TMS320F28377D: Recommended crystal parameters

Is there any specific crystal parameter required other than crystal Frequency?
how much load capacitance(CL) required in crystal?

Yes, you can find the crystal requirements in the data manual, screenshot below. The specifications of the crystal that you choose, has to meet this.

We are using FOXS200-20(20 MHz) crystal. It has load capacitance(CL)=20pF, and we are using(CL1=CL2=22pF) in our control card. Will it be ok?

Amtech,

You have to meet CL=20pF spec from the crystal manufacturer in order to get 20MHz or else you could see some frequency shift. CL1 and CL2 are in series so provided CL1=CL2, their equivalent capacitance Ceq is (CL1)/2 which from your example = 11pF. You also need to account for trace and parasitic capacitance, 3pF is a reasonable estimate but that of course will change depending on your board. This gives you about 14pF of Ceq. If follows that CL ≠ Ceq.

From the point of view of the electrical oscillator circuit in the F28377D, CL1/CL2 requirement is met since it's between 12pF-24pF.

So to answer your question: assuming the ESR requirement is also met, the crystal could oscillate and you could see a frequency shift.

Thanks for your valuable feedback.
If we are using above configuration, is there any harm to DSP?

No, it will not harm the dsp.

In our present design, we are using FOXS200-20(20 MHz) crystal. It has load capacitance(CL)=20pF, and (CL1=CL2=33pF) . This design has already been commercialized and running in field. So will dsp be harmed or any other harmful effect will be there, in already running systems?

CL1/CL2 is out of our specification which is 12-24pF. I can't tell you your design is ok because you are violating our specification.

Thanks for solving this query.
But still we have doubt that what will happen to already running systems in field, because as you have pointed out we have violated the cl1/cl2 limits. The main concern is that will dsp be harmed other than frequency shift?

Amtech,

We usually can't comment when customers use our devices out of spec because the behavior will be undefined as there are too many variables involved that we just can't tell you what will happen.

However in this case, i can explain a few things. As i mentioned before, the CL1/CL2 you use has to meet two specs: the quartz crystal's CL spec and the electrical oscillator's CL spec.

From the quartz crystal's perspective, CL1/CL2 being out of spec will shift the frequency which might be ok depending on your application. From the electrical oscillator's perspective, since CL1/CL2 has an influence on the negative resistance, too high or too low of CL1/CL2 will lower the negative resistance which can cause the crystal not to start-up and that is definitely more worrying than a frequency shift.

In my experience (not with this part) problems with the wrong capacitor values show up at temperature extremes. You might find that the oscillator won't start up at high or low temperatures. If you have the capability, before production starts up it's always a good idea to verify that your product will start up at both ends of its rated temperature. Sometimes the oscillator cap values need adjusting.

Lloyd

Hi Amtech,

Do you have any further inquiries on this or is it clear?

Amtech,

I'll be closing this thread. Feel feel to reply if your question is still not resolved or create a new thread if this one locks out due to inactivity.