[Ovmsdev] Another incident

[Chris Box]

On 2025-02-08 22:22, Dan Edwards wrote:

> If you google I-key system error , you'll find posts on issues with the 
> BCM, which have been resolved with cleaning contacts on bcm (if there's 
> been moisture issues), and others replacing 12v acc battery.

Thanks for the pointer. I drove the car for a year without OVMS and 
didn't have an i-key error then.  Did you experience switching to 
neutral while driving?

> Check your DTCs with leafspy, if it shows a bunch of DTCs including a 
> B2604 BCM Shift PN diag, then likely low 12v, try a 24 hr charge or 
> replace it.

I've now ordered a dongle to work with Leaf Spy.

On 2025-02-09 09:23, Michael Balzer via OvmsDev wrote:

> IOW, Chris, you should look for "dip" log entries well within a parking 
> / driving / charging phase, when the 12V level should be more or less 
> constant.

I've checked the logs for the 7th, and below is the one such dip I can 
see. It occurs towards the end of the drive home.

2025-02-07 13:51:03.584 GMT D (468184) events: Signal(vehicle.awake)
2025-02-07 13:51:05.424 GMT D (470024) vehicle-poll: Pollers: Queue 
SetState()
2025-02-07 13:51:05.424 GMT D (470024) vehicle-poll: Pollers: 
PollState(1)
2025-02-07 13:51:05.424 GMT D (470024) events: Signal(vehicle.on)
2025-02-07 13:51:06.464 GMT D (471064) events: 
Signal(vehicle.charge.12v.start)
2025-02-07 13:51:08.384 GMT D (472984) events: 
Signal(vehicle.aux.12v.blip)
2025-02-07 13:51:11.114 GMT D (475714) vehicle-poll: Pollers: Queue 
SetState()
2025-02-07 13:51:11.124 GMT D (475724) vehicle-poll: Pollers: 
PollState(2)
2025-02-07 13:51:11.124 GMT D (475724) events: 
Signal(vehicle.gear.forward)
2025-02-07 13:51:16.394 GMT D (480994) events: 
Signal(vehicle.aux.12v.charging.blip)
2025-02-07 13:51:56.394 GMT D (520994) events: 
Signal(vehicle.aux.12v.charging)
2025-02-07 13:56:50.754 GMT D (815354) events: 
Signal(vehicle.headlights.on)
2025-02-07 13:56:56.904 GMT D (821504) events: 
Signal(vehicle.headlights.off)
2025-02-07 13:56:58.964 GMT D (823564) events: 
Signal(vehicle.headlights.on)
2025-02-07 13:57:05.164 GMT D (829764) events: 
Signal(vehicle.headlights.off)
2025-02-07 14:01:54.384 GMT D (1118984) events: 
Signal(vehicle.aux.12v.charging.dip)
2025-02-07 14:02:00.384 GMT D (1124984) events: 
Signal(vehicle.aux.12v.dip)
2025-02-07 14:02:30.384 GMT D (1154984) events: 
Signal(vehicle.aux.12v.normal)
2025-02-07 14:03:36.394 GMT D (1220994) events: 
Signal(vehicle.charge.12v.stop)
2025-02-07 14:03:46.394 GMT D (1230994) events: 
Signal(vehicle.charge.12v.start)
2025-02-07 14:04:06.394 GMT D (1250994) events: 
Signal(vehicle.charge.12v.stop)
2025-02-07 14:04:09.954 GMT D (1254554) events: 
Signal(vehicle.gear.reverse)
2025-02-07 14:04:16.394 GMT D (1260994) events: 
Signal(vehicle.charge.12v.start)
2025-02-07 14:04:28.064 GMT D (1272664) events: 
Signal(vehicle.gear.forward)

At 14:04:29 I connected to home wifi.

> You should also have a look at the general 12V level behaviour, e.g. 
> using the 12V chart plugin or creating a chart from the MQTT metric.

I recorded the attached 12v profile on the 7th via MQTT. It was 12.2 at 
13:17, and briefly 11.4 at 13:43. The 14:02 dip is not visible. My 
battery was new in November 2023.

> When using V2/MQTT, consider raising the update interval to get a 
> higher time resolution.

What interval is sufficiently frequent? This is what I have set.

   updatetime.charging: 60
   updatetime.connected: 300
   updatetime.idle: 600
   updatetime.keepalive: 1780
   updatetime.on: 120

> Btw, also consider shortening the file logging sync period, 30 seconds 
> is rather high.

I've now set log flush to 1 second.

> On the 12V shutdown: even with 30 seconds log flushing, you would have 
> seen the log entries for a shutdown condition, unless you configured 
> the shutdown delay to 0 (default is 2 minutes).

I have no specific values set in vehicle.12v.*, so I'm using default 
behaviour. Which I think disables auto-shutdown.

> That is, if the 12V level wasn't suddenly getting drastically low, i.e. 
> too low within seconds to keep the ESP32 in operation. That could 
> explain the freeze. Not sure how low that would need to have been. I'd 
> check for another potential source of such a sudden voltage drop: a 
> mechanical/electrical issue with the OBD connector or cable.

How would the V3.3 hardware react when power is completely removed (12V 
goes to zero) and then restored (back to 12V)? I would hope it treats it 
as a cold start, and begins to boot up. But I didn't see that in the 
log. This also raises the question of capacitance to ride out brief 
interruptions. It appears there is 470uF between 12V/5V and GND, if I'm 
reading it right. I'm not sure what time period that could supply power 
for.

> If for example there is a lose wire in the connection, that 
> occasionally causes a short circuit, that would both explain the module 
> freeze and the car fault.

It would. In the Leaf the OBD port is plastic mounted on plastic, and 
not particularly robust. I haven't seen inside the cable.

Chris
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openvehicles.com/pipermail/ovmsdev/attachments/20250210/03228977/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 12V.png
Type: image/png
Size: 11662 bytes
Desc: not available
URL: <http://lists.openvehicles.com/pipermail/ovmsdev/attachments/20250210/03228977/attachment-0001.png>