The specs I've read on Li-Mn2O4 button cells indicate up to 1500 charge/discharge cycles. However, to obtain this kind of longevity, the depth of discharge must be limited to only 10%.
If the RTC battery is fully discharged over a period of 5 days as Uniden indicate (Acknowledging your comment about being able to leave the radio off for 2 weeks), a 10% discharge level would be reached after only 12hrs of the radio being powered down. From what you've said though, I think the more likely scenario is that a (new) fully charged RTC battery actually lasts up to 2 weeks on a full charge.
All good scenarios!
Do you think the Li-Mn204 5mAh sized cell is what is actually found in the radios or did you just take a guess from a similar sized example?
The cell in the radio is closer to 5/32nds to 3/16th inch diameter compared to the 1/4" diameter of the Panasonic LiMn I mentioned earlier. Thickness is similar though judging by eye alone.
I could not see it well enough to read any markings on it without removing the front panel on a 536. And the solder tab is often spot welded over any markings in many cases. My cell location is marked BT401.
I think it was you that saw a current draw spec on the VBAT line to the RISC chip between 0.6 and 1.0 mA. I read the datasheet and I'd swear it said 0.6 to 1.0 uA (micro) amps so that may help a lot if my eyes were seeing correctly!
If your calculations are correct, will Uniden will replace that battery in the 536 models now that they have a 3 year warranty for those that fell in the serial number range for the headphone jack noise issue. I'd bet they have a battery exclusion in the warranties.
I think the specs for the RTC VBAT input listed a decent low voltage before the RTC may need to be reset. I forget exactly but I think it was near 1.6 VDC or so. So anything equal or above should maintain the clock settings.
I can't recall looking at the discharge curve of a Li or Li-Mn cell but do they fall rapidly when they hit a certain voltage (like NiMH does) or do they fall gracefully towards zero like a Carbon Zinc or Alkaline cell does?
I think they would have been better off using something like a CR2032 in a socket similar to a computer motherboard.
Remember the fiasco with some of the Icom R71 era radios where the operating code was stored in a battery maintained memory board?
I think those used a CR2016 size Li cell but soldered in place. When they went dead, your radio no longer worked.
There are still some of those radios running today with the original factory cell that is now over 30+ years old! Way longer than the average 5 year life of the Li cells from that era.
What was found is that the owners that use them daily are the ones with cells that are still maintaining the OS memory. The very diode that was installed to prevent the cells from charging has a very small leakage current that is believed to be responsible for those 30+ year old cells to still work and hold the memory.
It was the owners that shelved the radios for years that found them to no longer work when they drug them back out.
Luckily, a few companies built drop in memory boards for the affected Icom models. Some even had jumpers that you set so it used the memory contents for your specific model so one board fit all models.
Mine still had the original Li cell in it when I swapped it out. It was still showing near 3.0 volts and was about 20 years old when I did the swap. I did a live swap with the memory board externally powered while I replaced the Li cell. I later replaced the board as the 3rd party boards extended the tuning range some. Of course one could build a programmer and reload the original memory board code (including extended tuning ranges) but just $40 dollars for a non volatile memory board sounded to be much easier! I still have my original memory board and the 2032 Li cell I soldered in place of the original 2016 is still showing over 3.0 volts event though it has not been inside a radio in 7 or more years now.
Anyway, Uniden could have used a 2032 in a socket that is only drawn from when the radio is off. I'd imagine a 2032 would run the RTC circuit for a long time compared to the tiny Li-Mn cell used now.
Of course a 2032 should never be charged but it never seemed to hurt any of the thousands of Icom's still running strong out there with 30+ year old 2016 Li cells!
Of course the reverse leakage current through the charge protection diode in the Icom's had to be very very small. So small that it never caused the 2016 Li cells to heat up and explode but enough that it allowed them to maintain voltage near 3VDC.
When you had mentioned in another thread that you thought the x36HPs used a super cap, I thought that was good news as they should have a much longer life when you factor in total or partial discharge/recharge cycles that some users may see in a year.
My weather stations use super caps that do discharge pretty darn low each night if I have cloudy days and the solar panel cannot provide the needed current. I've not had a super cap fail me yet. They also have a CR3 Li cell that provides operating voltage for extended periods.
Some of the super caps Davis used have failed they found. They believe it was caused by improper forming of the super caps leads when bending them at a right angle to insert into the circuit board. Another words, they bent the leads by hand instead of holding them with something so the leads did not move inside the caps body when being formed.
I guess in this case, time shall tell!