Watchdog timer accuracy

[lpcware]

Content originally posted in LPCWare by 1234567890 on Mon Dec 02 13:29:25 MST 2013
Hi,

datasheet for LPC11Cxx says: Accuracy of watchdog oscillator is +/- 40 % between -40 to 85 °C. Not such a big problem yet as I can calibrate (e.g. with the 1% IRC). But unfortunately there is no information how temperature and voltage influences the frequency (like the figure for IRC).
Has anyone driven tests or more information about the dependencies?
And is it stable without any changes in temperature and voltage (what I assume)?

[lpcware]

Content originally posted in LPCWare by embd02161991 on Fri Jan 31 11:09:24 MST 2014
Hi,

The watchdog oscillator is an RC oscillator . So as temperature varies , the frequency is going to deviate. frequency = 1/2*pi*RC.

As temperature increases , the resistance is going to increase . This results in decrease in frequency and vice versa. For the watchdog oscillator the variations is +-40% . The frequency spread for WDOSC is 9.4 kHz to 2300 kHz. Process variations would also affect the frequency in this range along with temperature . So it is pretty difficult to characterize the data for watchdog oscillator on basis of temperature alone. Hence , it is not present in the datasheet.

Thanks,
SK 

[lpcware]

Content originally posted in LPCWare by liaochengyi on Sat Dec 28 21:12:50 MST 2013
Hi,
The frequency change is depend on temperature not depend on time.

[lpcware]

Content originally posted in LPCWare by 1234567890 on Fri Dec 27 10:09:48 MST 2013
If anyone is interested in: I measured the watchdog oscillator at 18 °C at 2 LPCXpresso LPC11C24, bought 3 weeks and a year ago: The frequency differs only about 0.07 %!!!

Should be 3 MHz, was 3,000,1831 Hz (666.2 ns). Measured on the CLKOUT pin with an ultra expensive 12 $ USBee AX PRO LA clone from China with sigrok, 1 channel and 24 MHz.

But unfortunately it's of course non-linear. At 600 kHz the accuracy was 1.24 % (3,375.25 ns).

Of course that is not a scientific message, but that's good enough for me.