MSP430F5242: DCO accuracy at 24MHz

Hello Michiaki,

The customer is correct.  The datasheed shows the % drift in DCO frequency over temp.

It is possible to improve the accuracy by providing an external 32khz crystal with low drift (in ppm) on XT1 and configure the clock system to use the FLLREFCLK for the DCO.  This stable reference clock frequency will keep the DCO at the desired frequency.

Here is a TI link to MSP430 32-kHz Crystal Oscillators that might be helpful.

Here is link to the crystal that we use on our CapTIvate EVMs.  It is a +/- 20ppm part.

As a side note, here is an informative article written about uart accuracy and how accurate it needs to be.

Hi Dennis,

Please let me confirm the accuracy of DCO at 24MHz

Don’t you have the number of DCO accuracy at 24MHz ?

And if customer add the external 32kHz crystal, could they achieve their high speed communication around 900kbps ?

Thank you and best regards,

Michiaki

Hi Michiaki,

Unfortunately, this parameter is only tested at 1MHz.  If it's not in the datasheet, the spec doesn't exist.

Let's take a different approach.

 Below is a spreadsheet comparing the frequency deviation as a function of temperature for a common 32kHz crystal and the MSP430 DCO.

You can see from the table that using the 32kHz changes very little compared to the DCO.  When using the 32kHz to provide a stable reference clock for the DCO, the drift will be very small.

Yes, the customer should have no problem achieving the 900kbps running the DCO at 24MHz.