TMS320F28374D: External crystal oscillator does not work during chip burn-in test

Bruce(Shanghai) Liu said:

My customer found 3 out of 100 pieces failed to vibrate when the external crystal oscillator was tested during the burn-in test. 

What was the max burn-in temperature?

Bruce(Shanghai) Liu said:

We checked the crystal oscillator parameters selected by the customer, and there were indeed deviations from the specifications

What are the deviations? Are you referring to  CL1=CL2=33pF?

Hi Hareesh,

Thanks for reply,

Hareesh Janakiraman said:

What was the max burn-in temperature?

60℃

Hareesh Janakiraman said:

What are the deviations? Are you referring to  CL1=CL2=33pF?

Yes, I see max load capacitance is 24pF in the specifications and ESR is a little higher, Could this be the main reason? I see there is nothing wrong with the customer’s layout.

Bruce,

    If only 3 out of 100 boards failed burn-in, that means there is some marginality in the design. is there a reason why customer used CL1=CL2=33 pF when crystal datasheet calls for 20 pF and our datasheet calls for 24 pF? 

You can refer to the 6.12.3.4 XTAL Oscillator section in www.ti.com/lit/SPRS61. Even though it is the datasheet for a different  device, the same crystal oscillator is used in that device as well. The section states the below:

"The effect of CL on the crystal is frequency-pulling. If the effective load capacitance is lower than the target, the crystal frequency will increase and vice-versa. However, the effect of frequency-pulling is usually very minimal and typically results in less than 10-ppm variation from the nominal frequency."

So, marginal deviation in CL value impacts the frequency value, not the operation of the crystal oscillator itself.

Hi Hareesh,

How about ESR? Will an excessively high ESR cause excessive power and affect the operation of the crystal oscillator?