• State Resolved
  • Locked Locked
  • Replies 13 replies
  • Subscribers 22 subscribers
  • Views 3598 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

LMX2592: LMX2592 about the CAPCTRL question

Mickey Zhang
Genius 12080 points
Part Number: LMX2592
Other Parts Discussed in Thread: LMX2594

Hi team,

The customer is using LMX2592. He needs to know how to get the CAPCTRL value for LMX2592.

Compared to LMX2594, LMX2594 has the CAPCTRL value calculation formula on page 26 of the datasheet. But LMX2592 does not have

this calculation formula.

I notice there is a post that our engineer shows the CapCTRL for LMX2592. The link is: 

Our engineer shows the information as below:

           Fmin Fmax CapMin CapMax KvcoMin KvcoMax
VCO1 3550 4070      169       23         22             37
VCO2 4070 4530      157       15         24             40
VCO3 4530 5010      168        25        26             43
VCO4 5010 5570      160        7          30             43
VCO5 5570 6060      172       12         37             48
VCO6 6060 6590      173      13          41             54
VCO7 6590 7100      172     15           47             58

So, for LMX2592, can I use the above information and the  formula VCO_CAPCTRL_STRT = round (CCoreMin – (CCoreMin – CCoreMax) ×

(fVCO – fCoreMin) /(fCoreMax – fCoreMin)) that is on page 26 of the LMX2594 datasheet to calculate the CAPCTRL value?

Best Wishes,
Mickey Zhang
Asia Customer Support Center
Texas Instruments

  • 0
    TI__Expert 8065 points

    Hi Mickey,

    Yes, the formula from LMX2594 also applies to LMX2592. Please note that those values in the table are approximate values. There will be part to part variation of the exact frequency corresponding to each Cap value. But this is a good starting point. Is the customer planning to manually configure each VCO frequency instead of using auto-calibration?

    regards,

    Brian Wang

  • 0 in reply to Brian Wang
    TI__Genius 12080 points

    Hi Brian,

    Thanks for your help.

    Yes, the customer in planing to to manually configure each VCO frequency.

    Now he uses the above cores values and the formula from LMX2594 ,but the phase cannot lock. Do you have any suggestion for

    the CAPCTRL value?

    If he uses the auto-calibration, how to calculate the CAPCTRL value?

  • 0 in reply to Mickey Zhang
    TI__Expert 8065 points
    Mickey,

    That would be expected, those values in the table is a typical approximation. On different parts, the same capcode will be a different frequency, not the exact same in the table. The LMX2592 does not have the same function of readback for the CAPCTRL so cannot do the method of auto-calibration and just readback what the CAPCTRL value found is. To find the CAPCTRL manually for the LMX2592 you would need to:
    1. VCO_CAPCTRL_OVR=1 (allow setting CAPCTRL manually)
    2. VCO_SEL_FORCE=1 (allow setting VCO manually)
    3. VCO_SEL={the VCO you want to use of the 7 available}
    4. VCO_CAPCTRL={here you set from value 183 (lowest frequency) to 0 (highest frequency) and record the frequency of this CAPCTRL setting of the VCO you selected}

    Usually this is manually CAPCTRL question is asked when customers try to do a very fast calibration. Can you ask what calibration time they are expecting, and we can try the different calibration related settings for auto-calibration to have the fastest possible time (and let the internal frequency search algorithm do the CAPCTRL search for you).

    Regards,

    Brian Wang
  • 0 in reply to Brian Wang
    TI__Genius 12080 points
    Hi Brain,

    OK. Thanks. I will confirm with the customer.
  • 0 in reply to Brian Wang
    TI__Genius 12080 points

    Hi Brian,

    For LMX2594, if the customer set CAPCTRL first, then fast calibration, finally, PLL lock phase in the mode 2,

    he would like to know how much the lock time is.

  • 0 in reply to Mickey Zhang
    TI__Expert 8065 points

    Mickey,

    You mean the LMX2592 right? The calibration has many different steps. On a high level, it will search through the cores first, then at the right core search through the CAPCTRL steps, then spend time determining the best amplitude setting for phase noise. The fast calibration that we describe is about how to shorten each of these portions, and after optimizing these knobs, calibration times can go down to around the 30us range. If customer is using an evaluation module with our software, you can get the settings from them and we can create a file for them to try these settings that give the fastest calibration time.

    Regards,

    Brian Wang

  • 0 in reply to Brian Wang
    TI__Genius 12080 points
    Hi Brain,

    I mean LMX2594. If the customer set CAPCTRL first, then fast calibration, finally, PLL lock phase in the mode 2,

    he would like to know how much the lock time is.
  • 0 in reply to Mickey Zhang
    TI__Expert 8065 points

    Mickey,

    That would be the "partial assist" in datasheet section 7.3.6.1 VCO Calibration. Expect around 35us calibration time.

    Regards,

    Brian Wang

  • 0 in reply to Brian Wang
    Prodigy 30 points

    Hi,Dear Brian

     I'm the customer mickey said , haha,and can you help me answer a few questions please!!

    In lmx2594 ,the 35us VCO calibraion means what?If we need a total lock time (When the phase of locking point  is less than 10 degrees  ) less than 48us,can we  meet the requirement? And what should we do?

     

  • 0 in reply to WEI EASON
    TI__Expert 8065 points

    Hi Wei,

    A good plot to explain from is the "Figure 12 - VCO Unassisted Calibration" from the LMX2594 datasheet. You see it breaks down the different portions.

    Firstly, there is the Core finding, where it is trying to find the VCO core out of the 7 internal cores.

    Secondly, is the FCAL (frequency calibration) portion, where is searches through the capcode of the VCO to find what is needed for the target frequency. As you see here, there are different features in this part, one of which is you can give a predicted capcode knowing what frequency you are targeting. In doing so you will start the search closer to your final target frequency and thus cut down on calibration time. This I believe is what you are trying to use.

    Thirdly, there is the ACAL (amplitude calibration portion). Here the device searches through an IDAC code (setting the swing of the VCO) until if arrives at a good level.

    These three steps are what we mean by the calibration. After this is done, there is still what is called the analog settling time (getting down to the fine resolution frequency and phase steady state). This is determined by the loop filter bandwidth (the wider the bandwidth the shorter this time). Around 100kHz loop filter bandwidth is approximately around 20us or so analog settling time.

    Hope this helps!

    Regards,

    Brian Wang

  • 0 in reply to Brian Wang
    Prodigy 30 points
    Dear brian.
    Thank you very much.I am very thankful that you are considering my problem.I have the last question about lmx2594.
    Can i manually set the value of the vco idac when i using lmx2594 on partial assistant mode.Then i can decrease the calibration time for lmx2594.
  • 0 in reply to WEI EASON
    Prodigy 30 points
    En .I means that we set vco _daciset _force=1 and set the value of vco _daciset rather than set the value of vco_daciset _start when lmx2594 working on partial assiant mode.
    I am sorry for my english.
  • 0 in reply to WEI EASON
    TI__Expert 8065 points

    Wei,

    You understand this well. Yes, to reduce that amplitude calibration to the bare minimum you can force to go into manual amplitude selection with VCO_DAC_ISET_FORCE=1, and then set your own value in VCO_DACISET (you should be able to see the phase noise improve as you increase this value, but at a certain point it usually saturates I think typically around 300s). 

    Regards,

    Brian Wang