Battery Capacity Testing

[tinoale]

Hi,

I read Canion Wh out is correct and reliable.

So are we saying there is a simple way to measure battery available Wh by fully charging it, reset canion trip, turn the heater on while the car is parked, and leave it as is until the heater stops ? (BTW at what SoC does the heater stops ?)

If that's so simple we have a simple and easily repeatable method to compare our batteries capacities, don't we ?

 

[BlueLightning]

So are we saying there is a simple way to measure battery available Wh by fully charging it, reset canion trip, turn the heater on while the car is parked, and leave it as is until the heater stops ? (BTW at what SoC does the heater stops ?)
If that's so simple we have a simple and easily repeatable method to compare our batteries capacities, don't we ?

Yes, that's how I did, see post http://myimiev.com/forum/viewtopic.php?p=18034#p18034 but a full discharge is not helpfull and not to advise.
I just drew a graph Wh out versus SoC, you get the graph below showing steps because SoC resolution is of 0.5 %.
The slope of an average straight line or the ratio delta Wh out divided by delta SoC give the full battery capacity. "Delta" means a subtraction between two values, (Wh out 1 less Wh out 2) divided by (SoC 1 less SoC 2 ).

Why a full discharge is not to advise and not helpfull.
Full discharge is not good for a battery and anyway the car computer will forbid a 16.28 kWh discharge.
Also, around ten per cent SoC, as you can see on my graph below, the car computer stops the heater, by the way I didnt get the turtle.
Why wasting electricity?

I think Peugeot use the heater for their battery checking they do every two years, but I am missing their exact protocole.
For me, cells equilibration and SoC calibration http://myimiev.com/forum/viewtopic.php?p=18067#p18067 is necessary to get a proper curve Wh out versus SoC before a battery capacity test.
Peugeot or Mitsubishi might do a discharge below two bars and then make a full charge, before to proceed to this battery capacity test.

I use SecLog data to proceed this capacity test, but if you are more at ease to use the Wh out given by CaniOn on the right side of the screen to draw this kind of graph, be sure you pick up the Wh out only when there is a change of SoC otherwise the average curve slope will be wrong and so on the battery capacity.
If you are aware of this "step" or "0.5 % SoC resolution" problem, a single measure of "delta Wh out" divided by a "delta SoC" will give an accurate result, I did a discharge of around 1 kW-h for more accuracy but 0.5 kW-h would have given the same accurate result.
For my test, the curve's slope was 0.1624 kWh per % SoC, or 16.24 kWh for 100 % SoC which is the Battery capacity.

It's a pity Canion doesn't provide graphs versus SoC instead of time, by example, a simple curve like Volt versus SoC would give informations about the evolution of the battery State of life or when the battery calibration occurs.
State of Health Estimation of Li ion Batteries: Cycle Life Test Methods, http://komar.bitcheese.net/files/JensGroot.pdf See curves Voltage versus SoC or Voltage versus Capacity giving informations about Battery State of Life, pdf, page 97 or Paper page 85, Chapter 7 Cell Ageing Analysis, 7.1 Galvanostatic Voltage Profiles

Capacity Battery measurement

 

[EV4Taos]

Amp-hours seems to be the holy grail of battery capacity measurements for the Nissan Leaf and DIYs. Or at least top 2 or 3 anyway. Does the i-miev report amp-hours? I know how much work it is to analyze can bus data and it is easy for the Nissan Leaf community to map this out because they have 15 times more cars and people to do the work, but I thought I would ask just encase someone has run across a can bus ID reporting amp-hours of the battery.

J

 

[offgridQLD]

Amp hrs isn't a very good (accurate) way to measure capacity as it's just load (amps over time). (whrs) is way more accurate as it takes the voltage current (amps) x the voltage over time. Even though a lithium battery has a flat voltage over its state of charge compared to some chemistry the voltage still changes from 100% SOC to 0% SOC say roughly 4v to say 3V is a 25% variation that you don't account for in AH measurements.

Kurt

 

[BlueLightning]

Below workshop manual for Mitsubishi i-MiEV batteries capacity checking.
http://mmc-manuals.ru/manuals/i-miev/online/Service_Manual_2011_2012/2012/54/html/M154945500003901ENG.HTM Then clic on the blue arrow on top right

or from http://mmc-manuals.ru/manuals/i-miev/online/Service_Manual_2011_2012/2012/index_M1.htm Follow
54D-ELECTRIC MOTOR UNIT AND MAIN BATTERY
(Refer to GROUP 54D - Bttery Management Unit (BMU) and Main Battery - On-vehicle Service - Main Battery Capacity Measurement, Measure the main battery capacity.

I do not know how the "Main Battery Capacity Automatic Measurement" software works, may be that at time or SoC intervals, the charge is stopped, then a small discharge is done throught the heating system as I did, see http://myimiev.com/forum/viewtopic.php?p=19095#p19095 , and then charge goes on again.

MAIN BATTERY CAPACITY AUTOMATIC MEASUREMENT
note
Do not disconnect the battery cable from the auxiliary battery (-) terminal.
To improve the measuring accuracy, working at more than 10° C of the outside temperature is recommended.

1.Connect the diagnosis connector to the M.U.T.-III.
2.Perform the regular charging by connecting the regular charging cable.
3.Check that the charging lamp in the combination meter is lit and the battery is charging.
4.Check that the residual main battery level indicator shows 1 segment or more. If the residual main battery level indicator shows 0 segment, charge the main battery until the battery level indicator shows 1 segment or more.
5.Select the "BMU" from the System Select Screen of the M.U.T.-III.
6.Select the "Special Function" from the BMU Screen.
7.Select the "Main Battery Capacity Automatic Measurement" from the Special Function Screen and do it. If errors occur, restart from the step 2. When any malfunction has been found, perform troubleshooting.

8.Turn on the air conditioner under the following conditions.
<Air conditioner operating conditions>
Electric motor switch: LOCK
Air flow: largest
setting temperature: highest
A/C switch: ON
MAX switch: ON

note
Capacity measurement time can be reduced if executed while the power window is fully open compared to the execution while the power window is fully closed.
Check that the "Execution status" displayed on the M.U.T.-III is "Executing". If "Not Completed" or "Not Executed" is displayed, check "Error state" , then restart from the step 2.
Check "Automatic Capacity Measurement Result History 1" and "Automatic Capacity Measurement ODO Value(Result History 1)" displayed on the M.U.T.-III.
9.Check that "Capacity Measurement Discharge Current" is "ON". If "NG/Not Executed" is displayed, restart from the step 8.
10.Leave the vehicle until the charging lamp goes out.
note Even if the M.U.T.-III is disconnected during this, the procedures cannot be affected.
11.Select the "BMU" from the System Select Screen of the M.U.T.-III.
12.Select the "Special Function" from the BMU Screen.
13.Select the "Main Battery Capacity Automatic Measurement" from the Special Function Screen and do it.
14.Check that the "Execution status" displayed on the M.U.T.-III is "Completed". If "Not Completed" or "Not Executed" is displayed, check "Error status" , then restart from the step 2.
15.Check that "Automatic Capacity Measurement Result History 1" and "Automatic Capacity Measurement ODO Value(History 1)" are updated. If not updated, restart from the step 2.
caution Do not operate the electric motor switch or disconnect the charging cable during procedure 2 to 13. If make a mistake during the work or stop the charging inadvertently before the battery is fully charged, restart from the beginning. Under low temperature condition, full charge may be judged and charging may be stopped before the full segment on the battery level indicator is shown.

 

[EV4Taos]

What does it all mean? Context? Sorry brain not ready for reverse thought engineering this morning. How can this help I-Miev owners?

Below workshop manual for Mitsubishi i-MiEV batteries capacity checking.
http://mmc-manuals.ru/manuals/i-miev/online/Service_Manual_2011_2012/2012/54/html/M154945500003901ENG.HTM Then clic on the blue arrow on top right

or from http://mmc-manuals.ru/manuals/i-miev/online/Service_Manual_2011_2012/2012/index_M1.htm Follow
54D-ELECTRIC MOTOR UNIT AND MAIN BATTERY
(Refer to GROUP 54D - Bttery Management Unit (BMU) and Main Battery - On-vehicle Service - Main Battery Capacity Measurement, Measure the main battery capacity.

I do not know how the "Main Battery Capacity Automatic Measurement" software works, may be that at time or SoC intervals, the charge is stopped, then a small discharge is done throught the heating system as I did, see http://myimiev.com/forum/viewtopic.php?p=19095#p19095 , and then charge goes on again.

MAIN BATTERY CAPACITY AUTOMATIC MEASUREMENT
note
Do not disconnect the battery cable from the auxiliary battery (-) terminal.
To improve the measuring accuracy, working at more than 10° C of the outside temperature is recommended.

1.Connect the diagnosis connector to the M.U.T.-III.
2.Perform the regular charging by connecting the regular charging cable.
3.Check that the charging lamp in the combination meter is lit and the battery is charging.
4.Check that the residual main battery level indicator shows 1 segment or more. If the residual main battery level indicator shows 0 segment, charge the main battery until the battery level indicator shows 1 segment or more.
5.Select the "BMU" from the System Select Screen of the M.U.T.-III.
6.Select the "Special Function" from the BMU Screen.
7.Select the "Main Battery Capacity Automatic Measurement" from the Special Function Screen and do it. If errors occur, restart from the step 2. When any malfunction has been found, perform troubleshooting.

8.Turn on the air conditioner under the following conditions.
<Air conditioner operating conditions>
Electric motor switch: LOCK
Air flow: largest
setting temperature: highest
A/C switch: ON
MAX switch: ON

note
Capacity measurement time can be reduced if executed while the power window is fully open compared to the execution while the power window is fully closed.
Check that the "Execution status" displayed on the M.U.T.-III is "Executing". If "Not Completed" or "Not Executed" is displayed, check "Error state" , then restart from the step 2.
Check "Automatic Capacity Measurement Result History 1" and "Automatic Capacity Measurement ODO Value(Result History 1)" displayed on the M.U.T.-III.
9.Check that "Capacity Measurement Discharge Current" is "ON". If "NG/Not Executed" is displayed, restart from the step 8.
10.Leave the vehicle until the charging lamp goes out.
note Even if the M.U.T.-III is disconnected during this, the procedures cannot be affected.
11.Select the "BMU" from the System Select Screen of the M.U.T.-III.
12.Select the "Special Function" from the BMU Screen.
13.Select the "Main Battery Capacity Automatic Measurement" from the Special Function Screen and do it.
14.Check that the "Execution status" displayed on the M.U.T.-III is "Completed". If "Not Completed" or "Not Executed" is displayed, check "Error status" , then restart from the step 2.
15.Check that "Automatic Capacity Measurement Result History 1" and "Automatic Capacity Measurement ODO Value(History 1)" are updated. If not updated, restart from the step 2.
caution Do not operate the electric motor switch or disconnect the charging cable during procedure 2 to 13. If make a mistake during the work or stop the charging inadvertently before the battery is fully charged, restart from the beginning. Under low temperature condition, full charge may be judged and charging may be stopped before the full segment on the battery level indicator is shown.

 

[JoeS]

Even though we have a number of similarly-titled threads, I started this new one and brought a number of interesting posts over here from the CaniOn thread, as we're trying to keep that thread specific to the CaniOn App and not interpretation of the data seen by CaniOn.

For me it is a revelation that Mitsu uses the vehicle's heater as part of their capacity testing regime. Thank you BlueLightning for the details.

 

[JoeS]

For most of us, the i-MiEV's battery capacity is what it is and is perhaps not worth the effort to investigate further unless something suddenly goes awry.

Over on this thread, Energy vs %SOC or Wh/1%Soc, I had posed the following question, but had received no bites, so thought I would modify it slightly and re-post it in perhaps this more appropriate Topic.

What is it we're trying to achieve with this datataking?

Simply to ascertain if, and by how much, our vehicle's battery is degrading over time.

What parameter are we interested in?

Kilowatt-hours

What conditions do we need to control in order to obtain repeatable results?

Temperature and battery current; battery voltage is measured but not controlled.

How do we do that?

By performing controlled discharging and recharging tests while recording current, battery voltage, kWh, SoC, and temperature, using CaniOn.

How can we perform controlled discharge and recharge?

I am aware of only these:

Charging: using L1 or L2, with the ac charging current set by the EVSE. CaniOn shows the dc current flow but I am unclear whether this is current into the battery or current into the 'system'.

Discharging: the only static test I can think of is running the vehicle heater (wonderfully detailed by BlueLightning earlier in this thread), even though it may not be a steady-state load because the heater load 'cycles'. In the case of discharge, I don't think it matters where the current is going (e.g., some of it into the car's 'system'), as long as it is being measured coming from the battery at a fairly constant rate and converted to Wh.

Note that I am discounting discharge while driving, because current magnitude while driving is an uncontrolled variable. We may have identical kWh between tests, but they may be achieved using radically different currents.

I recognize that Lithium battery capacity is less sensitive to current draw than lead-acid batteries; however, if we are attempting to establish a baseline and then perform another test at a later date, then current should certainly be a controlled variable to ensure identicity between tests. I am also troubled that using the heater is not a consistent load because it 'cycles'.

Perhaps it may be worthwhile having a consistent methodology discussion so we can try comparing apples-to-apples when we get around to testing? I also agree with BlueLightning in that we do not want to overly-deplete our packs and thus will need to extrapolate the data to calculate full capacity. I rarely go below two bars, have never seen turtle, and won't sacrifice my battery in the 'interest of science'. :twisted:

BTW, has anyone in North America recently had a battery capacity test performed at a dealership with printed-out results? Just curious...

Or maybe I should just crawl back into my hole and forget the whole thing...

 

[Aerowhatt]

Perhaps it may be worthwhile having a consistent methodology discussion so we can try comparing apples-to-apples when we get around to testing? I also agree with BlueLightning in that we do not want to overly-deplete our packs and thus will need to extrapolate the data to calculate full capacity. I rarely go below two bars, have never seen turtle, and won't sacrifice my battery in the 'interest of science'. :twisted:

BTW, has anyone in North America recently had a battery capacity test performed at a dealership with printed-out results? Just curious...

Or maybe I should just crawl back into my hole and forget the whole thing...

A consistent methodology is exactly what is needed. It has to (IMO) be inexpensive and repeatable. I have some experience in this type of analysis over the years working on DC conversion EVs mostly 80's vintage gliders. I propose that what is needed is a fixed load for a measured time. The climate controls don't work for us because of cycling. So we need an exterior load or even more simply the cars lights. I often have often used a 1000watt inverter and a space heater.

equipment required:
1. 1000 watt 12 volt inverter. ($100.00 or so.)
2. fixed load such as a space heater or incandescent lights. (Most already have something)
3. Decent clamp on DC current meter (less than $100.00) used on the input side of the 1000watt inverter
4. Stopwatch app on a smart phone

Pros:
1. The DC to DC in the car will provide a fixed voltage at a steady load on the 12 volt system. So measurements of actual load usage is accurate and easy.
2. No special expensive gadgets or knowledge required.
3. Depth of main battery discharge for the test could be user defined. (as little as 2 or 3 bars drop measured would be useful scaleable information)

Cons:
1. tester would need to be attentive ending the time precisely when the target bar drops from the indicator.
2. The car would need to be in ready mode so the background usage of the inverter electronics and the DC to DC converter would not be measured by this test. (But . . . would that matter since this energy ("wasted" by the system) does not provide driving range anyway?).

Starting always with a full charge this would fairly quantitatively determine the energy per bar for any i Miev, comparable to others doing the same test profile.

How does that sound?

Aerowhatt

 

[JoeS]

Aerowhatt, using an inverter is a good idea for standardization and I don't think the state of health of the 12v battery would matter much for this test as the dc-dc would be doing most of the work. I've used both inverters with dummy loads and a PowerLab8 as well as a West Mountain CBA with good results. I think the i-MiEV's dc-dc is only good for about 50A so I wouldn't want to put much more than about a 500W steady-state 12v load on the system.

Although it will tell us the power and energy out of the i-MiEV, this inverter test method won't take into account all the i-MiEV's 'system' power consumers. For example, another uncontrolled variable that is significant is the coolant pump cycling. This increased current due to the pump is readily visible on CaniOn, and it is accounted-for in its energy measurement. Using CaniOn's average battery temperature reading should be sufficient, and I suspect this may go up slightly during the test. I think CaniOn would be needed for this testing in order to properly measure the traction battery draw.

Aerowhatt, I think the first accessory you might like for your new i-MiEV is an Android tablet and CaniOn and the high-speed OBDII bluetooth adapter.

The energy/bar is non-linear, with the first bar taking about 7.5% of the SoC. There's a table posted somewhere on this forum with the numbers for each bar that someone had measured on their car. At the bottom end there is also significant variation in the absolute SoC number shown by CaniOn (which is simply reading the i-MiEV's data) and is a function of a battery's state of health. Perhaps our 'standard' test should only use a couple of bars hovering around the middle - perhaps starting at ten bars and then recording the transitions from ten-to-nine-to-eight-to-seven bars which would give us two data points (for capacity at the ninth and eighth bar). At 500W, that alone could take over six hours…

For myself and with very good battery packs in each of my two i-MiEVs, this academic exercise will have to wait about five years when I get close to warranty expiration... (the car, and not me, hopefully)... :roll:

 

[PV1]

I was about to suggest using CaniOn to measure energy consumption instead of a clamp on meter for the same reason. I don't believe the coolant pump is uncontrolled, at it seems to cycle on and off based on time, not temperature. It doesn't matter if it's 10 F or 80 F, my coolant pump cycles at the same rate. Still, CaniOn would be best to capture the energy consumption of everything (as it does measure in/out of the pack, not the system. Flashing the headlights while charging can be seen in the amperage reading of CaniOn.).

For the discharge test, how about using one of these on the 12 volt battery while in READY? It'd provide a steady load and wouldn't be as wasteful as running a space heater for hours. I have one and use it with a couple of 12 volt panels as overcharge protection (you know, 'cause a real charge controller is too simple :lol: :roll: ).

http://www.ebay.com/itm/400-Watt-Power-Grid-Tie-Inverter-for-Solar-Panel-Wind-/130444131961

 

[Aerowhatt]

wouldn't be as wasteful as running a space heater for hours. I have one and use it with a couple of 12 volt panels as overcharge protection (you know, 'cause a real charge controller is too simple :lol: :roll: ).

Unless the space heater is helping to heat your home

Aerowhatt

 

[PV1]

Good point. I guess my workplace is getting to me. Instead of requesting the house temperature raised, half of the floor is running space heaters :roll: :twisted: on top of the AC running.

 

[Aerowhatt]

Good point. I guess my workplace is getting to me. Instead of requesting the house temperature raised, half of the floor is running space heaters :roll: :twisted: on top of the AC running.

Only in America!

Aerowhatt

 

[JoeS]

kiev just posted this Dealer Capacity Calibration as a read-only Reference Technical Article. Just thought I'd keep things linked together so we can find them.

http://myimiev.com/forum/viewtopic.php?f=19&t=2675

 

[Malm]

I expected that Carsten. Cells going to 4,07V (if they are all at similar voltage) at 100% or 4,105V at 100% its normal and many times don't means anything about current capacity.

I usually follow, when have the opportunity, what it charges from 50% SoC to 55% SoC. I see, in mine, something like 545 Wh.



I know measurements done on i-MiEVs that I believe still have a range of a new one, and they normally show a value of 620/625 Wh. If your i-MiEV charge that value then I believe it have the same range of a new one (to be sure you should see if you can go to 0,0% SoC with your car, or if at 5% SoC, cells are over of approximately 3,4V).

So, I believe my car has 545/625*100 = 87,2% of the range of a new one.

 

[RobbW]

I was about to suggest using CaniOn to measure energy consumption instead of a clamp on meter for the same reason. I don't believe the coolant pump is uncontrolled, at it seems to cycle on and off based on time, not temperature. It doesn't matter if it's 10 F or 80 F, my coolant pump cycles at the same rate. Still, CaniOn would be best to capture the energy consumption of everything (as it does measure in/out of the pack, not the system. Flashing the headlights while charging can be seen in the amperage reading of CaniOn.).

For the discharge test, how about using one of these on the 12 volt battery while in READY? It'd provide a steady load and wouldn't be as wasteful as running a space heater for hours. I have one and use it with a couple of 12 volt panels as overcharge protection (you know, 'cause a real charge controller is too simple :lol: :roll: ).

http://www.ebay.com/itm/400-Watt-Power-Grid-Tie-Inverter-for-Solar-Panel-Wind-/130444131961

So, I was just about to post an off-topic reply to this post from PV1. But I decided to start my own topic over in the Off-Topic forum instead: http://myimiev.com/forum/viewtopic.php?f=13&t=2835

 

[BlueLightning]

Red triangle points on this graph are showing me68’s cell 23 collapsing, blue and green triangle are me68 voltages values before the problem and after the change of cell.
Lowest Cell Voltage versus SoC by Blue I.ightning

Lowest cell voltage data are collected battery at rest.

If data are not collected battery at rest, then voltage becomes a function of amperage, and not of battery state.
Driving or heat on, higher is the amperage, lower is the voltage (1), so a battery shut down at 2.75 volt is there to avoid any damage to the lowest voltage cell.

Battery capacity will remain at its highest capacity if battery cells have the same voltage during charge or discharge.
So collecting the lowest cell voltage which must fit the red curve might be a more efficient way to check the state of health of the battery.
Red triangle points on this graph are showing me68’s cell 23 collapsing, blue and green triangle are me68 voltages values before the problem and after the change of cell.

At 23% SoC me68 had a battery shutdown “… 2,75V of cell 23, resulted in a shutdown of the battery from the Battery Management System, no matter which SoC was calculated” (4).
There was no “Turtle” warning before this shut down.
So Turtle is not voltage but SoC related and surprisingly a battery shut down is possible without warning, so again, thanks for Canion for providing us critical informations.

For a new battery, for a battery in a good State of Health, or recalibrated, Turtle will light up for some vehicles at 10.5 % of State of Charge and for other ones at 8.5 % SoC (2).
The battery will shut down for a 0% SoC (3). Malm with his “tricks”, heading now to a second earth circumference or 80 000 km, following the path of his ancestors, Magellan, Vasco Da Gama, and others, will show you irrational BMS behaviours (5).

(1) http://myimiev.com/forum/viewtopic.php?f=23&t=1789&p=17242&hilit=resistance+cell+battery#p17242
(2) https://forums.energymatters.com.au/alternative-vehicles/topic5388-310.html#p44207
http://myimiev.com/forum/viewtopic.php?p=12122#p12122
http://myimiev.com/forum/viewtopic.php?p=12698#p12698
(3) http://www.vehiculeselectriques.fr/topic13207-30.html#p181035
(4) http://myimiev.com/forum/viewtopic.php?p=23489#p23489
(5) https://www.youtube.com/watch?v=PkWW5w2e4XU

 

[kiev]

très bien, blue lightning--great collection of data and graph, plus the finding about the turtle.

 

[Malm]

There are two results of my i-MiEV that are in the red line. That are the real results, without tricks. And two false results, one with the car giving less 10% of SoC then it had in the reality, and one giving more 5% of SoC then it had in the reality. By the way, my turtle come to light at 9,5% SoC. Allways, since 2011, at 9,5% SoC.