Re: Consistency of DCO parameters across multiple samples of a given

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

Re: Consistency of DCO parameters across multiple samples of a given

David Feldman-4
Thank you for both quick replies - much appreciated.

I now understand that there is greater variability unit-to-unit than I assumed, however, this it not a problem, just requires that I include a mechanism for determining then storing in information flash the required unit-specific constants.

Thanks,

Dave

--------------------------------------------
On Thu, 6/21/18,  <[hidden email]> wrote:

 Subject: [msp430] Digest Number 5229
 To: [hidden email]
 Date: Thursday, June 21, 2018, 9:04 PM
 
 There are 2 messages in this issue.
 
 Topics in this digest:
 
 1a. Re: Consistency of DCO parameters
 across multiple samples of a given   
     From: Onestone
 
 2. Re: Consistency of DCO parameters
 across multiple samples of a given   
     From: Fedde Bloemhof
 
 
 Messages
 ________________________________________________________________________
 1a. Re: Consistency of DCO parameters
 across multiple samples of a given
     Posted by: "Onestone" [hidden email]
 onestone_apc
     Date: Wed Jun 20, 2018
 8:29 pm ((PDT))
 
 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
 >
 >
 >
 >
 
 
 
 
 
 Messages in this topic (2)
 ________________________________________________________________________
 ________________________________________________________________________
 2. Re: Consistency of DCO parameters
 across multiple samples of a given
     Posted by: "Fedde
 Bloemhof" [hidden email]
 fdover22000
     Date: Thu Jun 21, 2018
 12:15 am ((PDT))
 
 DCO is not accurate indeed. That's why
 TI programs DCO values during manufacturing. Some kind of
 calibration is inevitable.
 
 
 
 Messages in this topic (1)
 
 
 
 To unsubscribe from the msp430 group,
 send an email to:
 [hidden email]
 
 
 
 ------------------------------------------------------------------------
 Yahoo Groups Links
 
 
 
     (Yahoo! ID required)
 
 
     [hidden email]
 
 
 
 ------------------------------------------------------------------------