BMU Cell voltage smoothing

[HHL]

The vehicles sold here initially were supplied with an EVSE that had a 15A plug, the actual maximum current draw of this device was set at 10A!
Later, Mitsubishi supplied an EVSE witha 10A plug and this one allowed a maximum draw of 8A, so it takes longer to charge.
The reason they would have done this would be because most people do not have 15A sockets and the added expense of installing one deterred some buyers, so they opted to supply the 10A plug EVSE.......

OK, that's new to me. I thought all EVSEs coming with the car have a 15A plug here in AU?!
Anyways, the 15A plug was the first thing I replaced with a 10A plug so I could use it everywhere without an Ampfibian adapter...

I think even the new model now is supplied with the 10A plug, which only charges at 8A max, hence the longer charging time quoted in the specs..
So, are your cells still smooth ?? :lol:

 

[jaapv]

The procedure you are looking for is not the 'smoothing' one but the 'Capacity and Control Reset', and its the same used on i-MiEV battery pack to reset the history of the Control information for the BMU module.

You write this in response to who / what?

The procedure you describe IMHO is pure administrative (like resetting your MPG number) and should take no more than 2 minutes. The procedure performed against my car actually encompasses charging and discharging the battery at a very low pace and takes several (12?) hours to complete.

Another long lecture by my dealer Battery smoothing changes nothing in the battery itself. The car is discharged overnight to real zero using the air-conditioning (and a hidden menu setting) Afterwards it is recharged and the charging capacity of each individual cell is measured. These data are used to reset the charging algorithms for each individual cell to ensure that they will all take the same level of charge, resulting in a higher capacity of the complete pack. It is advised to do so every 100.000 km.
BTW, the car will indeed use some of the spare capacity to maintain range as the battery goes down over time.

 

[HHL]

The procedure you are looking for is not the 'smoothing' one but the 'Capacity and Control Reset', and its the same used on i-MiEV battery pack to reset the history of the Control information for the BMU module.

You write this in response to who / what?

The procedure you describe IMHO is pure administrative (like resetting your MPG number) and should take no more than 2 minutes. The procedure performed against my car actually encompasses charging and discharging the battery at a very low pace and takes several (12?) hours to complete.

Another long lecture by my dealer Battery smoothing changes nothing in the battery itself. The car is discharged overnight to real zero using the air-conditioning (and a hidden menu setting) Afterwards it is recharged and the charging capacity of each individual cell is measured. These data are used to reset the charging algorithms for each individual cell to ensure that they will all take the same level of charge, resulting in a higher capacity of the complete pack. It is advised to do so every 100.000 km.
BTW, the car will indeed use some of the spare capacity to maintain range as the battery goes down over time.

That's good info. Nice of your dealer to explain it. Seems they are a bit more "switched on" than the ones we have here......

 

[anko]

Another long lecture by my dealer Battery smoothing changes nothing in the battery itself. The car is discharged overnight to real zero using the air-conditioning (and a hidden menu setting) Afterwards it is recharged and the charging capacity of each individual cell is measured. These data are used to reset the charging algorithms for each individual cell to ensure that they will all take the same level of charge, resulting in a higher capacity of the complete pack. It is advised to do so every 100.000 km.

I got this document, that says prety much the same. Unfortunately, it is in Dutch:

BTW, the car will indeed use some of the spare capacity to maintain range as the battery goes down over time.

This is in line with what I experienced:
- For a long time the Ah number went down rapidly where the predicted and actual range were hardly affected.
- Full is still shown as 100%
- ICE engagement is still around 25 -30%
- Minimum value for Remaining Capacity Ah (shown by EvBatMon of PHEV Watchdog or my own tools) is much lower than before (< 8 Ah)

Based on my readings I have come to the conclusion that %SOC is defined as Remaining Capacity Ah / Current Capacity Ah (where the latter is the number from the health report, some of us receive after service)

What I do understand, however, is that my battery has been treated twice. Both times gain was little below 2 Ah and both times gain was gone within a month and a half. So, what is real then?

 

[HHL]

Hi Anko,

From what I understand looking what the graph and explanation show, the BMU should actually correct any deviation from the "trendline" as long as the battery is always charged from the mains. i.e no external fast charging.
Is that correct or did I miss something?

 

[jaapv]

Another long lecture by my dealer Battery smoothing changes nothing in the battery itself. The car is discharged overnight to real zero using the air-conditioning (and a hidden menu setting) Afterwards it is recharged and the charging capacity of each individual cell is measured. These data are used to reset the charging algorithms for each individual cell to ensure that they will all take the same level of charge, resulting in a higher capacity of the complete pack. It is advised to do so every 100.000 km.

I got this document, that says prety much the same. Unfortunately, it is in Dutch:

BTW, the car will indeed use some of the spare capacity to maintain range as the battery goes down over time.

This is in line with what I experienced:
- For a long time the Ah number went down rapidly where the predicted and actual range were hardly affected.
- Full is still shown as 100%
- ICE engagement is still around 25 -30%
- Minimum value for Remaining Capacity Ah (shown by EvBatMon of PHEV Watchdog or my own tools) is much lower than before (< 8 Ah)

Based on my readings I have come to the conclusion that %SOC is defined as Remaining Capacity Ah / Current Capacity Ah (where the latter is the number from the health report, some of us receive after service)

What I do understand, however, is that my battery has been treated twice. Both times gain was little below 2 Ah and both times gain was gone within a month and a half. So, what is real then?

It may be that there is one dodgy cell in your battery pack that deviates much quicker than expected.
The service sheet you posted neglects to mention that the procedures are per cell, and one cell can pull the whole battery down.

 

[jaapv]

That's good info. Nice of your dealer to explain it. Seems they are a bit more "switched on" than the ones we have here......

Maybe I'm lucky, but in general, I have found over the years that dealers are more than happy to natter to customers with a technical interest.

 

[elm70]

This is in line with what I experienced:
- For a long time the Ah number went down rapidly where the predicted and actual range were hardly affected.
- Full is still shown as 100%
- ICE engagement is still around 25 -30%
- Minimum value for Remaining Capacity Ah (shown by EvBatMon of PHEV Watchdog or my own tools) is much lower than before (< 8 Ah)

Based on my readings I have come to the conclusion that %SOC is defined as Remaining Capacity Ah / Current Capacity Ah (where the latter is the number from the health report, some of us receive after service)

What I do understand, however, is that my battery has been treated twice. Both times gain was little below 2 Ah and both times gain was gone within a month and a half. So, what is real then?

Any report on Ah is just a pure guess
EV range is a guess
SOC in % is a guess

Guess driven by questionable and not disclosed formulas.

So ... nothing bring any value if you want to analyse in deep the battery status in your car.

As far as I know there is no procedure for gain any capacity from "used" Lithium battery .. so this "smoothing" must be another voodoo ... possibly it is just a way to trick the questionable logic used for calculate the Ah and SOC of the battery pack.

The only real information is the voltage of each cells ... and the amount of current used for charge the battery (unfortunately this is also something to be trusted from the car, but normally this is relative accurate to measure)

What I did find out with my PHEV it is that in some condition I have a difference between the highest voltage cell to the lowest just a bit over 0.02v that is "a lot" ... when this happen most of the time the weak cell is always the same (in my case the cell "125") .. if you have reported degradated battery pack, I would expect you should be able to see something similar, since it is hard to believe all the cell ages in similar way

Per what I did notice, it is looking that the car, while left parked without charging, it does actively try to balance the pack ... and I'm afraid this means it does discharge the cells with highest voltage (this is really a waste of energy) .. this explain why sometime the expected range is less, sometime much less, then when the car has been left.

 

[jaapv]

I guess you didn't read my post. There are no changes made to the discharging side. The car adjusts the charging algorithms per cell to charge them all to an optimal level. It does so automatically, based on predictions in the firmware; but after some time the pattern can be brought into line with the actual situation.The whole battery will then perform to that level. Even if some cells could be charged to a higher level it would not help, as the discharge will be limited by the "worst" cell.

According to my dealer, the most gain he has seen was 5 Ah, normally it will be between 2 and 4.

 

[elm70]

I guess you didn't read my post. There are no changes made to the discharging side. The car adjusts the charging algorithms per cell to charge them all to an optimal level. It does so automatically, based on predictions in the firmware; but after some time the pattern can be brought into line with the actual situation.The whole battery will then perform to that level. Even if some cells could be charged to a higher level it would not help, as the discharge will be limited by the "worst" cell.

According to my dealer, the most gain he has seen was 5 Ah, normally it will be between 2 and 4.

Another long lecture by my dealer Battery smoothing changes nothing in the battery itself. The car is discharged overnight to real zero using the air-conditioning (and a hidden menu setting) Afterwards it is recharged and the charging capacity of each individual cell is measured. These data are used to reset the charging algorithms for each individual cell to ensure that they will all take the same level of charge, resulting in a higher capacity of the complete pack. It is advised to do so every 100.000 km.
BTW, the car will indeed use some of the spare capacity to maintain range as the battery goes down over time.

Right ... I did missed this ...

Still ... it does not make much sense.

"discharged overnight to real zero" .. this is not possible ... Lithium cells should be never send to 0V ... possibly the lowest value acceptable and for a short time is 2.5v .. still for me it is not even good to slowly discharge below 3.5v

Each cell can't be individually charged ... the same charging current cross the entire battery pack .. only the BMU can cause to discharge on request individual cell (normally with a relative small current)

All that it is needed to be done is:
- Charge all the cell to the equal max voltage .. that is defined to 4.10v
- Define the 26% and 30% SOC for an individual cell ... and this should be based on the rest-voltage of a single cell.

Question is which is assumed 26% and 30% SOC ... is this 3.77v and 3.80v ?

Or maybe .. the process above is in charge to define the voltage to be used for these two different SOC that are relevant for the EV usage of our PHEV

Anyhow .. there is some black voodoo happening in the guess algorithm for know the battery health and real SOC

 

[jaapv]

I guess you didn't read my post. There are no changes made to the discharging side. The car adjusts the charging algorithms per cell to charge them all to an optimal level. It does so automatically, based on predictions in the firmware; but after some time the pattern can be brought into line with the actual situation.The whole battery will then perform to that level. Even if some cells could be charged to a higher level it would not help, as the discharge will be limited by the "worst" cell.

According to my dealer, the most gain he has seen was 5 Ah, normally it will be between 2 and 4.

Another long lecture by my dealer Battery smoothing changes nothing in the battery itself. The car is discharged overnight to real zero using the air-conditioning (and a hidden menu setting) Afterwards it is recharged and the charging capacity of each individual cell is measured. These data are used to reset the charging algorithms for each individual cell to ensure that they will all take the same level of charge, resulting in a higher capacity of the complete pack. It is advised to do so every 100.000 km.
BTW, the car will indeed use some of the spare capacity to maintain range as the battery goes down over time.

Right ... I did missed this ...

Still ... it does not make much sense.

"discharged overnight to real zero" .. this is not possible ... Lithium cells should be never send to 0V ... possibly the lowest value acceptable and for a short time is 2.5v .. still for me it is not even good to slowly discharge below 3.5v

Each cell can't be individually charged ... the same charging current cross the entire battery pack .. only the BMU can cause to discharge on request individual cell (normally with a relative small current)

All that it is needed to be done is:
- Charge all the cell to the equal max voltage .. that is defined to 4.10v
- Define the 26% and 30% SOC for an individual cell ... and this should be based on the rest-voltage of a single cell.

Question is which is assumed 26% and 30% SOC ... is this 3.77v and 3.80v ?

Or maybe .. the process above is in charge to define the voltage to be used for these two different SOC that are relevant for the EV usage of our PHEV

Anyhow .. there is some black voodoo happening in the guess algorithm for know the battery health and real SOC

Well, it is always nice to hear from somebody who knows better... However, we may assume that Mitsubishi does know what they are doing. To zero may well mean down to the lowest safe level, for all I know, but Mitsubishi does claim that they monitor and control the charge to each individual cell, and why not? It is just a matter of design and fimware.
As for black Voodoo - isn't that just another name for digital technology to explain a shortfall in our understanding?

 

[anko]

During little over an hour I have retrieved and logged voltages for all 80 cells (628 cycli, so roughly 1 per 6 seconds) and put them in a graph ...



Hard to see, but these are indeed 80 different traces. Highest value: 4.071, lowest value 3.633.

 

[anko]

Average voltages per cell:

 

[jaapv]

During little over an hour I have retrieved and logged voltages for all 80 cells (628 cycli, so roughly 1 per 6 seconds) and put them in a graph ...



Hard to see, but these are indeed 80 different traces. Highest value: 4.071, lowest value 3.633.

And a couple of them spike quite a bit downward - probably the cause of your reduced capacity.

 

[anko]

During little over an hour I have retrieved and logged voltages for all 80 cells (628 cycli, so roughly 1 per 6 seconds) and put them in a graph ...



Hard to see, but these are indeed 80 different traces. Highest value: 4.071, lowest value 3.633.

And a couple of them spike quite a bit downward - probably the cause of your reduced capacity.

It is not a couple of them. It is all of them. But all at the same time(s). It looks as if you see 1 graph, but you are looking at 80 graphs on top of each other.

 

[elm70]

It is not a couple of them. It is all of them. But all at the same time(s). It looks as if you see 1 graph, but you are looking at 80 graphs on top of each other.

From the graph it is not visible the delta between the highest cell and the lowest cell .. this will be a nice information for know

At the end in any battery pack, with cells in series .. the capacity is equal to the smallest capacity of the weakest cell in the series.

The spike down . are due to the high IR .... and more or less are looking similar to my PHEV battery .. I have seen as low as 3.5xV under 40/50kw of load with less then 10km range left

What is strange .. is that you pack is not at 4.10v at any time .. but I guess this is due to pre-heating the car before start it

Another important information for me .. is to know the Voltage when 0km range of low km range is left ... but voltage after at least couple of hours of rest time ... this will provide the information which is the voltage level consider by the car for have 30% SOC

If this voltage value is high .. it means the logic in the PHEV is too conservative, and cause a reduced EV range .. I think Vtech has the code for re-set this value .. so by sending a special code, we can have "magic" extra capacity and range ... at the cause to allow lower voltage in the battery pack .. for me 3.80v is already too conservative ... real limit should be 3.75v .. also I don't get why to differentiate 26% with 30% ...

I have seen as much as 3.85v and 2km EV range .. for me is a big waste of real capacity of my battery ... luckily my daily trips are less then 30km or just a bit over .. so always inside my "poor" EV range even with pre-heat the car.

If your reduced EV range is causing to start daily the ICE only for make the last kms ... then .. maybe you should be the first which may have an economical advantage on trying to have an "extended" battery pack .. in parallel to the main pack .... I did estimated this should cost around 300/400 euros ... plus some skills on wiring the extra battery in parallel to the main one.

 

[anko]

From the graph it is not visible the delta between the highest cell and the lowest cell .. this will be a nice information for know

How not? I think it shows that delta is nearly 0. The second graph kinda backs this us, showing that the trip average voltage of the different cells is no more than 5 mV apart.

What is strange .. is that you pack is not at 4.10v at any time .. but I guess this is due to pre-heating the car before start it

You must stop guessing. You are to good at it

.. also I don't get why to differentiate 26% with 30% ...

At higher speed, more kW is used. So, more kWh from the battery will be consumed after the ICE was started, but before the generator starts to contribute. SO, at higher speed you need a bigger reserve. You can also observe that when the ICE is already warm, the low water mark for higher speeds goes down a bit.

If your reduced EV range is causing to start daily the ICE only for make the last kms ... then .. maybe you should be the first which may have an economical advantage on trying to have an "extended" battery pack .. in parallel to the main pack .... I did estimated this should cost around 300/400 euros ... plus some skills on wiring the extra battery in parallel to the main one.

Makes no sense, IMHO. As soon as I spend that money, I will be reassigned to a project 4 km further down the road. In other words, it will never be enough.

 

[elm70]

From the graph it is not visible the delta between the highest cell and the lowest cell .. this will be a nice information for know

How not? I think it shows that delta is nearly 0. The second graph kinda backs this us, showing that the trip average voltage of the different cells is no more than 5 mV apart.

If your reduced EV range is causing to start daily the ICE only for make the last kms ... then .. maybe you should be the first which may have an economical advantage on trying to have an "extended" battery pack .. in parallel to the main pack .... I did estimated this should cost around 300/400 euros ... plus some skills on wiring the extra battery in parallel to the main one.

Makes no sense, IMHO. As soon as I spend that money, I will be reassigned to a project 4 km further down the road. In other words, it will never be enough.

About the 5mv difference ... in my case I have seen even more then 20mv (but not sure it happen so often .. something I still have to investigate more .. I believe our PHEV while is sitting unused the BMU does balancing .. or better does discharge the cells with higher voltage)

Anyhow .. up to 10mv is ok .. above 20mv (but only in rest status .. under load it is not really relevant) it means some cells are significantly weaker then other .. or ... the balancing process at charge did not work good .. or the balancing at rest did screw up totally under every aspects

From your first graph .. looking at the "ending" ... it is looking your voltage per cell is 3.80v and you did not had high power request in the last X minutes of usage ... so .. your PHEV is fully using the capacity between 4.10 down to 3.80v .. your cells are balanced all the time .. so ... I don't think there is any magic possible to gain more capacity from your battery ... the smoothing process in your car .. is just wearing further the battery

PS: With ageing batter , so with higher IR ... using above 20/30kw power from the battery cause to waste a lot of energy , and as well, without a "stop" ... the car may believe that the battery has less capacity then what it is really inside .. since one effect of old cell with high IR .. is to bounce to higher voltage after few hours sitting unused ... that's why somebody see higher range after leave the car unused for some hours. So .. with old battery ... driving more "soft" does cause to have way more EV range .. from the graph above it is looking you drive relative "hard" in some part of your driving ...

 

[elm70]

If your reduced EV range is causing to start daily the ICE only for make the last kms ... then .. maybe you should be the first which may have an economical advantage on trying to have an "extended" battery pack .. in parallel to the main pack .... I did estimated this should cost around 300/400 euros ... plus some skills on wiring the extra battery in parallel to the main one.

Makes no sense, IMHO. As soon as I spend that money, I will be reassigned to a project 4 km further down the road. In other words, it will never be enough.

For me it make sense, if the daily car usage is always the same.

For example ... I'm doing every day no more then 30km + 15 to 30min pre-heating in winter .. that is around a need for 40km summer range

If my EV range will not allow me this ... either I will need to save on pre-heating .. or waste fuel and money every day.

Considering I'm saving 1 euro a day in fuel ... if my EV range drop, and I lost most of my fuel saving ... add battery in parallel for 350 euro (for ~1kwh battery) .. does make economical sense in my view

Replace the battery is more complicated and more expensive ...

Apparently with latest Tesla truck announcement ... it is looking they can make battery pack for less then 100USD per kwh .. this means at this cost replace our entire battery cost just 1200USD ... and for just 2000USD .. we can get a 20kwh battery pack .. with a much more range ... so .. maybe in a relative short time .. replace battery will make sense also for our "old" PHEV

 

[Harald]

...after the ICE was started...

Yes, it was clearly visible that ICE was in operation. The Netherlands don't have the height difference fot such long loading periods...
(How about tailwind situations for regeneration? )

It looks as if you see 1 graph, but you are looking at 80 graphs on top of each other.

Thanks, I've seen the different colors after zooming in to max. resolution.

...The spike down . are due to the high IR...

What dou you mean by "IR"?
I assume that the extreme spikes down will be due to high load for the battery during acceleration or similar.

Best regards, Harald