[Armin]

So after years of trying to figure out how to monitor all 10 subpack voltages simultaneously, I came across this $5.50 multimeter. With a piece of plywood, lots of wires and some tiewraps, I built this contraption.
Photo!
The connection to the battery pack is via Mike Dabrowskis pogo-pin fixture.

The aim was to try to understand recals at the subpack level. Of course this is not as nice as a real 10-channel chart recorder. Especially when one needs to watch the road... So I don't have any real hard data, but I can share some basic observations after driving around with this for a week:

One often mentioned theory was that recals are triggered when the difference between the highest and lowest subpack voltage exceeds a threshold. Well, this is not true. Or at least, it is not the only condition that triggers a recal. I have watched voltage differences as high as 1.2V under max assist at times, without a recal following. I have also watched a recal when parked in idle-stop (much safer to watch all those meters then), when the difference was less than 0.5V.

Also, while even after a recal, assist is still available using MIMA, this too will stop after a while. This second, hard cutoff appears to be based purely on Ah, not on battery voltage at all. (I drove untila a recal, then grid-charged the battery, but not enough to trigger an upper recal. After a few miles with moderate MIMA-assist, all assist was disabled.)

I can see different behavior in the sub-packs, though. The one connected to the meter at the bottom left is usually the lowest at no-load and during assist it drops faster than the others. But during charge, it goes higher than the others. (the photo was taken after I just connected my grid-charger) One way to describe this would be increased internal resistance in this sub-pack.

Another thought about recals was that they are there to prevent cell reversals. I haven't been able to observe a cell reversal, yet. Even after draining the battery as far as the hard cutoff beyond the normal recal.

 

[Mike Dabrowski 2000]

Cool idea
How close are the meters readings when you connect them all to the same subpack tap? Usually those cheaper meters are not too accurate.
I just got a gift of a box of analog edge panel meters, from someone and was thinking of setting them up in a similar way. The differential between subpacks would be easier to see at a glance on an analog meter, but the resolution would not be there.
High internal resistance on a 6 cell subpack could be a single cell that is comprimised?
Do you have a video camcorder that could be set up to record the meters,with enough resolution to read all of them?
You could forget about watching, and just drive while verbally discribing what is happening, then look at the tape to see what the meters were doing?

 

[Armin]

Hi Mike,

good point. I checked a couple of the meters when I installed them. They were within 1%, so it's probably safe to ignore the last digit on the display

I was orginally looking for panelmeters, too, but couldn't find anything that cheap! And most of them need power, which is not isolated from the measurement input. So you either need a bunch of isolated DC/DC converters or batteries. This seemed like the least-hassle solution for a non-recording system. But then, if you happen to have the panelmeters... They should at the very least make for a more compact design!

Video woulkd be cool, but remember the title of the thread...

 

[Mike Dabrowski 2000]

Even poor men can have a VCR.
The panel meters are 1 MA full scale, but the scale is either 0-30 or 0-100.
Have been gathering parts to make the opto isolated monitor that you and I spoke about several years ago. This could interface to the labview data acquisition card and program, so real time data logging could be done, but your system with a video recorder could capture the same information.
I have a nice MiniDV and have made a car mountable mounting system for it, if you want to get together and give it a shot.We all really want to know what causes recals.

 

[Insightful Trekker]

Not to discourage this very interesting thread but it has been reliably reported by at least one member who should be "in the know" that the Insight _measures_ total power I/Os.

That's at least one significant parameter for a recal.

HTH!

 

[Mike Dabrowski 2000]

John
We realize that, but are looking for a quick and simple monitoring solution.
If a comprehensive monitoring system were to be built, the AH in and out would be measured, as well as the pack temp.
http://www.99mpg.com/articles/batterypacksexpose/
The link shows the prototype of this system which has not progressed much since my pack just keeps on running with out a single classic recalibration .

 

[Insightful Trekker]

Understood. 8)

The above will "only" be able to identify sub pack voltage variations as they relate to a recal. Although I am wondering _IF_ this parameter will even trigger a recal event. Hopefully Armin will share his findings as he progresses. :!:

The other 2 "known" factors of temperature and AH won't be in the picture.

Sincerely,

 

[Armin]

Hi John,

are you suggesting that discharge capacity (Ah) is used to trigger a recal? That seems odd to me, since the common assomption is that Ah is used for the SoC indicator and a recal would only occur if some other condition signaled an empty battery while the charge integration-based SoC still thought it had some capacity left.

Among the ways that I think this could be done is the voltage difference between subpacks, as had been suggested earlier and as I seem to have read in the technical manual that someone had once posted on the net. And a lookup table of cutoff voltages as a function of temperature and current. That second one seems more likely from what I found so far, but it will be very hard to confirm without serious instrumentation as Mike suggested.

One of the thoughts behind my system was also my concern that I might be deep-discharging the battery with MIMA, if I continue to force assist after a recal. But I'm starting to relax about that one.

 

[Insightful Trekker]

Insufficient number of words on my part Armin. Sorry for putting you to the trouble of further inquiry.

One scenario causing a recal is when the total accumulated power I/O contradict the high or low voltage limits of charge. For others reading here - a potentially damaging condition when excessively repeated. IIRC the credible post reports that tables are not used. However, I would think that a temperature vs. voltage table would highly desirable. But its been awhile since I've extensively reviewed NiMH battery specifics in that detail. We know that IMA current is reduced or eliminated of the packs temperature is too hot or cold.

Sub pack voltage differences must be monitored due to the complexity of using a 120 cells in an array. Your posts in the past have clearly demonstrated your understanding on that topic. When / if a sub pack individually reached the 100% or 0% SoC voltage point then a recal would also be required.

Monitoring individual sub pack voltages for SoC would allow for "intelligent" battery array control. Further narrowing the overall used power I/O as individual sub-packs begin to show their age. And is a good explanation of the more fragile SoC that I am increasingly seeing as my Insight ages.

 

[Armin]

One scenario causing a recal is when the total accumulated power I/O contradict the high or low voltage limits of charge.

John,

sorry, but I might be a little slow this morning. I don't think I understand what you mean. But I haven't been following IC for a while , so I might have missed this discussion.

I always understood the common theory as this:
SoC is determined by current integration over time (is this what you call power I/O?). When a low (or high) voltage limit is reached while the SoC is not empty (or full) a lower (or upper) recal is triggered.

Now my speculation:
The voltage limit could be on the entire pack or subpack level. For subpacks, it could be relative to other subpacks or absolute. And it could be modified by temperature and/or current, with a compensation algorithm or lookup table.
I don't see what oher measured data the BCM has to trigger a recal. And I don't understand the term power I/O.

I do undersand the desirability for the BCM to monitor subpack voltages, ideally down to the cell level. However, I don't think much can be done with this information, other than limiting assist or triggering recals based on voltage thresholds (again possibly compensated by other parameters). There can't be any active balancing because of the high series resistors installed in the subpack voltgae taps.

So far, I haven't seen assist limited as the voltages drop. But this refers to the limits still in place with MIMA. MIMA does override the limit imposed based on low SoC.

 

[Mike Dabrowski 2000]

Armin and John
When I get down to 2 bars or less, assist is limited, and gets proportionately less as I approach zero SOC. I cannot override that with Mima. Once I get back over 3 bars of SOC, I can again get full assist.
The best explaination I have seen for the Insight SOC operation is this:
http://www.99mpg.com/Data/downloads/rel ... nation.jpg
This was copied from the large word document that Yves posted of a honda service tec training manual.

 

[Armin]

Mike,

that's exactly the diagram I had in mind! I didn't know you preserved it. Thanks!

This seems to support the theory of voltage cutoff based on sub-pack voltages. And I guess it must be current and temperature compensated. Hmm I can think of a few experiments to try to prove this, but need to find the time!

 

[Insightful Trekker]

Better late then never,

Yes Armin,

I/O - input/output - intergation function (a more precice for this point and mathmatical term). Sorry for the delay.