NHW11 Traction battery autopsy

You would be measuring between one terminal and the battery case (the sheet metal). You'll have to take one measurement between the positive terminal and the case, and then a second between the negative terminal and the case. You won't be measuring across module terminals, so there is no danger of damage if your megger current is low enough.

What kind of megger do you use?

Not necessarily. The relationship between cell voltage and state-of charge is not linear. While low modules, such as your modules in the 6-7 volt range, are almost certainly degraded, modules with "normal" voltages can just as easily be degraded. You won't know until you assemble the pack.

This fact is difficult for some folks to accept. Some time back, there was a thread in which a Gen 1 Prius owner sought advice regarding his vehicle's battery pack. The information he gave indicated that the pack was in need of replacement. Despite this, he was advised to undergo a tedious two-week process of charging and discharging the modules.

Not surprisingly, the process failed, and the guy stopped posting. So it goes.

You can try. You won't know how well it works, however, until you put the pack back in the car and drive it for some time. Sometimes it takes several months to identify a weak module.

 

Then it will work for your tests.

That would indeed be a nice way to test.

Load testing individual modules is better than picking modules by OCV, but in the end, the battery ECU remains the final judge.

 

Table tags are more elegant but code tags work ok for holding formatting together.

Code:

Cell voltages: 
1    7.39       
2    7.35       
3    7.25       
4    7.29        
5    6.07       
6    7.32       
7    7.29       
8    5.63       
9    7.21       
10   7.06       
11   5.04       
12   6.13       
13   6.02        
14   7.13       
15   7.16       
16   6       
17   7.09       
18   7.16       
19   7.11 
20   6.05       
21   7.09       
22   6.11        
23   7.19       
24   6.23       
25   7.33       
26   7.18       
27   6.95       
28   7.3       
29   6.04       
30   7.16       
31   6.14        
32   6.19       
33   7.38       
34   7.41       
35   6.23       
36   7.46       
37   7.45       
38   6.17      

I also charted your data and found an interesting pattern. With the exception of one cell, the cells are divided into two groups, one near 6 volts and the other near 7 volts. I don't have a clue why but found it interesting.

 

When judging battery pack health, it looks for differences in state of charge. It needs to watch voltage, current, and module temperature for a period of time to do this.

Yes.

 

The 38 modules are in series. The cables at midpoint connect to the orange traction battery interlock. The idea is that when the interlock is removed, the connection at the center of the battery is opened up, to make the battery safer when it is being serviced.

It's still possible for you to be zapped when you touch one of the traction battery cable connections while the interlock is removed. For this to happen, the system main relay in the path to the cable connection (there are three relays, one switching the negative side and two on the positive side, one of which bypasses a current-limiting resistor) has to be stuck in the closed position and there has to be a voltage leak from a battery module to the traction battery case. Hence it is good practice to use a multimeter to make sure that you don't measure battery voltage across the terminal connections or from one terminal to the battery case. If you measure voltage to the case then you know you have a ground fault (probably due to electrolyte leakage.)

A current sensing device provides battery current information to the traction battery ECU. The ECU also monitors voltage of the 19 module pairs. If it sees that voltage of one pair is below the others, then obviously one or the other of the pair is bad and a DTC is logged which shows the location of the pair.

Given the voltage readings you logged, it seems that your battery has several modules that have failed. You'd probably be better off getting rid of all 38 modules. 2G salvage batteries could serve as an inexpensive source of replacement modules, although jk450 has raised a cautionary note about the length being slightly longer which may lead to cooling issues during normal operation, and safety issues in the event of accident impact.

Each of the 38 modules is composed of six cells. Each cell has a nominal voltage rating of 1.2V. Hence, each module has a nominal voltage rating of 7.2V, and the entire Classic battery has a nominal voltage rating of 273.6V.

Normally, the Classic traction battery voltage will be somewhere in the 300+V range (as high as 320V, for example) depending upon its state-of-charge.

 

The busbars consist of the copper plates that fit within the long rectangular orange plastic holder, on both sides of the battery.

 

Yes, here are the precautions:

• Electrically insulated gloves (rated for 1000 volts) must be worn.
• Ensure that the bus bar modules are installed in the correct positions. If any of the bus bar modules are installed in the wrong location, it may cause a short circuit and may lead to severe electrical shock.
• Do not use power tools of any kind or damage to the terminals may occur.
• Use the Special 48 in·lbf T-handle torque wrench. Do not over-tighten the nuts beyond the torque specifications.

5.4 N·m (55 kgf·cm, 48 in·lbf)

• Wrap the ends of the tools with vinyl tape to insulate them.

I suspect my HV battery's electrolyte leakage were caused by over-tightening of the nuts by my dealer.

http://priuschat.com/forums/international-owners/6539-prius-in-singapore-16.html

Some pictures from Art's automotive if you decide to remove your HV battery:

Installing a Prius HV Battery

 

Tables are best handled by using:

• "["table"]" - without the quotes, this is the table tag
• "|" - is the seperator between each column
• <RETURN> - ends a line or row of data
• repeat for as many rows as needed
• "["/table"]" - ends the table

Thanks!

Everytime we have someone investigate these units, we gain valuable insights. No one person has all the answers and I appreciate your observations.

Sorry to be late but I've been working on a ScanGauge hack. <grins>

Bob Wilson

 

Sorry but I was busy on another project this weekend but you can see the advantage of a specific thread on the battery versus a mix. I want to briefly cover safety, NiMH characteristics, specifics of this battery and a "get well plan."

SAFETY

Before proceeding, remove the buss bar assemblies on one side of the traction battery. Carefully remove the terminal nuts and pry up and off the buss bar on one side. I would recommend the one without the sense wires. Once you have one side removed, you'll have 19 pairs of modules. The key is only touch one nut at a time . . . put one hand in a pocket makes a lot of sense.

NiMH CHARACTERISTICS

Please study NiMH data sheets and the Wiki article on NiMH batteries. The key is NiMH battery state of charge and capacity can not be told by a simple voltage check. The radio control (RC) crowd has a lot of excellent articles about NiMH batteries and the problems of series cells. Critical facts:

• 1.2 V - nominal voltage per cell, six cells -> 7.2 V nominal
• self-discharge - NiMH batteries lose their charge just sitting on the shelf. Use, the charge/discharge, is one of the best things that keep them healthy
• 6.5 Ahr -> ~2.2 Ahr in use - the Dept. of Energy fleet studies and my own show capacity decreases over time (actually it looks to be fairly quickly measured in months.)

Ok, so let's look at your battery voltages plotted using 1.2 V as the major scale:

You'll notice the grouping around 1.2 V. boundaries. Looking at this chart:

• one module looks to have two bad cells - the one closest to 5V.
• multiple modules have self-discharged so one cell is 'dead' - notice they are spread across the board. This is the signature of a well worked battery (think "Deacon's One Hoss Shay.")
• most modules show six cells - slowly discharging, a bunch are still show six good cells but self-discharge is not your friend

Before going any further, after removing the buss bar on one side, use a felt pen with one color to number all modules 1-38. This will be critical if you plan to rebuild the battery.

GET WELL PLAN

You mentioned cost is important but I think it makes sense to call Re-InVolt who rebuild traction batteries. I see they have listings on Ebay and I know they do good work. As a depot doing hundreds of traction pack rebuilds, they have the tools, experience, and inventory to build a quality pack. However, Lucisous Garage and Arts Automotive are alternatives located on in the San Francisco area. I also understand $1,700 for Re-InVolt's best rebuild is money that may not be available but they have NHW11 module packs for less. They may be able to negotiate a little off if you offer both packs as trade-in.

Now if you need to rebuild your pack, you're going to need:

• multi-cell, six cell, NiMH charger - the RC community has them and a key element is dV detection with temperature backup. It needs to support a discharge and charge cycle and report the battery capacity in Ahr. I'm using an MRC 989 but at $150, it isn't cheap (but cheap tools . . . ) With 38 modules to measure, you may want to go with two lessor chargers to cut the time in half.
• clamps and board - (see Prius Battery Photos) to keep gas from distorting the case if the NiMH battery charger vs module doesn't work out. I like to use a maximum of 1 A but this stretches the cycle time.

My research and testing indicates loss of water in the electrolyte dries it out and leads to a permanent cell failure. As the electrolyte dries, it concentrates the current in smaller and smaller areas through the plastic mesh, separator. When the ohmic heating gets hot enough, it melts and shorts the cell. This is unrecoverable and the module needs to go to a recycler.

There is a Toyota/Panasonic patent on module refurbishment that involves replacing the liquid electrolyte. I have done similar experiments just replacing the lost water (i.e., electrolysis and battery chemistry produces a small amount of H{2} and O{2} gas that seeps out of the relatively weak, NHW11 terminals.) The hard problem is resealing the modules which my testing indicates must be done by well controlled, plastic welding, more than something simple hand-tools can handle.

So the first thing is to survey the 38 modules with a charge-discharge test. Record the Ahr capacity of each module and rank them. You'll need the current battery modules surveyed too, 76 modules (although I don't have any hope for #11 of this current pack.) Once you've surveyed all modules, take the top 38 and rebuild the vehicle pack. Don't forget the self-discharge effect, you are working against the clock!

Now look carefully at the buss bars. You are likely to see a lot of corrosion. Re-InVolt specially preps the wires to minimize corrosion but you can buy new assemblies from Toyota for a couple of hundred dollars. Some folks have tried to remove corrosion using chemical and scrubbing. I wish them luck but having dealt with the electrolyte, KOH, it is wicked, vicious stuff.

I would want to look closely at the terminals and use pool pH testing to detect terminal leakage (your P3009.) It may make sense to select modules showing less leakage versus ones with good capacity today but evidence of higher leakage.

Doing the job right, I would expect it to take a couple of weeks to properly survey the modules, $150-300 in RC charger(s), and another $200 in new buss bar assembly. It will also take at least an hour each weekday morning and evening during the survey period and then the weekend too. Realisticly, it is likely to take more than a month.

You'll want to carefully wash each module, individually, to remove any traces of KOH. Also, you'll want all the modules to be balanced. Finally, I would put a topping charge on the modules before putting the whole pack in the car so they all start at the closest to identical state of charge. Don't rush as you want this pack to last for years (although by the end you may be ready to chunck the whole thing!)

Bob Wilson

 

Have you personally rebuilt a complete Gen I Prius battery pack, using the methods you describe above?

If so, what were the results?

 

The method cited in the thread you found mentions a "DVM", which is presumably referring to a DVOM, rather than a megger, which you are using. However, it contains several significant errors; Wikipedia is not necessarily an accurate source of information. Here's some experience from the field:

Luscious Garage | Blog | Toyota Hybrids P3009: High Voltage Leak Detected, internal to battery

Keep in mind that if you don't have a P3009 code, you don't have a high-voltage leak to ground issue.

 

Thanks,

A good plan, to compare and contrast both batteries.

I didn't think of this before but it makes sense to ground the probes to eliminate any latent charge. But if the display goes to 0 V. when removed, what I would expect to see, then there is no latent charge from the instrument.

The meter input impedance affects high resistance circuits such as an electrolyte leakage to ground. This is what Wiki reports:

[ame="http://en.wikipedia.org/wiki/Multimeter"]Multimeter - Wikipedia, the free encyclopedia[/ame]

When dealing with circuits with megaohms of resistance, the DVM is just another resistor that changes circuit behavior. When the values wander, you're in a region where we don't know what we're looking at. But when the values appear stable, it is a trusable value.

This sounds like a high resistance, RC time constant charging a capacitor.

	Column 1	Column 2	Column 3
0	1	5.14
1	1/2	2.34
2	2/3	-0.65	Here
3	3/4	-0.234	or here
4	4/5	-5.9
5	5/6	-6.4
6	6/7	-10.8
7	7/8	-11.1
8	8/9	-15.5
9	9/10	-15.12
10	10/11	-20.8	Decreasing
11	11/12	-19.09	Decreasing
12	12/13	-25	Decreasing
13	13/14	-23.29	Decreasing
14	14/15	-31	Decreasing
15	15/16	-29.2	Decreasing
16	16/17	-36	Decreasing
17	17/18	-33.8	Decreasing
18	18/19	-42	Increasing
19	19	-39.7	Decreasing

Code translated to a table.

You'll notice in this bank the voltage reaches a value closest to zero volts at the 2/3 and 3/4 point. This is the area I would look for a resistance to ground in the series.

Code:

     
20      2.13     Decreasing       
20/21   10       Decreasing       
21/22   -0.2     increasing       
22/23   -4       Increasing       
23/24   -10      increasing       
24/25   -10      Increasing       
25/26   -16      increasing       
26/27   -18      Increasing       
27/28   -20      increasing       
28/29   -24      Increasing       
29/30   -27      increasing       
30/31   -27      Increasing       
31/32   -29      increasing       
32/33   -30      Increasing       
33/34   -39      increasing       
34/35   -39      Increasing       
35/36   -40      increasing       
36/37   -40      Increasing       
37/38   -44      increasing       
38      -29      Increasing

I changed the font to Courier, a monospaced font and aligned the text by " " insertion.

These look isolated which is correct. There is a loose change but no 'smoking gun.' The other bank shows some evidence of resistance path to ground.

I still see a local minimum at the 2/3 group. As for the other bank, maybe at 38 but not a smoking gun.

Bob Wilson