Replacing the battery in an Apple IIGS

You’ll likely get a lot of opinions.

I like lithium batteries because the original ones lasted 30 years before I replaced them and most of them were still working when I did.

I don’t like alkaline batteries for this kind of use largely because I have a lot of flashlights that I use for my work and I commonly find leaking batteries.

For ROM 1 IIgs I solder a battery holder to wires long enough for it to rest under the power supply and for extra measure I put it in a small ziplock bag. If you were to use alkaline batteries I’d suggest doing this with them as well.

The two hardware versions, the ROM 01 and the ROM 03, use different means of attaching the clock battery. On the original (and "Woz signature" version) ROM 01, there is a 1/2 AA, axial leaded, LiSOCl2 (lithium thionyl chloride) cell soldered to the board. The location is barely under the power supply, so may be hard to see in a photo. The original cell used was made by Varta in all cases I've seen. The updated ROM 03 (with 1 MB RAM on board) IIgs uses a 1/2 AA cell holder, the type with a plastic clip cover that can be pulled out using a screwdriver. In the holder is fitted a plain barrel LiSOCl2 cell. The original was made by Maxell in every case known to me.

This type of cell lasts an extraordinarily long time in this application, often still holding charge after over 30 years. Unfortunately, like almost all battery types, they can fail and leak liquid electrolyte onto the equipment. The probability of failure increases with age and is higher for some brands depending on how they are constructed. Maxell cells have a particularly bad reputation for leaking. When this happens, the electrolyte attacks everything metal inside the equipment, including copper traces on the circuit board, leaving a mess that is very difficult to repair. You can see many photos online if you wish.

I reflexively remove any Maxell I see immediately. Since no desoldering is needed, this is easy to do. The soldered in Varta cells should also be removed (their lifetime is well and truly expired now) and you can simply cut them out with side cutters and take them to your local hazardous waste dump. Unlike some computers, the IIgs does not require a battery at all to work: all it is there for is to keep time and settings when the computer is switched off. Since many vintage machines are rarely used, it is pointless for them to be keeping time anyway. Simply remove the battery and don't replace it.

If the computer will be used regularly instead of stored, the replacement LiSOCl2 cell type is Tadiran TL2150P, SL-{350,750,or 850}/P, Saft LS14250CNA for the ROM 01, or the same units sans suffix for the ROM 03. The former have axial pins that are bent to be soldered into the board, and the latter are bare cylindrical cells that fit into the battery holder.

Some collectors make adapters to use different battery types. I don't think this is a good idea. Thinking that you are safer with moldering alkaline cells inside your computer instead of LiSOCl2 is a false comfort, because they are much more vulnerable to leaking and destroying the computer. Thionyl chloride cells are hermetically welded, whereas alkaline cells only have weak seals holding in their electrolyte. Adapting the computer to lithium manganese dioxide button cells (CR2032 type) also doesn't eliminate the possibility of leakage, and these cells are much lower capacity and lower performance (only a sixth the lifetime of the originals). At least they are cheap.

The only battery chemistry that won't leak and destroy the board is the BR type (lithium carbon monofluoride). These never leak and even if they did, their electrolyte is a corrosion inhibitor, so no damage would occur.

I have more than a dozen calculators, computers or notebooks which were destroyed by leaking batteries. I still keep them as spares to "harvest" ICs whenever needed. I also have all sorts of instruments with hidden batteries inside, such as Oscilloscopes, Spectrum Anayzers, etc. While it is convenient to have the instrument store and keep the last configuration, what good for is this feature if the leaking battery has destroyed the instrument ?

My solution is to remove any batteries from the PCB where they were soldered in, solder wires to the PCB, and run them to a battery holder which can be put into a small plastic tube that can be sealed ... such as the ones used for various pharmaceutical pills. I run the wire though the stopper (plug) of the tube and seal it in with epoxy glue or heat glue. The tube gets some velcro glued on and the counterpart velcro is glued to a place in the enclosure of the computer or instrument which in normal use positions is at the lowest point. After battery is installed in the tube, and the tube is closed with the plug,  the tube is fastened in place with the velcro. If the battery starts leaking - which they may do before they "die" and have no voltage left - the worst case damage is to the wires, which may corrode and need replacement. The tube always can be cleaned and reused.

Alas, notebook computers are too crowded to install such tubes or other sealed containers, so regular change of the button cell batteries inside is necessary - I do that on a planned schedule, and this includes batteries in every gadget I have, such as the 9V blocks in the small multimeter instruments. To my horror is discovered that some well known brands of batteries which were advertised to be "guaranteed leak proof" now also will leak. This is called "progress", I think. When they moved the production to China, they somehow must have lost the know-how to make them leak proof. Now it's not worth to pay the premium price for these brand name batteries. It's a fraud. Because they are not leak proof anymore, which they were several decades ago.

I have adopted this scheme more than 40 years ago for all the instruments and computers I have bought myself. There never was a case of permanent damage by leaking batteries in any of these. But those computers or instruments I got from a recycling place or as "parts only" off Ebay more often than not were destroyed by leaking batteries. Almost all collectible Hewlett Packard calculators I have bought used had damage from leaking batteries - these were the evil NiCd rechargeable batteries all these 1970s era LED display calculators had. I was able to salvage / restore most, which involves chemical baths to neutralize the corrosive emanations from the leaking batteries. BTW, this experience with these calculators alone was sufficient for me to never consider to buy an electric car.

It still eludes me why some makers of premium laboratory instruments opted for batteries. They could have used EEPROMs which were available at the time of the manufacture of these instruments. Maybe it was evil intent - a sure way to destroy that $25000 instrument after X years and render it worthless / unsalvagable. But price wise it's probably not worse than owning an electric car...

- Uncle Bernie

Most types of electrochemical cells evolve hydrogen as a byproduct of their reaction, which builds up pressure inside the casing. Eventually this excess pressure will cause either a leak or explosive rupture. This can be delayed greatly if a hydrogen absorber is included inside the cell. Do you know what makes an excellent hydrogen absorber? Mercury.

It was possible to make cells that were very nearly leak-proof using mercury or cadmium. But after these heavy metals were taken out of the production process, the cells are leak-proof no longer. The best cells, like the ones from Panasonic, are unlikely to leak during their expected lifetime, and the worst ones, like Duracell, frequently leak all over their packaging before leaving the store.

Which model you have? Here's what I did which is to use a wired clip connected to the original battery soldered leads. I glued the holder into a small project box which is hot glued to the shield. This way if the battery does leak (and I doubt it will, given the failure rate of lithiums is extremely low and just appears high because when it happens folks cry  loud) the bathtub will hold any electrolyte until it dries out and not cause any damage.

gs battery.jpg

All the parts I got can be found on mouser, and if you want part numbers I can get 'em for you.

If you have a ROM 3 would be different given the soldered holder. 

The 3 AAA batteries, did you misunderstand that? Three AAA would be 4.5V which is 30% higher than the board wants. Also if you go with A type cells you'd do better with AA over AAA because there's more poewer available with AA cells. The other thing to understand is some talk about 2 AA batteries which would provide 3V to the board. The original GSes shipped with 3V Varta cells, but Apple switched to 3.6V Tadiran after disovering 3V didn't provide good life. With the 2 AA(A) cells you run into a similar situation. This 3V battery configuration mans by the time that potential reaches the the clock chip at full carge is just a few tenth's of a volt above the the brown out voltage of the clock chip (what hold the parameter settings). While it's not a large voltage the batteries still have a lot of power to sorce  at ~2.3V before the clock reaches it brown out voltage. This just means the AAs should not preserve settings as long as a 3.6V lithium cell would. Please note, my numbers are from memory but should be close, it's been years since I looked into this but those numbers should be fairly close. 

This shows some of the stock batteries, the Varta came from my stealth, the Tadiran from my original GS, and the last one... that's my favorite it was found is a system I worked on for someone. Slide-on, and I really like the idea!All three are lithium 1/2AA cells, the Varta is 3V, the other two are 3.6V which was the norm after the first few months of production. The Tadiran cells were also used in many of the mac models though at least 92 and maybe longer. 

 

IMG-1577.jpg

Jedd D's solution is my preferred method.

You can buy any number of 3V 1/2 AA cells from various manufacturers, with various chemistries, a 1/2 AA cell holder that can be mounted off the board with leads connected to the original battery lead positions on the motherboard and you've got the best solution possible.  Even if there is cell lleakage it will not damage the board because it's mounted off the board.  Both ROM 0/1 and ROM 3 boards can be modified in this manner and it is perfectly fine and won't alter it's value because done properly they can always be reverted back to their original states.

Also note, when people say that "cell xx" leaks 100% of the time, they are really saying that it will eventually leak when it is depleted, and that occurs perhaps 10-20 years or more from now, mostly iin neglected or rarely used devices.

If you replace those cells as a matter of course every so often you'll never have a leakage issue.

baldrick wrote:

Also note, when people say that "cell xx" leaks 100% of the time, they are really saying that it will eventually leak when it is depleted, and that occurs perhaps 10-20 years or more from now, mostly iin neglected or rarely used devices.

I would add that storage conditions are a contributiing factor for the same reason electrolytic caps sometimes don't age great. 

We should all disregard those foolish comments about lithiums leaking because lithium cell have a very low leak failure rate. It is very rare, and we're talking fraction of a percent with more exposure you'll get more failures.  When lithium cells discharge to a specific voltaget there is physical damage to the cell and the voltage drops quick, like falling off a cliff quick. At that point the battery can't discharge they just sit there until they dry out.  I think the conventional wisdom is skewed because when one does leak someone cries loudly from the roottop and no one cheers when it doesn't; only one of these two gets attention.  I'm not saying it doesn't happen but it is very rare. 

Lithiums also don't leak because they are depleated if that was the case all of those in my picture would have leaked but none did.  Alkalines (not rechargable) on the other hand are notrorious for leaking and Duracell OEM batteries are the worst because they are designed to leak to prevent exploding!! Yup alkalines can explode when gasses released from discharging exceed the pressure the can can handle. Both of these conditions (leaks and explosions) are oftne the result of batteries used in series where one has a lower internal resistance caucing it to discharge faster and when it gets low enough the polarity can flip and that's a bad thing for the cell and very common. 

Austin_iigs wrote:

I have a ROM 01. Assuming your's is a ROM 01 too, that is incredibly helpful! If you wouldn't mind, where did you find the battery holder and wired clip? I'm assuming you used the Tadiran TL-5101/S 1/2 AA battery? Also, would 3d-printing a case to protect the circuitry from battery leakeage work well? Do you recommend anything else?

Mine is a ROM01 too, they moved and changed the mounting on the ROM03 closer to the power input which IMHO puts it in a more questionalble spot should it leak... but you can easily replace it! LOL

I guess it would help to just add the items I used in the event anyone else needs to know, all items were from Mouser and printing your own pan should be fine. I expect electrolyte won't affect ABS or nylon I'm not totally sure about PLA. 

These are the parts I used from Mouser:

Tadiran Battery - I'd get a 3.6V 1/2AA and Tadiran or Saft are good options. The 3V Varta is still an option, but I'd go with the 3.6V cell. 

Battery Holder - Note, this comes with a 2-pin plastic connector for the pins but it's got a nice center channel which can be easily cut with snips or knife then run to the battery leads you hopefully left soldered to the board.

The plastic case 

Will just need hot glue or similar (silicone) to afix the plastic tub to the metal shield, and similar to hold the battery holder in place. I went with hot glue because it can be easily removed should I want to do that. I may have used high-temp hot glue, but can't remember now and  regular hot glue melt point should be higher than anything found inside the case.