CC2540 On Chip Crystal

Hi Dan,

I forwarded your query to a HW expert, assuming you ask about CC2520 as the message body says. If you are referring to CC2540, as in your title, I suggest you move the post to the appropriate forum.

Best regards,

OD

Hello Dan,

The crystal frequency determines the frequency accuracy of RF signals from CC2520 and hence the RF performance.

The device can be programmed to transmit CW signal and the difference between programmed frequency and measured frequency is the frequency accuracy, If the difference is more than +/-40ppm, cyatal frequency needs tuning. This can be done by changing the value of crystal load capacitors and/or using FREQTUNE register until the freuqency accuracy is better than +/-40ppm. The procedure is explained in section 22 of CC2520 datasheet. Hope this helps.

Regards

 

 

Hello TI,

In follow up to the TI response:

The crystal frequency determines the frequency accuracy of RF signals from CC2520 and hence the RF performance.

The device can be programmed to transmit CW signal and the difference between programmed frequency and measured frequency is the frequency accuracy, If the difference is more than +/-40ppm, crystal frequency needs tuning. This can be done by changing the value of crystal load capacitors and/or using FREQTUNE register until the frequency accuracy is better than +/-40ppm. The procedure is explained in section 22 of CC2520 datasheet. Hope this helps.

My customer has these comments:

 

1. How do we program the device to transmit a CW signal (what is the register command)
2. If we measure the frequency on a Signal Analyzer, the frequency will be incorrect by some amount since the Signal Analyzer itself has an offset. Is it recommended to use a reference for the Signal Analyzer?

 

Also, there is an interesting line on page 30 of the datasheet:  “It is possible to feed a single-ended signal to the XOSC32M_Q1 pin and thus not use a crystal.”

 - Can your comment on this? Could I use a TCXO instead?

Thanks,

Dan

Hello Dan,

1. The register settings needed for generation of CW signal can be exported (register export option) from SmartRF studio by launching the control panel for CC2520 and programming the desired frequency in continuous TX unmodulated mode. In fact, SmartRF studio can also be used to program the device for this test.

2. Frequency counter can be used for this measurement. Many spectrum analyzers also have frequency counter with programmable frequency accuracy. Also, note that 40ppm accuracy on a 2.4GHz signal is in the order of 96KHz.

3. Yes, TCXO can be used to provided clock reference on Q1 pin of the XTAL. The use case and cost generally drive this decision.

Regards,

Hello TI,

I am a little confused by the 2.1 example schematic. The value of the components connected to RF_P and RF_N are different. I would think they would be the same because they are converting a differential output to a single ended output. Having the components have the same value would ensure balance. I am talking about C171 and C191. Could your guys shed some light on this?

Thanks,

Dan

Could you link to/ attach the schematic in question? I don't want to comment on the wrong schematic by accident.

Here is the schematic:

CC2520EM_2_1.pdf

 

As far as I have been able to find out the imbalance is mainly to compensate for different parasitics in the RF_P and RF_N path. The line inductances will be slightly different due to the surroundings are different.