Consistency of DCO parameters across multiple samples of a given part number?

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

Consistency of DCO parameters across multiple samples of a given part number?

David Feldman-4
I am working on a project that uses MSP430G2452IPW14 (14 pin TSSOP package.)

At initialization, I set MCLK and SMCLK sourced from DCO and set DCO to my chosen clock frequency (I did not rely on Info flash segment D calibration parameters, as those didn't yield the frequency I require.)

I was surprised to find upon building my second prototype of the project that the DCO frequency, using the same parameters as the first prototype, was off considerably (perhaps 5% or so.)

What should I expect in terms of DCO frequency consistency? Should I plan a calibration step for each device I use (i.e., find it's correct parameters to obtain the required frequency?) I do not have a crystal or other primary frequency source in the design -- the DCO is being used stand-alone with no discipline mechanism available.

Thanks,

Dave
[hidden email]
Reply | Threaded
Open this post in threaded view
|

Re: Consistency of DCO parameters across multiple samples of a given part number?

Onestone-2
The accuracy and repeatability of the DCO  clock for the pre-programmed
values is around 3%. I tested just 20 samples and found that typically
this was closer to 1%. Certainly on a range of 200 shipped units, using
a calibrated DCO for a 460kb UART, no complaints were received regarding
failed communication to a PC based program. But this is a small sample.
For any other value of DCO frequency I would expect much worse, given
that Ti rate some of their calibrated values as worst case +/-6% over
the full temperature range.  What prevents you from using one of the
calibrated frequencies, or even a cheap watch crystal. Not all, but
most, of my designs use a watch crystal, and still achieve very low mean
power consumption, and I design very tiny systems typically 1cm square
or smaller, so size shouldn't be an issue.

Al


On 21/06/2018 9:01 AM, David Feldman [hidden email] [msp430] wrote:

> I am working on a project that uses MSP430G2452IPW14 (14 pin TSSOP package.)
>
> At initialization, I set MCLK and SMCLK sourced from DCO and set DCO to my chosen clock frequency (I did not rely on Info flash segment D calibration parameters, as those didn't yield the frequency I require.)
>
> I was surprised to find upon building my second prototype of the project that the DCO frequency, using the same parameters as the first prototype, was off considerably (perhaps 5% or so.)
>
> What should I expect in terms of DCO frequency consistency? Should I plan a calibration step for each device I use (i.e., find it's correct parameters to obtain the required frequency?) I do not have a crystal or other primary frequency source in the design -- the DCO is being used stand-alone with no discipline mechanism available.
>
> Thanks,
>
> Dave
> [hidden email]
>
>
> ------------------------------------
> Posted by: David Feldman <[hidden email]>
> ------------------------------------
>
> To unsubscribe from the msp430 group, send an email to:
> [hidden email]
>
>
> ------------------------------------
>
> Yahoo Groups Links
>
>
>
>