Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Retrofit CCS compatibility onto earlier (NA) Model 3 - DIY approach

This site may earn commission on affiliate links.
Status
Not open for further replies.
@falcon Are the yellow wires in the correct position in your photo? Pinout appears to show Pin 3 empty and Pin 4 connected, but photo looks like Pin 3&5 connected vs 4&5.

One side (Gen3 port side) takes pins 21 & 22 as TEMP-IN-1 and TEMP-IN-2. The other side (Gen4 ECU side) puts them on pins 20 & 22 as INLET_NTC_1 and INLET_NTC_3 because INLET_NTC_2 is unpopulated on Gen4 (so they moved it). Hopefully you're referring to this wiring diagram :)

So with that, I'm fairly confident calling this my "final answer".
1645339489323.png
 
Whoops I’m sorry I didn’t realize I had pressed send on that question- after looking at it for a while I realized I had the socket connector flipped 180 (hence mentioning pin 4&5 instead of 20&22).

As an unrelated side note… lock wire can be a really versatile cheap de-pinning tool. It’s stiff enough and comes in various diameters.
 
  • Informative
Reactions: FalconFour
- Charge port (itself) (1490374-10-D)
- Charge port door/front (possibly 1501773-30-A but can't find earlier/later PNs so can't verify that's Gen4 or even if Gen4 is any different)
- Charge port to ECU harness (1551811-00-B)
- Charge port ECU (1537264-00-B)
- Charge port to battery harness (1501773-30-A)

These are the significantly different parts I can trace together... and honestly I'm not even sure if they're compatible with "Gen3" cars (e.g. if the mounting holes of the port to frame are the same - or if the routing of the 2/0 AWG orange charging cable mounts to the same points).

In fact, as I look at the charge port to battery harness, it looks to have a completely different connector for the battery 😳 So honestly, that may not even be possible!
View attachment 777565
For those who didn't quite get what was going on here, here's a picture from a Model Y teardown. Looks like when they ported over some of the Model Y changes, they changed things to this busbar system, which I guess required the new charge ECU

xgql1mpmos5cvyhkz8og.png
 
  • Like
Reactions: FalconFour
@FalconFour Did you encourter some charging station that didnt work with your CCS adapter ? Im wondering what would the difference in the PLC or canbus protocol between charger brand that would make some work and some dont.
So far, two stations that I wouldn't expect to work (because they're in development, or are complete crap and very old), haven't worked. Everything else (and I've tried quite a few), have worked.

EA (and their various station models) seems to work every time. ChargePoint Express 250 (CPE250) works every time. Tritium RT50 (at least I'm pretty sure it is) works with the reliability it's known for (that is: not perfect, not great - but same as on any other vehicle).

Basically, the experience so far has been "it works about as well as it works for any other car". CCS is MADDENINGLY complex (remember: it's PLC - which is a physical layer for Ethernet - and once that's established, it goes to IPv6 and HTTP and XML etc etc etc) and that all exists along-side J1772 signalling (+/-12v at 1khz PWM, signaling station/vehicle ready/charging/standby states in an analog manner) and honestly, the most amazing part of this setup is that it works at all - that Tesla crammed it into such a small part. CCS is a standard formed by committee, where the committee just went "okay, who wants what features?" and then just said "why not throw the kitchen sink in, too?". So they did. haha

I'm quite happy with its performance so far. Well, except for this long-standing issue: Did you know the 3 heats the battery (actively) constantly while DC charging at any speed or temp?
 
  • Like
Reactions: KevinT3141
So far, two stations that I wouldn't expect to work (because they're in development, or are complete crap and very old), haven't worked. Everything else (and I've tried quite a few), have worked.

EA (and their various station models) seems to work every time. ChargePoint Express 250 (CPE250) works every time. Tritium RT50 (at least I'm pretty sure it is) works with the reliability it's known for (that is: not perfect, not great - but same as on any other vehicle).

Basically, the experience so far has been "it works about as well as it works for any other car". CCS is MADDENINGLY complex (remember: it's PLC - which is a physical layer for Ethernet - and once that's established, it goes to IPv6 and HTTP and XML etc etc etc) and that all exists along-side J1772 signalling (+/-12v at 1khz PWM, signaling station/vehicle ready/charging/standby states in an analog manner) and honestly, the most amazing part of this setup is that it works at all - that Tesla crammed it into such a small part. CCS is a standard formed by committee, where the committee just went "okay, who wants what features?" and then just said "why not throw the kitchen sink in, too?". So they did. haha

I'm quite happy with its performance so far. Well, except for this long-standing issue: Did you know the 3 heats the battery (actively) constantly while DC charging at any speed or temp?
Thanks. Would you see an easy way to diagnostic a fail charge on specific charger ? I guess the answer is no but i still ask :p
 
Thanks. Would you see an easy way to diagnostic a fail charge on specific charger ? I guess the answer is no but i still ask :p
Not really. Stations and the car are completely opaque as far as all the complexity goes. I don't know any way to dig deeper, even with Tes-LAX/CAN info. I'd say, though, that knowing how CCS negotiates can be helpful to getting a station to work:
  1. you plug in, regardless of whether you've swiped your card or not, and negotiation starts (as👏long👏as👏the👏handle👏latch👏clicks. The handle latch is all too commonly overlooked. If the latch button doesn't "click" when inserted, it thinks you're holding the button down. If you get the "charging plug not fully inserted" notification on your phone, that's what this means! The latch button is pressed somehow)
  2. not exactly sure which is true: that the station signals B1 (+12v solid, functional, but not ready) and then begins SLAC (PLC signal-tuning to "get the Ethernet cable plugged in")... or whether it signals B2 (+/-12v PWM with 3-7% PWM duty, meaning "digital communication required") first. Either way, at this point, SLAC takes place in the high frequency RF realm over the existing low-frequency or DC pilot signal.
  3. after SLAC completes and the car<->vehicle have an established communication channel, they start talking about requirements and who they are. (this includes stuff like kWh-to-full, max charging rate, minutes estimated, etc etc)
  4. Finally, there's this {{{{LOOOOONG PERIOD}}}} where the car and station will wait for a thing called "contract authentication", as I recall it referred to. This, now, is waiting on you to swipe your card, etc ;) (The car actually shows when it reaches this phase - on the screen is a little notification saying something like "charging station requires authentication" - and that's how you know digital comms are established and it's just waiting on your card at the station - or at least, that's what the station has told the car).
  5. After swiping card, the two quickly chat about performing the cable test (station sends high voltage up the cable, but the car's charging relay isn't closed yet), and makes sure it's not shorted anywhere or leaking.
  6. Next, the car tells the station what voltage to set the cable to (the battery voltage), then when happy, the car's charging relay closes without a spark (since the voltage matches).
  7. The car tells the station to gogogogogogo and how many amps to pump to reach what target voltage, etc. And they keep chattering back and forth rapidly while it charges; the car telling the station what to do, many times a second.

Most often, it seems to break down at points 2 or 3 above -- the station isn't signaling in a way that the car understands (either the station or the car isn't "listening for" SLAC signals, so one side's saying "hello? hello? hello?" while the other isn't listening -- or the station is internally in a disjointed state where the CCS modem is asleep but the UI is saying it's establishing communication on the screen)... and you end up with this whole routine falling apart. Some stations do stupid things with the CCS process - they worked so hard at implementing the digital side of things, they didn't bother thinking of the analog (PWM) side. Lots of things "just worked" by sheer interoperability compatibility chance, but not actually the way it was intended to work - so it's shipped, it breaks later.

Or sometimes, around step 4, it's actually NOT waiting on that - the station may have been developed to have you swipe your card FIRST, and then it'll just magically, automatically have "contract authentication" passed in hard code. If you plug-first, the SLAC will time-out while waiting to swipe the card. So sometimes (admittedly very rare, and on badly-behaving stations), you'd need to swipe first, plug second. But the overwhelming majority of the time (and the way it's meant to be done), you plug first (so they can start chatting), swipe second (so the chatting & the authentication can occur in parallel), and it should work.

At this point, with holes in the standards, it seems like "every vehicle/station for themselves", and we're just hoping the de-facto standards crystalize sooner than later. But with my experience with that in-development station, it still seems we have a ways to go. 😂

(sorry, a bit of brain vomit there. CCS gives me nightmares.)
 
Last edited:
Not really. Stations and the car are completely opaque as far as all the complexity goes. I don't know any way to dig deeper, even with Tes-LAX/CAN info. I'd say, though, that knowing how CCS negotiates can be helpful to getting a station to work:
  1. you plug in, regardless of whether you've swiped your card or not, and negotiation starts (as👏long👏as👏the👏handle👏latch👏clicks. The handle latch is all too commonly overlooked. If the latch button doesn't "click" when inserted, it thinks you're holding the button down. If you get the "charging plug not fully inserted" notification on your phone, that's what this means! The latch button is pressed somehow)
  2. not exactly sure which is true: that the station signals B1 (+12v solid, functional, but not ready) and then begins SLAC (PLC signal-tuning to "get the Ethernet cable plugged in")... or whether it signals B2 (+/-12v PWM with 3-7% PWM duty, meaning "digital communication required") first. Either way, at this point, SLAC takes place in the high frequency RF realm over the existing low-frequency or DC pilot signal.
  3. after SLAC completes and the car<->vehicle have an established communication channel, they start talking about requirements and who they are. (this includes stuff like kWh-to-full, max charging rate, minutes estimated, etc etc)
  4. Finally, there's this {{{{LOOOOONG PERIOD}}}} where the car and station will wait for a thing called "contract authentication", as I recall it referred to. This, now, is waiting on you to swipe your card, etc ;) (The car actually shows when it reaches this phase - on the screen is a little notification saying something like "charging station requires authentication" - and that's how you know digital comms are established and it's just waiting on your card at the station - or at least, that's what the station has told the car).
  5. After swiping card, the two quickly chat about performing the cable test (station sends high voltage up the cable, but the car's charging relay isn't closed yet), and makes sure it's not shorted anywhere or leaking.
  6. Next, the car tells the station what voltage to set the cable to (the battery voltage), then when happy, the car's charging relay closes without a spark (since the voltage matches).
  7. The car tells the station to gogogogogogo and how many amps to pump to reach what target voltage, etc. And they keep chattering back and forth rapidly while it charges; the car telling the station what to do, many times a second.

Most often, it seems to break down at points 2 or 3 above -- the station isn't signaling in a way that the car understands (either the station or the car isn't "listening for" SLAC signals, so one side's saying "hello? hello? hello?" while the other isn't listening -- or the station is internally in a disjointed state where the CCS modem is asleep but the UI is saying it's establishing communication on the screen)... and you end up with this whole routine falling apart. Some stations do stupid things with the CCS process - they worked so hard at implementing the digital side of things, they didn't bother thinking of the analog (PWM) side. Lots of things "just worked" by sheer interoperability compatibility chance, but not actually the way it was intended to work - so it's shipped, it breaks later.

Or sometimes, around step 4, it's actually NOT waiting on that - the station may have been developed to have you swipe your card FIRST, and then it'll just magically, automatically have "contract authentication" passed in hard code. If you plug-first, the SLAC will time-out while waiting to swipe the card. So sometimes (admittedly very rare, and on badly-behaving stations), you'd need to swipe first, plug second. But the overwhelming majority of the time (and the way it's meant to be done), you plug first (so they can start chatting), swipe second (so the chatting & the authentication can occur in parallel), and it should work.

At this point, with holes in the standards, it seems like "every vehicle/station for themselves", and we're just hoping the de-facto standards crystalize sooner than later. But with my experience with that in-development station, it still seems we have a ways to go. 😂

(sorry, a bit of brain vomit there. CCS gives me nightmares.)
Thanks for the great reply. I should've ask you before trying to understand pages of IEC and J1772 combo standard documents.

For now i was able to :
155KW on a 150kw Electrify Canada station.

70KW on AddEnergie 100Kw station. Since my SOC was in the 50's and not preheated enough.

I maxed out a 50Kw Flo Station (Addenergie 50kw)

90Kw on a 125kw ChargePoint Express 250. SOC 55%.

I plan to try a 350Kw ABB charger or the 350Kw of electrify Canada (was offline tonight) with full battery conditioning and a low SOC like 5%>. Im aware i will probably max out at 255Kw cause the car is software lock but still want to try.

FYI the car is a plaid 2022 with the last Ukrainian CCS adapter.
 
Actually, you're looking at kW when the real limiting factor is amps (not shown on stations). These rated kW figures are often at higher voltages (e.g. the 800v cars). Your Plaid is a higher voltage system than our 3's (400v - often 350-395v with nominal around 370v I think), but not sure what the Plaid actually classifies as (supposedly 450v nominal is cited in the manual, thus "500v" relative to 400v). So you'll get more juice than the 3's will, just due to the higher voltage/same amps.

Pretty cool that you get such high numbers none the less!
 
  • Like
Reactions: android04 and wlee
Today I have bad news. I swapped my GEN 3 CP ECU NA version "092755-98-D", to EU version "1092755-82-B". Previously, I have tested to charge my car with a general J1772 charger with a J1772-Tesla adapter and a Tesla stock wall charger, they both work perfectly.

But I tested two Tesla SCs in TPC (Tesla Proprietary Connector, as you used in North America), the SC refuses to charge my car. The red light near the charge port was on when the charger was plugged-in. The screen shows a rare alarm code 'cp_a091' "charger plug incompatible" (I translated it from Chinese). Replugging the charger in does not help.
1647093331443.png


As I currently still have no CCS to TPC adapter, I cannot test Tesla SCs in CCS2 or third-party CCS2 SC.

It seems bad news to me. If I cannot use Tesla TPC SC, swapping to EU GEN3 CP ECU would be less meaningful. (I am not sure it can use Tesla CCS2 SC since I don't have the adapter now). I believe this is a software issue that can be solved by Tesla, but I cannot do anything as end-user now.

So may I ask you, have you tested your "TPC" Tesla Model 3 with Tesla TPC SC after swapping your CP ECU? does it still works?





 
  • Informative
Reactions: FalconFour
Today I have bad news. I swapped my GEN 3 CP ECU NA version "092755-98-D", to EU version "1092755-82-B". Previously, I have tested to charge my car with a general J1772 charger with a J1772-Tesla adapter and a Tesla stock wall charger, they both work perfectly.

But I tested two Tesla SCs in TPC (Tesla Proprietary Connector, as you used in North America), the SC refuses to charge my car. The red light near the charge port was on when the charger was plugged-in. The screen shows a rare alarm code 'cp_a091' "charger plug incompatible" (I translated it from Chinese). Replugging the charger in does not help.
View attachment 779815

As I currently still have no CCS to TPC adapter, I cannot test Tesla SCs in CCS2 or third-party CCS2 SC.

It seems bad news to me. If I cannot use Tesla TPC SC, swapping to EU GEN3 CP ECU would be less meaningful. (I am not sure it can use Tesla CCS2 SC since I don't have the adapter now). I believe this is a software issue that can be solved by Tesla, but I cannot do anything as end-user now.

So may I ask you, have you tested your "TPC" Tesla Model 3 with Tesla TPC SC after swapping your CP ECU? does it still works?
Hmm this might be the downside of the EU ECUs. I guess EU model 3s and Ys only speak CCS even at superchargers. US cars need to be able to do both.

Best bet might be South Korean gen 3 ECU, or wait for the US to get the retrofit kit.
 
  • Like
Reactions: FalconFour
Presumably South Korea. You could see if Tesla is willing to order 1092755-04-B for you

I google this part number and found nothing, then I use Yandex and found this sentence "PLC Gen3 P1092755-04-B Корейский поддерживает одновременно CCS и CHAdeMO, но Сколково всё равно мимо. Предлагаем назвать его ретрофит 2.0"

I use translator and it seems mean the south Korea version CP ECU P1092755-04-B supports both CCS and CHADEMO. I know EU version CP ECU does not support CHADEMO but I don't need it, so it is not a problem for me. To me, EU version CP ECU must enables me to use Tesla SC (both TPC and CCS2), otherwise, even only support CCS2 is acceptable, TPC SC is too crowded in my country.

So I guess use South Korea CP ECU might not help to my situation.
 
Well South Korean cars use the TPC. Unless South Korea has no tpc superchargers, I imagine they must support tpc supercharging.

I think your EU ECU will work fine with CCS. But it might not be capable of speaking with Teslas proprietary supercharging signals since they migrated away from that in Europe.
 
  • Like
Reactions: FalconFour
So may I ask you, have you tested your "TPC" Tesla Model 3 with Tesla TPC SC after swapping your CP ECU? does it still works?
Hmm. I always thought that swapping to the EU Gen3 ECU would be kind of an iffy proposition. I'm surprised it "took it" (didn't need config change?) but informative to know it messes up TPC Supercharging.

Here in the US, there's no Superchargers other than TPC and no public chargers other than Type-1 J1772 (or TPC, less commonly) for AC, with CCS and CHAdeMO on the DC side. Even that spread of standards is too complex for some people. 😂

So for me, with my NA Gen4 swap and "Bundle of Wires" adapter, everything works great - TPC Supercharging, CCS, CHAdeMO (though I just sold my adapter recently), and AC charging.
 
Me, musing over the fact that the charge port cable there appears to be in a rigid metal tube, and how it could be, in theory, fully sealed end to end...



Now I'm gonna have to dig into that a bit. haha

It's likely a solid conductor. One of the Munro Tesla dissection videos pointed out something similar on the Model S as changed from the Y and 3 they ripped apart. They noted it costs significantly less. They didn't go into the construction so I'm not sure how it insulates the current carrying portion.
 
Actually, you're looking at kW when the real limiting factor is amps (not shown on stations). These rated kW figures are often at higher voltages (e.g. the 800v cars). Your Plaid is a higher voltage system than our 3's (400v - often 350-395v with nominal around 370v I think), but not sure what the Plaid actually classifies as (supposedly 450v nominal is cited in the manual, thus "500v" relative to 400v). So you'll get more juice than the 3's will, just due to the higher voltage/same amps.

Pretty cool that you get such high numbers none the less!
So do the car ECU limit to 500A on CCS?

I saw Ryan huber getting 213Kw at 39% soc. I got arround 151Kw max at that SOC even tho i was preheated ( prior to charge i used a supercharger v3 near to preheat by entering in NAV).
Confirmed at the same time by Scan My Tesla app i had over 210Kw available of charger max power.

I tried another Electrify 350kw charger but it was not working. Trying a ABB 350Kw didnt help cause the cable is too thick and make a bad connection with the adapter.

I will wait for them to fix the 350Kw E.A and try again...
 
So do the car ECU limit to 500A on CCS?

I saw Ryan huber getting 213Kw at 39% soc. I got arround 151Kw max at that SOC even tho i was preheated ( prior to charge i used a supercharger v3 near to preheat by entering in NAV).
Confirmed at the same time by Scan My Tesla app i had over 210Kw available of charger max power.

I tried another Electrify 350kw charger but it was not working. Trying a ABB 350Kw didnt help cause the cable is too thick and make a bad connection with the adapter.

I will wait for them to fix the 350Kw E.A and try again...
No limit on the ECU's side in my experience. Far as it's concerned, it's a Supercharger, it seems (so, same BMS limits as Supercharging). BTW, you can sometimes use waypoints to place the Supercharger ahead of the CCS station navigation (CCS 1st -> SC second) and it'll still preheat. Not always reliable though. I monitor its behavior using Tes-LAX and I see it sometimes ignoring the SC (passive target=86f), but most of the time it treats it as a nav to SC (active heat target=130f).

Really, the stations just have frustrating and opaque limits not revealed in Tes-LAX (CAN data) nor the station, that I can see. So it's hard to tell - but I have seen 200kW at a 350kW EA station on my Model 3 - I think that's a bit over 500a, maybe 600.
 
  • Like
Reactions: jimbeault
Status
Not open for further replies.