How to initialize MSP430F5510 correctly?

[jazz 209]

From my point of view, big plus for MSP430 family is that external oscillator is not needed. Can't find info about DCO consumption, frequency temperature drift is 0.1 %/C and frequency voltage drift is 1.9 %/V. I didn't done application till now, that was sensitive to this. And for CDC USB example this is big problem, right.

 

MSP430F5xx also have:

1. Internal Very-Low-Power Low-Frequency Oscillator (VLO) with frequency temperature drift 0.5 %/C and frequency voltage drift 4 %/V .

2. Internal Reference, Low-Frequency Oscillator (REFO) with frequency temperature drift 0.01 %/C and frequency voltage drift 1 %/V (consumption 3 ?A).

 

BTW, XT1 consumption is <0.3 uA and XT2 consumption is 200-450 uA.

 

For all my projects, that are not using USB, there will not be XTAL2. It will be present only in USB projects, because USB module can't be run from internal clock generators.

 

If you want to run the system by XTAL2, no problem, this can be tested by simple hello world program, outside of USB_Dev_Package. Later when you solve the problem of "XT2 stabilization loop" (slac394 MSP430F550x MSP430F5510 C Code Examples, MSP430F550x_UCS_08.c), just go back and adapt original USB_Dev_Package code with your changes.

 

In all examples of USB_Dev_Package (AFAIK) DCO is used for uC clock (there is nothing wrong in this), and if somebody is not happy with this, and want to clock the uC with XTAL1/2, this can be done (even I don't see any reason for this), but don't blame TI if the things are not going by your wish...

 

Exactly this is the Problem - not everything is covered there and the examples provided in the USB_Dev_Package are very poor (minimalistic) programmed. Just copy-and-paste of the same source with some modifications for each example.

[SirZusa 34]

Let's concentrate on the problem: The Fault-Flag wont dissappear ...

Same issue after stripping all unnecessary lines of code

[jsolarski 94]

if the fault flags wont clear, then clear them in a loop and check them again, remember the settling time for xtal and dco.

[thanhtran 10]

There is a write up on TI site somewhere about the watch xtal giving OSC fault flag. I made a  water pump controller last week using a launchpad with a watch xtal. The watch crystal  was very sensitive to noise and present of an 120v line next to it. I didn't want to debug problem and didn't want the risk of overflowing my basement with water so I stopped using it. 

[SugarAddict 227]

Isn't the VCore supposed to be increased in steps and not just up'd straight?

[jazz 209]

By datasheet Core voltage is in relation with DCO, not XT1/2.

 

My MSP430F5510 using 12 MHz XT2 (board without XT1) as MCLK, without any problems. Setup is based on TI example (slac300d MSP430F552x C Examples, MSP430F55xx_UCS_08.c, "MSP430F552x Demo - XT2 sources MCLK & SMCLK").

 

Right now I am working with multiple MSP430F2xx target devices, P1.0 blink example, clocked by 1MHz calibrated DCO. And what I see is that calibrated DCO is not precise. They start synchronized, but after some time (1 minute), blinking synchronization is lost. Calibrated DCO can't be used for applications that requires precise timing. But for everything else is more than enough.