While you make a point about the dangers of explosions in short circuits, your other generalizations do not align with the relevant facts.
In searching the FAA/NTSB database, I find no instances of exploding ELT batteries, of any chemistry. To understand why this is so when Li-SOCl2 and Li-SOCl2,BrCl, Li-BCX chemistries are deployed in the field, it helps to understand how these batteries are configured and packaged. The cells themselves have built in short protection. In addition, the cells are wired directly to current limiting circuitry. The entire structure is factory mounted on a rigid plate that constitutes one side of the ELT box. Replacing the battery involves removing that side and replacing it with a new one. While these chemistries would probably provide a decent safety margin with a 10 year replacement cycle, the FAA mandates a five year cycle. (By the way, ELTs are not activated by a "bump" and they are not placed on standby until the final pre-takeoff checklist. The do have g-force, manual, and (for some) software activation options.)
Your assertion that these ELTs are "close to useless" deserves a closer look. While the 121.5 mHz frequency of the old ELTs is no longer monitored via satellite, it is monitored by almost every commercial airline in the sky. Since these birds are at FL30-40, they are a great network of ears. Thousands of rescues have been effected via the old system.
The new generation ELTs are monitored by satellite. Most encode GPS location to 25 meter accuracy. Even without GPS encoding, satellite doppler trilateration resolves location to a typical 100 meters. Once a satellite picks up a signal, it is indexed on the global database and relayed appropriately with full contextual information within 60 seconds.
Finally, while it is possible that an accident can render an ELT non-functional, these kinds of situations are outliers and not the base case you imply. By far the most common GA emergency goes to fuel management. In such situations, the pilot has ample time to activate the ELT while looking for a suitable off-field landing spot. The second most common type of emergency is related to weather; again, ample time to activate for satellite capture, logging of location, indexing on identification data, and launching an appropriate response. Even the in the very rate situation of ditching, the typical GA aircraft is above water for several minutes, sometimes much longer. Water activated ELTs are also available.
While not unheard of, the kinds of cataclysmic situations that would immediately render an ELT useless are rare. Experts judge that the 406 mHz ELT system is exceptionally sound in relation to its purpose. To date, ELT beacons have assisted in the rescue of over 28,000 people in more than 7,000 distress situations. Wiki indicates that in 2010 the system provided information used to rescue 2,388 persons in 641 distress situations. This doesn't sound like "close to useless" and the system is much more sophisticated than "yoo hoo."
Explosion isn't the only reason ELTs should not be shipped with batteries in them. Multiple incidents have occurred of ELTs being mailed in for service, certification, &c. (or even just being taken home by the aircraft owner), and being activated by some bump. The result is usually a C-130 circling over the postal facility in question, with a thoroughly disgruntled crew aboard.
Unfortunately, though they are mandatory, ELTs are close to useless for their intended purpose. They don't work if submerged, or the antenna's been torn off, (which is not unheard of in abrupt stops). The basic idea of yelling "Yoo-hoo, I'm over here" until somebody finds you is not the most effective way to deal with the problem.
Most of the new generation 406 mHz ELTs (emergency locator transmitters) required in general aviation aircraft now use lithium thionyl chloride technology as do some emergency backup supplies for avionics displays.
The comment about the delay in producing usable power after a period of disuse raises questions. How long is the delay and would be a good idea to cycle the ELT on and off occasionally for a short period? If so, does anyone know how often and for what duration would be optimal in relation to maximum battery life in the event of an emergency activation? Would a very slight drain in the form of capacitance be sufficent to maintain readiness? The fact that the technology has received FAA approvals suggest that they have determined this drawback to be managable but when you look into their processes, you see room for errors.
Thanks for that update. I have been watching the pioneering company for some time. I emailed them to ask what is the chemistry, since 80 milli-ohms impedance is typically not Thionyl Chloride domain, and it is why they built capacitors inside in older generations, to at least handle a bit of pulse current. Here is the link to the datasheet for the HE (high energy) version:
Some batteries need discharge balancing for series-connected cells, to prevent the weekest cell from taking full-load current into a very high internal resistance. This is one reason why balancing chips, such as S-8209BAA-T8T1y from Seiko Instruments Inc, have included that function. Sometimes, even at the beginning of the depth of discharge, a high-rate, fast-switching load demands dynamic currents faster than the battery chemistry can respond, which can overheat a given cell. To compensate for heating as well as low pulse-current ratings, companies like Tadiran put a parallel capacitor inside their Lithium Thionyl Chloride cells. However, everyone enjoys a good, safe poof/bang, now and then.
The mention of Lithiium Thionyl Chloride and expolsive reminds me of the following...
Many years ago, I was working for a company that made postage meters. In those days, you had to take the postage meter to the post office to add postage. The old postage meter had a power supply in the case but that required that there be an outlet available. We decided that we could make the postage meter operate off battery power for adding postage. Since the battery would not be user replacable, it had to have a long life so we made a pack of AA size Lithiium Thionyl Chloride batteries with a key operated switch so the battery woul donly be active when it was in the post office.
This was all very well until one day we got a report that one of our postage meters made a loud noise and emmitted smoke while in transit to a customer. The meter was returned to us for analysis where we found that the batteries had exploded and done an impressive amount of dammage to the inside of the postage meter, which included burning a hole most of the way though the metal case.
As part of our analysis, we decided we needed to determine what caused the explosion. Having seen the damage caused, we put the batteries under test in a sand filled metal trash can outside next to the back door and ran the wires and thrremocouples out. Our precautions were justified when we finally triggered an explosion. The trash can lid (which was tied down) was bent in half and the rest of the trash can was decidedly battered.
We decided to redesign the battery pack using the Duracell lithium camera batteries which were much safer since they had built-in thermal fuses.
Good points. Shipping poses special circumstances, since it has the highest likelyhood of damage that could cause a short -- which is why the shipping regulations ask for protection specifically for shorts and accidental activation. What I do is give consideration to all issues, including shipping, when designing enclosures. For example, sometimes the batteries already have massive protection from over-temperature, over-pressure, and shorts -- so I ONLY have to meet shipping regulations that are above the design criteria. In some cases because of this, the enclosure needs no extra protection, shipping or otherwise.
I also like your point to add messages and warnings.