BR2032 batteries looking superior to CR2032 in SRAM carts
That Nintendo bundled CR2032's doesn't mean they're the best. They're cheaper. Top tier Panasonic has CR2032 at 30 cents loose and $1 tabbed, versus BR2032 $1 loose and $1.53 tabbed apiece, at quantity of 20.
I found this helpful comparison page with the important bottom half pasted below. The "B" versus "C" indicates a different cathode and "R" means...round. BR has 3 things going for it:
- Much flatter voltage curve showing it will last longer to the ~2.7V cutoff of 90s SRAM in your carts and to 2.5V or less in modern (and replacement) SRAM.
- Less self-discharge at ~0.5% per year versus ~1% per year you see on CR battery datasheets. This is crucial in our world of expecting, with CR, saves to last 20 years minimum and 30+ plausible but not guaranteed.
- BR powering SRAM closer to its minimum voltage to retain memory greatly reduces SRAM's leakage current. Meaning SRAM draws less current from the battery.
The catch, besides price, is BR sure rides it close to that 2.7V cutoff. The graph shows a high ~3/150k = ~20uA draw, whereas SRAM draw is 1-5 uA*. Working voltage is higher at lower current. Batteries aren't ideal voltage sources.
Carts tend to rest slightly hotter than at datasheet 20°C (68°F), and heat ticks up the BR voltage. Unless we're talking a shack with no climate control, where maybe BR loses the save in the winter. I think you got bigger problems like condensation corroding solder joints.
I was going to buy Panasonic CR and BR batteries at the same time since my CR's are 3 years old, just to be fair. I already have a nanoamp current probe with data logging to measure SRAM draw at different voltages. Curious if anyone has other thoughts or ideas.
Yes, I know FeRAM / FRAM chips pin-compatible with SRAM is a thing and need no battery. Downside is costing 10x more. The chip itself can still die in 30-40 years and lose the save. Plus I like "the feel" of carts working how they did in the 90s.
