NHW11 Traction battery autopsy

I'm struck by the irony of two safety hazards, one well known and the other known to those who deal with these batteries. The lethal, series voltage has been well known although the relatively easy mitigation, remove the buss bars from one side, not being as well known. In contrast, those who handle these modules are aware of the KOH electrolyte and its ability to dissolve just about everything that it comes in contact.

The KOH mitigation of boric acid both to neutralize and work as an eye wash is something I don't remember seeing discussed before. But it is such a simple and sophisticated safety feature. Heck, I use dry boric acid to handle insect pests. But neutralizing KOH does not eliminate the conductive path.

Acids will neutralize KOH and make a salt, a conductive salt. With the pH neutralized, the organic and metal eating characteristics are significantly reduced but the salt trail remains. It may not be as aggressive in 'seeking a current path' but it remains a leakage current path. Now that the original battery modules are in a storage mode, I'm curious about the operational battery that reported the P3009.

The ScanGauge can be programmed to report the traction battery codes but unless I missed it, I don't think it can reset the codes. So repeating P3009s would be difficult to observe unless the 12 V. ground is removed long enough to reset the ECUs and wipe out the codes.

The ScanGauge can be programmed to read out the 19 voltage pairs. Knowing how leakage leads to unbalanced modules, it becomes possible to interpret module voltages and manage the existing P3009.

Speculation on my part, if you decide to refurbish or rework the P3009 traction battery, it makes sense to start or restart a fresh thread. Perhaps starting with the last P3009 code, a reset, and survey of the 19 module pair voltages. Then see if if the P3009 is coming back and the 19 module pair voltages are changing over time. With the insights gained from reading the traction base corrosion, we have another tell for where there are electrolyte leaks.

Bob Wilson

 

Great data!

Looking at the discharge capacity, my eyes see three groups:

• recoverable 7, 1, 9, 4, 36, 14, 34, 15, 6, 26, 37, 28, 25, 2, 33, 3, 23 - these look to be recoverable by adding water to replace the lost H{2} and O{2}. There is some minor electrolysis of water over time and if the seals work, the gas remains in the modules and eventually recombines, metal hydride catalyst, to water. But weak seals allow the gas to escape the modules.
• chemistry change 18, 17, 19, 27 - this looks like a transition in module chemistry. Perhaps some of the rare earths are going into solution and changing the metal hydride.
• dead or dying 20, 8, 29, 11 - I've seen the curious, 8V behavior in the last module I tried to rehydrate. I could not figure it out but suspect a different chemistry is involved.

I have more confidence in the discharge capacity than charge capacity. Your data also shows the weaker association of voltage to module capacity and health.

Although rehydration works, the leaky terminal seals has me stuck. The original H{2} and O{2} leaked out via these seals. Without some way to make the seals more effective, the rehydrated modules will lose their water too. But I'm thinking something that moderates the internal pressure would be a better answer. Reducing the mechanical stress on these weak seals might be the answer. Sealing the modules remains a problem.

For the rehydration ports, I tested JB Weld but am not sure it is the best answer. I notice someone else is using a rehydrated pack in operation but I think they are using stainless steel, sheet metal screws (an interesting approach.) The Toyota patent discusses plastic welding.

I was confused about the bubbles from the acetic acid until I remembered Doug reporting that KOH combines with CO{2} to become potassium carbonate. When the acetic acid reacts to become potassium acetate it releases this trapped CO{2}. KOH and acetic acid should not release any bubbles.

I really appreciate this data. When I was doing my initial studies, I was too focused on individual modules and didn't properly survey the whole pack. Great job!

Bob Wilson

 

Owch, this is a hard problem and I sometimes forget why we went to the swamp. How to evaluate the capacity of a module-pair in an operational traction battery. So with the ScanGauge, we can measure operational:

• traction battery SOC
• 19 pair voltages
• 19 pair internal resistance

First, take the ones on the bench, top group and put them back in a parallel, trickle mode or pair them in series and then parallel all pairs. Note, I am not sure if it makes more sense to pair strong-to-strong, or strong-to-weak in the pairing but I'm leaning towards strong-to-weak pairing so all series pairs will have similar, dV behavior during charge and discharge. I am sure that absence of charge or allowing self-discharge is death so they need to be kept 'alive.' As for the fubar or marginal modules, time to let them go towards the recycler (any Batteries Plus should be able to handle them.) The bench pairs of modules, in series are building the replacement, module pairs for the weaker pairs that are in the operational traction battery. Now the hard part:

• measure replacement pairs - internal resistances
• put pairs in parallel to sustain trickle charge

So the challenge is to find someway to measure module-pair capacity in the operational pack. Speculations:

• Measure SOC and 19 module pair voltages - using ScanGauge
• Force charge traction battery, needs to be done to a cold battery because of heat generated, target 75-80%
• Measure SOC and 19 module pair voltages
• Put in "R" with wheels blocked and discharge traction battery to bring SOC down, target 45-50%
• Measure SOC and 19 module pair voltages

The slope of these 19 module pairs should identify the weakest and strongest pairs. The goal is to replace the weakest as well as remove any electrolyte leakage paths. Identify the weakest pairs.

Speculation, during replacement activity:

1. have new buss bar assembly ready to go (I understand Re-InVolt uses nickel plated ones.) Anything that can be done outside of the vehicle down-time should be done and buss bars are relatively easy.
2. have replacement module-pairs charged to 80% or better . . . they will be balance charged during the down-time.
3. remove operational pack and mark the module pairs
4. removed the buss bars and replace the marginal pairs
5. remove any electrolyte paths, use vinegar to detect
6. install one side of buss bars (so there will be 19 series pairs)
7. wire the 19 pairs in parallel to equalize the charge voltage, give it an hour or so. You may want to use low ohm, ~2-10 ohm, current limiting resistors for the initial connection, a jumper, and then connect the parallel buss. Wait an hour or so for everything to settle.
8. install the last buss bars
9. re-install the traction battery

I agree. A P3009 is not good but if it doesn't come back, lets monitor it for now while keeping the replacement pairs in ready reserve. When module-pair voltages begin to exceed 0.3 V, repairs will soon be necessary.

Which is why I'm suggesting watch and see if the P3009 returns or a more severe error.

The key is making sure the pairs are at equal voltage and as well matched in capacity as possible. I'm thinking take the car down on day 1 and swap the module pairs by evening. Leave the 19 pairs in parallel over night with a trickle charge. On day 2, install the last buss bar, close it up, and reinstall.

There are not many who understand the risks. The first step to mitigation.

This one is solved by setting the mAhr limit on the MRC. It will only charge to ~112% of the set value. But you've already noticed the MRC is not terribly fast.

One option might be to get another bad, NHW11 pack and see if enough modules can be found to rebuild the bench pack. Then when it is a good as possible, swap the packs. But this will leave a lot more material left over.

Given the weakest, module pairs appear to show up in the middle, it may make sense to re-order the 19 module pairs so the strongest are in the middle and weakest on the ends. However, this is getting pretty advanced and assumes we know the relative strengths of each module-pair.

This is very advanced work, bleeding edge, so be prepared to monitor the pack health over say:

• 0 day - initial traction battery state
• 1 week - sanity check
• 2 week - trend checking
• 5 week -
• 10 week -
• 20 week -
• annual check - long term

Bob Wilson

 

Make the Prius READY, shift to D. Press hard on the brake pedal with your left foot so that the car remains motionless. Floor the accelerator pedal with your right foot.

Since the car cannot move and hence MG2 is stationary, the power produced by the gasoline engine is routed to MG1 which charges the traction battery. You'll notice that as the traction battery SOC rises, the gasoline engine RPM will decline although the accelerator pedal remains floored. This is because the hybrid vehicle ECU realizes that the battery does not need charging so it tells the engine ECU to slow down the engine regardless of the accelerator pedal position.

This process may overheat the transaxle so use it at your own risk.

 

Patrick pretty well covered it. However, I've not seen a significant transaxle temperature rise when I first tried it (I too was worried so I monitored MG1 and MG2 temperatures in a warmed up Prius.) However, the battery temperature will go up from the exothermic charging, ~9C in my last test:

I also noticed the charging current decreased over time during the forced charge.

This was part of a late night test where I force charged the Prius and then proceeded to drive in "R" up a hill around seconds 2720-2760. The goal was to get an idea of the Ahr capacity of the traction battery by measuring the power drawn in the reverse drive up the hill. Unfortunately, the hill was too small and a police car came by and started to show some curiosity in the guy backing up a hill at 2:00 AM. So I decided to act as if I'd missed my turn and took a right on to a cross street.

One of the proposed ways to measure traction battery Ahr capacity is to put a forced charge on it to bring it up to 80%. Then in reverse, drive up a hill and measure how far up before the car stalls, ~40%. The altitude change and mass of the car allows calculation of how much energy it took to reach the 'high point.'

Legend:

• Hi C and Lo C - there are four temperature probes and the Graham miniscanner will report the warmest and coldest probes. I don't remember trying to record which probes are where. Read the temperature in C from the left scale
• Volts (divided by 10) - read on the left scale. First readings, ~295 V. Last reading ~302 V.
• Traction battery Amps - read on the right scale.

Bob Wilson

 

My guess is that the traction battery voltage measurement is using the left scale x 10. The battery assembly has four temperature sensors, so Hi C and Lo C are probably showing you the highest and lowest readings in degrees Centigrade.

 

Patrick is right and I've edited in some details about the test and what I was trying accomplish.

Bob Wilson

 

You won't overheat it by force charging it.

 

I don't know. However, the earliest traction battery failures seem to be associated with hot climates and hilly areas.

AMEN!

The ability to initialize the unit and use the 9V to bring it into the house and plug-in to a wall wart . . . But gosh it is tedious.

Bob Wilson

 

Sorry if I missposted:

OK: MG1 > MG2 :: MG1 is between the ICE and MG2 and should be close to the ICE coolant temperature.

BAD: MG2 > MG1 :: MG2 is on the end and is cooled top, bottom, front, rear, and end by air. If MG2 is warmer than MG1, there is a problem or my wife is driving at speeds of +75 mph up and down the hills on I-65 to Nashville.

Thanks,
Bob Wilson