12v battery dead for second time

This morning I found the 12v battery dead on my kona electric - this is the second time with autopi connected. Last time I bought a 12v charger to keep in the car, this got me going today (just).

I noticed the autopi wasn’t fully pushed into the socket ( it is easy to knock ) - so maybe thats an issue, not sure.

Looks like emergency power off didn’t work


The car service is due soon, so I’ll raise the issue then ( maybe its not related to autopi ). But somewhat reluctantly, I expect I’ll move over to OVMS since it seems to be much faster and has specific support for electric cars.

1 Like

Hi Peter

Sorry to hear about your battery, it is really annoying to wake up to a car that doesn’t want to start.

So it seems that the battery was drained from 14v to 5v in about 7 hours, I haven’t done the math if this is possible if the device was running the entire time, but given the missing events between the two voltage readings, it looks like the device has not been running.

That you have a reading of 4.67v does not indicate that the critical level is not working.
As you probably know, the device is by default configured to wake up from a voltage change of +0.2v within 2 seconds.
And we have seen where the battery voltage can change at random times, which is likely happening because the car will charge or run some maintenance on the auxiliary battery, which causes the device to wake up.

This trigger will also wake up the device in this situation, where the battery voltage is low, but the safety cut-out is still active, so it means that the device will only be up for 4 minutes, before it is shut down by the safety cut-out - and in this period the device will log data as usual, which is likely why you see the above voltage reading.

All the events logged on the device, should be sent to the cloud, and you can see those in the events page, but the log on the device could have more details that could indicate if the device was actually running between the last voltage reading, and the low value, or if something else has caused the issues.

So if possible, we’d really like to take a look at the log from the device, maybe both the syslog and the minion log.
Please send it to support@autopi.io - then we can investigate what has happened.

Best regards

Okay, email sent.

Looking myself I see that there are a lot of nulls in the logs around that time.

Hi Peter and Malte,

sorry to hear that. I am also afraid, but never happened to me.

I think that these defaults are bad, so I changed them:
In the power settings “Critical Level Duration” default is 180s
In the power settings “Safety Cut-out Duration” default is 240s (I changed to 150s)
It is important that this value is lower than any sleep timer period, otherwise the device can power down normally before this condition is met.

when I use “power.sleep_timer enabled=false” will the Safety Cut-out also be disabled?


Hi Peter

As mentioned in the support ticket, it did not seem that the device was running in the period from the device logged the normal voltage, to the critical level voltage was logged, so I am inclined to say that this was caused by some external factor.

Do you by any chance have the line numbers for some of the null values you mention? - I can take another look to make sure I didn’t miss anything.

I see what you mean, but the battery critical level duration is not a sleep timer directly, and both of the durations (battery critical level, and cut out) result in the same outcome, so it doesn’t really matter which one of those two durations is the lowest, but the critical level duration will also create an event in the log saying that the critical level was reached, so it is best to keep the two durations set to the default values, to retain the event creation.

With your current values, it will go to hibernation a few minutes sooner, but it will not create the critical level event, so I’d say the defaults are best.
I think we will update the description to indicate the above, because it might be a tad confusing the way it’s worded now.

Regarding your question then no, the safety cut-out is not running on the same level as the sleep timers, so it will not be affected by clearing them with the power.sleep_timer enabled=false command.

Best regards

messages line 2482. Opening in vim I see -


That’s usually how the log file will look after it has been shutdown by cutting the power, like what happens if the safety cut-out voltage is reached.

Best regards

I use critical value limit to cut, it’s not intelligent but work . You know Hyundai found many battery fault and change on warranty

I had my car serviced yesterday and specifically asked Hyundai to check the 12v battery and charging - they have reported no faults. So its not clear what this “external factor” might be.

As it happens I’ve also been trying OVMS ( https://www.openvehicles.com/ ) - different hardware approach and only aimed at electric cars, but the same general idea. Its interesting to note that they are also concerned about battery drain on electric cars since they tend to have small capacity 12v battery.

There’s something seriously off with Autopi, and i think it’s being handwaved as “external factors”, misconfigurations and whatnot.

I’ve had mid-drive crashes of the entertainment system, several in a row in one instance, even.

I’m getting messages about increased 12V battery load during driving.

I’m getting heartbeat timeouts that apparently wake the autopi from scheduled sleeps after 3-4 minutes.

With the autopi out, i’ve experienced 0 of the mentioned issues, and the heartbeat thing started only after the last few updates, after i’ve started using the autopi again for a few weeks.

This is a mess.
I’ve pulled mine out of the car.
I’m giving up on it for now. Again.
Time to look at OVMS.

I never had issues with that, but I was afraid of having it.

The last 2 weeks my car (Hyundai Kona 2019) was parked at the airport and never used. It woke up exactly every 24 hours for exactly 20 min to charge the aux battery.

I have enabled the “Aux-Battery-Saver+” in the settings of the Hyundai (read the manual).

The results are OK, but it seems that 20min charging is not enough. The Aux_Bat_SOC increased about 4% during charge, but lost more during rest of 24h. So I got from 92% to 61% in 2 weeks standstill. That´s fine, but now I know that I would not leave the car unused for more than 2 weeks. I don´t know, if it is an AutoPi-issue, but it may be.

Aux Bat SOC in %
“2019-11-26T04:20:00.000Z” 89.0
“2019-11-26T04:30:00.000Z” 91.0
“2019-11-26T04:40:00.000Z” 92.0
“2019-11-27T04:40:00.000Z” 70.0
“2019-11-27T04:50:00.000Z” 72.5
“2019-11-27T05:00:00.000Z” 74.0
“2019-11-28T05:00:00.000Z” 69.0
“2019-11-28T05:10:00.000Z” 71.0
“2019-11-28T05:20:00.000Z” 73.0
“2019-11-29T05:20:00.000Z” 69.0
“2019-11-29T05:30:00.000Z” 71.0
“2019-11-29T05:40:00.000Z” 73.0
“2019-11-30T05:40:00.000Z” 67.0
“2019-11-30T05:50:00.000Z” 68.5
“2019-11-30T06:00:00.000Z” 70.0
“2019-12-01T06:00:00.000Z” 64.5
“2019-12-01T06:10:00.000Z” 66.5
“2019-12-02T06:20:00.000Z” 63.0
“2019-12-02T06:30:00.000Z” 65.5
“2019-12-03T03:40:00.000Z” 63.0
“2019-12-03T03:50:00.000Z” 65.0
“2019-12-03T06:40:00.000Z” 65.0
“2019-12-03T06:50:00.000Z” 66.5
“2019-12-04T07:00:00.000Z” 61.0
“2019-12-04T07:10:00.000Z” 63.5
“2019-12-05T07:10:00.000Z” 58.0
“2019-12-05T07:20:00.000Z” 59.5
“2019-12-05T07:30:00.000Z” 61.0
“2019-12-06T07:30:00.000Z” 60.0
“2019-12-06T07:40:00.000Z” 61.5

see also my script for power management:

1 Like

There is an interesting discussion about the e-Niro and OVMS here -https://github.com/openvehicles/Open-Vehicle-Monitoring-System-3/issues/275#issuecomment-559045312 worth a read.

The thinking is that if such devices as autopi/OVMS are polling ECU’s whilst the car is off, it can stop the ECU’s from going to sleep and therefore consume more power.

Hi Peter

Ah yes, that could explain some of the issues you have seen as well - thank you for posting the link.

It could be interesting to see if the device, without any enabled workers, still causes the vehicle to drain extra power, and/or if some specific PID’s are the culprit, or even if it’s when the device is simply connected to the OBDII plug.

@Manfred - What loggers do you have configured?

Best regards

2 loggers - 5sec intervall
9 loggers - 300sec intervall
4 loggers - 3600sec intervall
2 cronjobs with each 1 min intervall

Hallo Malte and Peter,
i should also report the dead battery on my Car. It hasnt been driven for 22 Days, and the voltage reached 3V in the end! I have recaived a new battery as the old one turned itself unusable, because of the critical discharge. Now, as the dongle was connected all the time and had access to WIFI, i assume to see some logs there. Can you help me out on how and where should i look for the logs both local and in the cloud? Im not blaming the dongle yet, as it needs to be proven first if it was the reason…