Battery Capacity Test Results

Discussion in 'Clarity' started by Central Penn, Jun 1, 2019.

Tags:
  1. DucRider

    DucRider Active Member

    Cell voltages DO vary with state of charge!
    [​IMG]
    All the 40xx.x voltages are showing in mV so 4.08 V is entirely reasonable.
     
    Last edited: Jul 30, 2019
  2. MrFixit

    MrFixit Active Member

    Good point -

    I am looking at the data sets more closely. There are two strings of 84 batteries (Cell 1-84 line A and Cell 1-84 line B). This are a total of 168 values.
    If I count the number of 40xx.x values in @sniwallof data, there are 200 of them.

    It is quite plausible that these are the battery voltages (mis-placed) and there are 32 other things that happen to be in the 4000 mV range.

    I'm going to attempt to make a spreadsheet for both @AlAl and @sniwallof data to compare line for line. Note, there are 366 entried in @sniwallof and 395 entried in @AlAl
     
  3. 2002

    2002 Well-Known Member

    I didn't compare line but line but I did toss sniwallof's cell voltages into a spreadsheet and summed them, they added up exactly to lines 35 and 36 on the report

    35 HV Battery Line A Total Voltage 342.8 V
    36 HV Battery Line B Total Voltage 342.9 V

    Sum Line A 342838.8
    Sum Line B 342885.2
     
  4. MrFixit

    MrFixit Active Member

    OK,
    Attached is a line-by-line comparison (Excel spreadsheet).

    Summary:
    Good news - The parameter descriptions match exactly (param #1 thru #366).
    The 'professional' capture from AlAl has 29 additional parameters (#367-395).
    I have highlighted (in RED) anything that looks 'funny'.
    There is a LOT more good here than bad...

    Wait - How do I attach a spreadsheet???
     
  5. MrFixit

    MrFixit Active Member

    Don't know how to attach a .xls, but here is a PDF version of the comparison...
     

    Attached Files:

  6. 2002

    2002 Well-Known Member

    What it's doing is duplicating valid data from some fields into fields where they don't belong. All of the data you have highlighted in red comes from other fields. The 4,000's are the most obvious, each of those numbers can be found in the actual battery cell voltage.

    Also notice the values in the erroneous data in lines 107-109

    107 1.4
    108 0
    109 14
    110 0

    Now look farther down at lines 363-366

    363 1.4
    364 0
    365 14
    366 0
     
  7. MrFixit

    MrFixit Active Member

    Yes !
    I added up both the 'A' string and the 'B' string for both @sniwallof and @AlAl in the spreadsheet and they both match the respective lines 35 and 36.
    I think we can conclude it is likely that the @sniwallof tool is properly reporting the cell voltages.

    This gives wonderful visibility down to the cell level. It would be very interesting to see this report for @jdonalds car since he has observed odd charging patterns and reduced range.

    I was looking at the manual for the Autel MX808, and as best I can tell, I think it has this capability. This tool is more like $350... A fair amount lower than the MK808BT that @AlAl bought. It doesn't have Bluetooth so it is hard-wired to the OBD2 connector.
     
    Last edited: Jul 30, 2019
  8. Clarity Dave

    Clarity Dave Member

    Very interesting! Those cell voltages provide insight into where Honda has set its maximum cell charge limit to trade off battery life vs capacity. Table 4 on this page...

    https://batteryuniversity.com/learn/article/how_to_prolong_lithium_based_batteries

    ...lists battery life in discharge cycles and available stored energy at various cell voltages, and includes these statements:

    "In terms of longevity, the optimal charge voltage is 3.92V/cell. Battery experts believe that this threshold eliminates all voltage-related stresses; going lower may not gain further benefits but induce other symptoms"
    "Every 0.10V drop below 4.20V/cell doubles the [life as expressed in number of discharge cycles] but holds less capacity."

    The pertinent entries from the table:
    4.20V = 300-500 discharge cycles vs 100% stored energy
    4.15V = 400-700 discharge cycles vs 90-95% stored energy
    4.10V = 600-1,000 discharge cycles vs 85-90% stored energy
    4.05V = 850-1,500 discharge cycles vs 80-85% stored energy
    4.00V = 1,200-2,000 discharge cycles vs 70-75% stored energy
    3.90V = 2,400-4,000 discharge cycles vs 60-65% stored energy
    At 3.80V the stored energy plummets; about 4.20V the discharge cycles shrink significantly as well.

    clarityowner12's scan shows a Cell Voltage Limit During Plug-In Charging around 4.14V.
    Central Penn's report shows HV Maximum Battery Cell Voltage around 3.48V at a 3% state of charge.
    AIAI's first report shows cell voltages around 3.89V (lines 146 etc.) at 79% SoC(line 72)
    AIAI's second report shows cell voltages around 3.93V at 83% SoC
    sniwallof's report shows cell voltages around 4.08V at 100% SoC

    All conclusions are tentative since not all Li-ion batteries are alike, but it does appear that charging to around 80% (close to the 3.92V cell sweet spot cited above) would increase battery life. Of course, other factors affect battery life (see Table 2 for the impact of depth of discharge, for instance).
     
    KentuckyKen and Robert_Alabama like this.
  9. 2002

    2002 Well-Known Member

    The BU website points out that high voltage (charging to high SOC) combined with high temperature (especially above 30°C (86°F)) is sort of a one-two punch that is worse than either individually. I suppose a compromise would be to charge to a lower percentage during summer than in winter, essentially giving up some of the "bonus" miles that we get in warmer weather, but not cutting back in winter, or at least not cutting back as much during winter when you would miss the additional miles even more.
     
  10. sniwallof

    sniwallof Active Member

    I missed some posts that look promising (just did a quick read). Update - Obviously that first post was part garbage. I'm getting closer to understanding the problem. In electrical power train live data, if I scroll up or down very quickly, I get cell voltages at first for unrelated items, then it settles down very next update and shows the correct numbers. I may be able to quiet it, by scrolling slowly.

    Here are a few samples from 93% soc down to 80% soc (always line 72). Most of these (at first quick glance) look okay. Use the time in the file name, the time at the top of the page is apparently the beginning time of each "session". For example, a number of them say 12:10. Sitting in the sun, A/C auto to 72F.

    There is a very convenient pre-loaded computer sharing program called "team viewer" free for personal use, which lets you couple the file listings of the unit with your computer to transfer files. For now, I'm using that system to a less critical pc (can't vouch for security, it's probably okay, but who knows).

    @AIAI, if you scroll very fast on the 980, do you see quick odd values that quickly return to the correct numbers? Maybe the 808 processor is slower.

    only caution on 808 vs. 808BT, is review that youtube video which claims the BK808BT is a newer version of the BK808. may be a non-issue.
     

    Attached Files:

    Last edited: Jul 30, 2019
  11. MrFixit

    MrFixit Active Member

    Very Nice !!!
    I think you are onto something.

    There also seem to be the extra 30 parameters the 960 has and the 808 doesn't. I don't think those extra parameters are especially interesting however...
    For reference, here they are:

    upload_2019-7-30_14-4-14.png

    Anyone see something here that they couldn't do without?
     
  12. sniwallof

    sniwallof Active Member

    He has the 980, top of the line! (the 960 was the first model I was considering recommended by an Autel tech, but it has that oscilloscope, probing camera, and built in camera, probably part of the much higher cost). Then AIAI told us we could probably do the same lists with a much lower cost 808.

    I noticed this too, the first list was 366 like mine, than a later one was 395. I may have those items too, and just don't know how to access them yet. I'm only printing electrical power train now. @AIAI, did you do something different between the 366 items and the later 395?
     
  13. 2002

    2002 Well-Known Member

    394 Remaining Charge Time

    I remember someone saying they missed having this which they had on their previous EV. Although the results are in seconds which means getting out your calculator, unless in the settings it lets you specify the units.

    Also the results seem a little strange, I thought maybe they are not from the car but estimated by the system, but I don't know where else "time until charging start" would come from other than the car.

    394 Remaining Charge Time 122760 sec = 34 hours
    395 Time Until Charging Start 153000 sec = 42 hours
     
  14. sniwallof

    sniwallof Active Member

    Here are a few almost random samples of other pages (there are many more). The PGM-FI is pretty interesting (14-26-55). I did each pdf print before I started scrolling, maybe that helps too.
     

    Attached Files:

  15. MrFixit

    MrFixit Active Member

    Look at this one (your last file):

    upload_2019-7-30_15-6-32.png

    This works out to 2.5 hours. This seems too low to represent cumulative run time (how many miles do you have?)
    Would be interesting to see @KentuckyKen compare his reading to his hour meter,

    We have been looking for a way to have the car tell us cumulative EV vs. HV miles... Need to comb through this data some more. Very interesting.
     
  16. MrFixit

    MrFixit Active Member

    Is this all the miles you have on your car???

    upload_2019-7-30_15-13-4.png

    Maybe the 2.5 hour engine run time is right?
     
  17. sniwallof

    sniwallof Active Member

    correct (2,060 ODO), I only do about 4,000 to 6,000 miles a year, including about three to five highway trips over 400 miles. Most of my driving is 3-20 miles, with a few 100 mile trips here and there. Semi-retired, work out of my home office.

    I just got this one in Feb. 2019, and I overlapped with a Bolt for a month, before I figured out, as fun as the Bolt is to drive, it made no sense to keep both of them.
     
    Last edited: Jul 30, 2019
  18. MrFixit

    MrFixit Active Member

    Just for reference, Here is a comparison of the MX808 to the MK808BT...

    Here is the MX808:

    upload_2019-7-30_15-47-17.png


    Here is the MK808BT:

    upload_2019-7-30_15-48-17.png

    Noticable Differences - MK808BT has an HDMI port, it has a Bluetooth connection to OBD2, has a slightly larger battery.
    Both seem to have the same processor.
    The description for the MK808BT states that is has "superior" special functions like Oil Reset, EPB, SAS, BMS and DPF.
    In looking at the manuals however, It looks like the MX808 has those too in addition to TPMS.
     
  19. KentuckyKen

    KentuckyKen Well-Known Member

    At 18 months and 11,200 miles I'm at 19.2 hours on the engine and have pumped a little over 17 gal. (BTW, if you pluralize gal, the sentence looks horribly wrong.)

    The only information I see useful on that extra 30 is the individual cell max and min. That tells me the cells are well balanced but I can live without that data to save some $.
     
  20. sniwallof

    sniwallof Active Member

    Recording ... not yet, not sure what it can do.

    I made a really nice color graph of three current parameters on a short drive. The slowest x-axis time scale x8 is something on the order of minutes. Problem is I cannot figure out yet how to record it, or the underlying data. I think this type of use might be pushing the limits of what the little system can do.

    Apparently some models have a screen print button. I have that button for the list page, but don't see it anywhere for the graph, especially for the merged graph of three traces with three y axis with values.

    I did somehow make this odd recording of a series of over a thousand frame shots of the "recorded" parameters, but it's pretty useless. You can push frame forward, and it literally just shows the next list page in time, with the next value.

    Data logging - does not mean what most of us might think - It's part of their system about you capture errors or problems and send information to the support engineers. Does not seem to have anything to do with conventional data acquisition.
     
    Last edited: Jul 30, 2019

Share This Page