MSP430G2553: Timer not working

Do you have a 32kHz crystal attached to XIN/XOUT?

If you do, you should probably check your solder joints.

If you don't, you shouldn't be checking OFIFG (it will be on all the time).

Yes , I have the 32kHz crystal connected and also checked the solder joints. Have not found any issues as of now with the setup. Just to be sure, I also tried switching the MCU on the board with another one but in vain. Any other possible reasons for this to occur?

That appears to be the correct recipe (as seen in Example msp430g2xx3_LFxtal_nmi.c):

https://dev.ti.com/tirex/explore/node?node=AJ5Bmm7V5ca8wCEc-tsvkg__IOGqZri__LATEST

On the F2 series, the LFXT will take over ACLK even if you don't clear OFIFG. If you don't do the loop, what result (blink rate) do you get at P1.7? With VLOCLK (failover) the blinks will be more like 3 seconds than 1 second.

I get no output both with and without the loop part. I used P1.6 LED to debug my code.

In the loop section,

 do{
         IFG1 &= ~OFIFG;                 
         for (i = 50000; i; i--);  
         P1OUT ^= BIT6;      // To indicate the loop running
    } while (IFG1 & OFIFG);

P1OUT &= ~BIT6;  // To indicate once the loop terminates

I did this addition and noticed that my debug LED on the PIN1.6 was not blinking at all, it was just ON. I did not turn off either which it would when it comes out of the loop.
So, as far as my interpretation, it enters the loop and gets stuck after the for statement.
It gave the same result even after replacing the while condition with a 1.

What could be the reason for such an output?

1) When I remove the loop, I get pulses on P1.7. They're a bit faster than I expected (1.5Hz?) I suspect due to oscillator failover, since I don't have a crystal. 

2) When I tried it, the compiler made the delay loop disappear, since it wasn't doing anything. You can declare i "volatile" or use __delay_cycles(). 

I found out that the delay in the loop was too less , so it was not perceivably blinking. The loop section is working now, only the OFIFG flag does not get cleared. 

I found a (G2ET) Launchpad with the crystal enabled. When I run this code OFIFG clears, and I get a pulse on P1.7 once per second. What result do you get if you remove the OFIFG loop?

I suspect your crystal isn't oscillating. Can you describe your board?

The code is not working even after the loop is removed.

I have attached the circuit for your reference.

I also tried replacing the crystal with another one from the same lot and still it is not working. Is there any way I can test crystal as we do for other components in electronics?

Thank you.

What is it that isn't working when you remove the OFIFG loop? How do you tell?

The program should do something (if you remove the loop) whether or not the crystal is oscillating, but it will be different, which will be a clue. 

One thought: Try extending that 200-cycle delay to blink the LED to something much longer, maybe 100000 (100ms). The current blink is not very easy to see.

Sorry for not being clear on that. I changed the ISR code to just turn ON the LED (P1.7) in order to avoid the confusions of this delay function.

After removing the loop , the MCU enters the low power mode and stops there. Tested that by making the led P1.6 to glow before it entered the low power mode and also measured the current consumption of the mcu.

Hm. If ACLK is running, the ISR should eventually be called. So maybe we're not looking at the crystal per se, and maybe not at something I can replicate here.

Two possible experiments:

1) Add "BCSCTL3 |= LFXT1S_2; // VLOCLK" somewhere up at the beginning. This will set ACLK=VLO and make the crystal irrelevant.

2) Replace the LPM3 statement with something like: "__BIS_SR(GIE); while(1) {}". This will keep it from going into LPM.

The code works after changing the clock source to VLOCLK.

If I am right, there must have been a problem with my external crystal.

I haven't seen the failover fail (completely) before, though I have seen it behave erratically. (That's why I know the LFXTIS_2 trick.)

One more thing to try: Add "P2SEL=0xC0; .// Run the LFXT no kidding" somewhere up at the beginning. This is the Reset value, and as far as I know the BSL doesn't mess with P2SEL, but it seems that we're into territory that is somewhat uncharted.

Are you fairly certain that 22pF capacitors match with your crystal? The 32kHz crystals I've worked with were in the 6-12pF range.

I replaced those 22pF caps with 8pF capacitors and also included the reset value in the code. Still getting the same issue. The code is stuck in the loop.

I used 8pF capacitor as it is in the range you had mentioned above and that was the only low value cap that was available in the stores. Capacitors in that range seems quite uncommon here.

Unfortunately, my experience is that it's not so easy to measure the crystal circuit directly, since a scope probe tends to throw it off.

You might try experimenting with the XCAP setting in BCSCTL3. [Ref User Guide Sec 5.3.4] It starts out at 1 (6pF), so you could try setting it higher or lower.

[Edit: I forgot to ask: What is the recommended capacitance for your crystal? Do you have a part number for it?]

Gave all 4 XCAP options a try, still the same.

I am sorry if this is a basic question, do we have part numbers for lower frequency crystals? Mine is just a silver cylinder with no markings on it.

I didn't think about the recommended capacitor value since the circuit I shared earlier was a perfectly working circuit. It worked both with and without the capacitor. So, I thought that it wouldn't matter much with few pF variation in the value of capacitor .

The cylindrical "tuning fork" crystals don't have part numbers on them (or at least my old eyes can't see them), so I keep mine in the (labeled) envelope they came in. The first one I found in my gizmo box is:

https://www.mouser.com/ProductDetail/ABRACON/AB26TRQ-32768KHZ-T?qs=%2Fha2pyFaduhLEYIPQrxx91mYdfM%252BEQEqp%252BSly4jQSF%2FvyFlf8D2xNA%3D%3D

I'm not sure I understand your second comment. If the circuit was working, what changed?

I actually saw a video where the the circuit was used without any capacitors for the crystal. Still the timer functioned with the crystal as its clock source. . Since the cap value didn't seem that significant, I  just used the value mentioned in the circuit diagram without searching for any other details.

I think the official board supported by TI uses a 12pF capacitor. So I suspect my crystal oscillator.

Have you tried it without any capacitors? (You might need to adjust XCAP.)

I just found an old G2 Launchpad (Rev 1.5) in a drawer which has a crystal but no capacitors (C21/C22) installed and (per DMM) no external GND connection and the oscillator seems to run fine.

Removed the capacitors and tried running the code and its still not working. I also tried out all the XCAP settings, nothing worked. 

The problem might be with the crystal, I think. Though replacing the crystal from the lot that I had bought didn't workout,  I will try it out with some new crystals as the whole lot could be faulty.

This is a good mystery, and I'm running out of things to suggest. Maybe a photo would be enlightening?

This is the test setup I was working on. I actually have disconnected the capacitor from the crystal (as a part of my last test) which is not be visible from this view.

This is the bottom view. Hope the image is clear... 

What do the DIP switches do? I don't see them in the schematic.

I am using a CP2012 breakout board for programming this. The female pins (on the extreme right) are used to connect the CP2102 board. Since a press on the reset button clears the code in the MCU while having the board connected, I have used the DIP switches to disconnect all the other pins other than +5v and gnd.