Prius traction batteries

[bwilson4web]

Hi,

After some PM discussions about Prius battery modules, it is time to discuss 'lessons learned' about them:

The upper module came from the NHW11 Prius, 2001-03, and the lower one from the NHW20, 2004-09, the ZVW30, 2010-current, and possibly other Toyota hybrids. Significant changes between the modules:

• Substantially stronger case - in particular the terminals
• Offset terminals - no mixing of modules
• Lower internal resistance - the NHW20 carries more current with less loss
• Identical Ahr rating
• Mold release - the non-conductive, silvery material

The only Toyota report about traction battery failures indicates the NHW11 ones have been running ~1%. Given a USA vehicle salvage rate of ~3%, salvage battery packs should last many years. In contrast, the NHW20 failure rate is at least one and possibly more orders of magnitude lower, 0.1%-0.01%. I bought a failed traction battery pack in 2007 to understand what causes these failures.

Research before starting this effort:

• Toyota patent on battery refurbishment - there is a well written patent that describes how the traction battery is weighted to measure electrolyte loss, a hole drilled to add a 6 mole, KOH solution, and plastic weld close the module. Then the battery is 'reactivated.'
• TIS battery sealing - there was an early problem with electrolyte leakage causing a ground-fault. Toyota issued a TIS that takes the traction battery out and applies a 'glue' to help seal the terminals.

Now NiMH battery chemistry is different than ordinary battery chemistry because it stores charge as hydrogen gas trapped in a metal hydride matrix. The other side uses two forms of nickel hydroxide to hold the oxygen:

The metal hydroxide, a spongy mass, is separated by a 'plastic cloth' from the nickel hydroxide coated electrodes. The KOH electrolyte carries the charged ions between the two electrodes. The permanent failure mode is the electrolyte gets too low and spot heating melts the plastic separator, a permanent short. Given each NiMH cell generates 1.2 VDC, this difference is the signature of a failed module that can not be repaired. The battery module has to go to the recycler.

Charging NiMH batteries generates atomic hydrogen that 99% enters the metal hydride matrix until discharge. But a little of the hydrogen and oxygen escapes the metal hydride matrix and nickel hydroxide electrodes. During charging, this pressurizes the cell, which means it has to be mechanically clamped:


The rates of gas generation are fairly low, these are not great internal pressures. But overcharging has to be avoided because if the metal hydride is saturated, the modules become a high pressure, gas container and bad things can happen:


There are several approaches to limit overcharging:

• pressure spike - the penultimate, a pressure switch in the cell or a compression load sensor stops charging when the pressure takes a spike. This is based upon direct measurement of excess gas generation. A load-cell would be an excellent signal source.
• dV/dt - just as the NiMH cell reaches peak charge, there is spike up followed by a slight drop in voltage. The slight drop can be detected by smart chargers like RC hobby units and charging stops. However, a stack of cells becomes a problem because they all won't reach the same peak charge at the same time.
• temperature rise - the excess H{2} and O{2} gas will recombine and generate heat. This temperature spike can be used to stop charging but by then, the cell is already overcharged. Cheap NiMH chargers use this and better hobby chargers use this as backup.

This survey of overcharge protection approaches is a short list of what I've learned from the RC community about NiMH batteries and Isidor Buchmann's excellent book, "Batteries in a Portable World." But I wanted to find out if I could replicate Toyota's battery refurbishment patent.

I was able to refurbish NHW11 modules BUT I do not recommend it because it does not fix the leaky terminals. A refurbish NHW11 module is just going to lose the H{2} and O{2} gas over time and have to be refurbished again and again and again. A better solution is to use NHW20 modules and given a 3% salvage rate, they should be readily available and less expensive than the NHW11 modules.

Bob Wilson

[Christ]

Thanks, Bob!

The PM's noted above were between myself and Bob, regarding how to properly charge and maintain the NHW20 battery modules that I've just gotten from a purchase on eBay.

Questions had been raised about proper charging procedure and safety everywhere I've posted about using the batteries for projects, and when someone posted a link to a site which contained the pictures above, I found it necessary to message Bob for further information. After some discussion, it was suggested that we take the discussion public, so that more people who might have relevant information could chime in and suggest a proper means of mounting, charging, and maintaining a single 7.2V 6AH module.

That said, fire away!

A few questions that I'd personally like to have answered are:

Whether or not the modules can be mounted on their side (using the ~25mm dimension as it's height) without potential problems.

Whether the bike's alternator will charge the battery directly, or need some sort of intervention process to keep from overcharging and destroying the battery.
(The bike's electrical system is 6V, the AC generator charges through a rectifier at a rate of ~7.5V 3.3A at 5k RPM, 8V 6A at 10k RPM.)

How to properly monitor the battery's SOC and voltage, since they're apparently non-linear. (Gauges? Complex charging circuit? VooDoo?)

Thanks to all in advance for suggestions, and hopefully a collective of knowledge and input on the topic can help more than just myself and Bob.

[dcb]

typically nimh bench chargers use peak detection, i.e. the charging system stops charging for a sec and sees if the voltage is still going up from the last time it checked, if I understand it right (probably not). They also may be connected to thermal probes in the pack to be on the lookout for thermal runaway.

I am not sure how you get the battery to play an active role in the operation of a vehicle with an ICE with this arrangement.

If you have a kick starter, I might suggest removing the battery alltogether and seeing if the bike runs well with a large capacitor in it's place.

But for a potential EVer, that info above is quite useful.

[Christ]

Can we run with the capacitor discussion for a minute? What size/type of capacitor are we talking about? I don't mind using a cap and a kick starter, it will save me weight on the battery, starter, and associated wiring.

Could I build a reliable bank of caps from old electronics that I can salvage?

If that becomes the case, I've got 28.8V worth of a 6.5Ah pack to use for something else entirely!

What is or where can I find the low-end voltage (at ~50% SOC) of the NiMH modules?

[dcb]

re: cap, Just, grab the biggest cap from your parts pile and replace the battery with it and see what happens. Get the polarity right (it should have a - somewhere on it) and make sure it can handle at least 6v, (most likely it can). Lets move it to your bike thread though.

[Christ]

What about the low SOC voltage of the NiMH module?

[bwilson4web]

Originally Posted by Christ

. . .
Whether or not the modules can be mounted on their side (using the ~25mm dimension as it's height) without potential problems.

This is an area I'm not sure about. I know KOH has amazing abilities to soak and flow. I suspect at an angle, say 30-45 degrees, it would be OK. The problem would be if the highest parts got depleted and current had to flow through that area. I suspect it won't be a problem as KOH is likely to keep the area saturated.

There has to be 'gas space' in the cell so the gas that is normally created and not absorbed by the metal hydride has a place to collect before it recombines as water. It may simply make part of the electrodes 'dry' and thus not pass current. In short, I'm not absolutely sure how flat they can be and not lose capacity.

Originally Posted by Christ

. . .Whether the bike's alternator will charge the battery directly, or need some sort of intervention process to keep from overcharging and destroying the battery.
(The bike's electrical system is 6V, the AC generator charges through a rectifier at a rate of ~7.5V 3.3A at 5k RPM, 8V 6A at 10k RPM.)

Motorcycle?

There are several parts to the problem. The first is getting enough power out of the AC signal. What happens is no charge passes from the AC generator when the voltages are lower than most of the power cycle. It would only take energy from part of the cycle, the peak voltages.

Now a silicon rectifier has about a 1.2 V forward voltage drop before significant current flows. Using Schottky diodes reduces the forward voltage drop, ~0.6 V. If using a bridge rectifier, there are two diodes in the current flow at any time so it needs 2.4 V additional AC voltage or 1.2 V if using Schottky diodes, on top of the AC voltage. Schottky diodes are used in some automotive alternators and possibly in this case.

Is the rectifier external so you might be able to measure the DC voltage drop across the part? Take the rectifier and a 6 V bulb and connect to the battery and measure the voltage drop across the rectifier and we'll know.

Now as to protecting the battery, a linear voltage regulator, not terribly efficient but a common part, LM350, 3 A. But 3 amps is more suitable for low power electronic power supplies. For example 3 A x 7.2 V = 21.6 W., not very much. Plus a linear regulator is going to eat power, 1.2 V x 3 A = 3.6 W. Cheap parts and simple circuit but inefficient.

If we are dealing with a high efficiency, system, I would recommend something called a synchronous regulator:

• MOSFET switches - very low voltage when ON and higher currents, 10-50 A parts are readily available.
• switches the MOSFETs ON per phase - requires clever drive circuits

Then the other option is to add a switching circuit to step up the voltage from the lower part of AC circuit. At 85-95% efficient, a buck-boost switcher would provide power over a significant part of the AC circuit. But power versions, 10 A are not trivial but easy enough to make. They often use a power MOSFET to drive the switching part that does the work.

Originally Posted by Christ

. . .How to properly monitor the battery's SOC and voltage, since they're apparently non-linear. (Gauges? Complex charging circuit? VooDoo?)

Toyota and others use what are called "columb counters" that track the current and voltage in and out of the battery pack. The reason is the battery voltage is a complex function of the temperature, discharge current and internal resistance. A columb counter doesn't care since it counts charge going in and out.

As for over charge protection, a load-cell or switch that opens if the pressure becomes excessive makes a lot of sense. It just opens the charger (or a relay or power MOSFET) to protect the battery.

Originally Posted by Christ

. . .Thanks to all in advance for suggestions, and hopefully a collective of knowledge and input on the topic can help more than just myself and Bob.

It helps if we have a better description of the application. Especially, the load motor or electronics and the charger or alternator. Understanding the whole system, especially the peak and average loads, will make it much easier to propose an efficient design.

Thanks for initiating the conversation. Hopefully others will like "stone soup" bring their skills to the conversation. <grins>

Bob Wilson

[dcb]

In short Christ, a hell of a lot of work to replace your small lead acid battery on your bike with a nihm

re: coulomb counter, that is pretty straight forward, just monitor the volts and current going into and leaving the battery, plus some fudge for "leakage" and internal resistance and integrate over time.

[bwilson4web]

Originally Posted by dcb

re: cap, Just, grab the biggest cap from your parts pile and replace the battery with it and see what happens. Get the polarity right (it should have a - somewhere on it) and make sure it can handle at least 6v, (most likely it can). Lets move it to your bike thread though.

I'd probably go with 12 VDC minimum. There will be a peak voltage from a rectified AC, 8 V * 1.414 ~= 11.31 VDC.

I do like the idea of using a cap to smooth the alternator output. However, the kick starter still needs voltage for the spark ... unless it has a magneto setup.

What make and model of motorcycle? We may find enough information to figure it out.

Bob Wilson

[Christ]

It's a 1978 CM185T TwinStar. You can find/download the factory service manual here:
http://pages.prodigy.net/klricks/CM185T.html

I believe the section detailing the factory specs on the AC generator and rectifier is on the webpage, so you may not need to download the FSM. I have it, it's a legitimate executable file.

Everything on the motorcycle is rated for 6V, and it's got a silicone diode rectifier to control voltage spike (I suppose... I'm not so good with this stuff).

I think DCB is right though, if we're going to talk about the motorcycle specifically, we should discuss it here: http://ecomodder.com/forum/showthrea...deo-10553.html

Although general discussion involving the modules as they apply to any other object should be kept here, so as not to muddy the topics together.