This is turning into a rehash of a thread which returns periodically in one of the amateur radio newsgroups. There have been pages and pages of data written on the subject of nicads, nicad memory, etc. which I will try to summarize as briefly as possible.
There are 2 main reasons for the early demise of nicads - overcharging and overdischarging. If you continue to charge a cell after it's fully-charged, the excess energy is converted into heat. Carried to extreme the cell can explode. That's why nicad packs intended to be quick-charged (ie, less that 2-3 hours) usually have a temperature sensor in series with the battery. This opens the circuit if the battery gets too hot. When cells are overheated, the vent may open and cause loss of some of the electrolyte, thus reducing the cell's capacity.
Overdischarge occurs when you discharge a cell below about 1v. In a pack of series-connected cells, one cell will always have slightly (or more than slightly if not matched) less capacity than the others. If the pack is discharged too deeply, the weak cell(s) will be reversed charged. They don't like this at all and just a few instances of reverse-charging may cause the cell to refuse to accept a charge at normal charging currents, usually due to growth of crystals within the cell which effectively short it out. The result can be a pack which is 1 or more cells short of the nominal voltage and which will quickly look 'discharged' to the device it's powering no matter how long it's charged.
Usually 'nicad memory' will be blamed for either of the above conditions. While the memory effect is possible, it's very unlikely. It's actually a voltage depression of .1-.2 volts per cell, which reduces the voltage of the battery pack enough to cause the powered device to think the battery is discharged. Voltage depression is largely reversible by discharging the individual cells in the pack to about .9 volts then charging slowly (14 hr rate). It may take a few cycles. Shorted cells can be restored by passing enough charging current to burn out the short. Warning: D cells can take 10 amps or more to do this; fireworks can result! Cells damaged by overheating due to overcharging are permanently damaged.
Choosing a battery these days is complicated. Besides nicads and lead-acid gell-cells, there are nickle metal hydride, rechargeable alkalines, and lithiums. I question whether rechargeable alkalines are a good choice for telescope-related applications where a dew prevention device is used. They are not as well suited for high-drain applications as gell-cells or nicads. They are good for infrequent uses because they have essentially the same shelf life as primary alkalines and because they loose a little capacity with each charge cycle. Thus their expected useful life of only 25 cycles or so. They also have unique charge requirements and thus require a special-purpose charger (=$).
Gell-cells can give a couple hundred cycles if used properly, and nicads are capable of 500 or more. Nicads which have been degraded and restored can still be useful. They tend to self-discharge quickly, sometimes in only a week or so if they're really in bad shape, but if charged and then used that night, they will deliver almost all their capacity. Surplus gell-cells are widely available at attractive prices, in a variety of capacities and voltages. Hamfest flea markets are a good source, but as another poster in this thread noted, take a voltmeter along.
There's a lot more to be said, but this is already longer than I intended. It's not hard to get wound up on the subject.
Roger Grady K9OPO c21rag@kocrsv01.delcoelect.com Delco Electronics Corp. Kokomo, IN "All information and opinions are strictly personal."