copies for $30. 

Peter Brooks

Palm Bay 

Hello Mark:

According to what I read recently the proposed new Semiconductor Fab in India will be concentrated on trailing edge technology.

I assume that cars older that 1975 ,that are still running, you have to have additives (lead) added to the fuel.

The other problem is getting the correct air-conditioning fluids (if you have never had A/C forget this).

Having worked for an IC manufacturer who made parts for cars, I have many stories, that I can not tell!

Peter Brooks

Palm Bay