SLAC matched, but no SDP answer or TCP connection
The modems paired, yet the car never finds the charger on IP or cannot open its TCP connection: a network key, IPv6, discovery or listening socket problem.
What you see
- CM_SLAC_MATCH.CNF is in the capture, so SLAC passed.
- Then either SDP requests from the car to ff02::1, UDP port 15118, repeated without an answer, or an answer followed by a TCP SYN that nobody acknowledges, or a SYN answered with a reset.
- The stage strip stops at SDP or TCP / TLS. The verdict is Never charged with SDP failure (the charger never answered) or TCP failure (no connection to the announced address, a SYN without SYN-ACK, or a reset before the first V2G message).
Likely causes
| Cause | Who | How to confirm in the capture |
|---|---|---|
| The modems never formed one network: a host did not program the network key from the match into its modem | Either | On a sniffer next to the car, the car's SDP requests and neighbour solicitations get no traffic at all from the charger side |
| The charger's SDP server is not listening on the powerline interface (bound elsewhere, started late, firewall on UDP 15118) | EVSE | SDP requests arrive correctly addressed and nothing answers; SD-01 fails, SD-02 counts the retries |
| Multicast is dropped: the interface does not accept ff02::1, or a bridge in the charger filters it | EVSE | As above; a unicast ping to the charger's link local address works while SDP does not |
| IPv6 not set up on the link: no fe80:: address, or duplicate address detection still running | Either | The SDP request comes from a source that is not link local, or neighbour solicitations for the same address come from both sides |
| SDP messages malformed: wrong port, wrong length, or a security byte other than 0x00 or 0x10 | EV or EVSE | SD-05 names the rule that was broken |
| The car repeats too fast or gives up too early | EV | SD-06 checks the 250 ms minimum wait and the 50 request limit |
| The answer names an address or port nobody listens on (a global address, the wrong interface, a stale port) | EVSE | A SYN to the announced address with no SYN-ACK or a reset; TC-01 fails. SD-03 fails when the car connects somewhere else |
| Security mismatch: the car asked for TLS and the charger answered without, or the other way round | Either | SD-04, and the security byte of both SDP messages in the Message tab |
What to check in this tool
- SDP and TCP/TLS lanes of the timeline: requests, the answer, the SYN and what followed.
- The SDP rows of the message table: the Message tab shows the requested security and transport, and the announced address and port.
- Checks tab: SD-01 to SD-06, TC-01, TC-03 for retransmissions on a weak link, and TIM-04 for the total time from the first SDP request to SessionSetupRes.
Fixes that usually work
- Bind the SDP server to the powerline interface and make sure it runs before SLAC finishes.
- Enable IPv6 on that interface with a link local address, and let duplicate address detection complete before answering (or configure it to be fast).
- Open UDP 15118 and the announced TCP port in the charger's firewall.
- Announce the link local address of the powerline interface, and have the TCP server listening on the announced port before the SDP answer goes out.
- Check that both hosts set the network key in their modem after the match, and that the modem accepted it.
- If the car asks for TLS, answer with TLS if the charger supports it. A car may refuse to continue without it.
The SDP response timeout (sdp.responseTimeoutS, 5 s) and the communication setup timeout (tim.communicationSetupTimeoutS, 20 s) in the profile are flagged for verification.
Sources
- ISO 15118-2:2014, SECC discovery protocol
- ISO 15118-20:2022, V2G20-137 to V2G20-161, V2G20-081
- RFC 4291 and RFC 4862, IPv6 link local addresses and duplicate address detection