Disaster!

This article was first published in 2005.

WARNING!! Overcharging a Prius high voltage battery can result in the battery pack catching fire.  If overcharged, the battery pack can also release toxic, corrosive fumes.  Charging currents over 0.5 amps should be undertaken with extreme caution and for short durations only.  Being electrocuted is a likely outcome if you do not know what you are doing with high voltage AC mains power systems and high voltage DC battery systems. Because the design of charger covered in this article uses the transformer as an autotransformer, no isolation is provided. The transformer must be fitted in a strong metal box with both the box and the transformer frame connected to mains earth. An earth leakage circuit breaker (ELCB or ‘safety switch’) should always be used.

Tonight I destroyed my AUD$6,000 spare high voltage battery pack and very nearly set my house on fire.

After developing a quick and dirty high voltage DC charger to get my NHW10-model hybrid petrol/electric Prius back on the road (see When One Mistake Can Kill), I’d decided to make a better quality charger with lots of protection fuses, a higher charging current - and to put it all in a neat metal box.

The need for the high voltage charger had come about because the Prius, off the road for 3 weeks to have a turbo fitted, had dropped in the charge level of its high voltage (288 volt) battery. As a result, the ECUs wouldn’t let the car start. The simple charger that I’d made had worked, and now I wanted the better one to keep my spare battery pack charged up.

Safety Breaker

However, before I could build the charger, I needed to make a safety breaker. The Prius battery pack has a removable circuit breaker inserted halfway along the series of 240 cells. It’s pulled whenever any work is being carried out on the high voltage system – but unfortunately the battery pack that came with the purchased half-cut was missing its breaker. (That’s not surprising since the first thing the guys stripping the wrecked car would have done was remove it!) Without the breaker in place, the battery pack couldn’t be charged.

The factory breaker consists of two metal prongs housed in an insulated carrier. The two prongs are connected together so that when they are pushed home, the circuit is completed. (The factory breaker is even trickier than that: it uses a handle that twists to one side, locking the breaker in place and at the same time activating a micro-switch that tells the ECUs that the breaker is fitted.) The breaker I was missing didn’t require the micro-switch deactivation facility: it just needed to complete the circuit. However, it also had to be very well insulated, otherwise an unwary person inserting the breaker could get fried. You see, other than by pulling the breaker, there’s no real way of switching the battery pack off...

I used my lathe to turn up two 7mm diameter brass prongs and then connected them with a brass strip. The assembly was then mounted in a carrier made from thick insulating acrylic. Thinner acrylic sheet was used to form a ‘U’ shaped handle that also provided another layer of insulation. That completed the breaker, but the big job was yet to come.

The Charger

For the previous roughie high voltage charger I’d used a bridge rectifier working directly across the 240 volt AC power supply, feeding the battery through initially a 3 ohm, 500 watt resistor (a high power light bulb) and then later, as the charging current dropped right away, with no resistor at all. But one of the problems had been that the DC voltage developed by this combination wasn’t really high enough to fully charge the battery (or so I thought!). What was needed was a way of increasing the mains voltage above 240V AC, resulting in a higher DC charging voltage.

This AC voltage increase can be achieved by using the secondary windings on a normal low voltage power supply transformer. What’s done is to wire the secondary winding in series with the mains voltage feed to the rectifier. If it’s a 240V primary, 12V secondary transformer, in this way a voltage of 252V can be achieved. This diagram, taken from a Silicon Chip magazine article at Circuit Notebook shows how it’s done. The secondary winding needs to be wired in the correct phase - otherwise the voltage will be reduced, rather than increased. If the secondary windings have an output of 30V (or two secondary windings can be placed in series), the output voltage can be 270V or higher.

In the case of the charger shown here, the high voltage design work was carried out by Robert S Edgar (my father!). He suggested the use of a transformer that has two secondary windings, each with 12 and 15V taps. By wiring the 12 and 15V secondary windings in series and then using the transformer in autotransformer configuration, an AC voltage of 267V is achieved from the 240V supply. This is fed into a 600V 10 amp bridge rectifier which feeds the high voltage Prius battery through five, 15-ohm, 10-watt ceramic dropping resistors wired in parallel.

Further charger additions include an internal cooling fan; three neon indicators to show the presence of mains voltage, output charger voltage and battery voltage; and fuses for each of these circuits. In addition, two pairs of shielded binding posts are used to provide easy ‘tap-in’ points for a Cat III (ie high voltage compatible) multimeter. One pair of terminals allows the measurement of charging and battery voltages (depending on whether the charger is on or off), and another pair allows the measurement of charging current (a switch bypasses these terminals when the meter isn’t present). The charger pushes about 2 amps into the battery, so with a pack this big, it’s a trickle charger – or so I thought.

Disaster

It was late in the afternoon when I finished building the charger. I connected it to the battery pack (using my protective high voltage rubber gloves to clip the croc clips directly to each end of the series string of cells), switched on the charger and monitored current and voltages. They were fine – the battery was charging at 2 amps. Considering this was meant to be just a trickle charge, I had intended leaving the pack charging all night.

But thank God the fire happened before I went to bed.

After checking on the battery each hour, I left it for a few hours. Then my wife and I started smelling something. I live in a Queenslander multi-storey wooden house on stilts; the ‘garage’ is the concrete pad directly under the house. That’s where the battery was charging.

I walked outside to find the house completely enveloped in a thick, acrid, rolling fog. I rushed downstairs and saw clouds of smoke gushing from under the woollen blanket I had thrown over the pack to stop stray animals sniffing the high voltage terminals. When I pulled the blanket back, I could immediately see a brilliant red glow deep inside the battery pack; as I watched, flames 6-8 inches long burst out. Thick white smoke was boiling from every opening.

I yanked on my high voltage rubber gloves – they were luckily still close by – and pulled the circuit breaker out of the back of the battery pack. I turned off the charger and pulled off the charging leads before man-handling the burning, corrosive high voltage battery pack (it’s very heavy but I didn’t even notice the weight) 10 or 15 metres to a clear area of lawn. As I turned it end over end, I could feel that the thick plastic casing had softened – and even through the gloves, the heat of the plastic was obvious.

The flames subsided but smoke kept pouring out. After 15 minutes or so, the smoke diminished, but hissing gas could still be heard periodically venting from the batteries.

An inspection showed the batteries charred and nearby wiring melted, the plastic case distorted. Salvageable will be items like the battery ECU, current monitor and circuit breakers. But as a going concern, the battery pack is stuffed.

But Why?

But why had it caught fire – especially with a charge rate that had by then dropped to only 1 amp? The reason is that I had made a huge mistake in calculating the capacity of the Prius high voltage battery.

The battery is made up of 1.2 volt Ni-MH D-cells. These are arranged in 40 sticks of 6 cells, each stick having a quoted spec of 7.2V and being of 6 amp-hours capacity. The voltage therefore calculates to 288V (40 sticks at 7.2V each) but what is the total battery capacity? I’d thought the same logic applied, so had determined it as 240 amp hours. But that is completely wrong. In fact, in a series string of cells, the amp-hour capacity remains the same for the total as for one cell – in this case, 6 amp hours.

So with a charge rate of 2 amps, I’d thought that I was charging at less than 1 per cent of the battery’s capacity, when in fact I was charging at 33 per cent of its capacity. Where with a charge rate of 2 amps I’d thought it would take 120 hours to charge the battery up from dead flat but in fact it would take only 3 hours.... Even with the charge rate having dropped to 1 amp, from dead flat the battery would be fully charged in 6 hours. And in this case, it wasn’t dead flat to start with.

The result was that the battery came up to fully charged no less than 20 times faster than I expected – and then I kept on charging it, and kept on charging it, and kept on charging it ... The battery pack got hotter and hotter, the cells venting noxious gases and the temperature rising as the plastic holders started to melt and then burn...

And how do I feel? Well, pretty stupid – it’s a massive, expensive and dangerous mistake to have made. It’s a disaster in that I have lost my battery pack (it’s hard to value but here in Australia, somewhere in the region of AUD$2,000-$6,000 is probably right) but what really makes my skin crawl is the fact that I had intended leaving the battery charging after I went to bed. Directly above the battery pack is a wooden floor; in fact my whole house is made of wood....

I will never, never again take for granted charging any sort of battery...