[Yaric]

This is the same what I reported in my thread a week ago.

12v battery died while on level 2 charging in garage...

Come in the morning, its dead. As I was reading this forum, I was prepared, unpacked booster and car become alive within 3 minutes. Last night I plugged car at 10pm with 48% SOC using original level 2 charger with 240v connection. Based on current SOC level of 60% car died somewhere in the...

www.ioniqforum.com

If you try plug you car while 12v battery charging, charging stops. BTW, 12v charging stops even if primary battery is not in charging act due to high SOC or schedule. Don't keep your IO5 plugged.
Having additional load such as dashcam increases probability of this coincidence and full discharge as result.

 

[BrickedKeyboard]

12v battery died while on level 2 charging in garage...

Come in the morning, its dead. As I was reading this forum, I was prepared, unpacked booster and car become alive within 3 minutes. Last night I plugged car at 10pm with 48% SOC using original level 2 charger with 240v connection. Based on current SOC level of 60% car died somewhere in the...

www.ioniqforum.com

If you try plug you car while 12v battery charging, charging stops.

Thanks. Looks like you reproduced the same bug. What you maybe didn't know was that the light saying charging was on stopped, but also stopped charging in actuality. This doesn't always happen. Per my battery monitor, you can plug the car in, the light indicating it's charging doesn't come on, but it is charging the 12v. Also another user above tried what we did, and didn't get the bug. Probably this behavior only happens sometimes.

 

[jeffs_fom]

Whilst I don't disagree with the urge to blame the control firmware for this problem, since firmware issues are indeed implicated, note that in any system where there are massive amounts of current and voltage flying about (relative to the minuscule voltages and currents in the microcontrollers, canbus transceivers, and other random electronic bits) the problems you lot are seeing could very well be caused by periodic software "upsets" due to induced transients in sensitive electronic circuits. Hardening circuitry against transients from battery chargers turning on (and off), high voltage AC-DC converters and other such things is not trivial. We have enough trouble as it is in IC powered cars from things like alternator load dumps and AC compressor clutch disengagements and other transients. Not having worked on EV's meeself I can only imagine the design nightmares in trying to make all this stuff reliable. Sometimes it's impossible to reliably defend against induced transients so the firmware has to be prepared to deal with such events. Then again automotive firmware isn't known for the highest quality...

 

[BrickedKeyboard]

Whilst I don't disagree with the urge to blame the control firmware for this problem, since firmware issues are indeed implicated, note that in any system where there are massive amounts of current and voltage flying about (relative to the minuscule voltages and currents in the microcontrollers, canbus transceivers, and other random electronic bits) the problems you lot are seeing could very well be caused by periodic software "upsets" due to induced transients in sensitive electronic circuits. Hardening circuitry against transients from battery chargers turning on (and off), high voltage AC-DC converters and other such things is not trivial. We have enough trouble as it is in IC powered cars from things like alternator load dumps and AC compressor clutch disengagements and other transients. Not having worked on EV's meeself I can only imagine the design nightmares in trying to make all this stuff reliable. Sometimes it's impossible to reliably defend against induced transients so the firmware has to be prepared to deal with such events. Then again automotive firmware isn't known for the highest quality...

It's a firmware bug. I specifically specialize in embedded message passing systems. To simply it : it doesn't matter if a message got dropped because of a glitch (transient, etc). Over a 5 hour period last night, 2 things were true:
STATE: TRACTION CHARGING
STATE: 12v CHARGING DISABLE
This combination of 2 states is invalid - it's a system design rule violation.
There are a number of designs where periodic resyncs of state, or comparing the coherency of current state, or resending messages if no reply received would have handled this issue. It's totally fine if this happens transiently. For example one simple design is that the firmware that handles the traction charging could periodically poll the system and get the state, and it's state machine will send a message to request the +12V charging enable if the present state shows the charging is off.
This kind of setup where a message will be resent over and over at the polling interval makes it immune to transient glitches, microcontroller reboots, etc.

You make the traction charging controller immune to reboots by having it derive what to do - "should it be charging" - by the hardware. If it gets reset and it checks the pins and the charging connector is inserted and voltage is available, and it's not charging, it should send the messages to begin charging by the same idea.

 

[Yaric]

Whilst I don't disagree with the urge to blame the control firmware for this problem, since firmware issues are indeed implicated, note that in any system where there are massive amounts of current and voltage flying about (relative to the minuscule voltages and currents in the microcontrollers, canbus transceivers, and other random electronic bits) the problems you lot are seeing could very well be caused by periodic software "upsets" due to induced transients in sensitive electronic circuits. Hardening circuitry against transients from battery chargers turning on (and off), high voltage AC-DC converters and other such things is not trivial. We have enough trouble as it is in IC powered cars from things like alternator load dumps and AC compressor clutch disengagements and other transients. Not having worked on EV's meeself I can only imagine the design nightmares in trying to make all this stuff reliable. Sometimes it's impossible to reliably defend against induced transients so the firmware has to be prepared to deal with such events. Then again automotive firmware isn't known for the highest quality...

Lets state facts. At this forum there are a big percentage of owners, who has this issue.
As a Bolt owner and visitor of GM Bolt specific forum I can say they dont have multiply threads about dead 12V battery, even most of their cars are 3 years old.No one discussing installation of monitor for 12V battery. Its just works!

Lets admit IO5 has either a design flow or a bug. And it may be fixed, hopefully as software patch.

 

[GrantL]

Please report to hyundai if you had issues along with your findings. If there is a problem it can create a recall software patch such as with the parking pawl recall.

 

[Robocheme]

So would the temporary fix involve not AC charging the traction battery when the yellow light on top of the dash is on (indicating that the 12V battery is being charged)?
Am I correct to assume that this doesn't happen with DC charging because it involves driving to the DC station, which would charge the 12V battery?

 

[BrickedKeyboard]

So would the temporary fix involve not AC charging the traction battery when the yellow light on top of the dash is on (indicating that the 12V battery is being charged)?
Am I correct to assume that this doesn't happen with DC charging because it involves driving to the DC station, which would charge the 12V battery?

So far that's the theory. The only thing we know is there is a firmware bug and a design rule is being violated. (The design rule is : when charging traction, power +12v charger). It's possible it fails occasionally when DC charging or when the light wasn't on but so far there are no reports of that. The failure rate is low or there would be more reports. (Think I found 4 so far on these forums plus the 2 YouTubers)

 

[Kacela]

If someone knows the Ioniq Guy or Out of Spec Review's screenames

The Ioniq Guy is @Corbin

 

[Trider]

I have two really weird spikes on two days, I also had one day where the battery never got charged the whole day and got down to nearly 12.0v how and when it charges are extremely weird.

 

[alexw]

I've found some other gumminess with the 12V battery saver logic on my EV6 over here: 12V battery self-recharge plot

Long story short: the manual states that "if the battery saver+ function activates ten times in a row, it shuts off assuming there's a fault with the 12v battery. Continue driving the car normally and restore the 12V battery to a higher charge condition to clear the fault"

That second part just doesn't happen. If you leave the car long enough (a couple weeks), battery saver + will have activated 10 times, then it never activates again.

This is all in my own experience, of course - others have parked their cars for months and come home to a working vehicle. And I haven't actually had the 12V battery DIE yet, I just noticed on telemetry that the recharge cycle stopped.

Yours is definitely exciting though. Obvious bug.

 

[BrickedKeyboard]

Long story short: the manual states that "if the battery saver+ function activates ten times in a row, it shuts off assuming there's a fault with the 12v battery.

What I am seeing is:
The higher level engineers at Hyundai who know battery management probably said "ok, we need to recharge the battery whenever it needs it, but not if the 12v battery is faulty and just wasting the energy". Somehow this translated into "10 times in a row" but implicitly it should have been 10 times a row over some window of time. Your data seems suggest whoever wrote the actual code just has it count to 10 and refuse to charge again until that count gets reset by starting the car. Or maybe there is a time component but it just isn't working, it's actually kind of annoying to measure time with a microcontroller and it's easy to mess up and get it way off.

 

[alexw]

What I am seeing is:
The higher level engineers at Hyundai who know battery management probably said "ok, we need to recharge the battery whenever it needs it, but not if the 12v battery is faulty and just wasting the energy". Somehow this translated into "10 times in a row" but implicitly it should have been 10 times a row over some window of time. Your data seems suggest whoever wrote the actual code just has it count to 10 and refuse to charge again until that count gets reset by starting the car. Or maybe there is a time component but it just isn't working, it's actually kind of annoying to measure time with a microcontroller and it's easy to mess up and get it way off.

That’s my read on it as well, yeah. Although “restarting” and “charging” are both NOT reset conditions for that, which I find bananas.

It’s also possible that 3W of quiescent draw IS actually way over spec, and 10 times in a row should theoretically be months and months or something. But I’d argue the scheme is still naive even if so.

 

[BrickedKeyboard]

That’s my read on it as well, yeah. Although “restarting” and “charging” are both NOT reset conditions for that, which I find bananas.

It’s also possible that 3W of quiescent draw IS actually way over spec, and 10 times in a row should theoretically be months and months or something. But I’d argue the scheme is still naive even if so.

Wait, what. So you're telling me that if you simply leave your car parked about 1-2 weeks, you are doomed to eventually have to jump your car no matter what, even if it takes a year for the conditions to be met. Interesting.

 

[alexw]

Wait, what. So you're telling me that if you simply leave your car parked about 1-2 weeks, you are doomed to eventually have to jump your car no matter what, even if it takes a year for the conditions to be met. Interesting.

that’s true for me, and presumably you, yes. I’m getting the sense that maybe that actually IS unusually high draw. It certainly goes against other experiences where people have left the car for 3 months and come back to 4% traction battery drain and everything working fine.

 

[BrickedKeyboard]

that’s true for me, and presumably you, yes. I’m getting the sense that maybe that actually IS unusually high draw. It certainly goes against other experiences where people have left the car for 3 months and come back to 4% traction battery drain and everything working fine.

Maybe our batch have a bug in the firmware they came with. Since there's not OTA updates, the fleet isn't being kept all with the same firmware.

Or they substituted a part due to the chip shortage and, well, whoops.

 

[Rex]

I'm hopeful (but not confident) that as the manufacturing process evolves, fixes and upgrades will be introduced to the process that make later builds less susceptible to these issues.

My car was built in April so I'm keeping my fingers crossed lol.

 

[MIkeS]

Long story short: the manual states that "if the battery saver+ function activates ten times in a row, it shuts off assuming there's a fault with the 12v battery. Continue driving the car normally and restore the 12V battery to a higher charge condition to clear the fault

There is no such statement in my UK manual.

 

[ArtSpot]

Here, page 60 (1-55) in the US manual.

 

[MIkeS]

Here, page 60 (1-55) in the US manual.

View attachment 43499

No sign of that statement in my UK PDF manual.