Quick HV Battery Reconditioning – DIY Grid Charger Results & Analysis

Synopsis

The goal was to find a method to quickly restore capacity without degrading the battery life in my 2007 Prius Gen 2 traction battery.

Detailed measurements from nine charge/discharge cycles are presented. The conclusion is that three or four “short” cycles followed by a single “long” cycle is sufficient to fully recondition the battery.

A short cycle charges the battery pack to 224V and then discharges the battery to 168V. A long cycle charges the battery to 240V (6500mAh) and then discharges the battery to 168V.



Introduction

I have a 2007 Gen 2 Prius with its original HV battery. I had read that it was possible to restore capacity to NIMH batteries by careful discharge/charge cycling.

Initially I thought I would use a “hobby” charger and cycle each module independently, but the more I read, the less attractive that option became. I didn’t want to have the car off the road for several weeks because of the slow discharge rates on these chargers.

It occurred to me that It would be easier to charge/discharge the battery pack as a whole and that led me to investigate so-called “grid” chargers.

I was wary about damaging the battery so I decided to follow the limits and advice set out in the “Panasonic Nickel Metal Hydride Handbook”, the GlobTek “Nickel Metal Hydride Battery Safety Notes” and the Battery University article “BU-408: Charging Nickel-metal-hydride”

[Panasonic.pdf] [GlobTek.pdf]

Key points from these resources…

- Excessive overcharging of a NiMH cell can result in permanent loss in capacity and cycle life. If a cell is overcharged to the point at which pressure begins to build up, elevated temperatures are experienced and can cause the separator to lose electrolyte. The loss of electrolyte within the separator (or “separator dry out”) inhibits the proper transport of hydrogen to and from the electrodes. Furthermore, if a cell is severely overcharged and excessive amounts of oxygen (gas) are evolved, the pressure may be released through the safety vent in the positive terminal. This removes elements from within the cell needed for proper function.

- Long term trickle charging (0.033 - 0.05C) of NiMH batteries is not recommended

- Nickel-based batteries must cool down on trickle charge. If warm, trickle charge is too high.

- Total charge time should be limited to 10 - 20 hours

- Charge at 0.1C (625mA) for no more than 16hr (10,000mAh)

- Up to 70% State of Charge (SOC), the charge input is stored with close to 100% efficiency.



- Extreme over discharge of a NiMH cell results in excessive gassing of the electrodes resulting in permanent damage in two forms. First, the negative electrode is reduced in storage capacity when oxygen permanently occupies a hydrogen storage site, and second, excess hydrogen is released through the safety vent reducing the amount of hydrogen inside the cell.

- A NiMH cell is considered empty when its voltage falls below 1.0 volt.

- When the cell voltage is between 0.8V - 1.0V, the recommended charge current is 0.1 – 0.2C (650 – 1300mA). Below 0.8V the maximum charge current is 0.05C (325mA)

- One of the biggest misconceptions regarding NiMH cells is that rated capacity is the capacity that will be received by the user. This is only true if the user charged and discharged at the same rates of current at which the cell was graded. Rated capacity has been defined by the International Electrotechnical Commission (IEC) in document #61436.1.3.4, as a charge at a rate of 0.1C for a period of 16 hours. This is followed by a discharge of 0.2C to a voltage of 1.00V per cell.

Polarity Reversal

When cells are arranged in series there exists the possibility of polarity reversal, when the whole series is discharged. Polarity reversal causes permanent cell damage.

It can happen because there is no way to guarantee that each cell in series has the same capacity. If one cell in a series reaches zero volts before the rest, it can reverse polarity, because current continues to flow through it from adjacent cells.

The Prius has 28 battery modules containing 6 cells each. The modules are arranged in 14 blocks, where each block contains two modules. The total number of cells in the battery is 168 (28 x 6) and they are connected together in series.

A 7.2v battery module consists of 6 x 1.2v cells. We don’t have access to the individual cells, so we set the minimum allowable voltage to 6V [(5 x 1.2v) + (1 x 0v)] to avoid polarity reversal.

The Prius monitors block voltages (two modules in series) so in that case the minimum allowable voltage per block would be 13.2V [(11 x 1.2) + (1 x 0V)]

Equipment



I used a 500mA constant current charger (LPC-150-500) with an output voltage of 150 – 300VDC. The cost including shipping to New Zealand was $40USD. This charger only accepts 180-240VAC input, so if you need universal AC input use the LPC-100-350 instead. The only modification I made was to add a rectifier diode (1N5408) to the positive DC output, to stop current from the battery feeding back into the charger when it was switched off.



To discharge the battery pack at close to 0.5C you need a ~70 ohm 800 watt resistor. I used an old electric convection heater (bar fire). Unlike a light bulb, it is designed to dissipate heat, not glow white hot, so the resistance of the heater element doesn’t change very much (2 - 3 ohm) with current.



Two digital volt meters (DVM) were used; one to constantly monitor the overall pack voltage and one with crocodile clips, to measure the individual module voltages.



An old PC cooling fan was used to blow air under the modules. It was connected to an old car battery which was on charge, so it could run continuously.

Method

Charge (Rest) -> Discharge (Rest)

Rest for a minimum of 1 hour.

Discharging

- The voltage of each module was measured, open circuit (OC), before connecting the load and every 10 minutes while the load was connected.

- discharging was stopped when the overall pack voltage reached 168V OR when any module had fallen to 5.4V.

Charging

- The voltage of the battery pack was measured before connecting the charger.

- The AC input of the charger was attached to a timer, to control the quantity of charge delivered. For example, it took 13 hours @ 500mA to input 6500mAh (20 hours @ 350mA).

- 6500mAh is the capacity of the HV battery when it was brand new (16 years ago!)

Results

[Results.pdf]




All of the battery reconditioning was completed by DISC-5. Three full “short” cycles followed by one “long” cycle.

In the cycles after DISC-5, every module was individually discharged at the end of each cycle and then charged with progressively increasing input charge. This resulted in virtually no improvement in overall capacity and appeared to degrade the state of the weakest modules.





Observations

It is impossible to predict which modules will discharge first from the open circuit voltages and even under load because the modules discharge at varying rates and follow different profiles.

An apparently good module can suddenly start losing voltage alarmingly fast (>0.2V/sec) below 5.7V, even before the total pack voltage has fallen to 168V.

Once the pack voltage fell to 190V, it would fall to 168V within the next 10 minutes. Using one DVM it took me about 3 minutes to measure and record all 28 module voltages. The pack voltage would fall from 180V to 168V in 3 minutes or less.



A pair of “two by fours” used as a ramp, make it much easier to to remove the battery from the boot (trunk)

Conclusions

To recondition a Prius battery pack, it only needs to be discharged to 168V (28 x 6 x 1.0V).

Individual modules will fall below 6V, some extremely rapidly, as the pack voltage approaches 168V.

It is not necessary to cycle every single module in the pack below 6V. The battery pack as a whole, is only as strong as its lowest capacity module. Doing so is likely to degrade the lower capacity modules.

Three or four “short” cycles followed by a single “long” cycle is sufficient to fully recondition the battery.

A short cycle charges the battery pack to 224V and then discharges the battery to 168V.

A long cycle charges the battery to 240V (6500mAh max) and then discharges the battery to 168V.

Short cycling ends when the estimated power output during the discharge phase is less than the power input during the charging phase. These are both directly proportional to time, so it is also when the discharge times have stopped decreasing for the same quantity of input charge. Another indicator that additional short cycles are not needed, is that the pack voltage does not rise during the resting phase between charging and discharging.

Q+A

Q: Do I need to monitor individual module voltages like you did?
A: No, I did the experiment so that you don’t have to, but I am happy if someone wants to duplicate my results. If you are interested in watching the discharge process and are doing this with the HV battery in the car, it may be possible to monitor the block voltages using an App and an ODB2 adapter, but I have not tried that myself.

Q: Do I need to remove the battery from the vehicle?
A: No, but I recommend that you do so at least once, to clean the bus-bars and check the battery voltage connectors for corrosion. You will also need to connect a voltage source to the battery cooling fan, so that it keeps running throughout the process.



Q: What about the HV isolation plug?
A: You are responsible for your own safety! The HV Isolation plug must be removed when removing the HV battery from the car or attaching the charge/discharge cables. It must be in place during the charge/discharge cycles.

Q: Can I follow this method if I already own a Prolong charger/discharger hardware?
A: Yes, but you will need to monitor the charging and discharging pack voltages, currents and times to ensure that the battery is not over charged or discharged.

Q: How is the input power calculated?
A: Knowing the constant Current Output (Ic) in mA of your charger and the Charging Time (Tc) in hours

Total Charge Input (mAh) = Ic x Tc (mAh)

Q: How is the output power estimated?
A: Voltage (V) volts, Current (I) amps, Load Resistance (R) ohms, Average Current (Iavg) amps

During the discharging Time (Td) in hours the pack voltage falls from voltage Vstart to Vend (volts)

Ohm's Law: I = V / R

Assuming a linear voltage decay as the pack discharges…

Average current (Iavg) = ((Vstart + Vend) / 2) / R

Estimated Output (mAh) = (Iavg x Td x 1000)mAh

Q: How long does it take to recondition the battery using this method?
A: Three short cycles plus one long one can be completed in two days with the equipment I used. Add another day if you are removing the battery, cleaning the bus-bars and reinstalling the battery.

Q: What should I do before driving the car?
A: Temporarily disconnect the 12V auxiliary battery to clear the Battery Management System memory. After the final (long) discharge cycle and rest, charge the pack to 224V. The car will complete the HV charging cycle and start learning the new State Of Charge on the first drive. Before starting the car, don’t forget to reconnect the 12V battery and re-install the HV isolation plug if you have removed it at any point.

Q: How does your car drive now?
A: It feels like a different car. It has more “get up and go” and the energy meter never drops below six blue bars. It’s too early to tell how much the mpg has improved.

 

I just wanted to show that "less is more" when it comes to battery reconditioning. It's easy to over do it and there are plenty of examples online of people doing just that, probably without realizing.

I am expecting to be able to maintain my battery's condition with a single short cycle once a year.

 

I think that I have proved that the opposite is true. Excessive discharging is unnecessary and damaging.

I did not know that. Are you saying that using the Hybrid Automotive method will tend to "weed out" rather than recondition, the weaker modules in batteries that are more than 7 years old?

It is very good now! All the battery energy bars are either green or one less than green (6 x blue) all the time. I haven't tried EV mode yet, but before the battery reconditioning, the car would manage less than 300 meters before switching back to the ICE with the battery in purple territory.

 

You are correct that the process could be sped up using greater charge currents.

Up to 0.2C between 0.8 and 1V per cell. Above 1V/cell up to 1C can be used until you reach 80% charge, above which the current should be dropped again to reduce unecessary heating. The problem I see with this is that the capacity of the cells are all different, so how do you avoid damaging the lower capacity (weakest) modules with a 1C charge rate applied to all the modules in a pack at once? I suspect that trying to use negative Delta V to determine when to stop charging a pack of 168 cells of various capacities in series wouldn't be reliable. That gets us back to a timed charge at a max of 0.2C or temperature sensors on every module monitoring Delta T.

At 500mA there was virtually no heating of the pack, so if you could find a constant current source at 1.3A (0.2C) and air cool the pack, you should be fine. That would safely cut the charging times by two thirds.

 

That looks like it would work. The A/AB versions also have a potentiometer to adjust output current.

 

Necessary?