Honda Insight Forum banner
1 - 6 of 6 Posts

·
Registered
2006 CVT 179K
Joined
·
53 Posts
Discussion Starter · #1 · (Edited)
My goal is to find a method for measuring usable capacity of the IMA battery in my stock '06 CVT with an OBD2C&C gauge.

ASSUMPTIONS:
1. The BAT gauge on the instrument cluster is too smart. It seems to learn the usable AH of the battery after a few recalibrations and scale itself to match. An interesting approach for sure, but it hides any real data behind 20 bars of unknown value.
2. The SoC reported by the car on my OBD2C&C guage seems to be an Amp-hour counter that gets reset by recalibrations.
(I've attempted to verify this by recording the time it takes for 30A of assist to lower the reported SoC by 1%. I'm seeing around 8 seconds regardless of starting SoC. Likewise, 30A of regen seems to bring the SoC up 1% in around 8 seconds. I'm going to assume it is basically an Ah counter where 100% = 6.5Ah)

SETUP:
I have been enjoying my OBD2C&C for a while, thanks @retepsnikrep.
I'm using param value 12 for IMA Battery State of Charge(SoC)
I'm using param value 13 for IMA Battery Amps.

METHOD:
1. Force a positive recal by revving to 4k untill charging stops, battery amps drop to 0 and SoC is reset to 80%.
2. Deplete the battery by driving with lots of assist until a negative recalibration occurs, no more assist and the SoC is reset to 25%. Record lowest reported SoC before the negative recalibration. I usually get down to about 50% SoC.
3. Repeat number 1, force a positive recalibration. Record highest reported SoC before the positive recalibration. I usually get the SoC up to about 50%.

RESULTS:
Usable assist: 30% SoC or 2.0Ah
Usable regen: 25% SoC or 1.6Ah

CONCLUSION:
In theory, a new pack would assist from 80% to 20% resulting in 60% SoC or 4.0Ah of usable capacity.
So, I've got about half the usable capacity of a new pack?

DISCUSS:
Am I on the right track here?
Is this method useful for measuring IMA battery capacity in the car?
What would you change?

FINAL THOUGHTS:
I expected usable capacity for assist to be less than usable capacity for regen but there's a few variables I cannot control for, especially when trying to discharge by using assist while driving. Also, charging at 30A produces a noticable temperature increase which could lower the threshold for a positive recalibration.
 

·
Registered
Joined
·
2,022 Posts
It seems that if the car encounters a cell imbalance such that one channel in the pack reaches some amount below the highest channel (a difference of 1.0V can lead to a code) it will then treat the pack as having reduced capacity, at least for a while. The IMA will then attempt to rebalance the pack by overcharging it, which may raise the SoC of the low cell/channel and bring it closer to the others. A situation where the reduced capacity is temporary is if the car were parked for a long time and the imbalance developed. If the pack was otherwise good, normal driving of the car (and the overcharging by the IMA) should help rebalance the pack and the reduction in capacity would no longer be needed.

So it seems your method could be accurately measuring "available capacity" which is the capacity available to you as a driver, given that the system may be dealing with a pack in which an SoC imbalance between channels was detected by the BCM.

If you are in fact dealing with a problem pack, the lowest SoC cell in the pack determines the SoC of the pack. Each cell, including the lowest SoC cell, may actually have the capability to deliver nearly full capacity, but can't because of the imbalance. (To understand this requires some amount of noodling of how series charging and discharging works). However, getting all the cells in the pack at the same SoC - and staying there while the car is parked - and restoring imbalance in SoC while driving - is the thing that the IMA isn't really capable of doing on its own when cell(s) have rapid self-discharge or can no longer deliver the amount of current demanded without an excessive voltage drop or have actually reduced capacity.

Your measurement method may be the best metric as far as non-invasive measurements with available tools go, if the problem with the pack is not temporary (otherwise good pack not driven for a long time) because the general experience is that a problem pack will develop the same problem eventually even if things like grid charging and discharging are performed. That is, even if external attempts to correct the problem are performed and the BCM reset, the "new capacity" is likely to only be temporary until the underlying problem resurfaces.
 

·
Registered
Joined
·
8,390 Posts
My goal is to find a method for measuring usable capacity of the IMA battery in my stock '06 CVT with an OBD2C&C gauge...
Maybe I'm not understanding exactly what you're trying to achieve, because your methods seem a little more convoluted than they need to be. Just charge the pack until it won't charge any more, reset your OBDIIC&C amp-hour counter, drive using assist until neg recal, read the amp-hour count (or calculate net amp-hours if you weren't able to do only assist during the drain). The result is usable capacity. If cells are imbalanced your measured usable capacity is based on the cell/s that charge 'full' first and discharge to 'empty' first, but you still have a measure of the capacity that's usable (or, if not usable in the strictest sense, a measure of what's actually used; the pack could have more capacity that's potentially usable if you did some manual stuff, circumvented the BCM's rote routines)...
 

·
Registered
2006 CVT 179K
Joined
·
53 Posts
Discussion Starter · #4 ·
Maybe I'm not understanding exactly what you're trying to achieve, because your methods seem a little more convoluted than they need to be. Just charge the pack until it won't charge any more, reset your OBDIIC&C amp-hour counter...
I would use it if I could. AFAIK my OBDIIC&C doesn't have an amp hour counter. Maybe it was removed from the firmware at some point? Maybe there's a different way to access it? I see it in the manual "Ah Wh Time counting display (Right)" but right takes me to the 2nd page of parameters. Right a 2nd time takes me to some other screen full of mostly zeros.
Rectangle Font Gas Paint Number

Right a 3rd time takes me to a (LTO?) pack monitor screen.
Temperature Font Gas Rectangle Auto part
 

·
Administrator
Joined
·
13,931 Posts
Ah counting was only available in early non CAN versions of the gauge firmware that used the 18F2620 PIC IIRC.
It's not been in the software IIRC for at least 5 years..

If someone has a version with Ah counting that works I suggest they post the version number/splash details and I will look in my archives or they read it from the PIC with the usb programmer and upload the hex for others to download.
 

·
Registered
Joined
·
8,390 Posts
hmm, didn't realize that. I've been using the same ancient version, "0.76" I believe, for ages. I've attached a config I had at one point, you'll have to change the extension to .hex...

Fly3cyl, the method you described then sounds about right - what you say about SoC percent vs. amp-hours is about how it is. Each percentage point equals 65mAh, though there's a few exceptions. I think BCMs that pos recal directly to 80% have a weird jump down from I think 72% to 65%*, so you don't see SoC percentage values between 72% and 65%, at least on the way down. And nominal SoC hangs at 72% on the way up, and 28% on the way down, at least on versions that pos recal to 75%...

* edit: OK, here's an image I have of a dataset that shows that weird jump down, it's from 72% to 65.7% here:
Font Parallel Pattern Symmetry Handwriting


hmm, while I'm at it, I have this ancient scatter plot I made when I was trying to figure out if indeed SoC moved linearly, trying to figure out what each percentage point equaled in mAh. You can see how most dots form a straight line, but there's outliers - and those are the hangs and such. What I did was record the change in nominal SoC and the total amp-hours for those changes, over multiple drives and various charge-state windows, and plotted. So for example, reset Ah counter to zero, drive and use say 25% charge state, stop and record the total Ahs for that 25%, rinse and repeat. Each datapoint reflects that procedure.
Rectangle Slope Plot Font Parallel
 

Attachments

1 - 6 of 6 Posts
Top