Supercharger for the Roadster (Elon said "No")

[Doug_G]

All the reports I can find for hard driving say the PEM cooling is more of an issue than the battery cooling for the existing roadster.
Walter

Actually, motor cooling is the big issue for racetrack use. Most likely the rotor windings get hot and there's no easy way to get the heat out of there since it's all on a spindle.

One thing you notice right away about the Roadster - the cooling system makes a LOT of noise when you are charging at high current. That HVAC system is really working hard to cool the battery. The Model S whispers in comparison.

 

[markwj]

I really doubt that they are making any design changes to the pack, aside from putting in different cells. In fact, I would not be the slightest bit surprised to hear that these will all be refurbished packs. You trade in your existing pack, they put in a pack that has already been refurbished with the new cells. Then they refurbish your pack for someone else's upgrade. That would allow them to offer the replacement pack at a much better price. I suspect if you were offered a refurbished pack for $20k or a brand-new pack for $40k that you'd be really happy with the refurb.

Yes. I agree. The upgrade, as outlined, should be able to be done purely at the sheet level (of course, with firmware changes). No changes to the inlet, internal wiring, cooling, charger wiring harness, etc.

I started to list out all the items that would be needed to be changed/added to add supercharging, but gave up after 20 bullet points. It is not just a new socket and cable. You've got high current electrical paths from the socket to a new supercharging module, new module to pack, pack/module to PEM, socket to PEM. Then, a whole new ECU to handle the supercharger protocol and bridge the CAN buses (no way the supercharger protocol CAN bus is going to be just shoved onto one of the existing buses). New DC isolators. Firmware changes in the BMS, VMS and VDS. Then, wiring harness changes for that new ECU. Cooling. Documentation in Service Guides. Certification. Engineering training. Oh, and remember that the European/Asian car-side socket is different than that used in North America (including handling the Model S three phase kludge that entails - which I have no idea how it could be handled in the roadster PEM without having to have different wall chargers for roadster vs model S).

SuperCharging on the roadster would certainly be nice to have - I just think that too much would need to be changed to be able to deliver it at a reasonable price. When the price is announced, I doubt whether Tesla will even sell 50 of these 3.0 upgrades in the first year, and at those kinds of quantities the engineering payback is just not enough.

 

[pharma5]

I'll support your effort by building you a Model S inlet capable of 90kW to mount somewhere on the Roadster.

Go Henry Go

It's totally doable and I think they should just do it. Offering a supercharging upgrade package makes a lot more sense than an aero bodykit (it's bad form to change the look a discontinued and rare car), and as the supercharging network expands, it makes more sense than even offering 400 miles range in a sports car.

+1 to all of this, and to Supercharging for Roadster. Every user is different, but for me it's very valuable to get access to the world's best EV fast-charging network, which has become (and will continue to be) a defining characteristic of the Tesla brand going forward.

If Tesla can do it, that's great. If they can't justify the resources, I hope they make good on their word to make certain of their intellectual property available to folks who want to use it in good faith. (How can you get more 'good faith' than a bunch of Roadster owners who want to supercharge on Tesla's network? If we need to give a $2k fee when it's ready - Done!).

I'm in the wrong field to contribute much technically, but happy to look at any legal agreements with experienced, non-lawyer eyes, or just to egg you guys on as a fan.:biggrin:

 

[TurboFroggy]

The Model S whispers in comparison.

Eh not entirely true. This summer we were supercharging in Corning, it was 115'F out and the fans/compressor in our Model S was absolutely roaring.

 

[pharma5]

Btw, my opinion on the desirability of supercharging is independent of the 3.0 upgrade, which is a good thing in and of its own right.

I am hopeful that the 3.0 will prove to be a very economical package, with items available a la carte, and perhaps avoiding time/cost problems of a full big-battery-for-range moonshot.

 

[Juanmax]

I drove a Leaf in 2011 non-stop 500km, by driving at ~100kph and Chademo charging after every hour approx. After the forth charge, the Battery was so hot that while driving the power was reduced. The chargers were doing 50kW (reducing the value as they approached 80% SoC).

The Kia Soul EV can charge up to 100kW, and only has an air cooled pack.

Some quick estimations:
The Roadster battery pack weight is about 450Kg
If I assume it to have a similar heat capacity to water (-> 0.45m³)
Volumetric heat capacity of water (steel is 950) of 1000 kcal/m³ °C
1kWh = 860kCal

Assuming a SC charging session of one hour@50kW will make our current battery full. And having Battery charging efficiency of 95% (which means losses due to internal resistance 5%). Makes a total of 2.5kW for one hour, so 2.5kWh.

With this amount of Energy, the pack will increase 2.5[kwh]*860[kCal/kWh] / (1000 [kcal/m³ °C]*0.45[m³]) = 4.7°C

This is assuming no cooling whatsoever. If you think that a normal AC Compressor can provide cooling for 1kW constantly (other assumption, I do not know what the Roadster can), then you can take ~2°C out of the final value.

Did I make any mistake in my calculations ? Of course they contain a lot of assumptions, but the real numbers should be withing +-30% I guess.

Take into consideration, that when charging on an HPC the cooling system has to cope with the heat generated in the PEM too, which is not the case of SC, since it bypasses it.

 

[Doug_G]

If they can't justify the resources, I hope they make good on their word to make certain of their intellectual property available to folks who want to use it in good faith.

This is a common misconception. Tesla never promised to release any intellectual property other than the patents, which are already public information. I can pretty much guarantee they won't hand over schematics, protocols, or software for anything they make.

 

[Saghost]

Yes. I agree. The upgrade, as outlined, should be able to be done purely at the sheet level (of course, with firmware changes). No changes to the inlet, internal wiring, cooling, charger wiring harness, etc.

I started to list out all the items that would be needed to be changed/added to add supercharging, but gave up after 20 bullet points. It is not just a new socket and cable. You've got high current electrical paths from the socket to a new supercharging module, new module to pack, pack/module to PEM, socket to PEM. Then, a whole new ECU to handle the supercharger protocol and bridge the CAN buses (no way the supercharger protocol CAN bus is going to be just shoved onto one of the existing buses). New DC isolators. Firmware changes in the BMS, VMS and VDS. Then, wiring harness changes for that new ECU. Cooling. Documentation in Service Guides. Certification. Engineering training. Oh, and remember that the European/Asian car-side socket is different than that used in North America (including handling the Model S three phase kludge that entails - which I have no idea how it could be handled in the roadster PEM without having to have different wall chargers for roadster vs model S).

I was thinking most of this would be easily handled by a drop in replacement charger module (high current paths, contactors, SpC CANBus.)

But I was forgetting that the Roadster uses a single integrated AC Propulsion-style Power Electronic Module for both the drive inverters and charging - which means to get that result you'd have to replace/upgrade the whole PEM. Certainly doable, but a bigger and more expensive project than I was envisioning.
Walter

 

[Doug_G]

But I was forgetting that the Roadster uses a single integrated AC Propulsion-style Power Electronic Module for both the drive inverters and charging

That was only true for the 1.5 Roadsters. They stopped using the AC Propulsion patent for Roadster 2.0.

 

[Saghost]

That was only true for the 1.5 Roadsters. They stopped using the AC Propulsion patent for Roadster 2.0.

Are they still using a single PEM to handle both tasks? If not, I think a replacement port and charger module for Supercharging should be fairly straightforward - they could even steal most of the parts from existing Model S systems.

I've been reading blog posts to make sense of the car's system, but most of those are from the original development, so I'm pretty fuzzy on the electrical differences between the various versions.
Walter

 

[pharma5]

This is a common misconception. Tesla never promised to release any intellectual property other than the patents, which are already public information. I can pretty much guarantee they won't hand over schematics, protocols, or software for anything they make.

Not a misconception - that's why I said "certain of". Access to patents is what they offered, which is agreement not to exclude. (Patents are exclusionary rights on making, using selling. Yes, they are by definition published, public documents). Other stuff is open to debate - maybe you have some non-public info there.

To put it another way, if a bunch of capable Tesla owners want to do-it-themselves (maybe with a bit of coaching, short of real trade secrets or data assets), do you really think Tesla will stand in the way? I doubt it.

 

[Doug_G]

Are they still using a single PEM to handle both tasks? If not, I think a replacement port and charger module for Supercharging should be fairly straightforward - they could even steal most of the parts from existing Model S systems.

I've been reading blog posts to make sense of the car's system, but most of those are from the original development, so I'm pretty fuzzy on the electrical differences between the various versions.
Walter

Yes, the PEM is handling charging and driving. But it's not using the motor coils as part of the charger circuit.

 

[Juanmax]

"CHAdeMO is the trade name of a quick charging method for battery electric vehicles delivering up to 62.5 kW of high-voltage direct current via a special electrical connector."

Via: CHAdeMO - Wikipedia, the free encyclopedia

I've read that CHAdeMO was originally designed for 100-kW service, but only delivers 50-kW at stations.

You read right, it can do 100kW.

http://insideevs.com/kia-installs-first-100-kw-chademo-dc-fast-chargers-europe/

Take into consideration, that when charging on an HPC the cooling system has to cope with the heat generated in the PEM too, which is not the case of SC, since it bypasses it.

I fooled myself, the PEM is air-cooled, forget that.

 

[stevejust]

I doubt whether Tesla will even sell 50 of these 3.0 upgrades in the first year, and at those kinds of quantities the engineering payback is just not enough.

That could be true and it will be interesting to see. Of course, it'd be really cool if we could also see how many Roadster 4.0 supercharging upgrades they could sell vs. Roadster 3.0 cells, retrofit aerokit and LRR tires. And at what pricepoint they could sell each.

I understand now, better, the factors going into re-enigneering the Roadster for supercharging.

I understand it's not worth Telsa's time to re-engineer the Roadster for supercharging. My problem was that I thought we were talking about a much more extensive overhaul to the old battery pack than what I suppose is coming. Now it seems the general consensus is we're talking about taking out all the bricks, putting the new cells in the old bricks, and then putting the packs back in the car.

 

[markwj]

Are they still using a single PEM to handle both tasks? If not, I think a replacement port and charger module for Supercharging should be fairly straightforward - they could even steal most of the parts from existing Model S systems.

I've been reading blog posts to make sense of the car's system, but most of those are from the original development, so I'm pretty fuzzy on the electrical differences between the various versions.
Walter

Yes. The cabling goes port -> PEM, PEM <-> ESS, and PEM <-> Motor. The PEM handles everything AC<->DC conversion related. Charging, driving and regen.

So, if you want to add supercharging without changing the PEM, it means a new separate module and a large number of changes to the high power cabling.

It could be done cleaner with a new PEM, but that is a major undertaking. The PEM is a massive piece of high power electronics.

Here is some bedside reading: Roadster Technology - Power Control | Tesla Motors

and some pictures from here: #34 Salvage auction - Page 11

View attachment 44154
View attachment 44149
View attachment 44151
View attachment 44150

- - - Updated - - -

I understand now, better, the factors going into re-enigneering the Roadster for supercharging.

I understand it's not worth Telsa's time to re-engineer the Roadster for supercharging. My problem was that I thought we were talking about a much more extensive overhaul to the old battery pack than what I suppose is coming. Now it seems the general consensus is we're talking about taking out all the bricks, putting the new cells in the old bricks, and then putting the packs back in the car.

I think it is even simpler than that. The bricks are assembled into sheets (11 of which slot into a pack). These sheets are metal encased self-contained modules. The capacity upgrade could be as 'simple' as sliding out the 11 sheets, and sliding in 11 refurbished replacements. Plus, firmware upgrades to deal with the new cells of course (both on the BMS control microprocessors for the sheets, as well as VMS). Building bricks (and sheets) with new cells is just engineering. The hard work will be all the firmware tuning to match the new cell chemistry.

 

[bonnie]

I think it is even simpler than that. The bricks are assembled into sheets (11 of which slot into a pack). These sheets are metal encased self-contained modules. The capacity upgrade could be as 'simple' as sliding out the 11 sheets, and sliding in 11 refurbished replacements. Plus, firmware upgrades to deal with the new cells of course (both on the BMS control microprocessors for the sheets, as well as VMS). Building bricks (and sheets) with new cells is just engineering. The hard work will be all the firmware tuning to match the new cell chemistry.

+1, this has been my assumption, too. That would be the simplest approach.

 

[pharma5]

From Jan 2014 investor deck. Still feels compelling.
Want the whole brand on it, long-term.
Any car with that T on it should be able to charge wherever that T is located.
(Thanks, hcsharp, for bringing that a big step closer with Can SR).

 

[bxr140]

I'd bet that they could put 16kW DC from a Supercharger into a Roadster via DC directly into the existing PEM. After all the inverter chops at a higher frequency; they don't need AC input. But I doubt they'll offer that because it will tie up Supercharger stalls with Roadsters that take hours to charge.

I'm certainly not smart enough to determine the feasibility of making Roadster compatible with a supercharger, but as a model S owner I wouldn't have a problem with roadsters tying up a SC stall for a few hours if the charge solution could only utilize a fraction of the supercharger's capability. It would actually be pretty cool to see a different kind of vehicle charging (probably good for marketing as well) and there are so few roadsters to begin with that the additional load on the network would practically be in the noise.

 

[tom66]

Charging a li-ion produces much more heat than discharging it. Just because the Roadster can discharge ~220kW peak and ~110kW continuously, does not mean it can charge as fast as supercharger rates, even with a better battery. Taking that heat out when stationary is difficult, especially given the Roadster is not designed to do it currently. So I do not see supercharging happening myself.

 

[dhrivnak]

I would be ecstatic if they just install a 70/80 amp HPWC. Now that the Roadster can use a Model S plug and I can envision times when a Model S could use them. I think that would be a lot cheaper than redesigning the Roadster electronics.