Battery degradation - real versus BMU-real?

[Woodman411]

I know some people have been monitoring the battery degradation issue closely with Watchdog and performing various DIY BMU resets. I know the BMU is aggressively lowering SOH and that electric range gets smaller and smaller. And I know Mitsubishi has been reluctant to address this. What is not clear to me though, is if anyone is experiencing actual real battery degradation that is significant? This may seem like a simple answer - Watchdog SOH. But when users have performed DIY BMU resets, the amps have increased and SOH degrades at similar rates as before, meaning the Watchdog SOH is showing BMU-lowered SOH, not actual SOH, right? I mean, if the battery actually degraded to what Watchdog SOH is showing, a BMU-reset would make no difference, but it does.

I'm asking, because I'm still unclear on this, and if Watchdog SOH is BMU-lowered SOH and not actual SOH, I actually don't mind it - it means the BMU is protecting the high-end charge of the battery and will help the battery last longer, albeit at the cost of electric range. But then I realize the original question remains - how can one measure actual battery degradation (versus BMU-lowered SOH)? And is this actual degradation significant? A recent article in Electrek (https://electrek.co/2019/12/14/8-lessons-about-ev-battery-health-from-6300-electric-cars) showed that for air-cooled BEVs like the Leaf, there is about a 4% annual degradation average. I am curious how the air-cooled Outlander PHEV compares, but for the reasons above, I'm unsure if we are seeing on the forums real SOH or BMU-lowered SOH.

 

[ThudnBlundr]

Watchdog can't conjure figures out of thin air - it can only report what it is told by the car, which only has a single, calculated SOH figure. The only way to get an exact SOH on a battery is to perform a lengthy test by emptying the battery to a set voltage on every cell, and then charging it so that each cell reaches the same voltage, and measuring the energy that each cell has accepted. It has to be at a cell level, as balancing wastes energy. Because most people don't want to spend many hours finding out exactly what their SOH is, every car manufacturer on the planet uses BMU to measure the battery usage and calculate the SOH based on an algorithm that they have designed. The manufacturers are thankfully cautious and calculate a pessimistic rate of degradation to protect the battery, but many seem to think that Mitsubishi is being overly cautious on later cars.

On our Zoe, they changed the BMS to calculate SOH differently. Our SOH went from 85% to 97% after the reset. So either Renault changed the software to decrease the rate of decline of SOH and so reduce warranty claims, or they realised that they were being overly pessimistic and changed the software to be more accurate. I don't know which I believe

 

[KWh]

Shouldn't the reported daily charge pattern be an indication for real SOH? My overnight charge is from depleted to full every day.

I know the SOC when pluging in will differ every day, nevertheless the charge from depleted to full is always between 10.5 and 11.5 kWh. Long term the trendline should decline when SOH drops.

Or is a declining trendline in charged kWh's only an indication for the BMU SOH as well?

 

[jaapv]

Not really - the car can only charge what the BMU tells it - that does not mean that the battery has no capacity overhead left - unused.

 

[HHL]

Watchdog can't conjure figures out of thin air - it can only report what it is told by the car, which only has a single, calculated SOH figure. The only way to get an exact SOH on a battery is to perform a lengthy test by emptying the battery to a set voltage on every cell, and then charging it so that each cell reaches the same voltage, and measuring the energy that each cell has accepted. It has to be at a cell level, as balancing wastes energy. Because most people don't want to spend many hours finding out exactly what their SOH is, every car manufacturer on the planet uses BMU to measure the battery usage and calculate the SOH based on an algorithm that they have designed. The manufacturers are thankfully cautious and calculate a pessimistic rate of degradation to protect the battery, but many seem to think that Mitsubishi is being overly cautious on later cars.

On our Zoe, they changed the BMS to calculate SOH differently. Our SOH went from 85% to 97% after the reset. So either Renault changed the software to decrease the rate of decline of SOH and so reduce warranty claims, or they realised that they were being overly pessimistic and changed the software to be more accurate. I don't know which I believe

There is no way the BMU can tell how much charge each individual cell has received and balancing is absolutely necessary in Li-ion battery packs. I think the BMU is actually quite accurate, the only time it will deviate significantly from the real SOH is if the car is exclusively charged from a rapid charger or used extensively in hybrid mode without regular charging. If it is charged from the normal mains connection all the time, the battery will get balanced at every charge and in my view stuffing around with re-sets is not achieving anything useful at all. So, if an app like the dog reads 80% SOH, then you better believe it. Not sure where people get the idea that you can have a 4 year old battery that has been in regular use and it can still be over 90% of its original capacity. Tell them they're dreaming!

 

[ThudnBlundr]

You can think whatever you want, but without facts it's just an opinion along with everyone else's.

Every BMU/BMS SOH reading on every EV is an estimation over time using some software algorithm designed by the car manufacturer. Even a small error or pessimistic setting will get amplified over time. As jaapv says, the battery will only charge to the level that the calculated SOH says is OK. And it is usually only charging from 25-30% to 100% of the restricted capacity set by the BMU. The only way to measure the SOH accurately is to bypass the BMU, charge each cell from empty to full, and measure the energy consumption.

For some reason, later PHEVs seem to have worse degradation than earlier ones. Our 2015 model still shows 18-20 miles in this weather after 60k miles

 

[Woodman411]

You can think whatever you want, but without facts it's just an opinion along with everyone else's.

Every BMU/BMS SOH reading on every EV is an estimation over time using some software algorithm designed by the car manufacturer. Even a small error or pessimistic setting will get amplified over time. As jaapv says, the battery will only charge to the level that the calculated SOH says is OK. And it is usually only charging from 25-30% to 100% of the restricted capacity set by the BMU. The only way to measure the SOH accurately is to bypass the BMU, charge each cell from empty to full, and measure the energy consumption.

For some reason, later PHEVs seem to have worse degradation than earlier ones. Our 2015 model still shows 18-20 miles in this weather after 60k miles

I'm wondering if this is the steady-state point that the BMU is targeting? In USA advertising (under "range": https://www.mitsubishicars.com/outlander-phev/2020#meet-the-phev) it shows only 22 miles of range, or 35 km, which is around where you're at. Is it accurate to say that BMU "degradation" is fast in the beginning, and then tapers off to a steady state and range?

 

[jaapv]

Not my experience - range degradation has been slow and steady (judging by daily commute) down to about 27 km after 6 years and after a recent dealer reset increased to about the level it was after two years of use. (20000-25000 km/year) i.e. about 35 km on a charge (outside temperature 5-10 ºC) I'm sure I'll do about 40 km in summer @ 20 ºC. 9was 45 when new)

 

[Woodman411]

Not my experience - range degradation has been slow and steady (judging by daily commute) down to about 27 km after 6 years and after a recent dealer reset increased to about the level it was after two years of use. (20000-25000 km/year) i.e. about 35 km on a charge (outside temperature 5-10 ºC) I'm sure I'll do about 40 km in summer @ 20 ºC. 9was 45 when new)

You're only 23% down on the USA advertised brand new range of 22 miles or 35 km after 6 years, which is about 4% loss every year. Assuming this loss is linear, in another 6 years, your electric range will be 20 km (I will be curious if the loss is linear). For whatever reason (eg. outdated NEDC ratings or over-zealous marketing), a longer advertised European electric range perhaps set unrealistic expectations, even though that range could be achieved initially. Perhaps Mitsubishi learned from this and set North American expectations appropriately. Because 35 km to 27 km in 6 years doesn't seem bad at all, sounds a lot better than 50 km to 27 km

 

[jaapv]

It IS a lot better, as my daily commute is 25 km one way, and I charge at work I expect the car will meet my needs for a very long time. I also expect the rate of degradation to slow down over time.
My dealer - who has over 100 PHEVs running- confirms that my experience is typical for the car. In fact, he thinks the batteries hold up a lot better than expected when the car was introduced.

 

[HHL]

You can think whatever you want, but without facts it's just an opinion along with everyone else's.

Every BMU/BMS SOH reading on every EV is an estimation over time using some software algorithm designed by the car manufacturer. Even a small error or pessimistic setting will get amplified over time. As jaapv says, the battery will only charge to the level that the calculated SOH says is OK. And it is usually only charging from 25-30% to 100% of the restricted capacity set by the BMU. The only way to measure the SOH accurately is to bypass the BMU, charge each cell from empty to full, and measure the energy consumption.

For some reason, later PHEVs seem to have worse degradation than earlier ones. Our 2015 model still shows 18-20 miles in this weather after 60k miles

The BMU knows EXACTLY how much energy goes into a battery. The charge cut-off on Li-Ion batteries is well defined and while a battery COULD be charged to more than the nominal value, i.e. charging to 4.2V/cell rather than 4.1, the BMU will stop the charge once the current has dropped to a pre-determined value, usually around 10%, at the set charge voltage. So, at every full slow charge, it can and will indeed re-calibrate. There are of course other factors like battery age that come into play, and this is where some (well educated) guessing would be involved, but nevertheless, you seem to give a lot more credit to people who stuff around reading data off the OBD port than the people who actually designed both the battery and management system.
This car has been on the market now for over half a decade, nobody can seriously suggest that if there were a significant error in the way the battery capacity is calculated, Mitsubishi or the supplier of the battery would not have addressed that by now.
Mitsubishi have indeed replaced some batteries that have fallen well below the expected performance, it would have been much less effort (and cost) to re-progam the BMU if there was a problem there.
Applications like the DOG or EVBATMON are useful tools but at the end of the day, they can only read what the OBD port provides and give an insight into the condition of the battery, especially when you want to buy a 2nd hand car. They cannot be used to determine that there is a design issue with the BMU!

 

[KWh]

This car has been on the market now for over half a decade, nobody can seriously suggest that if there were a significant error in the way the battery capacity is calculated, Mitsubishi or the supplier of the battery would not have addressed that by now

I don't know about the battery degredation (as there seem to be well founded arguments on both sides of the spectrum). What I do know is that assuming an issue will always be addressed after 6 years on the market is optimistic at best.

My previous car (2016 Skoda Superb tdi dsg) had dsg issues that where well known since introduction of the dsg gearbox. When I bought the car it was a brand new model and I optimisticly assumed the known dsg issues would be addresses, of course this was confirmed by the dealership before buying the car. Well, the rest is history... It was so bad, I decided to take my loss and give my bookkeeper a headache (and then bought the Outlander )

Of course this was a VW AG car and not a Mitsubishi. So this won't happen again (I assumed optimisticly). The wobbly seat in my MY2020 Outlander unfortunately proved me wrong...

I'm determined to stay optimistic by character, but the assumption that issues will always be successfully addressed after half a decade turns out to be naïve rather then optimistic

 

[HHL]

This car has been on the market now for over half a decade, nobody can seriously suggest that if there were a significant error in the way the battery capacity is calculated, Mitsubishi or the supplier of the battery would not have addressed that by now

I don't know about the battery degredation (as there seem to be well founded arguments on both sides of the spectrum). What I do know is that assuming an issue will always be addressed after 6 years on the market is optimistic at best.

My previous car (2016 Skoda Superb tdi dsg) had dsg issues that where well known since introduction of the dsg gearbox. When I bought the car it was a brand new model and I optimisticly assumed the known dsg issues would be addresses, of course this was confirmed by the dealership before buying the car. Well, the rest is history... It was so bad, I decided to take my loss and give my bookkeeper a headache (and then bought the Outlander )

Of course this was a VW AG car and not a Mitsubishi. So this won't happen again (I assumed optimisticly). The wobbly seat in my MY2020 Outlander unfortunately proved me wrong...

I'm determined to stay optimistic by character, but the assumption that issues will always be successfully addressed after half a decade turns out to be naïve rather then optimistic

Yes, that is one way to look at it, however, doing a small software upgrade or even replacing a little box that is easy to get at, one would think that would be done. Certainly far less trouble to fix than the DSG.

 

[Tai626]

When I used to work with Lead Acid batteries in UPS (Uninterruptible Power Supply), we have several non invasive tools to test SOH:

https://www.cdtechno.com/pdf/ref/41_7271_0512.pdf

Are they applicable on the LiIon? Maybe. But probably not installed in the car ($ and weight). I am sure that Mitsubishi has these tools in their test bench at lab.

Tai

 

[Tai626]

Mitsubishi Motors should talk to Mitsubishi Electric and develop a trailer like this:

https://www.mitsubishicritical.com/media/1992/sa-enl0055-bcl05-lithium-ion-battery-cabinet-data-sheet.pdf?param=1577023713507

It will triple the range albeit at cost of the entire car

This 900lbs batteries bank has all you are asking for: real-time SOC, real time SOH, etc..

Just need to pass the cooling/vibration test

Joking a part, people tend to forget how far we have stretch the technology so you could have EV in land available to general public at competitive price.

The trend is spreading to boat and airplane. We just have to embrace while we can. Let the next generation figure out the SoH...

Tai

 

[Keefy]

This car has been on the market now for over half a decade, nobody can seriously suggest that if there were a significant error in the way the battery capacity is calculated, Mitsubishi or the supplier of the battery would not have addressed that by now

I don't know about the battery degredation (as there seem to be well founded arguments on both sides of the spectrum). What I do know is that assuming an issue will always be addressed after 6 years on the market is optimistic at best.

My previous car (2016 Skoda Superb tdi dsg) had dsg issues that where well known since introduction of the dsg gearbox. When I bought the car it was a brand new model and I optimisticly assumed the known dsg issues would be addresses, of course this was confirmed by the dealership before buying the car. Well, the rest is history... It was so bad, I decided to take my loss and give my bookkeeper a headache (and then bought the Outlander )

Of course this was a VW AG car and not a Mitsubishi. So this won't happen again (I assumed optimisticly). The wobbly seat in my MY2020 Outlander unfortunately proved me wrong...

I'm determined to stay optimistic by character, but the assumption that issues will always be successfully addressed after half a decade turns out to be naïve rather then optimistic

Yes, that is one way to look at it, however, doing a small software upgrade or even replacing a little box that is easy to get at, one would think that would be done. Certainly far less trouble to fix than the DSG.

The thing with planned obsolescence is that its not about the cost to fix but the opportunity cost if they don't. A number of people (Probably more than you would expect) would do as above and cut their losses and buy another car. When you expand that behaviour market wide you can work out which option offers more churn of purchases. The industry knows that on balance if you can get a few more purchases because of planned obsolescence the you all make more money on parts and service and new car purchase and finance etc etc etc.

Someone with a lot of data and a very big spreadsheet has worked out its cheaper to screw the customer than fix the problem. That math has likely been true for 6+ years. We need to push back for the greater good and tip the balance on the math.

 

[greendwarf]

Well, my 2014 PHEV with 50k miles on the clock has today just done 26 miles South to North London and back on a full charge. I imagine that many owners are similarly satisfied with this sort of performance - meaning there is little actual demand for Mitsu to change anything. :idea: