Smart Car of America Forum banner

41 - 52 of 52 Posts

·
Registered
Joined
·
275 Posts
The SOC gauge is just a voltmeter. Where is this watt-hours counting coming from?

My disconnecting the 12v battery is the whole point. The "working hypothesis" hereabouts makes no sense to me, although the fear of a self-destructive hv battery is very real. Best to find out by doing experiments but not with the 12v battery; this will always be disconnected and kept on a float charger when I'm away on a trip.

My concern with the Smart HV battery is that the BMS has direct access to the module voltage at all times and through mis-design or component failure the BMS might somehow deplete the high voltage.
--
 

·
Registered
Joined
·
1,267 Posts
Discussion Starter · #42 ·
The SOC gauge is just a voltmeter. Where is this watt-hours counting coming from?

My disconnecting the 12v battery is the whole point. The "working hypothesis" hereabouts makes no sense to me, although the fear of a self-destructive hv battery is very real. Best to find out by doing experiments but not with the 12v battery; this will always be disconnected and kept on a float charger when I'm away on a trip.

My concern with the Smart HV battery is that the BMS has direct access to the module voltage at all times and through mis-design or component failure the BMS might somehow deplete the high voltage.
--
No, resting voltage in lithium cells is only a rough SOC indication - and when under a discharge load like driving, and under varying ambient temperature, the voltage is a useless indication of SOC. SOC meters are Coulomb counters - measuring watt hours in and out of the battery pack and comparing it to the total allowable capacity of the battery pack. There are also probably "calibration points" to keep the gauge accurate - for example, resetting to 100 percent when the battery pack reached a complete full condition when charging.

The hypothesis comes from at least one case here where hte 12v battery was removed from the car for the winter, and the HV pack was foiund to be totally dischargs in the spring. But yes, a malfunction of the BMS might have been the cause and the 12V battery removal had nothing to do with it. If I leave my smart in storage, I'll keep the 12V battery connected, but I will put a maintenance charger on it.
 

·
Registered
Joined
·
275 Posts
If something is draining the hv battery the SOC gauge will show this with two different readings. The SOC gauge itself is a voltmeter albeit driven by a logic circuit somewhere, the location of which would prove interesting but not crucial to anything.

I am disconnecting my 12v battery during extended periods of non use since it is the original from 5/2015 and it is therefore a liability. I am not encouraging others to do this. I am treating Daimler's hv battery just the same as I treat Chevy Spark, Tesla, Ford C-Max, BMW i3, and Honda Fit hv batteries and so far I've observed no differences.

My 'new' auction non-salvage 2014 ED has a similar history: left sitting in driveway for 6 months and the dealer has only one option, to replace hv battery. I opened this car's battery and find 10v per module all cells in balance, so presumably it can be saved. Until Daimler builds more ED batteries salvage is our only choice to keep these cars running.
--
 

·
Registered
Joined
·
1,267 Posts
Discussion Starter · #44 ·
Oh yes, of course the physical analog meter on the dash is liekly to be jsut a voltmeter with pwm voltage signal sent to it by the drivetrain computer. I thought that it would be understood that I was talking about mesuring the SOC itself.
 

·
Registered
Joined
·
275 Posts
24 days [trip #2] - Disconnected 12v battery negative post. Since this is OLD original battery from 5/2015 I put it on float charger.
- SOC = 60 % on dash gauge.
- Leave car unlocked.

Arrive home - hook up 12v battery which is fully charged.
- SOC = 60% on dash gauge.

Resume normal daily service.
--
 

·
Registered
Joined
·
275 Posts
26 day absence. DISconnected 12v battery negative post & put battery on trickle charger. Left car UNlocked so Sentry System is OFF and not likely to drain hv battery. SOC = 98% on dash gauge.

Arrive home find 12v batt @13.0v, connect it and hv is still @98%. Test drive car with headlights ON @47*F for 7 miles. End up with 86% SOC which leads me to believe that 98% reading was a holdover. Once the car cycles thru a trip then the gauge show a truer reading, ie there was a small amount of energy dissipation in the Li-Ion batt during those 26 days (no surprise).

Please, if someone is planning to store their Smart_EV for 3 months or more, go thru this routine in order to protect your car from a Bricked Battery and also to provide long-term data.
--
 

·
Registered
Joined
·
44 Posts
Yes you are more or less on the right track. Understanding the work of the BMS and battery in detail and also understanding the nature of the failures which appear, the trick is not to drain energy from the HV cells. There are several loops which drain energy. Some are coupled. So the balancing is done by a 91 Ohm resistor. Calculation will tell you a maximum of 45mA could flow on the highest charged cells in balance mode. If there is no command by the bms the cse will stop balancing after a short time frame.

So as you investigated, the root cause is a failure and ageing of the LV battery. The configuration in central gateway unit will cause a recharge by dcdc charger after closing the contactor's by the BMS. Problem is the BMS is controlled by the LV battery again. On brown out of the LV battery the BMS starts to be damaged and retriggered to start again and again. This cause the isolation measurement loop, which is powered by the whole serial voltage of all stacks to measure the current and voltage of the internal HV battery to drain current . On brown out the power supply of this function get uncontrolled an cause the famous P18051C but also the next level the current sensor could be killed here, but also later during charging process again.
There are more failure codes which clearly indicate what get wrong and what get broken inside the battery and BMS. But the root cause is the fading LV battery voltage. A hard switch off by disconnecting the LV battery will protect the HV battery if there is not already an hw defect on BMS.

Fell free to raise questions, I can tell you the whole story and also to overcome the broken state. In meantime there is a community in Germany, where we even repair stack by installing cells to used in a Smart ED3 before. Also the getting rid off the famous BMS errors is now semi public. There are a lot of wrong information out there. Some are copied and go around and copied again. If you take a deeper look into the system you will see how it fit all together and what is fake.Too bad some really famous information channel publish wrong and misleading information.
 

·
Registered
Joined
·
1,267 Posts
Discussion Starter · #48 · (Edited)
Fell free to raise questions...
This is a question and a comment. Most BMS I'm familiar with should not be draining any cells through the shunt resistors unless the cell voltage is at some maximum value. (usually 4.2 volts or 3.7 volts for LFP cells). At least that is how the the BMS's for other types of lithium battery powered things work (like the ones in my electric motor scooters). The purpose of the BMS is to top-balance the battery pack at the end of a charging cycle. So, what else is the Smart BMS doing?

It seems that there is too much complexity here - "too many cooks spoil the broth" to use an old English expression (maybe German too).

But yes, the hard switch-off seems to be the best solution for storage. Maybe this can be made easier by rigging a "storage" switch to the 12V battery circuit.

I remain amazed that Mercedes Benz probably knew about this defect and issued no kind of recall to fix it. I assume the BMS software revision in the 2016 service campaign had nothing to do with this problem?
 

·
Registered
Joined
·
44 Posts
Hi Yinzer,
to be honest, when you see the binary of the version before (...8300) and the version which where distributed afterwards 18.14 (0401) you would see no big area with changes.

The function you mention with the wording BMS is for the Smart located on the CSE PCBs. They are supplied by the stack voltage and get their commands by the main BMS part supplied by the 12 V battery. Even there is a small power supply for handling the isocoupler for the communication. So no 12V by KL30, no communication and no control which cause a idle state and no drain of energy at all as you expect by a BMS or in this case CSE.

The voltage of the three stacks in series is used via isolated relays controlled by the 12v battery supply of the BMS main part to measure the voltages and the current. These part is normally switched off and only started and triggered on measurent condition in active state. Here all cells are loaded by a smps to generate isolated 13,5V for the current sensor and the power supply ASiC from Bosch. If the three stacks in series do not have the necessary power ( remember if the voltage drops, the current increase) all stacks get drained empty.

So know we are at the point where the main BMS part come into the story. If the 12V drop there its just a trigger and trigger in the brown out area of voltage which trigger an oscillating and restart of the CSEs but special this isolated power supply of the measurement part of BMS. This oscillating could kill some switches which stuck closed, but also cause over voltage of the smps intermediate voltage area. The occurance of over voltage condition trigger the famous locked P18051C but also the local power supply ASIC can survive up to 40V ( typical fault in classical ice ecus). The used 12V intelligent battery sensor, which is used for charge measurement via current and parallel the voltages via CSE could handle only about 20V. So he die if not unplugged in fading voltage conditions of the stacks during discharge or charging.
This all is triggered by the fading 12V. So if you cut the 12V by disconnecting the lv battery the only chance to trigger the discharge or load of the cells in the stacks is gone.

This fits with you finding of behavior, if there is not already a damage of the sticking switches on the BMS.

That's the reason I stress to mention it is not only software and to delete failure codes, which are more or less difficult to delete, but also to check the hw on failures which are caused by the brown out of the 12V. Here the overall car system architecture if the smart run in problems by secondary and following failure modes is the real reason for discharging the cells by just triggering the restart again and again. Normally the BMS will stop to load the cells, which is indicated by also a failure entry, but the play with the power button (reset of 12V in intermediate zone) will brick this protection.
Hope this description would help you to understand in more details and benchmark and link together the common knowledge against some fake informations just copied together by some blogs.
 

·
Super Moderator
Joined
·
1,738 Posts
If the car's 12v battery is getting weak and is not noticed by the owner, how does a car that gets driven often fare compared to one that sits unused for weeks or months?

Len
2014 EV Coupe 21,500 miles with a new battery six months ago
2014 EV Cabriolet 21,000 miles with the original six year old battery
 

·
Registered
Joined
·
275 Posts
Excellent discourse! You guys spent a lot of time on this which is much appreciated. My technique of DISconnecting the 12v battery when car will be unused seems to be a workable solution for up to 50+ days so far, but that is my max and I don't plan on any longer absences. 60 days or more the owner should consider 1. Taking the car with him (pulling on a small trailer) or 2. Loaning or selling the car since there is a very good chance that the hv batt will be destroyed and thus the car becomes scrap value.

My scrap value 2014 Smart ED3 will receive new Li-Ion cells and will be driven by a Prius Gen3 inverter using open-inverter logic board. This car will only need to have 30 miles range max, truly a city car. Also will be driven @30mph max so it won't much matter if Daimler disables any of the stability/traction functions etc. Indeed with such a purposely small hv batt the car will be unable to venture out of the city limits thus no opportunity to exceed 35mph.

My 2016 ED3 will continue in stock form allowing 20 mile trips into the mainland to visit the big city.
--
 

·
Registered
Joined
·
44 Posts
The funny item on this is, there is a intelligent battery management sensor installed on the 12V battery. The actual values of current, voltage and temperature but also the internal resistance of the battery should be somehow available (could be read out in the CEPC read via LIN from the sensor). So best detailed information for a lifetime prediction. But nobody really use this values.
I just start a test on the CEPC configuration about the control strategy to recharge the battery.
There are 2 options which could be optimised. I even realise the car2go config had a different setting than the others. It seems to me this values have much more influence to the battery life and sudden death than the BMS firmware update in 2016.
So the voltage level when a recharge is triggered independed of the ignition is on or connected to the grid and also on of the cycle length.

I started with normal level and standard value, but i will change to higher value and shorter cycle. I will come back with measurement results.
In meantime to disconnect the battery as proposed by Vicent and add a charger for the 12V Battery longtime storage seams to me the most suitable solution.
 
41 - 52 of 52 Posts
Top