module pairing: match similar or opposites?

You want to pair the highest internal resistance with the lowest internal resistance. Probably high internal resistance one is the low capacity one and low internal resistance one is high capacity one. Prius doesn’t allow the battery to get so low that it should be problem that high capacity one is paired with low capacity one.

 

Not sure if I'm right or wrong, but on the ones that I've rebuilt I simply put them in order of capacities and paired them up. I put the worse ones at the front on one pack (near ECU) and on the rear in another. That was to see if it made a difference, but I won't know until both batteries fail. I hope that's not for a long time. No problems with either yet.

 

Have not done this yet. Based on posts I have read, I would try to pair strong + weak cells in such a manner that that assemblies voltage and voltage drop under load stays as stable as possible across the cell assemblies. I would be trying to avoid having the ECU detect the .3 volt difference among the cell assemblies under load which I understand triggers the HV battery problem codes.Good luck!

 

RyanG,

In the same boat as you but has been given guidance from those that knew more and have had lots of experiences than us which this forum has a lot of. I’m close to getting all my modules done and planning to do as follows:

1. Match series Resistance instead of module capacities. I will not match or pair modules based on capacities. So far, all my modules after the balancing process are within 500 mAh of each other.
2. Highest module resistance to be matched with lowest resistance module so that if possible, all my module pairs are the same or very close in resistance.

The voltage drop under heavy load across each block will be much the same. It is the voltage difference between the blocks under load that the ECU sees and issues a fault code if the voltage varies out of specification.

I can send you a picture of how I measure internal Resistance if you’re interested. Again, I learned it from the forum and gotten their blessings that it is the right way to measure it. I also experienced the wrong way of doing it.

Danny

 

No worries Ryan. Our experiences will make us more informed next time.
I'll be anxiously waiting for your new thread as to how your experience goes.
To success if not now, then later

Danny

 

.....

Prius NHW20 uses about 15% of the new battery designed capacity (Tested on My HV Battery Analyser).
Your goal is to make sure that within the operational range (and above) moduels are in conformance with each other....

Within that range - the load when EM starts and before ICE kicks in - last for few milliseconds and could be around 60 Amps (55-67Amps as observed during live tests with my Analyser attached).

Delta V on pairs (Vmax - Vmin) during those extreme loads is up to 0.5V (on brand new as well as on the second hand - e.g. 67% usable capacity remaining pack) without error codes.

Voltage drops at those extremes is below 7V per module - on brand new and second hand packs (as tested with my HV Analyser(s)).

FYI: By running test with my tools and having brand new pack as a reference - I know how much "usable capacity" is left in the second hand pack on the spot.

May I introduce to this forum my terminology:

- Module Remaining Capacity - measured from full charge to the discharge level of 6V (at specific DSCH current - usually at around 6 Amps)
- Pack Usable Capacity - down to the level when Delta V (on pairs) is below or equal 0.5V

Please note: difference in individual module capacity around 6v in a good pack is minimal.

I've seen working packs with no error codes which have few modules with less remaining capacity than other modules, but the usable capacity of the pack was still big enough. Less usable capacity of the pack slightly impacted fuel efficiency (owner's feedback - 20km/l vs 21 km/l).

You read success stories from people when they just change one or two modules - being lucky to get good matching ones and think they've fixed the pack. Technically speaking they did - temporarily - great for DIY person who have plenty of time.

But when garages do the same and sell the pack for the prime price with 3-6 months warranty where in fact their pack may have usable capacity of only 37% - I feel that it is unfair to unsuspecting customers.

Not to mention sellers of individual modules on eBay who claim - "good working module", "tested", without disclosure of the remaining capacity under specific load (e.g. 5-6Amps) and ask $40-45 for what may not worth even a 1$ (e.g. 2Ah left).

I've gave away once the whole pack to the guy who was building his own charger - it was only 1.3 Ah left and I told them capacity and advised that it is not good for being used on the car. Found the next week his intention was different from testing his charger - he put it on the auction - selling each module for $45 each claiming "tested, working" advising "change one or two modules and she'll be fine"...bla bla.

I am no longer giving away low capacity modules - full dischage to zero v and dump. There are people out there who would try to get a buck out of junk with no remorse....

 

Dan. Please can you also send the picture to me? In interested on how you measured the resistance.
@Danny Gonzalez.

 

Any updates here?

 

ISTM the answer to the thread-title question depends on your objective.

Pairing a strong with a weak is a way to make the ECU think the battery is made up of better-matched modules than it really is. This is good if your objective is to avoid seeing error codes shortly after your rebuild, but not so much if your objective is to have your rebuild last a long time. You can hide the mismatch of those two modules from the ECU because it doesn't have an extra voltage sense wire between them. But therefore the ECU will let them get repeatedly flogged with overcharge/discharge because it doesn't know better, which will shorten their lives and bring eventual error codes anyway, further down the road.

For long-term longevity, you want to avoid having that happen, which means pairing well-matched modules, so what the ECU can see about the state of each pair will really be an accurate picture of both modules in it.

Of course, to do that and not see codes right away, you could need to start with a better-matched set of modules all 'round, since you don't end up concealing the degree of mismatch from the ECU.

 

I think I may have maybe figured out why some people do the "pairing with opposites" style.

Based on this it seems that modules tend to fail in the middle. Which is also where it gets hottest.

So if you want your weaker modules to fail less, put them near the sides. And if you want them to get even less heat (perhaps), put them between some high quality modules, that won't generate much heat of their own (the most heat is generated when modules reach max, which weaker modules do more quickly).

I was thinking it was because a weak one "affects" its neighbor but turns out they don't affect their neighbor too much.

So by pairing them with opposites at the ends (with the middle being "closely matching"), you get the best heat system for the weak cells. So maybe the target is not actually to dupe the ECU, but to protect weak cells. And then it accidentally dupes the ECU?

 

Assuming this is the case, sorting based on "capacity" (putting weak cells at the end, nearest the fan), may not be a bad option, though weak cells heat each other still, the fan does a lot of cooling on the edge they end up on. And the "most" cooling on the edge nearest the fan, better by 1 degree only

It also seems to be the case based on this that the very edge cells get degraded slightly over time.

So maybe an alternative that doesn't dupe the ECU but still tries to "balance" weak with strong cells might be.

[ecu]
1. strong (for the edge)
2. strong
3. weak
4. weak
5. strong
6. strong
7. weak
8. weak
...
21 weak
22 weak
23 strong
24 strong
25 weak
26 weak
27 strong
28 strong (for the edge)

It's like balancing pairs, without duping the ecu? Just a hypothesis

 

I suppose it's also possible that the ECU is fairly conservative, enough so that if it's duped it won't be enough to matter and degrade the weak modules, and since they're in the cooler edges the under charge/over charge won't hurt them anyway. Would be interesting to see some studies comparing the ordering options, but most seem to agree that weak should be at the edges somehow

If that is the case then why not grab a little efficiency I guess?