[Armin]

A few months ago, there was an Insight engine (including IMA etc) on ebay. Matt Muelver, who is on the yahoo group, bought it.

He sold me the BCM last week (the box that monitors battery sub-module voltages and temperatures). I posted a picture here:

http://homepage.mac.com/kusig/PhotoAlbum2.html

The one thing that seems pretty obvious from looking at it: there is no active charge equalization. This surprises me somewhat. I still don't understand how they keep the cells from building up major differences over time.

The ten green single inline modules you see along the top are connected to the module voltages. Apparently, they provide galvanic isolation for each voltage input. On the other side of the isolation barrier, they have a power supply bus (6 lines, for some strange reason!), a signal bus (two lines) and one line each going to the micro.

My going theory is that each SIM has an A/D converter on the battery side of the isolation and optocouplers for the digital signals (probably an SPI or similar serial bus). This would make a lot of sense and explain the two bus lines common to all SIMs and the one chip select unique for each module. The overall size of the SIMs and the tiny traces going to the battery all indicate that there is no power supply for charge equalisation.

Some other facts: The micro is a Hitachi H8. The frame is ultra-light magnesium alloy (Honda crazyness!). The rest of the circuitry is pretty much temperature measurement, two-speed fan control, serial bus to MCM and a bunch of power supplies for different isolated circuits.

 

[Insight Driver]

When I can remember if I still have the link I read a paper by a research institute on the charging scheme for the Insight with their commentary that it is a very conservative one. I don't remember enough details but I do not recall NimH chemistry needing balancing like lead acid batteries do. Not having the facts at hand leaves this as just conjecture on my part. I hope to remember to look it up. I actually work on power conversion equipment so this stuff is right up my alley.

 

[Rick]

I dunno know, it seems that any type of battery pack with multiple cells in series would need an occasional equalization. And I at one point witnessed what almost seemed like an equalization charge while comming down a very long down hill grade. The trip to the top of this hill was a very long and gradual climb that could be done in 5th without any assist and then it dropped fairly rapidly for several miles. Going in to the down grade I had all but the last bar of charge showing.

I used a little bit of brake tap regenerative braking to slow me down at first, but it pretty quickly filled up the last bar, so at that point it was either ride the brakes or enjoy a twisty windy fun road at 85mph via gravity in fuel cut so I just let it glide. Just going without touching any pedal the meter showed 4 bars of charge. If I tapped the brake they went away. For about 3 or 4 miles, maybe longer I just coasted like that and it stayed showing 4 bars of charge while the meter was showing full charge. It finally did let up when it was satisfied that it was charged fully and there was still a few more miles of downhill ahead.

I could be wrong, but I sware it seemed like the car was putting an equalizing charge on the pack at that point. And considering how normally charge burns off while using assist, that top bar disappeared more quickly that it should have after that. I could be wrong, but it just kept charging for a while.

But, interesting question, does the car every equalize the charge on the pack. Perhpas recalibrations have something to do with it as well.

 

[Armin]

equalization

I do not recall NimH chemistry needing balancing like lead acid batteries do.

I dunno know, it seems that any type of battery pack with multiple cells in series would need an occasional equalization.

I go with Rick on this one: No matter what the chemistry in the cell is, there will always be minute differences between individual cells. (Although I do know that Panasonic provides specially matched cells in the packs made for Honda and Toyota.) The minute difference at the start will likely grow over time and make the weakest cell fail early. It's always the same: the weakest link determines the performance of the whole.

Rick,
I have seen what you described sometimes, too. The car usually follows a fairly predictable pattern of when and how much it charges the battery. But sometimes it goes off on a whim. There may be reasons of battery-maintenance for this.

However, by equalization I refer to a specific process: find the cell with the lowest charge and charge that one cell until it has the level of the others. Or reversed: find the cell with the highest charge and discharge it to level.

What I was saying is that apparently, the circuitry in the BCM does not support this.

 

[Insight Driver]

All my experience is in lead-acid chemistry. To equalize a large battery pack they are overcharged for a set period of time. The cells that go to full charge then just pass the current and take no more charge. The weakest cells finally fill up to full charge. The whole pack is overcharged, they don't equalize individual batteries.

What I remember reading about NimHD chemistry is that they would be optimally charged by regulating their temperature. What I mean is put in a current and monitor temperature rise and limit the current to maintain a specific charge temperature. This is not done in the Insight as far as I remember.

It could be (theorizing here) that the "whim," of the car's decision to charge equalize depends on measuring the individual cells. The other "whim," which determines when the batteries may be equalized is when their temperature is not too high or has cooled down after getting hot from a long discharge period previously.

I still haven't gotten around to finding the facts. My memory is not the best any more. I seem to remember the numbers 40 and 80. The Insight batteries are not allowed to go below 40% charge (zero bars on the display). The batteries are never allowed to be charged over 80% of capacity. It is for this reason the batteries should last as long as Honda warrants them to.

 

[Will M]

Lead Acid batteries become weak when the sulfate hardens on the lead plates if the battery is left at less than full charge for too long. This blocks further chemical reaction between the sulfuric acid and the surface of the lead. You break this hardened boundary by overcharging the battery, which boils the water at the surface of the plates, physically blasting the sulfate off of the plate surface.

Cells that don't need this simply boil water harmlessly, while cells that do need this boil the water and clean their plates. In the end, overcharged batteries become equalized and require the addition of water.

Perhaps nickel-metal-hydride batteries equalize by a different process. Perhaps the weakest cell is "hungriest", drawing down the overall voltage most and drawing the most amperage from the charge. If this were true, a battery of these cells would self-equalize during charging. I'm obviously speculating.

As for numbers, I remember 80%-20% limits, not 80%-40%.

 

[Armin]

equalizing by overcharging

Will M describes this process very well for lead-acid. One of the most understandable reason why this doesn't work for us is this: you cannot add water!

The overcharging in a lead-acid battery simply uses up water, which you can easily replenish. So just charge all cells until the last one is full and then top-off electrolyte on the others.

In NiMH (or NiCd), overcharging can cause excessive gassing which will, at the very least, deplete electrolyte irreversibly. If done too rapidly, it can also cause temperature and pressure to rise untill the cell explodes.

So no, I don't think the Insight equalizes by overcharging.