here i go again. IMA acting up

Discussion in 'Honda Hybrids' started by laurieaw, Oct 23, 2007.

  1. brick

    brick Answers to "that guy."

    I think you guys may be talking about two different things. I have regarded (and I think Mike is talking about) the 20-80% DoD as the absolute limits of the pack per the BCM. (On a Prius it's 40-80% IIRC). So when the gauge reads dead zero you're at 20%, full means 80%. MSantos, it sounds like you're talking about percent of usable range rather than percent DoD. Maybe?

    I like the idea of rigging a data acquisition system to record DoD, charge and discharge rates over time. It would be interesting to compare the varios Honda control schemes with what Toyota does. A Prius will keep the battery at one level all the time but there's always some activity registering on the MFD into or out of the battery even at a flat cruise.

    Something else I've heard: it has been suggested that batteries have a "break in" period of their own where efficiency increases somewhat during the first part of their life. Any truth to that?
  2. Fenrir

    Fenrir Well-Known Member


    I have a battery charger for AAA and AA NiMH and NiCd cells. It has a "refresh" function, which performs repeated full discharge/charge cycles on the cells to restore worn out cells. It has a display which shows the milliamp hours in and out while cycling. Guess what? It works! With each cycle, one can see the capacity of older cells increase, often to like-new condition.
  3. msantos

    msantos Eco Accelerometrist

    The figures I mentioned were obtained from Honda's own DVD based technical literature. I also have the paper volumes as well as some access the channel's tech bulletins.

    Tim may be right in that we may have been referring to different thresholds. The range I mentioned is the target (or optimal operating range) of 60% to 80%. The "managed range" is wider than that and it would represent the "absolute limits" as Tim mentioned (thanks Tim).
    I'm still looking for an explicit reference on the material I have available to me as to what the exact absolute range is. But as a "guess-timate" I would say that it begins at 40% to 50% (???) on the low end as well. I believe other HCH-II owners can/may also help me validate this figure as well through their own experiences.


    Last edited: Oct 25, 2007
  4. msantos

    msantos Eco Accelerometrist


    I can't say if I heard of that before.

    If I recall correctly, in all the literature I've been reading on the subject over the years, I have never seen any references to a change in the NiMH operational profile that indicated a positive trending break-in "period". All the data I have seen seems to imply the opposite when plotted as a function of time from the day of its manufacture.

    Then again, maybe it is time for me to send this question through to see what kind of answer we get back.


  5. Fenrir

    Fenrir Well-Known Member

    I read somewhere that NiMH batteries may take a few charge/discharge cycles to exhibit their full capacity. Not sure where- might've been on a pack of AA NiMH.
  6. Mike Dabrowski 2000

    Mike Dabrowski 2000 Well-Known Member

    I think Brick is correct.
    I was under the impression that 80-20% was the absolute range that the IMA will allow, not the range that the IMA tries to stay between so we were talking about different things.

    A related thing I have been thinking about lately:
    The SOC calculation is based on the bipolar current probe on the IMA battery pos leg. MIMA taps that sensor to generate the amp bargraph.

    The sensor is powered with a +12 and -12V supply from the BCM. Zero current is mid point between the two supplies, and corresponds to the neg leg of the standard 12V system.
    As part of the MIMA install, we measure the two supplies. Most are exactly +12 and -12, or 24V between them.
    I have seen some variation of this voltage on several Insights, with +11.8V instead of 12 on one recent Insights current detectors supply. If that slight offset was not corrected in software, it would bias the SOC slightly to the regen side, and could accumulate and cause the actual SOC to drift.
    Unfortunately I did not record which Insight it was so we could see if that car has recals or other battery related issues.

    Just got a call from a guy in CT with a 165K Insight that has an IMA light, and is facing some big repairs. He will drive up for this Sat hybrid workshop, and we may have another problem pack to play with to add to our information.
    If there was a requirement for cycling the NIMH batteries, it was probably done before installing in the car as part of the cell balancing they must do on a new pack.
  7. hobbit

    hobbit He who posts articles

    Part of the problem here, of course, is that the string is
    being monitored at a granularity of 12-cell blocks. [That's
    right for the insight/civic D-cell pack, right?] Like the
    Prius, as well. But if one CELL inside there starts to go
    south, that's going to throw off the whole measurement for
    that pair of submodules, and the BCM is going to throw a
    fit and start doing what it can to try and realign its own
    conceptual limits. A full cycle, similar to an "equalizing
    charge" done on big strings of 12V lead-acids in gear like
    big UPSes or PV systems, helps bring the cells in line with
    each other again, but it's a really brute-force way to go
    at it. On the bright side, slightly overcharging neighboring
    cells once or twice lessens the possibility of *reverse*-
    charging the one questionable cell, which will likely kill
    it. The only real fix here is per-cell monitoring and
    possibly even some clever electronics wrapped around each
    one to help re-eqalize on the fly when needed -- lithium
    chemistries need that sort of thing even more, since straying
    outside their limits is even more serious. But that's a lot
    more investment in control electronics, so here we are at
    the twelve-cell level.
    But the general inability to look in any closer leads to a
    lot of guesswork about what's really going on. Hopefully the
    take-the-pack-apart work that Mike and Ian are doing will
    turn up some useful info on "reconditioning". I'd like to
    see some of that actually go down to the cell level if
    possible -- with the Honda packs you can stick a probe through
    the tube and get at the outside of each cell's case. That
    would be a lot harder on a Prius subpack, but those have the
    advantage of somewhat updated manufacturing and seem able to
    keep the cells more aligned with each other in the first place.
  8. msantos

    msantos Eco Accelerometrist


    For example HCH-II has 11 monitored modules at 12 cells per module. And you are quite right... even if just one cell goes bad, it is enough to place the SoC anywhere in the map (depending on the failure's characteristics). Knowing which module that cell is in helps, but neither Auto manufacturer supports disassembly and repair on a per-module basis. So far, it still boils down to an "all-or-nothing" affair in that the service outlets simply replace the entire pack.

    Sometimes I wondered, if hybrids continue to proliferate at a good rate, perhaps there'll be a small market with few shops popping up here & there that will recondition whole battery packs - particularly for out of warranty vehicles... all on the account of replacing just a few bad cells. I would see reconditioned packs going for much less than $1,000 when bought as replacements, and a few hundred ($300-500) when repaired.


  9. xcel

    xcel PZEV, there's nothing like it :) Staff Member

    Hi All:

    ___No guarantee’s yet but Laurie’s HCH-I appears to be normal …

    ___Sean brought the SoC down to 1 bar from 15 while leaning on IMA heavily over the first 50 miles when heading down to Milwaukee late last night and early this morning. At 1-bar she was still assisting and regen-ing so the BCM appears to have lost track of actual SoC while IMA allowed both at a SoC that should not have been possible if the measured as shown per the SoC display was actual. That was weird? Anyway, from 1 bar, she came up to 17 of 20 bars and hung for the rest of his drive to Bradlee’s. 6x.x mpg to show for the effort. Sean drove from Bradlee’s to the meet at 70.0 mpg this morning.

    ___I took it out this afternoon and Assist was acting just as it should. A light touch and she blasts it out. Back off and bring her back online with a touch of accelerator and she comes up without a bar of assist while holding 17 of 17 bars of SoC. My 10 mile mostly highway segment yielded 72.6 and she showed 17 of 17 when we stopped. Bradlee and Sean pulled the 12V, forced a recal, drove it for maybe 5 miles of highway and she came up to 20 bars from 0 and held. One last hill dropped her to 19 of 20 and that is where she sits. We will know more tomorrow after Sean gets back home but everything appears to be working just as it should both before and after the forced recal.

    ___Good Luck

  10. MnFocus

    MnFocus hanging member :)

    *fingers crossed* I really am hoping this works
  11. Mike Dabrowski 2000

    Mike Dabrowski 2000 Well-Known Member

    Interesting test, it will be interesting to see the results.

    Just to keep things clear, what you are calling a forced recal, is really a computer reboot, nothing like a SOC quick drop recal with 12V active. The SOC does a positive reset up scale once it determines that it knows approximate SOC.

    I started a battery pack thread on the technical forum, as this is off topic here.
    Good luck with the car.
  12. xcel

    xcel PZEV, there's nothing like it :) Staff Member

    Hi Mike:

    ___IIRC, a recal is caused when the BCM sees a large enough difference between perceived SoC and a calc or map. A forced recal does the same. The BCM in each assumes/calculates the SoC is now at some minimum and forced charging occurs over the course of 3 - 5 miles to bring the pack back to a “known” high SoC condition. The SoC dance seen is exactly the same as I have seen in Randall’s and Justin’s Insight’s during their recal’s other then that one is forced vs. Honda’s own algorithms attempting to place the pack back into a known SoC. SoC drops to 0 and forced charging occurs to bring them back to this ~ known “SoC”.

    ___By all appearances, Laurie’s HCH-I BCM needed to more closely match actual SoC. The Assist at 1-bar is a telling sign that the BCM calc was skewed down the scale to some extent knowing how the Insight 5-speed and CVT’s as well as HCH-I’s w/ CVT’s act wrt a minimum 6 bar SoC when all assist disappears.

    ___Let us hope so for Laurie’s sake anyway?

    ___Good Luck

  13. laurieaw

    laurieaw Sorceress of the North

    THANK YOU BOTH!!! :)
  14. Right Lane Cruiser

    Right Lane Cruiser Penguin of Notagascar

    You are very welcome, Laurie!

    As far as I am able to determine, the problem was resolved. I played with it all the ~360mi home and I kept the majority of that distance with an absolutely full pack. It was only about 30F or so for the trip so lean burn was out, and the SOC stayed on 20 bars. It would drop out if any assist occurred, and I did run it down to 18 bars once. Background charging then brought it back up to 19 as it should and it didn't go back to 20 on its own... though I was able to regen back up to 20 and it did stay when I hit that value?

    Other than that, perfect behavior throughout. I told Laurie about this yesterday when I returned the car and conjectured that 20 would be unattainable after the SOC had been exercised a little lower than I had pulled it. I'll be interested to hear whether you are now unable to regen to 20 bars, Laurie?

    The car gets my special Sean seal of approval. ;) Not that it is worth anything much...

Share This Page