RTC using 18F4550

Guest · Posted: Wed Oct 21, 2009 11:16 am

Hello all,

I'm considering rolling my own RTC function using the "secondary" (TIMER 0) oscillator that is built into the 18F4550. According the the datasheet, this oscillator is often used with a 32.768 kHz watch crystal. Seems like a "nearly free" way to implement an RTC without using a dedicated clock chip (DS1307, DS1340, etc). I only need to have a fairly accurate clock/calender with the most basic functions. The challenge might be in the life of the backup battery as the PIC draws somewhere between about 2 and 35 uA while running off the 32 kHz oscillator. This is much higher than that required by a DS1307 for example.

Any thoughts or experiences are greatly appreciated.

bkamen · Posted: Wed Oct 21, 2009 12:23 pm

Consider using the 18FxxJ11 series that have built in RTC's and can do your time keeping function for you with that same 32kHz source.

I'd have to check, but I think the powerdown current is lower too because in the case of the J11's the CPU can shut down while the clock keeps running whereas the 18F4450 cannot.

Just a thought.

-Ben
_________________
Dazed and confused? I don't think so. Just "plain lost" will do. :D

PCM programmer · Joined: 06 Sep 2003 Posts: 21708

Guest · Posted: Wed Oct 21, 2009 1:04 pm

Thanks PCM. You're correct - it is Timer 1, not Timer 0. I'm familiar with the basics of an RTC implementation, however have not had any experience with trying to get a PIC (that has the secondary oscillator capability) to run at a low enough current to come close to rivaling a "real" stand-alone RTC chip. In order to do this, the PIC would have to be able to run 24/7 - even during a power loss. My application is not battery-powered which is a plus. I think the question comes down to being able to power the PIC from a coin cell when my board gets powered-down or loses power.

PCM programmer · Joined: 06 Sep 2003 Posts: 21708

If you're looking for battery backup circuits, this Tips 'n Tricks document
from Microchip shows the typical Schottky diode circuit:
http://ww1.microchip.com/downloads/en/DeviceDoc/01146B_chapter%202.pdf

Or, more suggestions:
http://www.microchip.com/forums/tm.aspx?m=398317

ckielstra · Joined: 18 Mar 2004 Posts: 3680 Location: The Netherlands

bkamen · Posted: Thu Oct 22, 2009 10:50 am

Interesting.

The PIC18FxxJ11 (I was looking) shows a spec of 850nA for running the RTC and another 1uA for timer1's oscillator (needed for running the clock - if I'm reading that right)

That's not bad at all.

And yes, software can make a lot of boo-boo's that up that consumption.

but at least they don't need cut-patch repairs to a PCB. :D

-Ben
_________________
Dazed and confused? I don't think so. Just "plain lost" will do. :D

PCM programmer · Joined: 06 Sep 2003 Posts: 21708

Actually it says the ds1307 battery backup current is 300 na typical:

FvM · Joined: 27 Aug 2008 Posts: 2337 Location: Germany

While 500 nA of DS1307 is a max value over the full temperature range, you have to consider worst case values of about 5 uA with "nanoWatt XLP" PIC18FxxJ11. DS1307 easily achieves 5 or 10 years operation with the smallest available coin cell, I'm not aware yet of any processor that can compete in this regard.

As an additional point, a separate RTC allows easy specific battery backup, much more auxilary circuitry is involved with battery supply of the processor. A suitable integrated RTC hardware would preferably provide a dedicated Vbat pin.