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.

CC2500EMK: CC2500 Calibration Solutions and FAQs required

Part Number: CC2500EMK

Dear All,

We have some very tight real-time constraints in our software.The reason for doing a "manual calibration" is that from IDLE to Rx/Tx with calibration, the state transition time is about 800 usec. whereas, from IDLE to Tx/Rx without calibration, its only takes 90 usec.

we have read in the datasheet that there are 3 ways to do calibration:

a) Auto calibration (default)

b) Calculating FSCAL1,2,3 upfront for calibration use

c) Calibration with Charge pump disabled

Our concern is

[1]  We cant afford auto calibration because it takes about 800us. If we go with "Option B":

a) How much time it will be required to do a calibration in this case ?

b) Do we have to give SCAL strobe for this option B ?

c) How much temperature variation we can afford with this option B ?

d) Do we have to give manual calibration if we go with "Option B"

 [2] If we go with "Option C": [ We think that this is the feasible way ]

a) If we disable charge pump what are the disadvantage we get by doing this ?

b) If we use SCAL to calibrate will we again have to wait for 800us to get calibrated like in case of  "Option A"

c) How often manual calibration have to be done if we go with "Option C"

Our architecture requirements:

Number of Channels : 0 to 24

Frequency hopping used: Yes

Range: 100 m

Note:

Please provide the answer with respect to above mentioned points.

Regards,

MHS

  • Hi
    Answers are as per your numbering nomenclature:
    [1]Answer: Having tried this in one of my application I found out that you only have to do this for every channel that you want to calibrate your radio on and save those values in an array and before switching channels you have to assign these values back to FSCAL registers.FSCAL is a read/write register. You may only do it at the startup of your application and only do It if and when temperature/environmental situations change drastically. This can reduce your calibration time significantly but the PLL stability will still not be as good as in case of auto calibration ON with charge pump current enabled. You will have to manually use the SCAL strobe whenever you want to calibrate.
    a) How much time it will be required to do a calibration in this case ?
    Answer:79-80us but please find an oscilloscope and verify it yourself using GPIOs as test points.
    b) Do we have to give SCAL strobe for this option B ?
    Answer: Yes every time you want to calculate FSCAL values on a certain frequency channel and calibrate afterwards.
    c) How much temperature variation we can afford with this option B ?
    Answer: You can set a threshold in your application but since you can manually calibrate anytime you want, why are you concerned about this?.You should re calculate the FSCAL values every 2 to 3 degree temperature change. Use the temperature sensor available on the SOC to sense the temperature continnously
    d) Do we have to give manual calibration if we go with "Option B
    Answer:
    You keep asking the same question again and again.Yes sir you are calibrating manually so you have to.
    [2] If we go with "Option C": [ We think that this is the feasible way ]
    a) If we disable charge pump what are the disadvantage we get by doing this ?
    Answer: Compromised PLL stability hence reduced RF performance because you might have to increase the receiver filter bandwidth to accommodate the frequency offset in your selected channel which will increase noise floor of the receiver hence the poor radio performance.
    b) If we use SCAL to calibrate will we again have to wait for 800us to get calibrated like in case of "Option A"
    Answer: Please get your concepts right, this is an automatic calibration mechanism and you don't have to do calibration manually in this case but to your question the calibration time will be 80us but it will be done on its own.
    c) How often manual calibration have to be done if we go with "Option C"
    Same answer as 'b' because manual calibration will not be required.

    Summary and Personal Experience:
    Please don't compromise on the PLL stability and do calibration automatically with charge pump enabled and when you are turning on the HS XTAL OSC, please make sure you give it enough time to get stable before you move on with your application.
    Why do you need fast frequency hopping in a 100m range application any way?
    I would recommend frequency agility over frequency hopping if interference/cognitive solution is a problem anyway. If this is a two way communication application, Frequency agility is a more stable method and will give best radio performance.
    Regards
    Ali

  • Thank you Ali Awan75 much appreciated.
  • You are welcome!