BMW i3 CCS/charge port controller

[Jack Bauer]

@ EV_Builder re the current control that sounds great! Anything would be better than my cave man hack at this stage:)

[EV_Builder]

Do you have windows somewhere? Maybe by using VMware?
Then I make a GUI app to generate the table values.

[Jack Bauer]

Wow. Yep have dual boot with win7.

[Jack Bauer]

annnddd there's 40kw:)

[Jack Bauer]

@davefiddles I seem to have a bug in state 6 where my crude little current control routine is no longer being called. worked yesterday before your mods. Could you have a look? Thanks:)

[Jack Bauer]

GAAHHH! ignore me Dave! It's a Damien special : signed value in an unsigned variable.

[tom91]

Damien, have you tried requesting current while the requested voltage is the same or only a little over battery voltage?

I wont how nicely the offboard chargers respect the output voltage and thus act in a true CV/CC supply method.

Think its time you invest in some water cooled CCS ports https://zqmachine.en.alibaba.com/produc ... 37c35IpZjr

[Jack Bauer]

Wow that's mad even the cables are liquid cooled. Not sure my little 22kwh zoe pack would be too happy at those power level. Not tired letting the charger control Cv. Will try tomorrow. Did get my nice zeroev rgb button fitted though:)

[mikeselectricstuff]

Wow that's mad even the cables are liquid cooled.

Yea - the annoying whine on your Ionity vid is the coolant pump - sounds like a crappy windscreen washer pump!

[EV_Builder]

Damien, have you tried requesting current while the requested voltage is the same or only a little over battery voltage?

I wont how nicely the offboard chargers respect the output voltage and thus act in a true CV/CC supply method.

Think its time you invest in some water cooled CCS ports https://zqmachine.en.alibaba.com/produc ... 37c35IpZjr

Isn't it just that you send a voltage limit and a current limit? A bit like your LAB powersupply? A pot for current max and a pot for voltage max?

So cranking up the voltage to 500 doesn't make a difference if the current is 10AMP max because the pack will limit it.
I think we should set voltage limit to max voltage of pack. And play (limit) only with the current.
Wrong?

[tom91]

Isn't it just that you send a voltage limit and a current limit? A bit like your LAB powersupply? A pot for current max and a pot for voltage max?

So cranking up the voltage to 500 doesn't make a difference if the current is 10AMP max because the pack will limit it.
I think we should set voltage limit to max voltage of pack. And play (limit) only with the current.
Wrong?

So far the Constant Current region been proven very well. Im just suggesting to check the constant voltage section, this is to ensure there is a saftey net in case the request for current causes battery voltage overshoot.

[bji900]

Just set the CV lower than max pack voltage so you can test it. That way If it overshoots no worries, you have time to shut it down.

[EV_Builder]

Just set voltage to 5 volts higher then pack and ask 30amps and make an amp curve. The amps should throttle down then. My bet is that it goes full steam to the volts and stops ramping.

So it's more voltage limit then charge voltage.
In order to reach a voltage over de pack you need amps anyway and you only got a limited amount.

[Jack Bauer]

Little bit more today ...

[Jack Bauer]

Well folks, what to do with your newly discovered CCS freedom...ROAD TRIP! That exactly what I did yesterday taking the goose on a 220km jaunt requiring 2 stops for some CCS action. Or so I thought! Day before I went to my local Ionity (works perfectly) and then to a local Efacec QC45 "50kw" charge point. Decided to go for the money shot and asked it for the full works : 125A please...nope shutdown. Tried again, nope shutdown. Ok let's sneak up on it...40kw,41kw,42kw,43kw,44kw all go so far, 45kw Shutdown! Seems its the worst kept secret that the 50kw rapid chargers in Ireland are actually 44kw despite what they say over comms and on the rating plate. But it's all good. We're used to this interesting interpretation of standards from the Chademo era.

On to the road trip. Phase 1 : I arrive at my first stop. A shiny new Circontrol Rapition 50. Damn that's a sexy name. I plug in. The state machine rolls onward hapily and I am rewarded by the now familiar thunk of closing contactors from the rear as we hit state 6. I'm about to crack a smile of hapiness when another thunk followed by the deaded Pilot : static. No amount of coaxing or messing about with the timings in the state machine make any difference. I'm stuck on 6.6kw ac.

Phase 2 : Another Circontrol, another fail. More 6.6kw ac. A 3 hour trip takes 7 hours. This time I capture the webui of a fail event. It is a VERY different affair to either Ionity or Efcaec. This I can observe :

Circontrol does not ramp cable test. Just boom! 500v then boom 0v. Likewise precharge.
Circontrol does not mirror cable test voltage or precharge voltage over comms. We only know it via contactor voltage measurement.
If you look carefully at the video (we close contactors in state 5 and begin current ramp in 6) I see 1 amp very breifly.

Now given our "50kw" Efcaec units shut down with a "charge error" warning if we ask them for the current they claim to have available it gives me an idea : Suppose its the other end of the spectrum with Circontrol : They can't give 0 (or close to 0) amps....

This is the most fun I've had since the model 3 bms:)

[Jack Bauer]

Set out for some more testing. Ah the excitement. First stop another Efcaec QC45. Works fine. Well up to 44kw anyway:) Then onto another network who use Tritium Veefil units. Connects perfectly but will only give me 8 amps no matter how much I request. So that's weird. Sure is great we have a standard for fast charging. I'd hate to see things otherwise

[EV_Builder]

We need to know how a i3 manages at those same spots....

[Jack Bauer]

Considering the fun I've had getting logs so far that's not going to happen unless I buy one and do it myself:)

[EV_Builder]

Well I bet that either when we compare your failed logs to working ones there must be a difference. Or all i3's have the same issue. Maybe there are multiple limits send over the bus for example station/cable limits and session limits?
Maybe we can find the max. Amperage on another place in the log?
I mean we do agree; that the hole point of communication is talking about capabilities so I don't think they pass their requirement testing when telling us 50kw and failing at 44kw or maybe we are to aggressive and need to ramp up slowly and then it starts us telling new limits?

Working on the table stuff

[johu]

Veefil was reported as not working with DIY CHAdeMO controllers also

Decided to go for the money shot and asked it for the full works : 125A please...nope shutdown. Tried again, nope shutdown. Ok let's sneak up on it...40kw,41kw,42kw,43kw,44kw all go so far, 45kw Shutdown! Seems its the worst kept secret that the 50kw rapid chargers in Ireland are actually 44kw despite what they say over comms and on the rating plate. But it's all good. We're used to this interesting interpretation of standards from the Chademo era.

That happens on many chargers in Germany as well. Most only allow 113A, some just 100A. 50 kW yes, for a battery pack that accepts 113A@440V. Don't know of one...

[CCSknowitall]

There are limits that are sent at the beginning of charge, and limits are sent every current request confirmation from the station.

And yes, the standard is a bit… imprecise. Best I can tell is this: there’s flags for various limits like charger max current, power, voltage, etc. and there’s also numerical values for these things. Sometimes cars only care about the values, others use the flags, others… who knows. Same with the station caring about car values.

That cable check is probably fine. Different manufacturers produce the required voltage differently and at different voltages, anywhere from 400-500v.

For the chargers shutting down at 44kW, you’ll probably need to look at the flags for “maximum current achieved”, having the other two would be good too. I do know of one manufacturer that uses “maximum power achieved” flag even if you only hit the current limit. But, I don’t think you have them in Ireland so I will refrain from saying who they are.

Not sure what’s going on with Tritum. You likely have a parameter set on your side that nobody else has cared about yet. Or that site was under some sort of demand response event or something silly?

Audi e-tron has a pack voltage around 420-450v, and can max out a lot of lower amperage chargers. Also the Porsche Taycan has a dc/dc converter (so their high voltage pack can use older low voltage fast chargers) which can adjust its input voltage to extract nearly the maximum from a 50 kW machine.

[EV_Builder]

Considering the fun I've had getting logs so far that's not going to happen unless I buy one and do it myself:)

Or you/we rent one or someone with one drives to you or...

[Jack Bauer]

Problem is most i3 owners at least here in Ireland don't like the idea of messing with their car:)

@CCSknowitall any thoughts on my new best friends at Circontrol?

[EV_Builder]

Damien which variable is your request voltage?
Only reason I can think of 8 amps is if it's the minimum requestable.

[Jack Bauer]

CCSI_Spnt in the i3lim.cpp library. Maps directly to current request can msg to the lim. Starts at 0 and increments by one amp every 200ms if :
1)CCSI_spnt<CCSI_limit (from web ui)
2)Vbat<Vbat_limit (from webui)