If I Ran The Zoo: Designing a charge connector for the future

[KarenRei]

So, I was thinking about the issue of all of the competing charging standards. CHAdeMO appears to be on its way to being eclipsed by CCS Combo, although it's probably not going down without a fight. Tesla has two separate charge connectors, a two wire connector for the US and a Type 2-based connector in Europe. CCS has the bully pulpit with more influence in European governments, but Tesla has more power with respect to mass deployment. Tesla is now in CharIN, meaning it will also have some influence over CCS.

When it comes to new charge connectors, there's always a push for backward compatibility. There's likewise always a push for higher powers - but cables aren't allowed to become bulky. Europe (rightly) loves 3-phase for AC, so Tesla's US connector is a non-starter. Tesla (rightly) loves simple connectors that are compact and not overly complicated. How to unify all of these things? Here's my attempt.

Backwards compatibility: The connector is based on the DC port of CCS combo, and matches it dimension for dimension. A car with this connector that also has the old Type 2 socket above for backwards compatibility can accept an old CCS charger (the backwards-compatibility port can be removed from future models once old CCS combos are phased out). A charger with this connector can still power unequipped cars, so long as it has a (clipped on) Type 2 plug as well (which can be removed once old vehicles are no longer on the road in significant numbers).

Simplicity: The cable is as small and simple as it could possibly realistically be, with no needless pins or bulk. It nonetheless allows for DC, AC single phase, and AC three phase (no ground like the Tesla US superchargers, although one could be added if desired). The extra pin for three phase ensures that the cable only fits in in one orientation (must be spring-loaded or actuated on the plug end to avoid getting in the way when connecting to old CCS combo ports). Data is handled optically.

Power: While the pin sizes stay the same, the addition of coolant channels allows the charger to run coolant down its cable - thus keeping cable temperatures down and allowing much more power to be run through the same size cable. If the car wishes to utilize the external coolant, it can route it through its battery pack. If the car wishes not to use coolant, it can directly loop back the inflowing coolant to the outflowing coolant so that it returns straight to the cable. Older chargers (or new ones with no coolant reservoir) need not provide coolant; like with charging, it is a parameter negotiated with the vehicle and limited by temperatures and supply. No attempt is made to provide coolant to vehicles equipped with old Combo ports.

What do you think? Just some idle thoughts I felt like sketching out.

 

[Xenoilphobe]

If I ran the zoo. They would have an SUV or a pickup and sell a million of them forcing the world to adopt the Tesla standard

 

[apacheguy]

IMO, the liquid cooling of the cable itself is a good idea. Tesla had one site that used this design concept before scrapping it. I don't understand why.

The ability to send coolant to the vehicle I am not as excited about. What if one vehicle had air bubbles in the coolant system and subsequently transferred it to other vehicles that connected to the same reservoir. Then all those cars would need to visit a service center to have their lines bleeded.

 

[KarenRei]

If I ran the zoo. They would have an SUV or a pickup and sell a million of them forcing the world to adopt the Tesla standard

When you say "the Tesla standard", which Tesla standard do you mean, US or European?

Tesla's US standard doesn't support three phase. There are good reasons to allow three phase AC (it's easy to supply at higher powers and in more remote locations than single phase, which is why it's more popular in Europe). On the other hand, Tesla's European connector, while more compact than CCS combo, is still a Frankenstein's monster, their attempt to keep their port as small as possible while still meeting European requirements (7 pins).

Also, realistically, whatever CharIN says is to be the next standard probably will go big. CCS has government support, even though Tesla has a big influence on-the-ground due to their volumes.

 

[KarenRei]

IMO, the liquid cooling of the cable itself is a good idea. Tesla had one site that used this design concept before scrapping it. I don't understand why. The ability to send coolant to the vehicle I am not as excited about. What if one vehicle had air bubbles in the coolant system and subsequently transferred it to other vehicles that connected to the same reservoir. Then all those cars would need to visit a service center to have their lines bleeded.

Ignoring that there's no reason that onboard and offboard coolant must be mingled (aka, the vehicle could air-purge its system before charging, and the charger could air-purge it afterwards before the vehicle restores its internal supply).... given that current Tesla vehicles' maintenance schedules doesn't call for a coolant swap for four years (aka, coolant "going bad" is not a common occurrence); given that glycol is easily filtered, and that the amount of coolant in a single vehicle is small compared to the total that would be in a reservoir; given that if there were any worries about coolant quality, there are lots of ways to analyze the chemical and mechanical properties of a fluid that cost far less than a six-figure supercharger; given all of that, I wouldn't have concerns. Externally supplied coolant has huge benefits beyond just cooling the cable, in that its volume is unlimited and it can be prechilled by an industrial chiller - not merely to ambient, but to as low of a temperature as is safe to charge at. Faster heat removal, unlimited by onboard hardware = faster charging. And external coolant were to take off, the mass and cost of onboard cooling systems could be reduced, making vehicles lighter and cheaper.

But even if there are initial concerns with external coolant, having the ability for the vehicle to decide whether to use an external coolant supply or not is "futureproofing". If the vehicle has no interest, the inflow ports simply need to route directly to the outflow ports.

 

[arnis]

3-phase requires L1, L2, L3, N, PE (full current, aka same size as else), PP (proximity pilot), CP (control pilot).
All together 5 full-size pins and 2 small. I see less pins on drawing.

I see some coolant cross-sections, with no walls around.
Also I see no way coolant can cool the high power pins as it is not touching them and there is a gap between. It's not the cable that overheats. It's the mating surfaces of two sides.

It's not possible to purge coolant with pressure. There will lots of it left due to uneven cross section in the piping.
Also I see no seals with that design. And no locking mechanism, for electrical and for coolant part.

Like Elon said: "Everything works in PowerPoint".

Also how about glycol sticking to the seals after drying up (and dust contamination during drying). They must be manually washed before next plug-in cycle. Or there must be pre-seals or shutter to avoid that.
Eager to see another design.

 

[KarenRei]

3-phase requires L1, L2, L3, N, PE (full current, aka same size as else), PP (proximity pilot), CP (control pilot).
All together 5 full-size pins and 2 small. I see less pins on drawing.

Three phase does not in any way, shape or form require those things. Three phase needs nothing more than three lines. Now, as noted in the original post, a ground could be added opposite the third pin if not supplied elsewhere**, for safety. Concerning Type 2, CP and PP are implementation-specific datapins. I included two "components" (pins or optical) for data and 1 for sense, without specifying more detail. The point was not to create some sort of full 3d model to send off to production, the point was to show a general way in which the combo port could be adapted to a general purpose, high power charging port without becoming any larger.

I see some coolant cross-sections, with no walls around.

Gray is walls. The shade of gray indicates the height.

Also I see no way coolant can cool the high power pins as it is not touching them and there is a gap between

This is an isometric view from the exterior, not a full 3d model. The cable and connector can route coolant as needed on either side. A brief section of pins not touching coolant is irrelevant (and they're actually quite close to the coolant regardless); copper is a superb conductor of heat, and will transfer it to where the wires are losing heat to the coolant. The normal limitation on the rate of heat loss in a wire is the rate at which heat transfers to the air, which is vastly slower than the rate at which heat moves through copper.

It's not the cable that overheats. It's the mating surfaces of two sides.

The cable absolutely will overheat if left uncooled. Ever looked up what sort of cable is needed to carry a "mere" 200A (90kW) DC indefinitely without cooling? It's 2/O AWG - 9mm/0,34" thick, 67mm² cross section, 0,66kg /1.45lbs per meter length. Tesla wants to make chargers to charge semis. Chargers that make 350kW look like "childrens' toys". And there's a large population that refuses to buy an EV until charge rates come down, which means far more power than 90kW. How exactly do you plan to make cables thick enough for that without making them unbendable and too heavy to hold?

As mentioned, however, a couple centimeters of copper pins is nothing in terms of heat transfer. Copper's thermal conductivity is 400W/m-K. The pins are about 5mm across, and if we say the pins are uncooled along 4cm, that's a U-value of 0,5W/K. Copper's electrical resistivity is 1.68e-8 ohm meters. Over a length of, say, 4cm uncooled, the resistance is 5,52e-10 ohms. From P=I²R, the heating across the connector for, say, 800A, is 0,00035W. Compare the rate of heat flow to the rate of heating. It's a complete non-issue.

It's not possible to purge coolant with pressure. There will lots of it left due to uneven cross section in the piping.

Is your impression that air purges of fluids out of plumbing systems are not actually already something that's extensively done in industry? Not that it's even essential, but....

Also I see no seals with that design. And no locking mechanism, for electrical and for coolant part.

That's because it's a front-on isometric drawing. Just like virtually every connector pin layout drawing. I'm not sure what about seeing a pinout makes you expect to see a full rendering.

Also how about glycol sticking to the seals after drying up (and dust contamination during drying). They must be manually washed before next plug-in cycle. Or there must be pre-seals or shutter to avoid that.

All of which is not the purpose of a pinout diagram. Even the concept of whether it's better to seal off coolant at disconnects or conduct an air purge is beyond the scope of a pinout diagram. But if you're trying to argue that there's something unrealistic about designing fluid lines for connect and disconnect cycles in real-world usage, I'm going to strongly have to disagree with you, as that's extensively already done.

Like Elon said: "Everything works in PowerPoint".

The funny thing is, Elon does extensive work on concepts with "powerpoint-level" detail in the early phases of his projects.

If you're suggesting that a power-point level detail presentation is a substitute for actual engineering - of course it's not. But you don't start trying to engineer something until you know what it is you're trying to engineer.


** In my original post, I had been operating on the principle that the Tesla supercharger lacks a ground pin. It's hard to get data on the pinout of the Tesla supercharger; however, I am now almost certain that it is grounded (which is IMHO a relief).

 

[arnis]

It's never the copper that is problematic on the mating surfaces. It's "everything else". aka contamination.

Middle center pin is ground. One of the big ones also has no potential voltage. 0V.
Ground is a must.
3-phase requires neutral. Tesla is using potential between phases and neutral,
not potential between two phases. Semiconductors and PCB are not rated for 500V (aka 400V AC nominal).

CCS already decided to go up with voltage. First step 800V, up to 1000V.
Same amps. Liquid cooling from charger til plug.

 

[McRat]

The USA-based SAE made three major screwups when thinking about EV standards.

3-phase power should have provisioned, and voltages from 100 vac to 600 vac should have been spec'd.

This would have allowed both common residential power as well as common commercial power.

Our power for normal buildings is:

120v 1ph residential.
240v 1ph residential - 2 phases 180° opposed of 120v measured across L1 and L2.
277v 1ph - single pole of 480v 3ph WYE common commercial building.
208v 3ph - commercial, what any 2 poles measure, L1/L2/L3 each 120v.
480v 3ph - commercial, again measured across 2 poles of 277v.
600v is just a standard US wire rating.

To really get a perspective on how dumb it was to exclude 3ph:

A 3ph 480v 32 amp onboard charger would not weigh or cost significantly more than a 32 amp 240v charger.

Big deal? It charges twice as fast? Not even close.

240v x 32a 1ph = 7.7 kW
480v x 32a 3ph = 26.5 kW

No expensive large DCFC needed for fast charging. Just a normal commercial building with a small 32a EVSE.

By allowing a 600 volt peak system voltage, you could use an inexpensive transformer at a house to make a synthetic lightweight on-board Fast Charger system for homes without 3ph power. 600v x 32a = 19kW from a 80 amp 240v draw with a lighter weight on-board charger, and a thinner, cooler EVSE cable. 80 amp charging without the drawbacks of car price, car weight, or heavy cables that get hot.

 

[McRat]

It's never the copper that is problematic on the mating surfaces. It's "everything else". aka contamination.

Middle center pin is ground. One of the big ones also has no potential voltage. 0V.
Ground is a must.
3-phase requires neutral. Tesla is using potential between phases and neutral,
not potential between two phases. Semiconductors and PCB are not rated for 500V (aka 400V AC nominal).

CCS already decided to go up with voltage. First step 800V, up to 1000V.
Same amps. Liquid cooling from charger til plug.

There are two kinds of 3 phase in the US, WYE and Delta. Delta has no neutral. You can run 3ph motors with it, but you can't make 277vac 1ph.

 

[AudubonB]

Karen and McRat - and any other cognoscenti:

1. Regarding 3-phase in the USA - can you tell me the reason 3ø charging is not possible as things exist today? Is there not a non-Frankenconnector that could permit 1ø 240V AC charging at a domicile, and 3ø 480V (etc) when at a SpC?

2. Regarding charge-cooling: there is zero reason to have any concern about mixing precious bodily fluids. Ultra-high performance heat exchangers are compact and not so heavy as to create an important weight penalty for it to be either integral with or adjacent to a vehicle's battery pack. When I ran a solar heat-capture system, my Alfa Laval brazed-plate heat exchangers were good for 200,000 BTU/hr, weighed 12kg, and looked like this:

Making use of a heat exchanger has the further benefit of permitting different coolant chemistries in the environment of a vehicular battery pack and in that of a base station, including, inter alia, the possibility of the station's fluids being kept at, for example, -40º.

 

[KarenRei]

3-phase requires neutral.

Three phase does not require neutral. Three phase requires:

* Three wires if all phases are guaranteed to be balanced and usage will be balanced
* Four wires if all three phases are to be used but there may be imbalances in the supply
* Five wires if phases are potentially to be used individually

See, for example, NEMA L15 through L17 sockets for example of 4-wire 3-phase.

CCS already decided to go up with voltage. First step 800V, up to 1000V.

How does that do anyone any good? What EVs use 800-1000V packs?

Same amps. Liquid cooling from charger til plug.

When did CCS adopt liquid cooling? I must have missed that. Is there a new charge socket that I'm not aware of?

 

[KarenRei]

2. Regarding charge-cooling: there is zero reason to have any concern about mixing precious bodily fluids. Ultra-high performance heat exchangers are compact and not so heavy as to create an important weight penalty for it to be either integral with or adjacent to a vehicle's battery pack.

That's actually a great notion. The manufacturer could choose to integrate a heat exchanger into a vehicle or not. You'd get drastically faster cooling of your onboard coolant via a heat exchanger with charger-supplied coolant than with air cooling. And you're right, heat exchangers are not prohibitively bulky or expensive.

 

[AudubonB]

Thank you...I do make a habit of trying to offer only (sometimes) a great notion.

I will, however, leave it to the fluid engineers to determine the appropriate heat exchanger size. And while my proffered -40ºF (or even -40ºC) may not be the optimum level, some Very Cold™ number, with a base station reservoir large enough to service a fully occupied SpC site, should be considered. This makes all the more sense when considering Best Practices for a semi-truck charging location.

 

[KarenRei]

I ran some numbers and the heat flow on your heat exchanger comes in right about what one might expect as the waste heat for a next-gen car charger (although a semi would probably want higher powers, perhaps double the heat flow)

12kg and a simple design vs. the mass and cost of the fans, louvres and radiators of the current system designed to cope with unassisted supercharging... that's a clear win.

 

[arnis]

When did CCS adopt liquid cooling? I must have missed that. Is there a new charge socket that I'm not aware of?

Cooled cable and plug. No transmission to vehicle.
Today's EV batteries (20-100kWh) are not capable to charge considerably faster no matter the thermal management.
Newer battery chemistries will also likely have better charging efficiency. Therefore the need for off-board cooling is reduced.

Instead of glycol loop it is much cheaper to have 800V DC charger.
Instead of glycol loop friendly vehicle is it much easier to have two 400V batteries, or one 800V battery.

 

[KarenRei]

Cooled cable and plug. No transmission to vehicle.

That would simply be a manufacturer design decision. CCS specifies the connector. If you're referring to something else, you need to be more specific.

Today's EV batteries (20-100kWh) are not capable to charge considerably faster no matter the thermal management.

The two primary issues that limit battery charging rates are ion mobility and heat removal. There are many ways to improve ion mobility with relatively little sacrifice, where it's even a limit at all. Heat removal however is bounded by coolant temperature and surface area of the cells. Cooling to ambient vs. cooling to highly chilled, and the rate at which coolant can be supplied, represents significant difference in cooling rates. Cooling becomes more important during equalization - as cells start to top up, excess energy that cannot be stored in a given cell gets converted to heat.

The amount of heat going into a battery pack is massive. At 100kW and 90% efficiency that's 10kW, about 2/3rds of the energy that a M3 consumes cruising at highway speeds. Increase that several times over, and drop the efficiency from higher charge rates and pushing harder on equalization? That's a massive amount of heat that you need to get out of the pack.

You know the most common cause of persistently slow supercharging? A stuck louvre. Again: it's about how fast you can take the heat out.

Newer battery chemistries will also likely have better charging efficiency.

Not necessarily. The biggest selective factor for Tesla is not charge efficiency, not energy density, not power density, or anything like that. It's cost per kWh. Everything else is secondary. They'll take a drop in all three to get cheaper cost per kWh.

Instead of glycol loop it is much cheaper to have 800V DC charger.

To repeat, a 800V DC charger is worthless to a ~400V EV.

Instead of glycol loop friendly vehicle is it much easier to have two 400V batteries, or one 800V battery.

Exactly how many individual cells are you picturing in the battery? There's a balance between voltage, number of cells, and degradation. Loss of a single cell is more significant the fewer cells you have in parallel. The more you want to run in series (for higher voltage) the fewer you have in parallel. Decreasing cell formats to increase the number of cells means a higher part count and lower energy density.

This beyond all issues of increased insulation requirements with higher voltages.

 

[arnis]

800V DC charger can charge anything between 0 and 800V. Including all vehicles that go up to 400V (from 300V and up).

Tesla S/X big packs have 7104 cells, 74 parallel cells connected in chain of 96. To get up to 400V.
All we have to do is have 37 parallel cells in a 192 chain. We will still have 7104 cells.
It is easy to cut the 192 chain in half and have a contactor switch them from series to parallel, so we have
37 bricks soldered together to 96 long chain. And two of those chains are shunted with contactor.
We get back to the original voltage. No need to change the drivetrain. BMS slightly changes.

 

[KarenRei]

800V DC charger can charge anything between 0 and 800V. Including all vehicles that go up to 400V (from 300V and up).

Yes, but it can only charge a 400V vehicle at 400V, and thus the charger is operating at half capacity. That extra 400V is worthless to essentially all real-world EVs.

Tesla S/X big packs have 7104 cells, 74 parallel cells connected in chain of 96. To get up to 400V.
All we have to do is have 37 parallel cells in a 192 chain.

And then a failed cell means 2,7% degradation instead of 1,3% degradation.

 

[Skotty]

Nix the coolant connection. Coolant for the cord, if there is to be any, should be self contained within the cord and the charge station. Coolant for the plug in the car, if there is to be any, should be self contained within the car. But there should be no exchange between cord and port. Too much potential for leaks and negative consequences.

Heat sensors should be part of the system (this is a more generic comment not directly related to your plug concept). Power should be throttled at certain temperature readings, and some app feedback or other indication should be provided when power is being throttled by plug heat. Maybe that already exists. I don't really know.