S12 / MagniV MicrocontrollersのトピックRe: Clock modes

Clock modes

Roelof — Mon, 22 Jan 2007 19:31:53 GMT

We are using the MC9S12DT256. Normally it is has a colpit clock circuit, which is working without any problems at 16MHz.

Now we need to have a large clock signal amplitde to synchronize it with other ICs on the same board. So I want to run it in Pierce-clock mode. So I changed PE7/XCLKS_not to ground and placed two capacitance from XTAL and EXTAL to ground (15pF) and parallel to the Crystal a resistor of 3MOhm. And I removed the series capacitance.

But now the processor is runing on a low clock frequency (internal clock), and I measure no clock signal on EXTAL and XTAL pins. I also see no DC voltage on these pins (VDDPLL is above 2V).

So my question is:

- Is our processor able to run in Pierce mode?

- Do I need to change more than the clock circuit, and pull-up/down resistor connected to PE7? (some code perhaps?)

- Should there be a DC voltage on the EXTAL and XTAL pins?

Re: Clock modes

Alban — Mon, 22 Jan 2007 20:01:31 GMT

Hello Roelof,

Yes, S12 work on Pierce mode as you described.

/XCLKS is sampled during RESET to select oscillator mode.

I don't see any code to change. The only thing to change is the oscillator.

If you are using the PLL, check your XFC filter values (S12(X) PLL Calculator Software).

Yes, the oscillations are not centered on zero, there is a DC bias.

Your capacitors could be a bit low (I usually see ~22pF) but it depends on your crystal characteristics.

Also, you can reduce the feedback resistor to 1Mohm.

Finally, you will probably need a damping resistor in series.

Cheers,

Alban.

Re: Clock modes

Roelof — Mon, 22 Jan 2007 20:17:28 GMT

Hello Alban,

So two things are worrying me:

- Why do I not measure any DC voltage?

- You mention the PLL. At this moment my PLL inputs are not connected. I have placed a resistor of 0 ohm between XFC and VDDPLL and a capacitance of 220NF between VDDPLL and VSSPLL. Do I need to do anything with the PLL to get the Pierce oscillator running? I downloaded the software, so I can try it. Thanks.

How can I deteremine the damping resistor?

Roelof

Re: Clock modes

Roelof — Mon, 22 Jan 2007 20:35:22 GMT

I increased both capacitances to 25pF, but I still measure nothing (no AC nor DC signal).

How can I get a DC voltage on the XTAL pins?

I used a pull down of 1kOhm on PE7. When this is pull-up I have a DC voltage of 2.5V on the XTAL pins.

Re: Clock modes

Alban — Mon, 22 Jan 2007 20:55:20 GMT

Hello,

I've been flagged something on Pierce availability on early DT256.

Can you please give me the mask set so I check ?

The mask set is 0X00X format with X = letter and 0 = digit.

The series resistor is determined by Negative Resistance Margin (NRM) or Safety Factor measurements.

If you have very high oscillations, they cause noise (EMI).

If oscillations are too low, the oscillator may not start.

NRM is taken by adding a resistor is series (damping) and see at which value it stops oscillating.

Then you can take the Safety Factor by dividing this resistor value by the resistance in the system.

Thanks,

Alban.

Re: Clock modes

Roelof — Mon, 22 Jan 2007 21:01:24 GMT

Hello,

The mask is:

2K79x

I've tried random a resistor of 47E, 100E, 1k. So I will look into the NRM.

I also tried different crystals (5MHz, 16MHz).

But still it worries me that there is no DC component.

Thanks,

Roelof

Re: Clock modes

Alban — Mon, 22 Jan 2007 21:21:17 GMT

Hello Roelof,

You use an early mask set which does not support Pierce configuration.

oscillator - 16-Bit Microcontrollers - Freescale Forums | Discussion Forums | html
... C parts - mask sets 0K36N, 0K79X, 1K79X and 2K79X don't support the Pierce crystal configuration ...

Also, the mask you are using is discontinued:

TSPG MCD16 EOL OF SELECTED 0.25UM S12 | Documentation | htm
... nbsp; Barracuda2 : 0K36N, 0K79X, 1K79X, 2K79X ...

Sorry to have mislead you at the beginning...

Alban.

Re: Clock modes

Roelof — Mon, 22 Jan 2007 21:51:07 GMT

Hello,

Thank you for this quick answer.

Indeed there is an old processor used on this board.

On the new board type 0L01Y is used, so maybe I can use or try it in the future.

So if I understand correctly: I have to find another solution on the short term (the first batch of this board).

I think the issue is solved.

Again thank you,

Roelof