Super Capacitor with Real Time Clock

[Kian]

Hi all,

I am designing a real time clock circuit (RTC) and using a super capacitor as a back up power supply. The RTC I am using is PCF8563, and I am using a 3.3V super capacitor (DMS3R3224). The connection is as shown in the image below:

Is this the right way to connect up the super capacitor to the RTC? Any problems? VCC is 3.3V too. Also, I noticed that the super capacitor is drawing a lot of current from VCC. I presume that it is charging up. How do I calculate the time it takes to be fully charged? And how do I calculate the time it take to fully discharge if VCC is taken away and the super capacitor is powering up the RTC.

 

[jjw]

Time to discharge t=U×C/ I
where U is voltage change, C is capacitance and I discharge current.
For example PCF8563 takes 0.25uA, works from 1V then with 0.22F capacitor
t=( 3.3V-0.6V-1V)×0.22F/0.25uA ~1.5 million seconds ~ 17 days

Edit: The leakage current of the capacitor also affects the discharge time.
With a 1uA leakage + 0.25uA the discharge time is about 3.5 days.

 

[Nigel Goodwin]

I really hate the name 'super' capacitor, as it encourages people to think they are particularly useful - when in reality it's just a very poor imitation of a battery.

Anyway, as far as this circuit goes it looks 'OK', but it will only give a fairly short backup time, and the capacitor will probably fail after a couple of years. They were used extensively in cheap VCR's for this same purpose, and it was an EXTREMELY common fault - I've replaced hundreds of them.

 

[JonSea]

Seems to me I read the recommended procedure to charge a supercap is through a current-limiting resistor to prevent excessive current flow and overheating.

 

[schmitt trigger]

Nigel
I would be very much interested in your experience related to these devices.
Why do you think was the reason for the failure rate? Improper charging!

 

[Nigel Goodwin]

Nigel
I would be very much interested in your experience related to these devices.
Why do you think was the reason for the failure rate? Improper charging!

No, I think they are just fairly useless devices.

 

[dr pepper]

Might be a good idea to put a current limit resistor inline with the diode that charges the cap, otherwise the current rushing into the cap on power up could be very high if its discharged, 220 ohm would be a good start.

Heres a charge time calculator, the cap will never be fully charged with a series resistor, but it will get pretty close.