12V battery

[hobbit]

Patent US9413182B2 seems to be about the "battery care" option, and to
me helps indicate why the Kona's implementation does more harm than
good. I'm still slogging the others ... Thanks for those, I'd forgotten that
going after a carmaker's patents is often really useful and does much to
answer "what were they thinking". But sometimes the theory gets poorly
put into practice...

_H*

 

[KiwiME]

Noting the reference to the "IBS information", intelligent battery sensor, regarding the item on the negative post, the evidence shows that it's not just a current shunt. There's a "EPCU" comms wire to it aside from +12V which I think it also might use as voltage data. It clearly supplies its own ground.





Some info on this type of item:
https://www.batterypoweronline.com/...-to-the-success-of-future-automotive-designs/
https://www.hella.com/hella-sg/assets/media/J01213_Intelligent battery sensor (IBS) 12 V.pdf

Gotta love this part regarding customer-attached loads such as a dashcam:

1. The startup performance of a vehicle and the durability performance of a sub-battery can be secured through the periodic charging of the sub-battery when the vehicle is left unattended for a long time. For example, discharging of the sub-battery due to an increase of private electronic parts such as a black box, i.e., the increase of a vehicle dark current, can be prevented, and the durability performance of the sub-battery can be improved due to a reduction of a deep discharge.​

The management of this "sub battery" is way more complicated than I thought it might be and a challenging read.

 

[hobbit]

... Except that the AMS fuse is the thing on the *positive* terminal block,
not the shunt on the negative side [which we already know isn't documented
anywhere someone's been able to find. Your connector above appears to
be for remote voltage feedback sensing that goes back to the LDC, so
it knows what's going in case there's a voltage drop from high currents
between LDC and the battery. The Prius has such a thing too, more
necessary there since the main LDC wire runs the entire length of the car.

Hyundai definitely *way* overthought this, and I think the implementation
is going to come back and bite them long-term. That patent is someone's
EE thesis gone horribly wrong. On the bright side, they do get credit for
liberally instrumenting the system.

I did some interesting testing today, proving that the current shunt really
is one and really produces that 7E2, 2102, W/V figure. Depending on
which side of it I connect the negative of a heavy load [charging one of the
Yeti boxes, about 30A once it ramps up] the OBD2 either "sees" it or doesn't.
LDC current increases in either case, to make up for the additional load.

_H*

 

[hobbit]

Oh wait, I get it now, that + wire from the AMS fuse must go over to
the negative-side module, if it's the architecture shown in the PDF,
but it is sent deep into the harness to make the trip over. Hmmm.
LIN data coming out of it?? Fascinating, maybe it's time to get a
scope on that sucker.

Yesterday I pulled that connector with the car powered up. No
errors popped up, no DTCs were thrown, but the voltage settled
to 13.5 or so. The LDC must have a "fallback" strategy...

_H*

 

[hobbit]

Today I got a scope on the battery sensor data lead, and once I convinced
my scope to stop *crashing" while trying to do a bus decode, got a few
reasonably good captures. It's indeed LIN at 19.2 kbaud, just like you [Kiwi]
found. Thanks for inspiring this investigation, btw.



The grey lead does indeed seem to go over to the AMS fuse at the positive
terminal [ohmed out]. The sensor draws 10 mA when active and 1.4 mA when
asleep. In the absence of bus activity, or below some threshold current, it
goes to sleep in about 15s but wakes up briefly about once a minute,
probably to send some "I'm here" packet. The data line normally sits at
the 12V level, so it's probably a simple pull-down transmitter.

_H*

 

[KiwiME]

That is all very interesting and the system certainly seems to have all the information it could want. If we can just figure out what it's doing...

Meanwhile, data from a short drive, similar to other data, noting that after 10.5 minutes of initially charging the aux battery, the system eventually just seems to leave it alone, ensuring that no significant current passes in or out - I can only assume by precisely regulating the system voltage to suit, despite the noisy current draw by 12V peripherals.

But in between is 4 minutes of weirdness, aux batt discharge from about 630 to 1080 seconds with very stable voltage. Could it be perhaps an SoC check?


Numbers are seconds, ticks are minutes.

 

[KiwiME]

Hobbit, if you haven't seen this, a relevant video from John Kelly (Weber Auto channel) on the Prius DC-DC. The only significant difference is that the charge voltage is a bit lower. At the end he mentions the two main modes of operation we're seeing.

 

[hobbit]

Clearly, I need to watch more of the Weber videos ... although the, ah,
stateliness of their pacing is only outdone by Jack Rickard's stuff.
The nuggets require a lot of patience to obtain...

Anyway, his tests and observations are almost identical to what I was
trying. What he missed was holding a couple of power window switches
up after they're closed ... that brings locked-rotor current of the window
motors into play, but you don't want to do that for too long. I got my LDC
current close to 100 amps in short bursts that way.

_H*

 

[Toolworker]

Clearly, I need to watch more of the Weber videos ... although the, ah,
stateliness of their pacing is only outdone by Jack Rickard's stuff.
The nuggets require a lot of patience to obtain...

If you view it on YouTube, click the settings and increase the playback speed - without changing the pitch!

Or click the "..." underneath the video. One of the options opens an automatically generated transcript on the right. Clicking on a line in the transcript jumps the video to that point.

Thanks for motivating me to look for these valuable features.

 

[KiwiME]

Back to the graph I posted above and noting the section from 630 to 1060 seconds which is while driving and directly after the charging event at around 14.7 V, it seems that the LDC imposed a controlled discharge on the aux battery at constant voltage, 12.7 V, something it can easily do with the help of the IBS on the ground lug. Clearly the discharge current is used to run the car's 12V systems, so at first the LDC contributes nothing to the car itself and then slowly comes back on line.

I added a curve for the cumulative discharge (x10 for clarity) energy where the peak is 3.16 Wh. That's just current x voltage integrated over the duration, specifically each second contributes VxA/3600 to get a total Wh of 3.16

What is this for?

So, next I worked out the same for the charge event that started a few seconds before TP kicked in and logging commenced, from 1 to 630 seconds. Again, x10 for clarity plus inverted to positive so it didn't mess up my graph, a total of 3.15 Wh. The OMG moment is that they are the same values.
I'm going out for a coffee now. When I get back I expect someone to explain to me what's going on

 

[hobbit]

Well, I imagine that what went on is that you ingested coffee, and
presumably had to give it back later on in roughly equal volume
minus small losses here and there. Kind of like battery charge.

_H*

 

[KiwiME]

Well, the battery didn't really have to give it back and it normally doesn't since the LDC supplies all 12V system power. I can only assume that the LDC is testing the battery's performance as normally it leaves it alone after a charge. Today I will measure the energy added during the timed aux battery saver event which is scheduled to start at 1:10 pm and normally runs for 15 min.

 

[KiwiME]

Here is the cumulative energy (in blue) delivered during an "aux battery saver" event measured on my stock Kona EV with TP at 21°C. The drop off at the end is when I opened the door to retrieve the phone and unplug the OBD dongle. Charge voltage is 14.56 V.
To put this energy in perspective the 9.7 Wh charge power would cover a current drain of a 27 mA continuous over the 24 hour period, matching closely the sleep current that has been previously measured by others. That calc is 9.7 Wh x 0.85 / (12.5 V x 24 h) = 0.027 A, where battery efficiency is taken as 85%.
I can only assume that the charge duration is contingent on the charge current reaching zero.
In theory the 45Ah battery should be able to supply almost 2 A continuously during the 24 hour gap but if the voltage drops below the acceptable minimum level required to support the electrical system, the aux battery saver won't save you.

Ticks are minutes, numbers are half seconds.

 

[KiwiME]

Documenting the schedule of the aux battery saver, for what it's worth.
There is no traction battery charging at any time during this test.
It starts on 9 Mar based on 24 hours after the end of the last trip the previous day, about 12:40.
It ends about 15 min later.
The small jiggles are from moving the car slightly, checking the charge level or limiter setting, or opening a door.
It continues every 24 hours after the end of the last aux battery saver charge event, here 12:55

In this case I opened and closed the door just before it started to insert the OBD dongle

In this case on 12 Mar I made two trips, about 5 min each at 9:00. The aux saver continued on the same
schedule, presumably because the trips were not long enough to complete a full battery charge.
Had they been sufficient no charge would have occurred today and the schedule would have certainly
changed to about 9:30 starting the next day.

And today, the 13th it continues on the same pattern,
noting that the battery voltage stays in the range 12.8 to 13.0.

 

[hobbit]

So it doesn't seem to impose any extra load, however small, in between
top-ups, i.e. overhead to keep track of the timing? I wonder what your
resting quiescent drain is at full rest, and why people seem to be having
so much trouble with this. But it may also be the "bluelink problem".

If you go to measure the quiescent with an ammeter in series, the hood
latch needs to be clicked closed with a screwdriver shaft or something
or the car may never fully go to sleep. The switch is in the latch mechanism.

_H*

 

[Toolworker]

It was my turn in the barrel today. Tried to turn on my one month old car and - stone cold dead, with the 12v. battery measuring 2.8v. And 64 kWh in the main battery, but not a drop to drink. What's wrong with this picture? Oh, yes, the Aux Battery Saver was enabled. Fat lot of good that did.

Hyundai Roadside Assistance promptly sent the AAA, who "jump started" the car with a portable pack. My first drive was to buy one of those suckers for myself. Never again.

When the car came back on, the climate system started running. I wonder if I inadvertently started it up when I was testing the Bluelink app to see if I had the problem everyone else is complaining about. I don't think I turned climate on, but that app has a mindless of its own. Could that have kept running until it drained the battery? The car was sitting for three days, plugged into an EVSE that's enabled from 11 pm to 10 am.

Should I replace the 12v battery, since it was completely drained?

The 12v battery was dead again this evening. The service tech was totally clueless after the first incident, so I'm not hopeful about taking it to him again. Any suggestions?

It had been sitting for two days, plugged into the EVSE that, as I mentioned, is enabled from 11 pm to 10 am. Fortunately, I now have that little 12v jumper pack and was able to start up. When it came on, the heat was on, just like the first time.



I wonder if something is turning the heat on and then it drains the 12v battery. I did use the app but only to check the charge on the home screen. Never used the climate control.

I thought, "Great! My neat BM2 battery monitor will show me exactly what happened and when." Well, no such luck. Looks like when the 12v supply goes dead, the monitor loses its stored data. Guess they never heard of flash memory. Now I know why they call it a BM number 2.

Funny enough I installed a CTEK bluetooth battery monitor a few days ago. I got up this morning and it tells me battery is 5.8V and its -22C.

Do you know if your CTEK retains data when the 12v battery is completely dead? Is there another monitor that will do that?

 

[Toolworker]

I wonder if something is turning the heat on and then it drains the 12v battery. I did use the app but only to check the charge on the home screen. Never used the climate control.

After playing with Remote Start in the app, I see that the 12v battery is being charged while climate control is running remotely, and it times out and turns off after about 12 minutes. So I can't duplicate this using the app.

I changed the Remote Start temperature in my app from 72 to 69. The next time the battery goes dead, I'll see if the climate control comes on at 72 or at 69.

Anyone have any ideas about what's causing this?

 

[Toolworker]

Would that charge from Aux Battery Saver at 15:00 be enough to give a healthy 12v battery that much of a boost?

More about the discharge from 3:00 to 8:00 is posted here.

 

[KiwiME]

You can see that the resulting end voltage from the charge is less than it was before being drained after 3:00 hours, so "no". The charge event not only seems to oddly ramp (current limiting?) from 14.0 to 14.6 ish over the 20 min approx duration but it seems that it doesn't take into account the reduced starting voltage or the clearly-lower SoC by means of extending the duration.
Unless the aux battery saver starts another event before 24 hours is up (bet it won't) there is no way that's sustainable over several days and therefore that EVSE issue is likely the reason for the occasional failure.

 

[Toolworker]

Does Aux Battery Saver dis

You can see that the resulting end voltage from the charge is less than it was before being drained after 3:00 hours, so "no". The charge event not only seems to oddly ramp (current limiting?) from 14.0 to 14.6 ish over the 20 min approx duration but it seems that it doesn't take into account the reduced starting voltage or the clearly-lower SoC by means of extending the duration.
Unless the aux battery saver starts another event before 24 hours is up (bet it won't) there is no way that's sustainable over several days and therefore that EVSE issue is likely the reason for the occasional failure.

Thanks for your reply.

What I was specifically wondering is, has the 12v battery been damaged by several full discharges? Looking just at the 20 minutes of aux battery saver charging, it raised the voltage from roughly 12.1 to 12.8. Would it be able to do that with a healthy battery?

Looks like when the 12v supply goes dead, the monitor loses its stored data.

Does anyone know a way to record the BM2 battery monitor data in the event the 12v battery goes dead?

I tried leaving a spare smartphone on a charger next to the car, with the app running, the phone in Stay Awake mode (developer option) and the graph displayed. But the app doesn't record the data from the monitor continuously - you have to go back to the home screen and then to the graph, and only then does it update.