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

[FalconFour]

Well, some success I charged at 3 different CCS stations today. Each of them had the same behavior and the same reported charge port temperature that I suspect is limiting it...

https://twitter.com/x/status/1490090481279127556

It starts throttling around 170f reported (likely 70f actual - outside temp is low 60s). When stopped, the adapter is still cool to the touch, but lightly warmed at the Tesla end.

I really wish I could get my hands on that replacement charge port. I suspect that the thermistors are different. I'd still want to fix this by tweaking the pull-up resistors in the ECU itself, as that's where the fix would need to be done. That way people won't need to replace their old port to get it working. Just some tweaks to the Gen4 ECU that retains all the necessary safety.

A little birdy informed me that this thermistor thing isn't an unknown issue (thus not specific to my car), and the fix is usually to just apply a fixed resistor offset (on the sensor side) to it to keep it happy. I don't particularly like that solution since it'd defeat the charge port temperature sensing in case of an ACTUAL hot port. Thus I'm trying to go for fixing the circuit at the correct side

More to come! This is my weekend for playing with it.

 

[FalconFour]

OK, so... pull-ups on board: 10k. There's also a 10k resistor between that and the pin on the microcontroller. Here's what it looks like, in effect:

Oddly, I measured my charge port thermistor at... 13k At about 62f. Odd thing is, when a thermistor is at 74f, they're generally at "nominal" resistance, so close to their rated value (e.g. a 10k thermistor will read 10k at 74f). That's an awful large value difference to be a 10k. Also, being NTC, being colder means they'd have lower, not higher, resistance.

(I can assure you, the port read proper temps on the Gen3 controller No problems there, thus no suspicion of the sensors being bad)

Raising the pull-up resistance is also more challenging a mod to do than lowering it. To lower it, simply add another resistor. To raise it, well... you have to remove and replace that resistor. It's about 1/4 the size of a grain of rice.

hmm. Now to go look at the Gen3 and see what its pull-up looks like...

 

[FalconFour]

... well, that was a fun evening. But I have a happy conclusion!

TL;DR: The thermistors should be grounded on pin 6 (combined - a secondary ground path), not pins 14 and 19 (system ground) as with the Gen3. Moving the pins & splicing the wires/creating an adapter harness will be necessary.

So, I tried digging into the Gen3 but found myself stopped by a terrible foe: conformal coating I don't want to dig through that. Pins didn't trace cleanly out to obvious resistor/capacitor sets either, so that was ugly. Didn't want to get into it.

Instead, I started poking around with a YOLO approach. An NTC thermistor decreases its resistance when it gets colder, increases when it gets hotter. Baseline is 77f/25c = its "nominal" resistance = 10k. So... if I'm measuring 13k and I want it to read 10k, I'd add... 47k. I added 47k, and you know what it did? IT INCREASED THE REPORTED TEMPERATURE.

Stumped, I kept digging for more things I could test to learn what's going on. There's an unused thermistor (#3) that's reported in Tes-LAX, that I can test sanity on. So, I ran a 10k resistor to it. It reported even higher...! Something like 179f.

(note: CP_pinTemperature3 here)

That's when it dawned on me: these things aren't grounded at all to the same circuit they're being sensed from. It's reading NOISE! On the wiring diagram, there's a THERM-4 pin (pin 6) that had a pin, and a wire. What could that be? It's not labeled "analog-in" same as the others, it's labeled "signal-in". Hint, hint. So, I re-routed that 10k fixed resistor between pins 6 (THERM-4) and 22 (INLET_NTC_3).

Gen4's wiring:

Compare to Gen3:

Please excuse the crudity of this model. I didn't have time to build it to scale or paint it.

Like a sign from the heavens... it reads 77.7f on the dot.

(note: CP_pinTemperature3 here)

Mystery solved. I'll work on rewiring mine tomorrow, and then go test some charging (hopefully at full/sustained speed this time). Then, I'll start looking into a way to build an adapter harness to simplify all of this (the "cover" sensor, and rerouting the thermistors) to make this a drop-in installation.

 

[FalconFour]

Went to rewire the pins today. Spoiler: it's not pin 6. I moved the ground pins for therms #1 and #2 to pin 6, and all 3 (including #3 which was working) now reads -67f. Dead.

I don't know what the heck is going on anymore. I'm so close to giving up on this. It's like this very simple thermistor circuit is somehow obfuscated just to make life a living hell.

A thermistor circuit is typically just as simple as that schematic snip I posted earlier. But here, no. I can't figure out why, but all 3 thermistors tend to move in unison when I make one change to a single resistor. That's not how a thermistor works...

Wish I knew anyone else to collaborate with on this. Trying to decipher this mess is so frustrating alone when I run into a dead end like this.

The alternative seems to just buy the EU Gen3 ECU. Cleaner, less BS. You only lose CHAdeMO adapter support. And you wouldn't need to replace the whole charge port

 

[Bunklung]

Went to rewire the pins today. Spoiler: it's not pin 6. I moved the ground pins for therms #1 and #2 to pin 6, and all 3 (including #3 which was working) now reads -67f. Dead.

I don't know what the heck is going on anymore. I'm so close to giving up on this. It's like this very simple thermistor circuit is somehow obfuscated just to make life a living hell.

A thermistor circuit is typically just as simple as that schematic snip I posted earlier. But here, no. I can't figure out why, but all 3 thermistors tend to move in unison when I make one change to a single resistor. That's not how a thermistor works...

Wish I knew anyone else to collaborate with on this. Trying to decipher this mess is so frustrating alone when I run into a dead end like this.

The alternative seems to just buy the EU Gen3 ECU. Cleaner, less BS. You only lose CHAdeMO adapter support. And you wouldn't need to replace the whole charge port

You have been doing great and we have all been following along. You clearly have the skillset. If I could offer a second set of eyes, I would, but it's slightly out of my skillset You would think this would be s simple circuit. Engineers usually don't try to make things more complicated for little gain.

 

[Xebec]

FalconFour - when the CCS chargers report high temperatures and start throttling -- does the temperature stay at "170F" (70C) or does it keep climbing and further throttling? How long have you left running?

Where are you seeing 170F reported?

Thanks.

 

[FalconFour]

Progress and anti-progress... kinda just got to a workable point and stopped 'til I can come up with more ideas.

So, top part here is "outside" the ECU, and bottom part is what I see on the board (capacitors omitted). Above is what I hacked together to get this result:

It's all still so wrong. At this point I've just disconnected the thermistors from the harness (pulled the wires out of the connector to get them out of circuit). I'm just trying to get to a point where I can put a 10k resistor in, and have it admit that it's 77.7f as it should be. I can't seem to find a configuration of "pulls" that results in that reading.

Most curiously, they all have a weird tendency to affect one another, as if it's some kind of chain - which is where I got the unholy idea in the schematic above (which, amusingly... still didn't work for a damn except to reach low enough values to keep it happy enough not to throttle DC charging rate).

I really need to get my hands on a Gen4 charge port, so I can see what the heck kind of unholy madness they performed with these thermistors. I could even offer instructions for probing resistance, if anyone has access to just the 24-pin connector there.

Given that Gen4 also has bugs (not limited to mine, but things I see other new Teslas doing) that annoy me as well - specifically, not re-opening the charge door if it gets stopped when trying to close -- I'm really leaning towards just trying to find an EU Gen3 ECU to get me CCS with my existing port. Gen4 just sucks

But I still feel like I should push on with this Gen4 decoding, so people can choose to use the more readily-available Gen4 instead of paying out the nose for a dwindling supply of EU Gen3 boards...

 

[FalconFour]

FalconFour - when the CCS chargers report high temperatures and start throttling -- does the temperature stay at "170F" (70C) or does it keep climbing and further throttling? How long have you left running?

Where are you seeing 170F reported?

Thanks.

Yeah, it tends to just stay there (173...174 ish), and the kW dumps down to 16-24 (lowest I think I'd seen was 16). I've camped out about half an hour.

I use Tes-LAX with a CAN breakout harness to read all this stuff. Scan My Tesla (with the same harness) would probably do the same.

BTW (just to recap earlier info), these same temperature readings used to report sane values when I had the Gen3 ECU installed. Switching to Gen4 (and adding a 10k resistor to spoof the "cover" being closed - a cover that doesn't exist on this older port) caused all this thermistor weirdness.

 

[Davidjess]

Does 2020 June M3 have CCS support??

I was disappointed when Tesla insisted on a diag labor fee for my charge port not opening automatically. Then I was disappointed again when I found out they replaced it without asking me. And a third time when it was revealed that they disposed of my old part in a non-recoverable way, also without my permission. I did 1-star them for that breach of basic trust expectations. But now, there is a potential plus. If they upgraded my CCS support then now I have something to be happy about. Unfortunately, it probably already had it, and I may never know.
All is well the ends well?

 

[FalconFour]

I did 1-star them for that breach of basic trust expectations.

If it's warranty, of course you can't ask for the part back. If you didn't have to pay them for anything, then really, the way they handled it is perfectly legit. If you DID have to pay for the replacement part, then getting the old one back is expected.

Unfortunately, it probably already had it, and I may never know.

Check the "Software->Additional Vehicle Info" page for CCS adapter support. That's where you'll find it.

At this point, though, it seems that either "you have Gen3, and thus no CCS, or you have Gen4, and thus optional CCS". But that "CCS adapter support" line in Additional Vehicle Info is the final arbiter of whether you "do" or "don't". Whether you can easily upgrade still has a lot of grey area, only resolved by looking behind the trunk carpet for the 24-pin connector harness and how it's wired.

ITT: I'm trying to break those rules and put a Gen4 on a Gen3. It's just those dang thermistors at this point...

 

[FalconFour]

I did a thing today. haha

https://twitter.com/x/status/1491659365589258242

Peak steady charging rate was 190kW (it ramped up as SOC% increased - because as SOC increases, voltage increases as well, while it seems current-limited at 500A). Still, the port sensors aren't working right and my band-aid is far from an appropriate fix. I was periodically checking the temperature, and after charging, I will say the adapter pins were reasonably (not alarmingly) hot! I made it from 8% to 70% in just 20 minutes or so (and sadly... nearly 20 bucks). CCS really can be Supercharger-like with the right car and the right station (most here were 150kW max - probably 80-120kW on a Tesla).

Still waiting for the port from Tesla, though ... honestly I may be waiting for the wrong thing. Maybe I ought to be sourcing a known Gen4-harness charge port to analyze. Need to get back on the ball there.

 

[rrolsbe]

I did a thing today. haha

https://twitter.com/x/status/1491659365589258242

Peak steady charging rate was 190kW (it ramped up as SOC% increased - because as SOC increases, voltage increases as well, while it seems current-limited at 500A). Still, the port sensors aren't working right and my band-aid is far from an appropriate fix. I was periodically checking the temperature, and after charging, I will say the adapter pins were reasonably (not alarmingly) hot! I made it from 8% to 70% in just 20 minutes or so (and sadly... nearly 20 bucks). CCS really can be Supercharger-like with the right car and the right station (most here were 150kW max - probably 80-120kW on a Tesla).

Still waiting for the port from Tesla, though ... honestly I may be waiting for the wrong thing. Maybe I ought to be sourcing a known Gen4-harness charge port to analyze. Need to get back on the ball there.

Must have been about 45 cents/KWH? High, but in a pinch would be much better than L2 charging!

 

[wws]

I did a thing today. haha...

Nice. Is that the new EA site at the back side of Valley Fair mall? I just drove by there a couple days ago.

 

[FalconFour]

Nice. Is that the new EA site at the back side of Valley Fair mall? I just drove by there a couple days ago.

Yup yup. Same site at this guy used for the Lucid testing:

- maybe even the same station

Really, really nice place! I know it's a "flagship" site, so it's rather unique (kinda like Tesla's Kettleman City is a "flagship", not the normal style of installation), but my god, it's wonderfully designed. Stalls are marked with kW right on the ground. I could find the 350kW station quickly as I was driving through.

 

[jrobillard]

I did a thing today. haha

https://twitter.com/x/status/1491659365589258242

Peak steady charging rate was 190kW (it ramped up as SOC% increased - because as SOC increases, voltage increases as well, while it seems current-limited at 500A). Still, the port sensors aren't working right and my band-aid is far from an appropriate fix. I was periodically checking the temperature, and after charging, I will say the adapter pins were reasonably (not alarmingly) hot! I made it from 8% to 70% in just 20 minutes or so (and sadly... nearly 20 bucks). CCS really can be Supercharger-like with the right car and the right station (most here were 150kW max - probably 80-120kW on a Tesla).

Still waiting for the port from Tesla, though ... honestly I may be waiting for the wrong thing. Maybe I ought to be sourcing a known Gen4-harness charge port to analyze. Need to get back on the ball there.

@FalconFour Where did you get the CCS adapter? I have someone here in Canada (Montreal) selling them. They showed me your video to prove it's working. Thanks!

 

[BigNick]

Yup yup. Same site at this guy used for the Lucid testing:

- maybe even the same station

Really, really nice place! I know it's a "flagship" site, so it's rather unique (kinda like Tesla's Kettleman City is a "flagship", not the normal style of installation), but my god, it's wonderfully designed. Stalls are marked with kW right on the ground. I could find the 350kW station quickly as I was driving through.

That CCS1 plug just looks so bulky to use compared to the Tesla plug, almost looks like you'd have to reinforce the charge port on the car to handle its weight.

 

[wws]

That CCS1 plug just looks so bulky to use compared to the Tesla plug, almost looks like you'd have to reinforce the charge port on the car to handle its weight.

In fact, with the Chevy Bolt EV, it is common to have to prop the plug a bit in order for the connection to initialize correctly.

Eric Way also did a review of the site:

 

[FalconFour]

@FalconFour Where did you get the CCS adapter? I have someone here in Canada (Montreal) selling them. They showed me your video to prove it's working. Thanks!

Here:

Tesla US CCS1/CCS2 (400A)

Due to war and huge internal demand, export orders likely will NOT be executed any time soon. We will do our best to restore the manufacturing supply chain, for each order placed. We will contact each order owner when it`s ready to ship. Our apologies. Tesla US - ccs1 adapter...

evhub.shop

It's... a bit spicy in price. I'm also just borrowing it to test for this project, tbh Might buy one later and perhaps share with the local Tesla groups...

Honestly, my faith in Tesla is pretty low lately... don't like their new directions and bunny-chasing (I'm in pretty high agreement with that recent Ben Sullins video regarding Tesla's "why" - and I had been, a while before then). They're probably too distracted to ever get the "official" CCS adapter produced for the rest of the world - at least any time soon, it feels like.

 

[jrobillard]

Here:

Tesla US CCS1/CCS2 (400A)

Due to war and huge internal demand, export orders likely will NOT be executed any time soon. We will do our best to restore the manufacturing supply chain, for each order placed. We will contact each order owner when it`s ready to ship. Our apologies. Tesla US - ccs1 adapter...

evhub.shop

It's... a bit spicy in price. I'm also just borrowing it to test for this project, tbh Might buy one later and perhaps share with the local Tesla groups...

Honestly, my faith in Tesla is pretty low lately... don't like their new directions and bunny-chasing (I'm in pretty high agreement with that recent Ben Sullins video regarding Tesla's "why"). They're probably too distracted to ever get the "official" CCS adapter produced for the rest of the world - at least any time soon, it feels like.

I am part of the lucky one that got my SR+ 2021 with the GEN4 stripped ECU... so it's saying CCS not installed. I am trying to get the *real* Gen4 and swap my ECU. I live about 10 minutes from a Tesla Service center. I try this morning to access the service mode at my house but it didn't work. I drove to Tesla and it worked in the parking lot. So I will most likely swap it here drive to Tesla to get the update re-installed.

 

[wws]

... Honestly, my faith in Tesla is pretty low lately... don't like their new directions and bunny-chasing (I'm in pretty high agreement with that recent Ben Sullins video regarding Tesla's "why" - and I had been, a while before then). They're probably too distracted to ever get the "official" CCS adapter produced for the rest of the world - at least any time soon, it feels like.

I'm guessing they don't want the service hassle of doing a bunch of retrofits for those of us with older cars right now. Of course that wouldn't preclude them from selling adapters to those whose cars already have CCS support.