'5438 vs. '5438A Timer Register Differences

John Westmoreland43

Hello All!

I have a question regarding the '5438 vs. the '5438A - I have some code that has been running fine on the '5438; received my '5438A samples recently from TI so I have changed the app; including the HAL, to the '5438A. The performance is poor - the app is running sluggishly. Kindof reminds me of when you accidentally use the same timer for the LCD code as for a generic timer interrupt.

I compared the includes in CCS - didn't see any changes in the TIMER Register definitions - just wondering if there's something I am missing - are there any Timer changes between the two - other than being able to run the '5438A at 25MHz? Are there any architectural differences between the two that could make proper set up of the timer(s) different?

Seems like there are several enhancements to the '5438A other than just the massive reduction of errata(um) that look interesting - TI doesn't seem to be 'bragging' about those just yet; perhaps they should.

TIA,
johnw

over 15 years ago

0 Priya Thanigai over 15 years ago

TI__Mastermind 26380 points

John,

There are a few significant differences between the F5438 and the F5438A devices - but the timer architecture is the same.

Have you already taken a look at the Migration Guide?

http://focus.ti.com/docs/prod/folders/print/msp430f5438a.html#technicaldocuments

Specifically what is the system & the timer doing? clock source/ capture or compare/ system frequency/ LPMs?

Regards,

Priya

0 John Westmoreland43 over 15 years ago in reply to Priya Thanigai

Guru 13915 points

Hello Priya,

I will review that document - thanks for pointing it out.

The app is an RTOS Demo - running on the MSP430F5438-EXP board - timers are just for controlling LCD and general RTOS tick IRQ.

Thanks,
John

0 old_cow_yellow over 15 years ago in reply to Priya Thanigai

Guru 58965 points

Migrating From MSP430F541x/F543x to MSP430F541xA/F543xA (Rev. B) (PDF 123 KB)

0 John Westmoreland43 over 15 years ago in reply to old_cow_yellow

Guru 13915 points

From the migration doc:

Existing UCS code should require no modification for proper execution.

I should probably check that code to make sure.

Thanks,
John

0 Jens-Michael Gross over 15 years ago in reply to John Westmoreland43

Guru 227245 points

There is a difference in the UCS module. On the A version, entering LPM(4?) will disable the XT2. On the non-A, XT2 stays enabled on LPM4. There was a different thread only few weeks ago about this.

So if you're using LPMs, it might be that your SMCLK/ACLK and therefore the timers fallback to the REFO or VLO clock temporarily, which would explain the slowdown.

0 John Westmoreland43 over 15 years ago in reply to Jens-Michael Gross

Guru 13915 points

Hello Jens,

Evidently something like this is what is going on - will go through the init code.

Thanks,
johnw

**Attention** This is a public forum

MSP low-power microcontrollers

MSP low-power microcontroller forum

'5438 vs. '5438A Timer Register Differences