Lets work out the Tesla Semi-Truck Technical Specs

[JKolodziejski]

Not might. Definitely thermal limit. I'm surprised that it can do THAT much. I think that is due to air intake cross section area
being many times larger than on Model S (JKolodziejski, can you estimate areas for air intake between Model S and BAMF?). Plus 2 more motors and all of them more efficient.

I'll see what I can do

 

[JKolodziejski]

Obviously the truck has relatively massive cooling to charge a 1MWh battery fast.

I'd assume 1000 kWh for the 500 mile version. 750 kWh for the 300 mile version.

Little frustrating to see that so many are still assuming the truck has as much as one megawatt hour. For cost reasons the Semi isn't likely to use anything other than 2170 cells (in production at least), and I am sure it's not physically possible to have more than 800-850kWh from the geometrical analysis I did earlier in the thread. Have a look at it.

Can send Solidworks files if anyone cares, but this site doesn't allow ".sldprt"

 

[mongo]

Little frustrating to see that so many are still assuming the truck has as much as one megawatt hour. For cost reasons the Semi isn't likely to use anything other than 2170 cells (in production at least), and I am sure it's not physically possible to have more than 800-850kWh from the geometrical analysis I did earlier in the thread. Have a look at it.

Can send Solidworks files if anyone cares, but this site doesn't allow ".sldprt"

Like the write up you did. Don't get frustrated (easy to say), the best we can do is analyze the data we have available and present our data, assumptions, calculations, and conclusions. Time will tell if they were right, not popularity. A little annoying that Tesla knows already...

I'm in the 4x200 kWh group with you. 400*1.5= 600kWh charge in 30 minutes, 300kW per pack, 1.5C.

I'm wondering if the open pack structure in the center of the group is for cooling. Flat bottom of tractor could be planar radiator. Phase change intrapack heat transfer? Or are they going all out with oil flooded packs?

 

[dhanson865]

I read that too but can't make sense of it. Model S has 9.73:1. (what could it be on M3?) Given the 50% bigger wheels on a semi, the motor speed would be same as Model S at approximately 15:1. One would think the motors would need to spin much faster in the truck than in a Model S to give enough wheel torque. What am I missing?

2 or 4 times the motors = 2 or 4 x the torque. Surely 2-4x the torque is enough?

 

[Farmertom]

2 or 4 times the motors = 2 or 4 x the torque. Surely 2-4x the torque is enough?

Your moving 28x the load (dragging over half of it) with almost twice the diameter wheels. a typical highway truck is putting over a thousand ft/lbs torque to each wheel under full power in high gear, the torque will multiply as he gears down on hills and such.

 

[Farmertom]

JKolodziejski , that is quite the analysis you made up there, impressive.

Is that right you calculate there are twelve times the number of cells in the truck vs the Model S?

 

[JKolodziejski]

I'm wondering if the open pack structure in the center of the group is for cooling. Flat bottom of tractor could be planar radiator. Phase change intrapack heat transfer? Or are they going all out with oil flooded packs?

I thought those were black spaces at first but it looks like it could have been dark coloured/non-graphic'd enclosures for the cooling.

Note that the truck graphic doesn't have as much detail as Kman's "ninja" video earlier in the thread if i recall, of the motorised back section, similar to the usual drivetrain graphics on the Tesla website.

To be precise with references:

The "rectilinear C-shapes", enclosing what I assume now to be three-layer modules, for want of better terms do make sense as conductor-convectors - and they do not cover the truck sides in the driver's left-right direction.

So perhaps the packs can be slid out sideways and connected from the left-right sides as they are snug against the sides. Would make sense with manufacturing lines.

 

[mongo]

I thought those were black spaces at first but it looks like it could have been dark coloured/non-graphic'd enclosures for the cooling.

Note that the truck graphic doesn't have as much detail as Kman's "ninja" video earlier in the thread if i recall, of the motorised back section, similar to the usual drivetrain graphics on the Tesla website.

To be precise with references:

View attachment 261515


The "rectilinear C-shapes", enclosing what I assume now to be three-layer modules, for want of better terms do make sense as conductor-convectors - and they do not cover the truck sides in the driver's left-right direction.

So perhaps the packs can be slid out sideways and connected from the left-right sides as they are snug against the sides. Would make sense with manufacturing lines.

Side connections would fit in with charger plug.
Voids could be wire routing to allow packs to all be the same vs rotated.

 

[JKolodziejski]

JKolodziejski , that is quite the analysis you made up there, impressive.

Is that right you calculate there are twelve times the number of cells in the truck vs the Model S?

Cheers Farmertom.

To answer, depends on which Model S model you're talking about.

I'd say a factor of 10 explains everything reasonably well - think an 80kWh historical average with 110kW-135kW Supercharging, as compared to 800kWh with 1200kW Megacharging, 1.5C as Mongo earlier said.

Since each module is smaller in it's dimensions than a Model 3 or S pack judging from Semi's cab floor plan, I would be heavily tempted to assume it would be 50kWh which is nice maths for multiples, but the cooling system may have some depth making the layer's planar dimensions (width and length) representative of slightly more. I can't imagine the layers having less than 63 kWh as then it approaches 1.5 kWh/mile for 500 miles, which I'm not too comfortable with.

 

[scaesare]

Adding cooling capacity won't help much as rotor doesn't touch cooling jacket. And coolant passthrough has not been confirmed (patented).

Teardowns have revealed that the Model S rotor is indeed liquid cooled.

 

[Farmertom]

Cheers Farmertom.

To answer, depends on which Model S model you're talking about.

I'd say a factor of 10 explains everything reasonably well - think an 80kWh historical average with 110kW-135kW Supercharging, as compared to 800kWh with 1200kW Megacharging, 1.5C as Mongo earlier said.

Since each module is smaller in it's dimensions than a Model 3 or S pack judging from Semi's cab floor plan, I would be heavily tempted to assume it would be 50kWh which is nice maths for multiples, but the cooling system may have some depth making the layer's planar dimensions (width and length) representative of slightly more. I can't imagine the layers having less than 63 kWh as then it approaches 1.5 kWh/mile for 500 miles, which I'm not too comfortable with.

Thanks, as you may have read I'm trying to determine an approximate tare weight for this unit which it appears may be over 25,000lb based on a battery weight for the model S of around 1200lb. I'm sure this would work for Walmart as I believe most of their loads are quite light but trying to use these units for fixed weight loads based on the average truck could render them uncompetitive.

I wonder what the efficiency of these powertrains is. External cooling is lost power.

 

[Olle]

Thanks, as you may have read I'm trying to determine an approximate tare weight for this unit which it appears may be over 25,000lb based on a battery weight for the model S of around 1200lb. I'm sure this would work for Walmart as I believe most of their loads are quite light but trying to use these units for fixed weight loads based on the average truck could render them uncompetitive.

I wonder what the efficiency of these powertrains is. External cooling is lost power.

Model S battery pack weight cannot simply be multiplied for capacity to reach the weight for this pack. For example, Model S has a 6 mm shield covering the entire underside of battery case, in fact, the case weights almost as much as the cells. You won't need this between each layer.
One way to approximate the weight is the 0-60 acceleration, 5 sec alone and 20 sec with 80,000 lb GVW. This equation gives you a tractor weight of 20,000 lb.

 

[electracity]

This equation gives you a tractor weight of 20,000 lb.

One article said the prototypes are carbon fiber but the construction of the final vehicles is undecided.

 

[Farmertom]

Model S battery pack weight cannot simply be multiplied for capacity to reach the weight for this pack. For example, Model S has a 6 mm shield covering the entire underside of battery case, in fact, the case weights almost as much as the cells. You won't need this between each layer.

Although you may not need quite as much support for each layer your still going to need substantial strength to hold all those cells in place, you have to remember your also dealing with a rougher ride and far more hours on the road.

One way to approximate the weight is the 0-60 acceleration, 5 sec alone and 20 sec with 80,000 lb GVW. This equation gives you a tractor weight of 20,000 lb.

That's true... didn't think of that. Factor in exaggerations... Might be close.

 

[mongo]

Although you may not need quite as much support for each layer your still going to need substantial strength to hold all those cells in place, you have to remember your also dealing with a rougher ride and far more hours on the road.



That's true... didn't think of that. Factor in exaggerations... Might be close.

Vertical strength goes up as the square of height, so a pack that holds a triple stack has ~9 times the strength with 3 times the load with same sidewalls. (all else being ignored).

 

[JKolodziejski]

Thanks, as you may have read I'm trying to determine an approximate tare weight for this unit which it appears may be over 25,000lb based on a battery weight for the model S of around 1200lb. I'm sure this would work for Walmart as I believe most of their loads are quite light but trying to use these units for fixed weight loads based on the average truck could render them uncompetitive.

I wonder what the efficiency of these powertrains is. External cooling is lost power.

Looking at typical pack energy densities:

• 85kWh Model S pack => 140+ Wh/kg

• Model 3 LR pack => 150+ Wh/kg

• 100kWh Model S pack => 170 Wh/kg?

• Alta Motors (dirtbike manufacturer) pack => 180 Wh/kg ( Alta Claims Its Redshift MX Has Battery With Highest Energy Density, Better Than Tesla Model S & X )

• Bosch lithium-metal trial => 263 Wh/kg ( Bosch Targets 50 kWh Battery That Weighs Only 190 Kilograms, 15-Minute Charge To 75% )

• LG Chem NCM 811 => 410Wh/kg, 963 Wh/L (http://pushevs.com/2017/10/03/battery-technology-whats-coming-soon/) Note the rather direct "will produce"... very nice.

I would assume 190-200 Wh/kg of 2170 cells with today's tech for such a large, generally slower accelerating vehicle, and perhaps 200-220 Wh/kg by mid-2019 Elon Time (2020) so 4.8-5.2 kg/kWh would be realistic with low cost 2170s.

I'll assume 4.8kg/kWh to 5.2kg/kWh for simplicity's sake with my pet favourite absolute capacity of 4 modules x 3 layers of 68kWh = 4x 204kWh = 816kWh = 800kWh @ 98% usable fraction (Model 3 and S are 97% and 96% respectively).

That means a battery mass of 816 x 4.8 to 816 x 5.2 = 3,920-4,240 kg or 8,620-9,330 lbs.

Assuming the battery mass fraction of the truck is closest to 35-45% of overall mass (original Roadster has 38%, and this is a tractor unit after all...)

Heaviest end of range - 4240/0.35 => 12,000 kg or 26,600 lb.

Lightest end of range - 3,920/0.45 => 8,700 kg or 19,200 lb.

I think it's roughly 4100kg at 40% of mass:

Roughest middle - 4100/0.40 => 10,000 kg or 22,000lb.

A bit heavier than the average US semi? Sounds nice and boring to me.

 

[scaesare]

, in fact, the case weights almost as much as the cells.
One way to approximate the weight is the 0-60 acceleration, 5 sec alone and 20 sec with 80,000 lb GVW. This equation gives you a tractor weight of 20,000 lb.

In the initial 7,104 cell 18650 based packs, the case accounted for only about 25% of the mass. With the newer 8,256 cell packs I assume the case mass accounts for an even slightly lesser percentage.

 

[Farmertom]

Looking at typical pack energy densities:

• 85kWh Model S pack => 140+ Wh/kg

• Model 3 LR pack => 150+ Wh/kg

• 100kWh Model S pack => 170 Wh/kg?

• Alta Motors (dirtbike manufacturer) pack => 180 Wh/kg ( Alta Claims Its Redshift MX Has Battery With Highest Energy Density, Better Than Tesla Model S & X )

• Bosch lithium-metal trial => 263 Wh/kg ( Bosch Targets 50 kWh Battery That Weighs Only 190 Kilograms, 15-Minute Charge To 75% )

• LG Chem NCM 811 => 410Wh/kg, 963 Wh/L (Battery technology: what's coming soon? - Push EVs) Note the rather direct "will produce"... very nice.

I would assume 190-200 Wh/kg of 2170 cells with today's tech for such a large, generally slower accelerating vehicle, and perhaps 200-220 Wh/kg by mid-2019 Elon Time (2020) so 4.8-5.2 kg/kWh would be realistic with low cost 2170s.

I'll assume 4.8kg/kWh to 5.2kg/kWh for simplicity's sake with my pet favourite absolute capacity of 4 modules x 3 layers of 68kWh = 4x 204kWh = 816kWh = 800kWh @ 98% usable fraction (Model 3 and S are 97% and 96% respectively).

That means a battery mass of 816 x 4.8 to 816 x 5.2 = 3,920-4,240 kg or 8,620-9,330 lbs.

Assuming the battery mass fraction of the truck is closest to 35-45% of overall mass (original Roadster has 38%, and this is a tractor unit after all...)

Heaviest end of range - 4240/0.35 => 12,000 kg or 26,600 lb.

Lightest end of range - 3,920/0.45 => 8,700 kg or 19,200 lb.

I think it's roughly 4100kg at 40% of mass:

Roughest middle - 4100/0.40 => 10,000 kg or 22,000lb.

A bit heavier than the average US semi? Sounds nice and boring to me.

LOL, thank you. In trucking weight is money until you reach cubic capacity, then you figure out how to increase that. Yes the average truck tractor weighs in at around 17,000 to 18,000lbs these days. Some loads a specifically designed for a typical truck/trailer tare weight, if you can't meet that they won't load you.

 

[JKolodziejski]

LOL, thank you. In trucking weight is money until you reach cubic capacity, then you figure out how to increase that. Yes the average truck tractor weighs in at around 17,000 to 18,000lbs these days. Some loads a specifically designed for a typical truck/trailer tare weight, if you can't meet that they won't load you.

Thing is, if the loads in question are divisible, like foodstuffs, appliances, materials, (and not a product designed to absolute weight limit, like a turbine genset optimised to the lowest unladen weights) there's an easy way around that issue.

That kind of optimisation is to minimise cost assuming all trucks are of a certain reasonable average of efficiency, aero resistance and considering the national weight limit(s).

Since the Tesla Semi is significantly more efficient (especially considering regeneration down hills), it's entirely possible they could specify the slightly higher unladen weight, use slightly more Semis than conventional trucks, and the running costs, at least, would still be much less anyway.

Or they could simply specify less batteries, given that the Semi page does after all say "300-500 miles" range. Also, repetitive loads (e.g. for supermarkets) tend to travel short distances from dockyards & factory areas for, again, yes, costing reasons.

It's just my opinion, but I think the difference in the comprehensive running costs, from a business perspective, is more than substantial enough to account for the smaller relative inadequacies of range and weight, which means that it is upfront cost that really matters in the marketability of the Semi.

 

[jkn]

Motor requirements have to be somewhat higher, because the truck must be torque limited for some portion of the acceleration run (impossible to deliver maximum power in the first second since it would require infinite torque.) With that caveat, your approach looks good.

It's interesting that you aren't getting closer agreement between the 0-60 and hill climbing power requirements - it sounds like Tesla may have understated the hill climb capability, possibly in interest of round numbers.

I also noticed limited torque. arnis made calculations with 28.6m/s = 60mph. I calculated with 26.8 m/s and constant acceleration from 0 to 12 mph (1.49 s), constant power 700 kW 12 to 60 mph (17.9 s) -> 700 kW is enough. I ignored air resistance, but it is small compared to acceleration force.