Gen II Prius Individual Battery Module Replacement

"matched" on what basis? voltage? Capacity?

Where did you get the replacement(s)?

 

In this particular case, surface voltage and from wrecked Prius batteries.

 

You were amazingly lucky. Whatever charging/discharging you did may have helped.

This is why I'm grilling you. The following is a plot of voltages and capacities for 28 modules from a 2006 w/154k miles:



As you can see, the voltages at the ends are a little lower, one is clearly bad at about 6.4V. The rest are all between 7.60 and 7.69.

Discounting the failed module, capacity ranged from 977-3963mAh

Note the following:
7.60V module was 3231mAh
7.69V module was 2026mAh
7.45V end modules were both around 2400mAh
Worst module at 977mAh was 7.66V.

One can only conclude that there is no reliable correlation between resting voltage and capacity on modules in the as-removed condition.

I'd like to think that anyone looking at the above would arrive at these conclusions:

1. The pack suffers from severe imbalance.
2. Simply replacing the 6.48V module would not achieve long-term results due to #1.
3. A long term repair would have to include balancing efforts along with potential replacement of additional modules.

Additional info:
The 7.60-7.69 swing in resting voltages is higher than what is typical.
The 0.015-0.025V drop at the end modules is greater than typical.

Once's chances improve the tighter the voltage bunch is, say within a total swing of 0.03V, and when the end modules are typically about .05-.10 lower than the other modules.

Steve

 

Great presentation of data here. Replacing modules on the premises of voltage is a shot in the dark. Measuring capacity is the only reliable way to measure the quality of cells.

SAMSUNG-SM-N900A ?

 

This is why I'm saying that all the supposed matching that people are doing often doesn't mean anything.
If you are doing this for yourself and don't have the right tools, do what you can but don't sweat the details.
HBS is an example of that. HBS didn't do any sort of complicated analysis. Just grabbed a cell and threw it in, and it lasted longer than some of the builds where people are spend weeks doing all sort of contortions. If you are going for speed, just do what you can and know that you will go back in and re-fix any issues again later.

 

Thanks.

It's not "the only reliable way," but it's an important part. More is needed. Measuring high-current voltage drop to establish outliers is also important. I have seen a number of modules that will give you good capacity at low currents. When you load them up with something more, they crap out. There are examples in this thread of using high wattage 12V bulbs for a set time period to a decent load on the module.

 

It would be interesting to see a graph with load testing vs. capacity vs. voltage.

I still think that measuring capacity properly is the most reliable way to approach a reliable hybrid battery repair.

As posted previously, 2 55w bulbs (headlights for instance) wired in parallel make for a decent load. Measure delta v over 60 seconds.

SAMSUNG-SM-N900A ?

 

I have hundreds... actually... probably close to 1000. They are very boring. Here's one:



You can continue to argue with me if you like. I'm speaking from experience and data.

Steve

 

The big boys each have their own "secret process".
You can do whatever you want.

Consistency is the real key.
Do the same thing over and over. Patterns arise.
Get rid of those that don't fall within the usual range.

 

Quoted for truth. Bolded for emphasis.

 

Why does capacity only go up to 4800 on that graph? I'm reading it on my phone and the resolution is piss poor

SAMSUNG-SM-N900A ?

 

Because that test only yielded ~4900mAh as you might have surmised by the voltage reaching 6.0V and 4800 was the highest even hundred number posted.

 

Can you post a graph that yields higher capacity?

SAMSUNG-SM-N900A ?

 

I have many graphs and videos showing those graphs - posted on my site
Testing Complex - HYBRIDS
and on my YouTube channel:
www.hybrids.co.nz - YouTube
I am doing that for you guys so that you know what you may expect from some rebuilders and looking for any suggestions of what else can be captured and published (like resent request from 3 rebuilders to automate scanning modules for dates).

Measuring capacity standalone is one thing. Good to know but not enough. To get the pack of modules to work their behaviour under load has to be the same - e.g. Delta V on pairs has to stay within the margin [I suggest 0.5V on pairs and 0.3 on individual modules] max within more than 1 AH band. Preferably over 3Ah band (could be 4.7AH URC @6A - in one of my old published tests). The terminology I have introduced a while ago is: URC - Usable Remaining Capacity of the pack (URC is not applicable to the individual module).

Imagine you have a team of navy seals - all working in sync. Their individual strength and abilities could be different, but they are trained to perform in sync. You can't "train" Prius batteries (although one professor here thinks you can if you slowly discharge them to zero).
Your aim is to find and match those modules which would perform in sync (i.e. their voltages will drop simultaneously under the same load and rise simultaneously under the same charge current within desired URC).
One batch of modules could be of lower capacity than the other - if the assembly (the pack) is showing minimum Delta V on pairs during charge/discharge witin desired URC - you are sorted. The bigger the URC the longer your battery last.

With my HV Battery Analysers I am measuring Delta V on packs of individual modules (e.g. 20 at a time) which makes testing more precise.
And final quality control test is performed on the assempled pack on pairs to capture URC of that pack.

My SOP (Standard Operating Procedure) include balancing but that is rarely required. That can be done if the need be in a fast and easy way by control dischanrge of individual modules down to 7.8V (or any other chosen threshold) without dismantling the pack. Still have to edit and publish that video which I've captured some time ago.

 

I can, but I won't. They basically look exactly like that except the numbers go higher.

 

Quick analysis and constructive feedback for S Keith graph:

Current dropped from 20A to 0 (zero) at the cut-off voltage of 6V on the module.

Capacity captured at 7V (that is what I am mostly interested in) under the changing load from 19A down to 11Amps was about 2.7AH.
That capacity from 7.8V down to 7V would be tested higher and the capacity from 7V down to 6V would be tested much lower if the constant load were used or if the load were changing not that dramatically.

Capacity captured within the range of 7V down to 6V in that test of yours was about 2.1AH (4.8 – 2.7) which is higher than normal.

I would suggest that the errors occurred in capturing capacity in those tests. Those errors (I think) are caused by the sampling rate of the testing equipment which has not been adjusted (not following) the huge change in the load current. As long as you can run those tests simultaneously for all modules in the whole pack under the same load – that would not matter, but if you are doing them one at a time those graphs woudl not be that easy to compare.

FYI: Capacity stored in the Prius battery modules within that voltage range 7 down to 6 is minimal. Could be about 0.1 – 0.5Ah in a working packs and can be neglected in bulk tests. I’ve done hundreds of tests (many are published) conducted under electronic load of 6A. Under electronic load current stays constant during discharge. That provides better representation, accuracy and individual test results are hence comparable.

FYI: In the live test I’ve conducted on brand new pack in Prius and on the very used pack the threshold when the MTG kicks in after discharge was the same – slightly less than 7V.

Measuring below 7v is of no use for us (only for demo purposes) as that is the area where Delta V increase and that is the area which is not used by Prius.

 

Ok. All that bragging made me think you were sharing something besides a pretentious attitude. Thank you for the 1 low capacity graph

SAMSUNG-SM-N900A ?

 

That was a pertinent one I had readily available on my Dropbox. I'm not being pretentious at all. I'm not here for your pleasure to do your bidding.

You're welcome.

 

I know that feeling. You show them picture which means nothing to many. Some people will think it is a GPS route to the nearest wine store..
So much remids me my day jobs in corporate Telco Hi-Tech world which suppose to only employ smart and tech-savvy individuals. Where I shared the knowledge how-to use tools readily available for peers and were going extra mile to demonstrate what is readily available for them for free - just use Help (F1) button and you'll be fine. But instead they just want one big magic button they will press and results for their business/financial/operational reports will magically populate on their screen. Some think that I am that magic button and did not want to make an effort at all to even think about doing somethign themselves....One told me yesterday - "your site is too complex and info is too difficult to digest" why you don't just come with simple Android app with one button and sell it for $5 on Google Play? I smiled as I do not want to upset my clueless friend... Sure we can do something with one magic button if the need be....

 

Asking for a graph with higher capacity wasn't a lazy solution to a problem. I don't have a problem that I am trying to solve here. I was merely trying to participate in a discussion as this is a discussion forum.
Bragging about having close to 1000 charging graphs was a pretentious move in my eyes considering the fact that most people are reading this entire post to learn how to solve a problem themselves.