Tesla Wall Connector load sharing protocol

[scaesare]

I'm looking to do something like this myself. I can't find any work by others I could build on. I already have latest version of HPWC.

@wk057 might be willing to share some experiences.

I seem to recall there being some discussion of once a pilot signal throttled the car down mid-charging session, it didn't want to come back up, but that may be old data.

 

[gregd]

I had a closer look at this. Only had the file “tesla hpwc log.txt” (see above, thanks ctr) to work from.
I have no access to any Tesla hardware (yet) !

So I ‘m not sure about the methodology used, the timing(s) & direction of those signals ...
I rearranged the data, but let the “0000” packets in. They appear important.

Have a look at the attached files. (Sorry but TMC-forum only takes TXT- or PDF-files)

Some stuff jumped out (mark the colour coding):

• The words “c0fe”, “fec0” + “7801” reappear in every telegram.
• In the scenario “Slave Reconnected”, all telegrams are identical.

Happy deciphering !

I'm just getting into car-hacking, so don't have the depth yet to decypher. But, if I were a Tesla engineer, I would use a protocol based on something I knew. CAN Bus. If we looked at these as CAN Bus messages, would they make more sense? They're not going to use standard PIDs, of course, but perhaps the structure of CAN messaging would provide a framework for decoding? Just a thought...

 

[Drewflux]

I think someone in the euro sub forum has worked it out. Cant find the thread though. I had started to read it but was having trouble with the translation

 

[Ulmo]

I just got this email from Cohere:

Thank you for your email. We do not yet deliver in the US, but coincidentally I am moving to the Bay Area next week. I had the opportunity because my wife got a job in Silicon Valley. I will start business development for Cohere in California. And my focus will be mainly on large projects with multiple charging stations, but in time we also wish to deliver to households.

In Europe we are compatible with the Tesla Wall Connector. I hear the unit is the same as in the EU, but this has not been officially confirmed. Also, we are running our new feature Solar Charging right now in beta with a number of customers. We plan to launch this new feature officially in the short term. So there is a good change that it will work in your case. We are still in the process of integrating a new kWh-meter to measure the main grid connection. The kWh-meter we use in Europe is unfortunately not [yet] UL certified.

One of the first tasks for me is to make Maxem suitable for the US and find a number of beta customers. If you like we could explore if we can make a test case in your home. But please note that this will be in a bèta phase since we are not even officially located in the US yet. To see what is possible, I have some additional questions for you:

• Where are you located?
  
• What size grid connection do you have?
  
• What size of MCB do you have on the Tesla Wall Connector?
  
• What size of MCB do you have on your solar panels?

I look forward to your response!

Best regards,

MCB means breaker. (Stands for "Miniature Circuit Breaker"; apparently people who use that term would consider $14,000 breakers in SuperChargers that use 8 foot cabinets as "normal sized"?)

The email for them is m, coussement, but at the domain name, as you know, of "cohere, ... eu, and please note that you must take out the spaces and replace commas with a single dot and turn the many dots into one dot and use the right at sign and yes that two letter thing is part and sorry to make this difficult but trying to get away from the rob't.s

 

[sorka]

I thought I read that it did. I'm pretty sure that @wk057 is controlling the charge current for his cars, but maybe he hacked the HPWC itself. (or maybe I am confusing it with him just starting/stopping the charging and he wanted to control the current.)

When the Z apocalypse comes, I'm camping out at Jason's house. By then he should have solar powered laser mounted turrets using his S batteries as power sources at night. He'll have retrofitted auto pilot sensor hardware and hacked the firmware to create an unbreachable defense barrier.

 

[dlt]

Anyone had any success with this? I have two european wallchargers and a four-wire usb rs485 module, and some raspberry 2-wire modules. I can do some tests too, but just wanted to check if someone have already figured it out...?

 

[IanAber]

Is the anyone within a couple of hundred miles of Gainesville Florida with two or more chargers who would be willing to let me come and sniff the network. I only have one charger and car so I can't see the traffic between chargers. I want to control the available current in my Solar installation so I need to decode this stuff. It would olny take an hour or so if you are willing.

 

[IanAber]

Alright, I tried in RS485 4wire mode (haven't tried RS422 mode yet) and received some stuff. Baud rate is 9600

Does anyone see the packet pattern here? Afaict its at least two different types in master mode, even more in slave mode.
Except for slave mode the packets are only send when reconnecting the HPWC or pressing the reset button, nothing continuous (although I havent tried very long). I also did not connect the car yet, I'm pretty sure something is being sent as soon as charging starts.

I have looked at the output from my single charger. It looks like the packets are 17 bytes long. They start with c0 f? and end with c0 fe.
This is the start up sequence from the single charger I have.
c0 fc e1 03 8c 49 00 00 00 00 00 00 00 00 b9 c0 fe
c0 fc e1 03 8c 49 00 00 00 00 00 00 00 00 b9 c0 fe
c0 fc e1 03 8c 49 00 00 00 00 00 00 00 00 b9 c0 fe
c0 fc e1 03 8c 49 00 00 00 00 00 00 00 00 b9 c0 fe
c0 fc e1 03 8c 49 00 00 00 00 00 00 00 00 b9 c0 fe
c0 fb e2 03 8c 49 00 00 00 00 00 00 00 00 ba c0 fe
c0 fb e2 03 8c 49 00 00 00 00 00 00 00 00 ba c0 fe
c0 fb e2 03 8c 49 00 00 00 00 00 00 00 00 ba c0 fe
c0 fb e2 03 8c 49 00 00 00 00 00 00 00 00 ba c0 fe
c0 fb e2 03 8c 49 00 00 00 00 00 00 00 00 ba c0 fe

If someone with two or more could run some tests I would love to work with you to figure this out. If there is anyone in Florida with two chargers that would let me connect my MacBook to get some data that would be terrific. Please let me know.

 

[widodh]

Did somebody already reverse engineer this protocol and publish it somewhere?

I'd like to put a HPWC (3-phase, EU) in Slave mode and control it using a RPi.

Is there anything out there yet?

 

[llavalle]

IIRC the wires car communicate with the new HPWC using SWCAN (single wire can)... and this is translated / forwarded to the other HPWC on the network.

This works because instead of using J1772, they are using a proprietary protocol...

 

[widodh]

IIRC the wires car communicate with the new HPWC using SWCAN (single wire can)... and this is translated / forwarded to the other HPWC on the network.

This works because instead of using J1772, they are using a proprietary protocol...

Sure? Since the loadsharing also works with non Tesla EVs at our office where we have two in master slave.

They communicate using RS485

 

[davewill]

Sure? Since the loadsharing also works with non Tesla EVs at our office where we have two in master slave.

They communicate using RS485

Yours is actually the first report I've had of someone who has successfully used HPWC load balancing with non Tesla cars. I figured it must work since Tony's been selling them for a while and no one's complained online, but...

It's fairly certain the the HPWC communicates with the car to get SOC info as it has been observed that the load sharing gets balanced so that two Teslas that are connected at the same time finish charging together. For non-Teslas I would guess that they fall back to a simpler algorithm like splitting the available amperage evenly...which is entirely satisfactory.

 

[widodh]

Yours is actually the first report I've had of someone who has successfully used HPWC load balancing with non Tesla cars. I figured it must work since Tony's been selling them for a while and no one's complained online, but...

It's fairly certain the the HPWC communicates with the car to get SOC info as it has been observed that the load sharing gets balanced so that two Teslas that are connected at the same time finish charging together. For non-Teslas I would guess that they fall back to a simpler algorithm like splitting the available amperage evenly...which is entirely satisfactory.

Remember, I am in Europe, so the US situation might be different.

Our HPWC setup is a single 3-phase 400V/32A circuit which provides a maximum of 22kW.

Both HPWCs are connected to that circuit and set in Master<>Slave setup.

The maximum current is set to 32A and when I'm the only one charging I get 32A (3-phase) over the Pilot Signal to my Model S.

Yesterday I arrived at the office and a BMW 330e was charging. I plugged in and my Pilot Signal / Dashboard showed 26A and my car started charging at 26A. The 330e was probably charging at 6A at that moment.

This setup has been working flawlessly for the last 7 months with a Renault Zoe, VW e-Golf, Audi A3 e-Tron, BMW i3.

The two HPWCs are connected using a CAT6 cable where we use two wires (1-pair) and that's what they use to communicate.

So what I want to figure out is how a HPWC works in Slave mode. Since you should be able to fake a "master" with some custom software and create your own Smart Charging system. The HPWC is fairly cheap.

The (Dutch) people from Maxem seem to have figured it out: Maxem® | Tesla Wall Connector

So it's possible, we just need to figure out the protocol and have it documented.

 

[cbutters]

Yours is actually the first report I've had of someone who has successfully used HPWC load balancing with non Tesla cars. I figured it must work since Tony's been selling them for a while and no one's complained online, but...

It's fairly certain the the HPWC communicates with the car to get SOC info as it has been observed that the load sharing gets balanced so that two Teslas that are connected at the same time finish charging together. For non-Teslas I would guess that they fall back to a simpler algorithm like splitting the available amperage evenly...which is entirely satisfactory.

I set up a load sharing config with 2 wall connectors (US 240V single phase) recently:

I'm currently charging my bmw i3 using the JDAPTER on the Slave Wall Connector with no issues. Model X gets delivered next week so I'll be able to test and see how the load sharing works with a non-tesla car in US load sharing setup. I'm encouraged to hear that other people have done it (although in europe) which makes me less scared to actually try it.

 

[widodh]

Right now there is a BMW 330e and a Porsche Cayen Plugin charging on our two HPWCs without any issues.

They don't draw the full 32A together so probably not a lot of loadbalancing going on there.

 

[widodh]

See the manual of the EU HPWC: https://www.tesla.com/sites/default...installation-manual-en-EU-v2.pdf?201612081439

See page 6: "Load sharing"

and then page 27 about the daisy chain.

I wonder how I should sniff the traffic between a Master and Slave to figure out how this works.

 

[K-MTG]

I plan to do this later in the year when my 3 comes..will have to sell my current wall connector (old version)

 

[widodh]

I asked the Dutch people of Maxem today if they would be willing to share the information. Negative. They say they are under a NDA and can't share anything.

I don't know if they got the information from Tesla or that they reverse-engineered it themselves.

 

[robertvg]

It's fairly certain the the HPWC communicates with the car to get SOC info as it has been observed that the load sharing gets balanced so that two Teslas that are connected at the same time finish charging together.

Interesting because that is actually the first time I read this. For all reports I have seen about users having master/slave setups with 2 Tesla's (in Europe) it simply splits the amperage equally.

Edit: looked up comments from September last year from a user with a M/S config, Tesla confirmed that the available power is distributed evenly over all HPWC's with a connected car....even if the car was already full and charging was stopped

 

[widodh]

Interesting because that is actually the first time I read this. For all reports I have seen about users having master/slave setups with 2 Tesla's (in Europe) it simply splits the amperage equally.

Edit: looked up comments from September last year from a user with a M/S config, Tesla confirmed that the available power is distributed evenly over all HPWC's with a connected car....even if the car was already full and charging was stopped

Not what I am seeing. I see 26A and 6A to the other car. The hpwc seems to measure the actual current going to each car