Is there any data for small size supercap aging?

[LinuxHata]

Hello.
In several of my projects I'm using 5.5V 0.47F supercaps made by Eaton, to back-up DS3231 RTC.
Supercaps are charged via 1N4148 and 330 ohm resistor from 5 volts.
I have several devices, which were made 2-3 years ago, and when they were fresh, supercap charge was enough for 3-4 days of standby. Now it is about 2-3 hours. The total full discharge/charge cycle for these 2-3 years would not exceed 10-15 times, temperature was always below 40C.

So it is normal them to loose capacity that much, considering non-existent abuse?

[Jwillis]

What's the standby voltage? Did you retest outside of device?

[CaptDon]

We did extensive thermal cycling and HALT testing on boards with supercaps. I have never seen more than a 20% apparent reduction in holdup time and capacity. I am curious if you can measure the current draw in standby and see if your 1N4148 has developed leakage or if the Dallas chip is a fake and has internal leakage more than it used to? Going from 2 days to 2 hours is unimaginable for any quality parts. Ours were not Eaton but sadly I forget who did make the ones we used.

[selcuk]

Isn't that due to using 5.5V capacitor with a 5V supply? I believe there is a lifetime estimate on the datasheet such that 1000 hours 50% loss with 5V, or 10% loss with 3V etc. But you may not get data for 3 years.

[LinuxHata]

Voltage on the capacitor is around 4.6 volts, so I guess this should be perfectly safe.
I'll try to measure current consumption and will report back.

[CaptDon]

Ouch!!!! 50% loss after 1000 hours with 5vdc on a 5.5vdc rated supercap?? I sure won't buy any of their product! If the loss is really that bad I would be looking into rechargeable cells like the what were they, the BR version of a CR2032?? I forget the rechargeable designation for those. I did not see those insane degradations in our HALT testing. Funny when you work with test engineers....They decided subjecting our boards to 150C was a way to accelerate life testing, until the bag style supercaps and several other things exploded right off the board!!! Then you try to argue with these chuckle heads saying you can't test that way, and their answer is "This is how industry does it". Well, then you guys clean up the smoking mess!!

[coppice]

Hello.
In several of my projects I'm using 5.5V 0.47F supercaps made by Eaton, to back-up DS3231 RTC.
Supercaps are charged via 1N4148 and 330 ohm resistor from 5 volts.
I have several devices, which were made 2-3 years ago, and when they were fresh, supercap charge was enough for 3-4 days of standby. Now it is about 2-3 hours. The total full discharge/charge cycle for these 2-3 years would not exceed 10-15 times, temperature was always below 40C.

So it is normal them to loose capacity that much, considering non-existent abuse?

How warm do these things run? At 20C a 1N4148 isn't too bad, but warm it to just 40C or 50C and the reverse leakage in many samples is horrendous. These diodes get used in all sorts of places where their leakage is a problem. Engineers do too much of their design testing at room temperature, and don't spot all the cases where warming things up has a really bad effect.