Tesla Model 3 revealed

[98elise]

EricE said:

I don't think any production Model S on the market at the moment has the hardware required for fully autonomous driving. I'm not an engineer but they'll at least need some type of LIDAR sensor for that. I suspect it will be an option for the Model 3, maybe even a retrofit option announced at a later date.

How do you drive without LIDAR? An autonomous car just needs to be able to sense as well as you or I

I don't know what Musk is planning, however he appears to deliver on things he says he will do.

[greygoose]

babatunde said:

walm said:

Ouch - a trusty Zero Hedge take-down: http://www.zerohedge.com/news/2016-05-05/tesla-tum...

I quite like Zero Hedge, however according to them all stocks and shares and indeed the world economy is a ponzi scheme and we sell everything buy GOLD and keep in under our beds, cause come the armageddon gold will be the one true God.

You can't value a growth company the same way as an established company, economics 101, while I suspect the market cap of Tesla has got ahead of itself the GM's of the world have billions of legacy debt and structure

Edited by babatunde on Thursday 5th May 19:38

True, the pension and healthcare costs at GM must be immense. I see quite a few Teslas about and could be tempted by a smaller one as a daily driver.

[The Vambo]

98elise said:

EricE said:

I don't think any production Model S on the market at the moment has the hardware required for fully autonomous driving. I'm not an engineer but they'll at least need some type of LIDAR sensor for that. I suspect it will be an option for the Model 3, maybe even a retrofit option announced at a later date.

How do you drive without LIDAR? An autonomous car just needs to be able to sense as well as you or I

I don't know what Musk is planning, however he appears to deliver on things he says he will do.

Musk response, when asked if it would be possible to move toward greater levels of autonomy in the Tesla Model S without adding hardware:

Musk said:

There’s more that can be done, but the sensor suite is not the full autonomy suite. For full autonomy you’d obviously need 360 cameras, you’d probably need redundant forward cameras, you’d need redundant computer hardware, and like redundant motors and steering rack. For full autonomy you’d really want to have a more comprehensive sensor suite and computer systems that are fail proof.

That said, I don’t think you need LIDAR. I think you can do this all with passive optical and then with maybe one forward RADAR… if you are driving fast into rain or snow or dust. I think that completely solves it without the use of LIDAR. I’m not a big fan of LIDAR, I don’t think it makes sense in this context.

We do use LIDAR for our dragon spacecraft when docking with the space station. And I think it makes sense [in that case] and we’ve put a lot of effort into developing that. So it’s not that I don’t like LIDAR in general, I just don’t think it makes sense in a car context. I think it’s unnecessary.

[AmitG]

walm said:

Remember we are talking about Moore's Law type products (batteries, sensors, etc...) not really the traditional 7-year model cycle of regular cars.

Not sure about this bit. I think a lot of people are mistakenly applying Moore's Law type thinking to EVs. Mobile phones and computers get twice as fast, twice as small and twice as efficient every couple of years, and have very fast product cycles, so the same thing is going to happen to EVs right?

Personally I don't think it will. There will be continuous improvement in batteries, but we aren't going to see massive changes unless there is a radical breakthrough in battery design on the horizon, which can be made suitable for automotive use. And Toyota - who are not infallible, but who presumably know a thing or two about this stuff - reckon there isn't.

(I know that Moore's Law is strictly about transistor count not performance, but it seems to have become best known as a performance yardstick...)

[98elise]

The Vambo said:

98elise said:

EricE said:

I don't think any production Model S on the market at the moment has the hardware required for fully autonomous driving. I'm not an engineer but they'll at least need some type of LIDAR sensor for that. I suspect it will be an option for the Model 3, maybe even a retrofit option announced at a later date.

How do you drive without LIDAR? An autonomous car just needs to be able to sense as well as you or I

I don't know what Musk is planning, however he appears to deliver on things he says he will do.

Musk response, when asked if it would be possible to move toward greater levels of autonomy in the Tesla Model S without adding hardware:

Musk said:

There’s more that can be done, but the sensor suite is not the full autonomy suite. For full autonomy you’d obviously need 360 cameras, you’d probably need redundant forward cameras, you’d need redundant computer hardware, and like redundant motors and steering rack. For full autonomy you’d really want to have a more comprehensive sensor suite and computer systems that are fail proof.

That said, I don’t think you need LIDAR. I think you can do this all with passive optical and then with maybe one forward RADAR… if you are driving fast into rain or snow or dust. I think that completely solves it without the use of LIDAR. I’m not a big fan of LIDAR, I don’t think it makes sense in this context.

We do use LIDAR for our dragon spacecraft when docking with the space station. And I think it makes sense [in that case] and we’ve put a lot of effort into developing that. So it’s not that I don’t like LIDAR in general, I just don’t think it makes sense in a car context. I think it’s unnecessary.

That makes semse. We are able to drive with passive 180 degree optical (at best), with limited rear view. We don't have LIDAR or RADAR and are easily distracted. We also get tired and emotional.

A car just needs to be a bit better than we are before it becomes acceptable.

[Tuna]

98elise said:

That makes semse. We are able to drive with passive 180 degree optical (at best), with limited rear view. We don't have LIDAR or RADAR and are easily distracted. We also get tired and emotional.

A car just needs to be a bit better than we are before it becomes acceptable.

We are also confused by optical illusions and have a few million years development of optical processing over a car. It's certain that computer vision will catch up, but then you have to make safe and sensible decisions based on 'fuzzy' input. Just from a legal perspective, a car has got to be utterly fail-safe before it will be let on the roads. The issue here is as much "can you prove it's reliable?" as "can you do full autonomy through computer vision?"

From a software development perspective, anything with real-world sensors is a nightmare to test. How can you be sure you've covered all likely scenarios correctly when nearly every junction on every road is subtly different from every other one?

We kill a few thousand people a year in cars. Just one autonomous car killing someone will cause a huge outcry. A bit better doesn't cut it.

[Tuna]

AmitG said:

walm said:

Remember we are talking about Moore's Law type products (batteries, sensors, etc...) not really the traditional 7-year model cycle of regular cars.

Not sure about this bit. I think a lot of people are mistakenly applying Moore's Law type thinking to EVs. Mobile phones and computers get twice as fast, twice as small and twice as efficient every couple of years, and have very fast product cycles, so the same thing is going to happen to EVs right?

Moore's law says twice as fast every couple of years. It's taken 20 years for rechargables to double their energy density, despite very heavy investment in the technology by Apple and every laptop and phone developer in the world. The two problems are that energy density has a hard physical limit of less than 4 times the current materials, and that even at current densities, the failure modes are quite dangerous.

Similarly for optical sensors, camera manufacturers are already hitting limits set by hard physics - the number of photons you need to hit a site to reliably detect light from a scene. As they're based on similar technologies to chips, optical sensors have benefited from much the same progress as Moore's law, and just as with current chips, that law is 'running out' as atomic level scales are reached.

There's no place for motors to go - their efficiency is very close to the limits already.

[RobDickinson]

Tuna said:

There's no place for motors to go - their efficiency is very close to the limits already.

By your own post we can store up to 4 times more power. So thats a P90D that can do 1200 miles on a charge or have 1/4 the battery size/weight. Or a 800 mile model 3.

[Tuna]

RobDickinson said:

By your own post we can store up to 4 times more power. So thats a P90D that can do 1200 miles on a charge or have 1/4 the battery size/weight. Or a 800 mile model 3.

Yes, that's the limit we could get to - pretty good if we did. If batteries followed Moore's law, we'd be there in 4 years.

Trouble is that unlike chip processes, it's not just a question of improving what we already have. It's a question of discovering entirely new battery chemistries and materials, and then finding a way to productionise them without them costing the earth of being too delicate. If you read the engineering press, hardly a week goes by without a University announcing they've discovered a new nano-material or chemistry that would allow faster charging, higher density batteries. Not one of them yet has got out of the lab, and it speaks volumes that Tesla have bet big on the current LiIon process.

[RobDickinson]

Batteries plainly dont follow Moore's law, so why bring it up?

Moore's law isnt even a real law just a guestimate that we are falling behind on with silicon now. Its a rough rule of thumb applicable to an entirely different technology.

We can build useful BEV's now. With 200-300 mile ranges that charge up in 30-60min. Those are good enough for the large majority of the worlds daily drives.

Cost (of battery packs) continues to drop, with the gigafactory breaking the $190/kwh level and likely to push $150 for the model 3.


Interestingly our (NZ) government has just announced some cool EV initiatives. NZ as a whole generates 80% of our power from renewables ( hydro , thermal and wind) and we have enough spare capacity to power all our vehicle miles.

[98elise]

Tuna said:

98elise said:

That makes semse. We are able to drive with passive 180 degree optical (at best), with limited rear view. We don't have LIDAR or RADAR and are easily distracted. We also get tired and emotional.

A car just needs to be a bit better than we are before it becomes acceptable.

We are also confused by optical illusions and have a few million years development of optical processing over a car. It's certain that computer vision will catch up, but then you have to make safe and sensible decisions based on 'fuzzy' input. Just from a legal perspective, a car has got to be utterly fail-safe before it will be let on the roads. The issue here is as much "can you prove it's reliable?" as "can you do full autonomy through computer vision?"

From a software development perspective, anything with real-world sensors is a nightmare to test. How can you be sure you've covered all likely scenarios correctly when nearly every junction on every road is subtly different from every other one?

We kill a few thousand people a year in cars. Just one autonomous car killing someone will cause a huge outcry. A bit better doesn't cut it.

It doesn't need to be uterly fail safe, nothing is. It needs to be better than a human then its a safter option. We accept a level of casualties already. We could stop pretty much all road deaths if we all drove at 10mph, but we know that just not practical.

There will be a period when autonomous cars will still need a driver ready to take over, but I don't think it will be long after that when a steering wheel becomes optional.

[98elise]

RobDickinson said:

Batteries plainly dont follow Moore's law, so why bring it up?

Moore's law isnt even a real law just a guestimate that we are falling behind on with silicon now. Its a rough rule of thumb applicable to an entirely different technology.

We can build useful BEV's now. With 200-300 mile ranges that charge up in 30-60min. Those are good enough for the large majority of the worlds daily drives.

Cost (of battery packs) continues to drop, with the gigafactory breaking the $190/kwh level and likely to push $150 for the model 3.


Interestingly our (NZ) government has just announced some cool EV initiatives. NZ as a whole generates 80% of our power from renewables ( hydro , thermal and wind) and we have enough spare capacity to power all our vehicle miles.

GM say the Bolt is $145/kWh and I thought Tesla had said it will be in the $130/kWh range for the Model 3. I know they are aiming for $100/kWh by 2020.

[Blaster72]

I wonder if the battery range on the Model 3 will be improved by the time it actually hits the roads.

I see the BMW i3 is getting upgraded Samsung batteries which in REX version will be pretty much all most people need.

Range for normal i3 (at best) is going up to 123 miles and the REX a max of over 200 miles without a stop.

If things keep improving like this we could see what I really would like to see before I look at buying a normal price EV - 300-400 mile range.

I know the Model S is pretty good but it's also flipping expensive.

If the Model 3 could get 300 miles from a charge even in winter that would be a pretty attractive proposition I think.

http://www.autocar.co.uk/car-news/new-cars/bmw-i3-...

[WestyCarl]

I think for 90% of people the current Model S range of 230+ miles is more than adequate.

What would be more useful is to increase the number of super changers around.

[babatunde]

Oh and for those who think Tesla's grand plans are pie in the sky I give you "Faraday Future" and their $1billion car factory that will be building..... well they will get to that in a bit, in the meantime:

The building will cover 3 million square foot on 900-acre land of which FF recently finalized the purchase. The automaker expects the development of the ambitious project will create approximately 4,500 direct jobs over a 10-year period......

complete with babble speech:

The facility’s exterior will be a clear calling card of the FF look and feel ....... The interior will offer ergonomically and economically adaptable workspaces in open, color-coordinated environments throughout that integrate with the overall FF culture.


Billion US dollars you say, oh yes now why aren't they building that in Ireland

[Blaster72]

WestyCarl said:

I think for 90% of people the current Model S range of 230+ miles is more than adequate.

What would be more useful is to increase the number of super changers around.

Longer range will surely attract many more people, those who don't have parking at their property for example. Being able to go 300 miles and just charge when parked up doing the weekly shop would be very appealing for many households.

I personally would love a range of 300 miles at worst but no EV can do that yet. Let's see what Tesla can do between now and Uk Model 3 launch.

[RobDickinson]

98elise said:

GM say the Bolt is $145/kWh and I thought Tesla had said it will be in the $130/kWh range for the Model 3. I know they are aiming for $100/kWh by 2020.

http://m.teslacentral.com/tesla-q1-2016-earnings-transcript

On battery costs

Colin Langan, UBS: If I could ask a follow-up. Obviously, cost is going to be an important factor when 3 launches. I think you've indicated that your body costs per pack are now under $190 per kilowatt hour. How do you think that compares to the industry? Where do you think it will be by the time the Model 3 is launching?

Elon Musk: We try not to comment on individual component costs; that stuff's fairly proprietary, it's kind of like giving away our playbook. I think it's pretty obvious that we will exceed anyone else in scale of economies with the Gigafactory. We're very confident in Panasonic's ability to execute on that front. I just don't know anyone who in terms of intrisic costs is going to be close to what the Gigafactory can produce on a cost-per-kilowatt-hour basis.


Honestly I've no idea how Chevy will get $145 when Tesla are ahead in scale , should be about 50pc of the entire world production.

Either way costs are heading down way faster than predicted

[Blaster72]

I hadn't realised quite how much Panasonic were funding the Tesla Gigafactory. I thought it was all Tesla.

Scary numbers $$$$

http://www.wsj.com/articles/panasonic-will-bet-big...

[TransverseTight]

Regarding autonomous surely IR is the area to focus. Surroundings aren't usually hot, and therefore stay still. Things that are warm, like people and cars stick out like a sore thumb in IR. If you also get movement you know there's something to avoid. Visual is there to find the road, but even in IR tyres leave a heat print on the road... You see it even in the rain when watching police, camera action where the helicopter is chasing someone, Specifically noticeable when the run the camera in inverse... So hot is black and background is white.

The mobile eye system in the i3 is a bit basic, it's only VGA resolution, but it works, sort of. It's Achilles heel is early morning and late evening sun, just as it clears the horizon. It's will disengage when there's sun shining directly into the lens. What I I was trying to work out how with a single lens can they equate distance. Must use some clever geometry algorithms to work out angle to the vehicle in front in the vehicle plane. Eg. - it knows it's height above the road based on where it's fitted in the car, it knows the angle to the base of the car in front based on which pixels are detecting it - so it can derive the linear distance from the camera using trigonometry.

[TransverseTight]

Mr GrimNasty said:

"He [Musk]
As I've repeatedly said, subsidy sucking showman fail, forgive the pun, we are watching a slow car crash.

Lots of big 'renewable' type companies are hitting the end of the road. "No commercial value" in the technology. The one-time largest PV manufacturing company is about to go under, amongst others.

The EU has spunked 1Trillion of our money on green subsidies and CO2 emissions have gone UP. That's what happens when political interference imposes a poor engineering solution. The free market would have exploited fracking and gas in preference to coal and oil and useless windmills and PV, and CO2 emissions would have come down dramatically

When you keep making statements like this you need to show where you are getting your numbers.
Just googled for EU CO2 emissions and came up with this...

http://ec.europa.eu/eurostat/statistics-explained/...

Definitely looks a downward trend there. As usual I'm happy to use newer more accurate figures if you have some.

Who's this big PV company? My own figures tell me PV is worth me investing in. Except that my roof faces the wrong way so I haven't. Not interested in investing in funds, I want to invest directly Note - even with no subsidy its got a return. I'm looking at building a car port instead but it needs planning permission and I haven't got costs for the build yet. Might not work out cost effective as the build will probably cost more than the PV!

Edit to add €1trillion sounds loads, but over what timeframe? That's about €2,000 per person based on EU population of 500 million. We as households spend just over €1trillion per year on buying, and running "personal transport". I.e. Mainly cars. http://ec.europa.eu/transport/facts-fundings/stati... (Page 27).

Edited by TransverseTight on Saturday 7th May 01:56