DC Charging for the Roadster

[Eberhard]

With the upcoming CHAdeMO-charging Station, offering up to 50kW or even the new Tesla Quickcharger with up to 90kW, this all make it to be of interest to the current Roadster Owner to take advantage for the DC-charging too. There is only a minimum on hardware required, most is done by the software and the communication via CAN-bus. The Roadster already owns a CAN-bus- What is needed to do? Put a CHAdeMO socket in the trunk, wire to the ESS and CAN-Bus. Attache a micro-controller to the socket to convert CAN-bus commands to control-signal for the socket? This can be don quite cheap, because no expensive charging hardware is needed. Will cost no more than an UMC-Cable which is doing a similar job on the AC side, simulating an EVCS and switch the power on/off.

What to you thing? charging the Roadster to 80% within 1h on 50kW?

[dhrivnak]

It would be great to charge at these levels. I have found 3 fast DC chargers in the area but have yet to find an AC station more than 30 amps. AT 22 miles per hour the J1772 are marginal at best

[sprediletto]

I would love quick charge capability for the Roadster. During the holidays, I drove 380 miles from Atlanta, GA to Ponte Vedra Beach, FL. The 5 hour stop at an RV Park in to recharge (220A/40V) was simply too long. Despite free WiFi and nice people to talk too, I was itching to get back on the road.

[dpeilow]

If they don't allow it on the S160, the same applies to the Roadster.

[Eberhard]

i cannot see any technical reason not to do it. Or we have to do it by ourself.

[richkae]

Even if you had to charge at less than 50kW to preserve the battery, being able to plug into CHAdeMO or other QC has lots of value because 16.8kW J1772 is very rare, and its unlikely anybody other than a Tesla owner is going to install it. ( All of the 70amp 200+volt charging in Washington and Oregon was installed/facilitated/donated by Roadster owners )
Limiting your charge rate to 27kW is the same C-rate as the 40kW Model S with twin chargers ( 20kW ) and still charges the Roadster to 80% in about an hour and 35 minutes.

[Eberhard]

The biggest concern is keeping the temperature of the pack below a certain level. But the current can be restricted to any level. You start with the recommended current and drop it when temperature rise to much or the voltage reach the final level.

[dpeilow]

Yes I agree (on the business thing).

Lithium-based Batteries Information - Battery University (http://batteryuniversity.com/learn/article/lithium_based_batteries) and Charging Lithium-Ion Batteries - Battery University (http://batteryuniversity.com/learn/article/charging_lithium_ion_batteries) says that LiCoO2 18650 cells should be charged at 0.8C.


Charge efficiency is 97 to 99 percent and the cell remains cool during charge. Some Li-ion packs may experience a temperature rise of about 5ºC (9ºF) when reaching full charge.


For the Roadster I make that a charge power of 48.5 kW.

[rsquared99]

At a recent EV event in Phoenix, a representative of one of the companies planning on putting DC charging along highway routes between US cities told me that they we're putting 50 amp AC outlets along side their DC stations for Roadster use. He said the current Roadster batteries "could not take" DC charging. He had his personal Roadster there and he and I were giving rides that afternoon, so I took the answer to be an informed one and didn't question it. I wanted to know why not HPC power levels and he said too expensive to justify the limited use. So, take it for what it's worth. I'm not technically compentant enough to know what kind of battery characteristics would prohibit charging at the DC levels being discussed.

[Eberhard]

Yes I agree (on the business thing).

Lithium-based Batteries Information - Battery University (http://batteryuniversity.com/learn/article/lithium_based_batteries) and Charging Lithium-Ion Batteries - Battery University (http://batteryuniversity.com/learn/article/charging_lithium_ion_batteries) says that LiCoO2 18650 cells should be charged at 0.8C.

For the Roadster I make that a charge power of 48.5 kW.

Thats fine with 50kW from CHAdeMO and gives 80% charge within 1h. Teslas 90kW wouldn't be faster.

[surfingslovak]

At a recent EV event in Phoenix, a representative of one of the companies planning on putting DC charging along highway routes between US cities told me that they we're putting 50 amp AC outlets along side their DC stations for Roadster use. He said the current Roadster batteries "could not take" DC charging. He had his personal Roadster there and he and I were giving rides that afternoon, so I took the answer to be an informed one and didn't question it. I wanted to know why not HPC power levels and he said too expensive to justify the limited use. So, take it for what it's worth. I'm not technically compentant enough to know what kind of battery characteristics would prohibit charging at the DC levels being discussed.
I can nearly guarantee you that DC charging with 0.5C will not harm the cells and might be even beneficial in the long term. This has nothing to do with the specific profile and chemistry of the cells. 0.5C is generally considered as the sweet spot for charging, and I've included the results for 800 charge cycles for commercially available LiMnO4 cells below. Depending on the specific characteristics of the cells used in the Roadster, even higher C-rate might be possible without any detrimental effect. However, this might require some investigation and preferably also a lab test using the same batteries Tesla put in the Roadster.

3880

[Norbert]

However, this might require some investigation and preferably also a lab test using the same batteries Tesla put in the Roadster.

3880

Yes, would be really interesting to have these numbers with a higher resolution for all the cell type(s) Tesla is using for Roadster and Model S.

A bit peculiar that for cycle numbers 100 and 150 the sweet spot is at 0.25C, while for higher cycle numbers it is 0.5C

[Eberhard]

0.5C means 2h charging time. Still a huge timesaving.

[stopcrazypp]

I've included the results for 800 charge cycles for commercially available LiMnO4 cells below.
3880
I'd be careful about taking the results from LiMnO4, Tesla's cells are LiCoO2 and there's a huge difference between the two.

[stopcrazypp]

Before getting all excited about the possibility of DC on the Roadster, I wanted to ask if any of you already know if the Roadster has an accessible (this is important, because if it's buried somewhere, you can't use it) contactor that's connected to the battery. Without it, safe DC charging is not possible (it doesn't come standard on the Leaf either, it's only included if you opt for the CHAdeMO socket). And I'm not sure how easy/cheap it is to source a high voltage, high current contactor.

Just having a 20kW onboard EVSE doesn't necessarily mean the car has components to handle 50kW DC charging.

[surfingslovak]

I'd be careful about taking the results from LiMnO4, Tesla's cells are LiCoO2 and there's a huge difference between the two.
Yes, the chemistries are different, but I would not say that the delta was huge in this context. 0.5C is generally considered safe for a wide range of cells and chemistries. I would challenge you to prove me wrong on this issue. If possible, I would prefer if you wouldn't take my words out of context. My post clearly says that further investigation is necessary.

What this report shows is that 0.5C charging is not only safe, which we would expect anyway, but potentially even superior to slow charging. This was news to me, and I like to use this paper whenever I see the claims that 120V trickle charging was best for battery life. That's is not necessarily the case, and your results might vary. And this should be the takeaway from this report. Likewise, I would not like to see DC quick charging being generally dismissed based on unfounded claims that it's somehow inherently hurtful to the batteries.

Unfortunately, I only have this report available, since I was doing research for the cells Nissan is using in the Leaf. I'm pretty sure that a similar report can be found for LiCoO2 cells, and I would encourage you to look for one.

Before DC charging for the Roadster is considered, technical feasibility aside for the moment, the exact properties of the cells used will need to be examined. This should include a bench test to confirm the expected results. Having said that, I would be surprised if 0.5C was found to be detrimental for cell life.

[stopcrazypp]

Yes, the chemistries are different, but I would not say that the delta was huge in this context.
I'm saying that because LiMnO4 cells typically have significantly better thermal stability, power density, and cycle life than LiCoO2 cells (as much as 2x or more). Without the relevant data for LiCoO2, I wouldn't use the assumption they are in any way similar in charging characteristics.

[dpeilow]

I'd be careful about taking the results from LiMnO4, Tesla's cells are LiCoO2 and there's a huge difference between the two.

I posted the LiCoO2 number up thread a few minutes before...

[stopcrazypp]

I posted the LiCoO2 number up thread a few minutes before...
The link you gave explores a completely different area than surfingslovak's chart. It's mainly examining different cut-off voltages, not how different charging rates affect the cycle life of the cell (which is the issue we are discussing). The only relevant point was this:

Manufacturers recommend charging the 18650 cell at 0.8C or less. Charge efficiency is 97 to 99 percent and the cell remains cool during charge.
http://batteryuniversity.com/learn/article/charging_lithium_ion_batteries
That's not particularly useful because it doesn't tell us the impact of charging faster, or how much life you gain (or lose) by charging at a slower rate. Nor does it tell us the "optimal" charging speed ("0.8c or less" is different than saying "0.8C" period). And it doesn't provide data to back that claim.

What I mean is we need an identical or similar chart as surfingslovak's for LiCoO2 cells before coming to a conclusion about the impact of different charging rates on battery life.

The most critical points we have to examine is roughly:

Tesla Roadster 56kWh
120V@16A = 1.92kW = ~0.03C
240V@16A = 3.84kW = ~0.07C
240V@32A = 7.68kW = ~0.14C
240V@70A = 16.8kW = ~0.30C
400V@125A=50.0kW = ~0.89C

Model S 40kWh
120V@16A = 1.92kW = ~0.05C
240V@16A = 3.84kW = ~0.10C
240V@32A = 7.68kW = ~0.20C
240V@42A = 10.0kW = ~0.25C
240V@83A = 20.0kW = ~0.50C
400V@100A=40.0kW = ~1.00C
400V@113A=45.0kW = ~1.125C
400V@125A=50.0kW = ~1.25C

[Eberhard]

Before getting all excited about the possibility of DC on the Roadster, I wanted to ask if any of you already know if the Roadster has an accessible (this is important, because if it's buried somewhere, you can't use it) contactor that's connected to the battery. Without it, safe DC charging is not possible (it doesn't come standard on the Leaf either, it's only included if you opt for the CHAdeMO socket). And I'm not sure how easy/cheap it is to source a high voltage, high current contactor.

Just having a 20kW onboard EVSE doesn't necessarily mean the car has components to handle 50kW DC charging.

The PEM has connectors to the ESS as well to the charger (and AC-Motor). The cables to the battery can handle the short term load of 700-800A (225kW) and continues load of 40kW. It will work.

[richkae]

Assume that someone could build such a box, that cleanly inserts into the car ( I assume between the battery and the PEM ) and provides you with a CHAdeMO plug ( that lives under the trunklid ).
Also assume this box provides you a maximum of 40kW charge rates from a CHAdeMO plug and you could get it built and installed for $5000.

How many people would want one?
What about if it was $10000?

[stopcrazypp]

The PEM has connectors to the ESS as well to the charger (and AC-Motor). The cables to the battery can handle the short term load of 700-800A (225kW) and continues load of 40kW. It will work.
To make it more clear, I'm not talking about "connectors". I'm talking about a "contactor", which is basically a switch that makes it so you don't have any live current on the DC pins until the DC connection is established. It is required for any type of DC charging, but is not necessarily present in every single BEV (esp. one you can connect to on the outside)
http://en.wikipedia.org/wiki/Contactor

In particular, you need one that can handle ~50kW of power to use CHAdeMO at near full power. Given the Roadster has a 16.8kW onboard charger, it may have one that can handle at least ~20kW of power, but I don't know how much higher it may go or if it even exists and can be connected to. And I don't know if there are contactors between the PEM and the batteries, what kind of power rating they have, and if you can access them.

I know you can probably source a CHAdeMO socket for fairly cheap (and the other electronics to do the hand shaking), but I'm not sure how much such a contactor will cost.

[Eberhard]

Assume that someone could build such a box, that cleanly inserts into the car ( I assume between the battery and the PEM ) and provides you with a CHAdeMO plug ( that lives under the trunklid ).
Also assume this box provides you a maximum of 40kW charge rates from a CHAdeMO plug and you could get it built and installed for $5000.

How many people would want one?
What about if it was $10000?

If this would include the CHAdeMO charger - thats great. But only the socket? should not be more expensive then the HPC or UMC.

[Norbert]

That's not particularly useful because it doesn't tell us the impact of charging faster, or how much life you gain (or lose) by charging at a slower rate. Nor does it tell us the "optimal" charging speed ("0.8c or less" is different than saying "0.8C" period). And it doesn't provide data to back that claim.

It also seems to refer to a "typical consumer Li-ion battery" in general.

[hcsharp]

To make it more clear, I'm not talking about "connectors". I'm talking about a "contactor", which is basically a switch that makes it so you don't have any live current on the DC pins until the DC connection is established. It is required for any type of DC charging, but is not necessarily present in every single BEV (esp. one you can connect to on the outside)
http://en.wikipedia.org/wiki/Contactor
...
You might need much more than just access to the contactor. You have to work with the software and hardware. For DC charging you would pretty much have to charge the cells in series, which means that unless the firmware was upgraded to communicate with our new DC charger, we would have no way to know if we were overcharging (above 4.15V) one of the bricks. If you were certain that your cells were already well balanced, you might be able to get away with charging to about 80% capacity to be safe.

If all we had was access to the contactor and lucky enough to get a firmware upgrade from Tesla, could we still do this? The only thing we know for sure about the hardware is that it can handle charging in series for very brief periods at rates up to 40kW. That's what regen does. You can't assume that since it can discharge at 200kW that the same circuit can charge at 40 or 50kW for extended periods like 1 or 2 hrs.

[GSP]

The high voltage contactors would have to be inside the ESS to allow disconnecting the ESS terminals for safe transport, vehicle assembly, and servicing. I think there is a basic ESS schematic posted elsewhere on this site that showers the contactors.

The cables from the ESS to PEM could be replaced with a three-way cable (probably $1000 from a good outfit such as MTG Moltec). This cable would be connected to a separate box with contactors for the ChadeMo connector. This box with contactors will be required to keep high voltage disconnected from the ChadeMo connector except when charging. Contactors can be purchased from Panasonic, GigaVac, KiloVac, or others. The box will also need a microprocessor with mutiple CAN transceivers to communicate with the ChadeMo charger and one or more vehicle CAN networks. Count on at least another $1000 for the box.

The box will need software. A roadster and a ChadeMo charger will be required for development and testing. This will not be free. Changes to the Roadster's firmware may also be required. That could be REALLY expensive if reverse engineering if required.

This project sounds like a worthwhile follow on to the J1772 conversion project and the OVMS project.

GSP