• State Resolved
  • +1 person also asked this people also asked this
  • Locked Locked
  • Replies 6 replies
  • Subscribers 80 subscribers
  • Views 670 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

MSP430FG4618: Issues with crystal on XT2

Peter Wolf
Prodigy 20 points
Part Number: MSP430FG4618


Clock for MSP seems to have wrong frequency

Description: On our Transmission Module we use a MSP430FG4618 (D300) for controlling input signals and communications between an BluetoothModul and a PIC Microcontroller over UART and with an accelerator chip over I2C.

We use the XT2IN and XT2OUT pin  on the MSP430 to source the clock with the crystal (G1) ABMM2 – 7,3728MHz – E2 – T from Abracom (CL = 18pF, EMR max. = 150 Ohm)

The two capacitors C307 and C308 have a capacitance of 39pF.

In a new design (not yet produced) we want to change the capacitors C307 and C308 to 36pF.

Issue: After start up the MSP seems to switch to the frequency of the crystal but when measuring a 1ms period of a flashing LED the period is not correct. We try to measure the MCLK.

Probe C1 (10:1 ratio, Input impedance 10MOhm || 12pF): XTIN over C308

Probe C2 (10:1 ratio, Input impedance 10MOhm || 12pF): MCLK

The clock module of the MSP (MSP430x4xx Family User’s Guide p. 290):

In the Family User’s Guide we find this quotation:

5.2.4 XT2 Oscillator

Some devices have a second crystal oscillator, XT2. XT2 sources XT2CLK

and its characteristics are identical to LFXT1 in HF mode, except XT2 does not

have internal load capacitors. The required load capacitance for the

high-frequency crystal or resonator must be provided externally.

The XT2OFF bit disables the XT2 oscillator if XT2CLK is unused for MCLK

(SELMx ≠ 2 or CPUOFF = 1) and SMCLK (SELS = 0 or SMCLKOFF = 1).

XT2 may be used with external clock signals on the XT2IN pin. When used with

an external signal, the external frequency must meet the data sheet

parameters for XT2.

If there is only one crystal in the system it should be connected to LFXT1. Using

only XT2 causes the LFOF fault flag to remain set, not allowing for the OFIFG

to ever be cleared.

 

To use the XT2 pins for connecting the G1 crystal comes from a former design which is already in the field and is running without issues.

I

  • 0
    TI__Guru 71216 points

    Hi Peter,

    I would suggest rather than connecting a scope probe to the XT2_IN or XT2_OUT pins, which can affect the oscillation frequency, enable the MCLK or SMCLK signals to output on their assigned GPIO pins and measure there.

    Now, it is possible that if using parallel caps that are to large, this can pull the resonant frequency off by a measurable amount.  

  • 0 in reply to Dennis Lehman
    Prodigy 20 points

    Hi Dennis,

    thanks for your replay. I already tried to measure MCLK. You can see the result in the above screenshot from the oscilloscope. There is no clear signal. I tried the same measure with a modified VLOCLK- signal (about 400kHz) and only got a saw tooth like signal, I assume that the risetime of the MCLK is too slow for a rectangular signal. This seems to be strange and I do not have an explanation for that.

    I assume that I have to change the clock-signal to XTIN and XTOUT. Or is there a workaround to delete the OFFIG flag when I only have a crystal on XT2?

  • 0 in reply to Peter Wolf
    TI__Guru 71216 points

    Hi Peter,

    What is your status?

  • 0 in reply to Dennis Lehman
    Prodigy 20 points

    Hi Dennis,

    what do you mean with "status"?

    I have some further informations about the MCLK and SMCLK. I found out that a filter was in the circuit. After desoldering this filter I got a nice square signal form the MCLK with the right frequency of 7.3728MHz. The issue with the fault flag is not solved. The MSPs CPU seems not to work with the 7.3728MHz but with a clock of about 1/8th and the OFIFG cannot be deleted. 

    When we only have one crystal in the system, do we have to connect it to XT?

  • +1 in reply to Peter Wolf
    TI__Guru 71216 points

    Hi Peter,

    Sorry for the confusion - I was asking if you had any updates, meaning what is your status.

    Ok, glad to hear you figured out what was wrong with the clock waveform. Yes, if using only one crystal, it must be LFXT.

    Now, when you see IFG1.OFIFG = 1, what are the states of theother oscillator flags in the FLLCTL0 register?

  • 0 in reply to Dennis Lehman
    Prodigy 20 points

    Hi Dennis,

    these are the other flags:

    There seems to be no other fault flag.

    Much of the design I copied from a predecessor application. They used the MSP430FG4618IZQW. In use the MSP430FG4618IZCAT. They put the crystal, a Jauch 7,3728MHz on XT2 and for about 10 years they had no problem with it. We looked at the registers of the old design and the OFIFG-Flag was set too, but non the less it was running. Has something changed, so the OFiFG-Flag can no longer be ignored?

    Best regards

    Peter

**Attention** This is a public forum