openinverter forum

I recently got myself a new KIA EV6, and I'm fascinated with the inconsistency of ultrafast public DC charging in my country. I started logging my charge sessions with CarScanner.

I have been reading pyPLC project github and this forums, it is probably best open source of knowledge of how CCS protocol really works. This might be off topic but I would like to ask your experience with public DC chargers, if you can explain or at least guess what might be going on in the following sessions.

alpitronic Hypercharger

https://blu.oook.cz/ev6/20230723_nyrsko_pre_75
This is the simplest possible session, just works at the nameplate power level.

https://blu.oook.cz/ev6/20230716_cejkovice_eon_75
This station is labeled 75 kW and most probably have single power unit, yet it advertises 98 kW in early dialog. Once the charging starts, it corrects to exact 75 kW. Is the 98 kW administrative power limit of the whole unit, set according to the grid connection, allowing for simultaneous CCS and AC 22 kW? And once the power delivery begins, the unit finds one power unit and thus sends 75 kW maximum?

https://blu.oook.cz/ev6/20230716_pre_mo ... mlov_20-75
This is another HYC_075 but limited to exactly 20 kW. Administrative limit? Hotline assured me it is supposed to be 75 kW and is not aware of any limitations (well well).

https://blu.oook.cz/ev6/20230720_hodonin_omv_pre_75-150
Now this is confusing. There are two power modules, and the cable is 400 A in reality, but only recently upgraded, so probably not configured yet in the station. Nevertheless, it sends 150 kW in early dialog, and once it starts charging it swings wildly up and down (84 kW actually displayed on instrument cluster at one point) until it settles at 75 kW. What is going on here?? This is a two stations installation but the second station was not in use at the time, so there should be no load balancing.

ABB Terra 500

https://blu.oook.cz/ev6/20230721_vyskov_eon_175
Here is a session on a single cabinet ABB Terra 500, ie 175 kW. Why it decided to limit to 164 kW after init? Is it 175 kW is switching capability of installed hardware as provisioned, but on charging start it detects one module faulty? Also all HV ABB units report 908 V while nameplate has 920 V. Need to do more sessions on these single cabinet 175 kW installations.

Ionity

https://blu.oook.cz/ev6/20230723_beroun ... _rapidgate
Tritium Veefil-PK, four dispensers, one power cabinet, Ionity site. Sends 256 kW power limit, then it replaces the value with calculated car max current*actual voltage, that is normal. Then couple times during the session it limits to 167 kW and then allows full power again. This coincided with three dispensers being in use. When two cars were charging, 256 kW. Three cars charging, 167 kW. I would expect more fluid site power sharing from Ionity, this feels like strict thirds out of 500 kW site limit, maybe? (At the end of the session the unexplainable 3 minutes of EV6 requesting minimal draw. Sometimes it does it. Not a charger fault.)

Also, is there a way to sniff CCS protocol from EV6/e-GMP cars?

Regarding the behavior of the chargers, I think only the manufacturers / application engineers can explain what is going on. I agree, that the behavior is often strange and surprising.

lucistnik wrote: ↑Tue Jul 25, 2023 6:37 pm Also, is there a way to sniff CCS protocol from EV6/e-GMP cars?

I do not have the details, but let me assume: Yes, but maybe it's a longer way. It could look like: Find the module where the powerline modem sits. Hope, that a QCA7005 is there (if not, provide some fotos of the used chipset). Find on the PCB of this module the test points of the SPI between the modem and the controller. Add a Saleae logic analyzer to the SPI. Using the Saleae tool, record an SPI log during the charging session. Export the SPI trace. Use the https://github.com/uhi22/Ioniq28Investi ... rpreter.py to convert the SPI trace into an pcap trace. This contains the traffic on ethernet level, and can be inspected with wireshark. Use https://github.com/uhi22/pyPLC/blob/mas ... nverter.py to decode the pcap, to see the content of the EXI messages. And finally use https://github.com/uhi22/pyPLC/blob/master/scope.py to plot the signals graphically.
This exercise is explained here: https://github.com/uhi22/Ioniq28Investi ... i-analysis

Don't know how you obtained the data for each EVSE, but if I understand it correctly the under current column you have "target" which is set by the EV and "delivered" which is reported by the EVSE. From this, it is look like all chargers that you collect data from seem to behave as it should. As I can see that when EV set target for current, the EVSE try to match it accordingly. Why does the EV not set target like that, is another complex story (SOH, temp, user setting, ....)

CCS charging work based on mutual agreement of both sides (EVSE & EV). At the beginning, both side announced its own limit and expect the other to respect that. During charging, EV set request and EVSE will try it best to meet the request, if external factor affect the EVSE delivery capability (such as beig throttle by grid) it will announce the new limit accordingly.

If you manage to obtain the pcap file for charging session, look for:
- ChargeParameterDiscovery (Req/Res)
- CurrentDemand (Req/Res)

they are the actual dialog between EVSE and EV that carry parameters I mentioned.