Supercharger protocol for diy CHAdeMO adapter

[arg]

I think this kind of "standard" is totaly absord, why not use the pin of the type 2 plug to also do DC (2 pin for + and 2 pin for -), as Tesla do !??

The standard allows for that (DC over the same connector as is used for AC), but the current is limited in that mode. For higher current DC, the standard adds the two extra pins; Tesla solved this problem by a connector which is not the standard connector but is compatible with it - Tesla's connector has longer pins.

On the face of it, Tesla's connector is a better solution, though we don't actually know the full specifications (only guesswork by watching superchargers in action), nor the relative costs.

In your link, it's said that the Tesla is 100% compatible on the protocol (but not physical). If it's TRUE, a CCS adapter for the Tesla is just a physical thing, CSS socket <---> Type 2 plug, to send the 2 DC power pin of the CSS plug to the 4 power pin of the Tesla Type 2 socket

Unfortunately, not quite so trivial. There's two things such an adapter needs to solve:

• Locking: the connector is required to be locked to the vehicle during charging. With an adapter, this means the adapter needs to lock to the car and the plug on the cable needs to lock to the adapter. Probably this can be done with a clever mechanical design (maybe the adapter has to be put on the cable first, and has a spring-loaded locking pin that is compressed as the adapter is inserted into the car). But it's not easy, as you have to make it foolproof while at the same time having low enough insertion force that the user can get the adapter on/off without the leverage that you usually have with one end anchored to a heavy car.
• The car needs to know that the adapter is present, and that it is definitely the right kind of adapter. The protocol calls for the EVSE to offer, and the car to select, which pins the DC is delivered on. Normally, the Model S will request it on the core pins (as it expects to be connected to a Supercharger, or to a lower power standards-based charger). Plainly, the car must not request power on the external pins if the combo plug has somehow been connected to the car without the proper adapter, as the DC pins would then be hanging in free air and so represent a safety hazard. Maybe you can't actually mate the combo connector with the car due to the connector keying, but the car still needs to know if the adapter is present to make the right selection.

So ironically, while the CHAdeMO adapter is a much more complex thing, it is actually rather easier to make (ie. it's a lot of work, but its clearly defined and you just need to sit down and do it), while the CCS adapter requires solving new problems and almost certainly requires Tesla to add specific support in the car itself.

There's probably more hope of the charger manufacturers providing an additional 'head' on the charger for DC on the core pins: since we are already seeing dual-head chargers with CHAdeMO and high-current AC (aimed at Renaults), and there's talk of adding a third head for CCS, the high-current AC head could easily supply DC as well, so adding Tesla support without needing any more 'heads' than they are already committed to. Since these chargers are typically supplying only 50kW, there's no need for the higher current offered by the extra two pins.

 

[nlc]

The standard allows for that (DC over the same connector as is used for AC), but the current is limited in that mode. For higher current DC, the standard adds the two extra pins; Tesla solved this problem by a connector which is not the standard connector but is compatible with it - Tesla's connector has longer pins.

On the face of it, Tesla's connector is a better solution, though we don't actually know the full specifications (only guesswork by watching superchargers in action), nor the relative costs.

Ok, thank for the precision

Unfortunately, not quite so trivial. There's two things such an adapter needs to solve:

• Locking: the connector is required to be locked to the vehicle during charging. With an adapter, this means the adapter needs to lock to the car and the plug on the cable needs to lock to the adapter. Probably this can be done with a clever mechanical design (maybe the adapter has to be put on the cable first, and has a spring-loaded locking pin that is compressed as the adapter is inserted into the car). But it's not easy, as you have to make it foolproof while at the same time having low enough insertion force that the user can get the adapter on/off without the leverage that you usually have with one end anchored to a heavy car.

The locking on car side is already done, just have to implement the locking of the CHAdeMO plug in the adapter

The car needs to know that the adapter is present, and that it is definitely the right kind of adapter. The protocol calls for the EVSE to offer, and the car to select, which pins the DC is delivered on. Normally, the Model S will request it on the core pins (as it expects to be connected to a Supercharger, or to a lower power standards-based charger). Plainly, the car must not request power on the external pins if the combo plug has somehow been connected to the car without the proper adapter, as the DC pins would then be hanging in free air and so represent a safety hazard. Maybe you can't actually mate the combo connector with the car due to the connector keying, but the car still needs to know if the adapter is present to make the right selection.

As I work on the CHAdeMO adapter, the adapter must detect the Model S wants DC on it's pin, and begin transaction with the CHAdeMO charger.
The big problem is to find the right protocol the Tesla uses. If what we can read on some topic is true, it use the standardized SAE DC signaling (GreenPhy PLC over the pilot line), but is this information really confirmed somewhere ?
Also, Tesla probably uses the official pinout for DC on Type 2 socket, but need to be confirmed too.

So ironically, while the CHAdeMO adapter is a much more complex thing, it is actually rather easier to make (ie. it's a lot of work, but its clearly defined and you just need to sit down and do it), while the CCS adapter requires solving new problems and almost certainly requires Tesla to add specific support in the car itself.

"Just" need the true and complete SAE DC signaling the continue my work...

There's probably more hope of the charger manufacturers providing an additional 'head' on the charger for DC on the core pins: since we are already seeing dual-head chargers with CHAdeMO and high-current AC (aimed at Renaults), and there's talk of adding a third head for CCS, the high-current AC head could easily supply DC as well, so adding Tesla support without needing any more 'heads' than they are already committed to. Since these chargers are typically supplying only 50kW, there's no need for the higher current offered by the extra two pins.

This will make the charger more and more complicated. Everything would be easier is they all do how Tesla do, have a specific Tesla socket which can support standard Type 2 plug (For AC), and modified Type 2 plug (for high power DC). Tesla can pass more than 300A on its Type 2 socket/plug !!!

 

[matbl]

There's probably more hope of the charger manufacturers providing an additional 'head' on the charger for DC on the core pins: since we are already seeing dual-head chargers with CHAdeMO and high-current AC (aimed at Renaults), and there's talk of adding a third head for CCS, the high-current AC head could easily supply DC as well, so adding Tesla support without needing any more 'heads' than they are already committed to. Since these chargers are typically supplying only 50kW, there's no need for the higher current offered by the extra two pins.

We already have dual head chargers with Chademo and CCS DC here (no AC at all). So there's an easily accessible public charger capable of 50 KW DC from two different heads but the Tesla can't charge from either.

This will make the charger more and more complicated. Everything would be easier is they all do how Tesla do, have a specific Tesla socket which can support standard Type 2 plug (For AC), and modified Type 2 plug (for high power DC). Tesla can pass more than 300A on its Type 2 socket/plug !!!

Actually the industry decided on the Type 2 CCS with the two extra pins. That's what is standard and that's what Tesla should have used!
That's also why you will see these pop up all over the place soon (already started here) since this is what will be used for BMW i3, VW e-Golf, Mercedes B-class electric, Audi A3 e-tron, etc, etc...
And that's why I think a CCS DC adapter is what's really needed.

But maybe not in France since you opted for Type3 instead...

 

[nlc]

Actually the industry decided on the Type 2 CCS with the two extra pins. That's what is standard and that's what Tesla should have used!
That's also why you will see these pop up all over the place soon (already started here) since this is what will be used for BMW i3, VW e-Golf, Mercedes B-class electric, Audi A3 e-tron, etc, etc...
And that's why I think a CCS DC adapter is what's really needed.

It's not because they are choosed that standard that this standard is good
If the Type 2 connector can be modified to accept standard Type 2 plug, and also specific high power DC Type 2 plug, it was a better and cleaner way.
The Type 2 combo is really big and ugly !!

But maybe not in France since you opted for Type3 instead...

Don't talk me about that
France choosed to enter Europe, but France want stay on Type3 while the Type 2 has been choosed as a standard for Europe :scared:
We all need a Type2 to Type 2 and Type 2 to Type cable to be sure we can recharge on public AC charger....

Here in France I believe there is not a lot of Combo2 charger, but there are CHAdeMO, it's why I want to build an adapter.

Need to find the SAE DC signaling protocol, any idea where to find it ?

 

[matbl]

It's not because they are choosed that standard that this standard is good
If the Type 2 connector can be modified to accept standard Type 2 plug, and also specific high power DC Type 2 plug, it was a better and cleaner way.
The Type 2 combo is really big and ugly !!

Need to find the SAE DC signaling protocol, any idea where to find it ?

Check:
http://www.google.se/url?sa=t&rct=j...3DZTqUBXNi0sFwDdRzwXt2w&bvm=bv.62922401,d.dmQ

Note that it is an old draft. The published version will cost you some money...
ISO 15118-2:2014 - Road vehicles -- Vehicle-to-Grid Communication Interface -- Part 2: Network and application protocol requirements
ISO/DIS 15118-3 - Road vehicles -- Vehicle-to-Grid Communication Interface -- Part 3: Physical and data link layer requirements


And for open source software:
Welcome to the OpenV2G project

 

[nlc]

Ok thank you that will be a good starting point.
I found a GreenPhy PLC module, waiting for quote, will be mandatory to try to communicate with the Model S...

- - - Updated - - -

Looked your first document, really complicated to just send a current consign to an external charger

 

[matbl]

It's an iso standard... Typical comittee product...

 

[nlc]

This morning I done some tests. When the Model S tells the wallbox it is connected, I sent a 5% duty cycle on the pilot signal PWM.
The Model S asks for charging (curious !?), thus my wallbox closes its relay.
And after approx 1s, I can see with the oscilloscope that the Tesla send a modulation on the pilot line. Need to try again later to try to measure the modulation signal

 

[matbl]

And after approx 1s, I can see with the oscilloscope that the Tesla send a modulation on the pilot line. Need to try again later to try to measure the modulation signal

Isn't this the Homeplug greenphy modulation then?

 

[nlc]

not sure yet, modulation seems very low frequency
need to measure again, this morning is was raining I stopped investigation

 

[matbl]

Here's a crazy thought. Get a homeplug av adapter. Dlink or netgear or something. Take care that it's homeplug av or av2 compliant though since greenphy is a subset of av as far as I can tell. Should get you an ethernet connection which you can easily sniff with wireshark...
Plus, they're cheap. You can get a single one for 25-30 EUR.

Of course you need to open it and decouple its power supply from its power line communications so you can power it normally while hooking up the PLC part to the pilot line and ground.

Or I might be totally wrong...

 

[nlc]

Ahahah I had exactly the same idea last this night !!! But I wonder if we can found a compatible PLC box.

 

[nlc]

Found this interesting document : http://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=4&cad=rja&uact=8&ved=0CEoQFjAD&url=http%3A%2F%2Fwww.qca.qualcomm.com%2Fwp-content%2Fuploads%2F2013%2F09%2FHomePlug_Green_PHY_whitepaper.pdf&ei=HaZBU8nANvOW0QW9goBQ&usg=AFQjCNFVAUI3Pci-x-AEVgEVAi98jgz8oQ&sig2=hukn2KB-2EtRgEqgjiSYcg

You are right, HomePlug AV and GreenPhy are interoperable (page 9) !!
Now I need to find a homePlug AV transmitter...

- - - Updated - - -

Ok it seems a lot of PLC are homePlug AV, I will buy one tomorrow.
At home we already have 2 PLC boxes, to connect our xDSL modem and the TV box. But if I remember, these device must be paired, don't know if I will see ethernet traffic if I just connect it as is on the pilot signal.

 

[Johan]

Forget this question I just bought the CHAdeMO specification

Since you bought the specification, are you not violating the legal terms by doing what you are doing now? I remember reading about some people who had made an "open source Chademo" charger and they were extremely diligent in pointing out that they had not bought the official Chademo specs since then would then have been in violation legally.

Here is a video of it:

 

[VolkerP]

Interesting project! Have 2 ideas to contribute.

1) Locking of a CHAdeMO or CCS plug to a DIY built adapter is not required IMHO. The cable is attached to the fast charging station, you don't own it -> no need for theft protection. The plug/socket interfaces are designed that communication pins are separated before power pins when disengaged -> no need to lock for electric safety (prevent arcing), just user experience. I could live without in a DIY project.

2) tap in on the green phy comms: the charge port connectors are mechanically accessible from inside the Model S trunk. There is a removable piece of trunk carpet right there. You could hook up a protocol apparatus that is invisible, portable, in the dry, mobile, powered by model s. You could drive to the next supercharger (I know, that would out of France ATM) and "listen" to your Model S supercharging.

 

[arg]

Interesting project! Have 2 ideas to contribute.

1) Locking of a CHAdeMO or CCS plug to a DIY built adapter is not required IMHO. The cable is attached to the fast charging station, you don't own it -> no need for theft protection. The plug/socket interfaces are designed that communication pins are separated before power pins when disengaged -> no need to lock for electric safety (prevent arcing), just user experience. I could live without in a DIY project.

The need for locking isn't just for touch-safety: the arcing problem is much worse with DC than AC, and I wouldn't trust the reaction time of the pilot monitoring to be fast enough to avoid the problem. The standards (both SAE/IEC and CHAdeMO) also provide for detection that the car's internal contactor has got welded shut and use the connector lock to prevent you unplugging in this case which would leave the car-side connector live to battery voltage: presumably the increased risk of arcing with DC made them think that this fault condition isn't so unlikely that it can be ignored.

Obviously with a DIY project there's nobody watching over your shoulder to see you obey the rules, but for something like this that you are going to be using out in public (and with somebody else's charger), it's a rather different situation from something you use only in your own garage. Even if nothing goes wrong, there's a high chance of being 'caught' using it and having to justify yourself, and in this situation being able to say truthfully that the DIY device is built to the same (or better!) level of safety as commercial devices is important.

However, in the CHAdeMO case it's a non-issue: the connectors on both sides of the device already have provision for locking, and the adapter is already a powered device that can operate the latches.

My comments up-thread about locking being a problem were talking about a hypothetical CCS->Tesla adapter which is hoped to be just a mechanical part with no electronics. To achieve that, it's necessary that closing of the latch on the car side also has the effect of latching the charging cable to the adapter. This can probably be solved with some clever mechanical design - fit the adapter first to the charge cable, then a hinged part in the adapter needs to close before you can insert the adapter into the car and so latches the charge cable - but it's not easy since it's got to be robust, small enough to fit in the clearances around the socket on the car, and (if you take the adapter-to-cable-first approach) it needs to be possible for the user to separate the adapter from the cable using only 2 hands. Given that some users may be struggling to hold a heavy DC charge cable in one hand, this may be tricky. It will be interesting to see how easy-to-use Tesla's CHAdeMO adapter turns out to be: they have a nice big handle for you to grip it by, and you might be allowed to fit the adapter to the car first.

 

[nlc]

Will see the locking in a second time. On the CHAdeMO socket, if you disconnect the plug, data signal are open before DC pin, normally the charger shutoff the power instantly (and/or open it's contactor).

Today I bought a couple of TL-PA2015P, they say HomePlug AV compliant. I opened one, they use a QCA6410 chip.
I am not sure it will work because these device must be paired, thus I think there is really a very little chance I will see something.

I think the pairing is for cryptographic reason ? Because I don't see any other reason to do a pairing if homeplug is a standard PLC protocol !?

- - - Updated - - -

I answer to myself, from the documentation of TL-PA2015P PLC, the pairing button is for configure the cryptographic AES128 key between the 2 device.
Really don't know if it will be able to pass on the ethernet side the Tesla packets (which are probably not encrypted because it need to talk with all charger).

 

[nlc]

Yesterday evening (and a big part of this night...) I made a big forward step
I learned a lot about the low level layer of the HomePlug AV/Green Phy protocol, and found a lot of very useful opensource tools (open-plc-utils) which permit to manage, configure and supervise the QCA6410 used in the PLC boxes I bought yesterday, through the ethernet port of my computer.

I carefully read all the documentation and begun to study the source code of these tools, and I see very useful thing, as for example parameters we need to configure in the Q6410 to have it work on PEV side or EVSE side ! Also found a very useful utility which permits to emulate an EVSE charger with the computer, and thus communicate with the car through the HomePlug/Ethernet gateway I bough yesterday.

At this point it remains some obscure things I don't understood yet, as what NMK (network management key) to use to be able to communicate with the car.
This key is the AES128 key which permits to encrypt PLC data, and all PLC device on the network must share the same NMK to be able to discuss on the network.

In other word for now I don't know how the EVSE charger pair itself to the car to be able to initiate network communication. Need to study again the source code of these tool to try to answer this question.

I opened one PLC device yesterday, to connect it on my wallbox it will be easy, I can directly connect mains pin on Pilot signal and earth GND signal, and I just need to power the PLC device with 12V, I seen where to put the 12V. (Because the PLC device will not be able to power itself form the mains pins !!)

 

[nlc]

That's OK I understood how PEV and EVSE connect them-self and exchange the right NMK to be on the same network :biggrin:

All the startup PLC session connexion between the car and my computer (through my PLC ethernet device) running the EVSE emulator should work out of the box if I can successfully convert my PLC device to connect on the control pilot signal of my wallbox

After that the next step will be the higher level protocol, but the openV2G project will probably helps a lot :smile:

- - - Updated - - -

Since you bought the specification, are you not violating the legal terms by doing what you are doing now? I remember reading about some people who had made an "open source Chademo" charger and they were extremely diligent in pointing out that they had not bought the official Chademo specs since then would then have been in violation legally.

This is a totally different approach : they made an open source CHAdeMO charger, thus I think maybe it's not be possible to release anything opensource if it's based on official spec ?
In my case I want to build a product for myself and maybe other, buy the spec is not a problem

 

[nlc]

Some important news :smile:
This morning I tried to synchronize my PLC device reconfigured as EVSE DC charger, supporting SLAC protocol (the first low level protocol to connect to the car and exchange the NVM key for higher normal PLC protocol).
It don't synchronize. So I carrefully analysed pilot signal with my oscilloscope. Can see the PLC modulation signal of my PLC device, a modulation from the Model S, but it's not an HomePlug AV/GP modulation !!

Thus for now my conclusion is that the Model S has a proprietary communication protocol with supercharger, not based on Homeplug Green phy PLC communication over pilot signal.

Will post some screeshoot of the pilot signal to show the Model S modulation when the EVSE send a 5% duty cycle.

Done some measuring, don't know exactly how they encode data, but it seems there are 32us bit, thus a 31250Hz carrier

- - - Updated - - -

A supplementary detail, you will see on screenshoot that it seems the Tesla can force a level 1 on the pilot signal when the EVSE send -12V, but it cannot force a 0 level when the EVSE send +12V (during the 5% duty cycle ON level).

Thus I can see the 5% ON time from EVSE with +12V perturbates the Model S signal because it cannot force a 0 level. Thus I think when the DC session begins (before the modulation the Tesla asks for charging so we can detect the Tesla begins its session), we probably need to force the signal pilot PWM to 0% (-12V), to not perturbate the Tesla signal and be able to receive it correctly. In this case the pilot signal maybe work as an half duplex serial port. This would be a clean and beautiful way to change the pilot signal behavior and do bidirectional data communication :smile: