• State TI Thinks Resolved
  • Locked Locked
  • Replies 12 replies
  • Answers 3 answers
  • Subscribers 21 subscribers
  • Views 462 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

Original question:

LMX2820: LMX2820

LMX2820: LMX2820 lock time

Eddie Chou
Expert 6180 points
Part Number: LMX2820

Tool/software:

Hi team, 

My customer is evaluating LMX2820 and the loop filter parameters are based on the LMX2820EVM default values (C1=470pF, C2=68nF, C3=2.2nF, R2=68.1Ω, R3=18.2Ω).. 

the simulated result shows a Lock Time of approximately 88 µs, including the impact of VCO calibration. 

However, in customer HW measurement, the time from SPI transmission completion (CS# goes high) to LD (Lock Detect) pin transition from LOW to HIGH, and it is approximately 171 µs.

customer would like to confirm, 

  1. Is measuring the Lock Time based on LD falling and then rising edge a recommended and reasonable method?
  2. If we would like to achieve a faster Lock Time (e.g., close to the 88 µs simulation result), would it require specific hardware settings (such as enabling Fast Mode or Instant Calibration)?
  3. Could you kindly advise the appropriate hardware configuration to achieve closer alignment with the simulated results?

Also in below application note, it states a method to reduce calibration time to below 5us,

https://www.ti.com/lit/an/snaa342/snaa342.pdf

Can we shoe customer how to evaluate the locking time?  and below are additional questions from customers. 

  1. The Lock time includes VCO calibration time and analog lock time, am I right ? So the method is only to reduce VCO calibration time, analog lock time depends on hardware loop filter.
  2. Faster lock setting? GUI set? (as below 4.3 state)
  3. Can we just check LD pin for total Lock time for LMX2820 to lock? How GUI start to initiate both VCO calibration and frequency change for check total lock time?

Thanks & Regards

Eddie

  • 0
    TI__Guru** 107660 points

    Hi Eddie,

    How many registers did you program during the lock time test? How fast were you writing the register?

  • 0 in reply to Noel Fung
    TI__Expert 6180 points

    Hi Noel, 

    Customer use TI’s LMX2820 EVB and press Calibrate VCO button in GUI. If change frequency then press Enter, the total program time too long and not good to check lock time. Customer need our suggestion how to test the fast lock time through TI’s EVB and how to configure LMX2820 to do fast lock.

    SPI clock: ~128KHz

    Thanks & Regards

    Eddie

  • 0 in reply to Eddie Chou
    TI__Expert 6180 points

    Hi Noel, 

    My I get team's feedback?

    Thanks & regards

    Eddie

  • 0 in reply to Eddie Chou
    TI__Guru** 107660 points

    Hi Eddie,

    Total lock time = programming time + calibration time + analog lock time.

    Right, analog lock time is proportional to the loop filter design.

    User can shorten the programming time with higher SPI speed.

    In LMX2820, calibration time can be eliminated with the InstCal feature.

    To use this feature, a one-time calibration is required. After that, changing frequency will not do calibration anymore. 

    Below is the procedure to use InstCal. 

    Initialization register setting
    1. Set DBLBUF_PLL_EN, DBLBUF_CHDIV_EN, DBLBUF_OUTBUF_EN, DBLBUF_OUTMUX_EN = 1
    2. Set DBLR_CAL_EN = 0; INSTCAL_SKIP_ACAL = 0
    3. If VCO doubler is required, set INSTCAL_DBLR_EN = 1, otherwise set this bit to 0
    4. Set INSTCAL_DLY = T x fosc (in MHz) / 2^CAL_CLK_DIV, where T = 2.5 x CBIASVCO / 0.47µF. CBIASVCO is the bypass capacitor at pin 3
    5. Configure other registers to lock to 5.65GHz without any calibration assist
    6. Set INSTCAL_PLL_NUM = 2^32 x (PLL_NUM / PLL_DEN)
    Programming
    7. Vcc power up LMX2820
    8. Program all the registers, LMX2820 should lock to 5.65GHz
    InstCal calibration
    9. Program INSTCAL_EN = 1
    9.5 run Index routine
    10. Program R0 (with FCAL = 1), calibration will begin
    11. Wait 100ms
    12. Program R0 (with FCAL = 0) to complete the calibration
    13. If LD pin did not turn HIGH, program RESET = 1 to reset LMX2820. After LMX2820 is reset, repeat Step 8 to 12
    Changing VCO frequency
    14. Program INSTCAL_PLL_NUM, PLL_N, PLL_NUM and PLL_DEN (if their value have change)
    15. Program R0 (with FCAL = 0) to change VCO frequency

    9.5 Index routine
    1. Program R106[10] = 1 (0X6A 0400)
    2. Program R107[5:0] = 0 (0x6B 0000)
    3. Program R108 = 0 (0x6C 0000)
    4. Program R109 = 0 (0x6D 0000)
    5. Program R106[9:0] = 514 (0x6A 0602)
    6. Program R106[11] = 1 (0x6A 0E02)
    7. Program R106[11] = 0 (0x6A 0602)
    8. Repeat step 5 to 7 for R106[9:0] = 515 to 519
    9. Program R106[10] = 0 (0x6A 0207)

  • 0 in reply to Noel Fung
    TI__Expert 6180 points

    Hi Noel, 

    Can we find InstCal API function in TICS-Pro? or can we provide InstCal script (*.tcb) file for customer execute in TICS-Pro?

    Thanks & regards

    Eddie

  • 0 in reply to Eddie Chou
    TI__Guru** 107660 points

    Hi Eddie, 

    There is no API in TICS Pro.

    I have listed all the necessary steps, simply follow the procedure and it should work.

    For the index routine, we can create a Burst Mode .tcb file.

    https://e2e.ti.com/cfs-file/__key/communityserver-discussions-components-files/48/2820index.tcb

  • 0 in reply to Noel Fung
    TI__Expert 6180 points

    Hi Noel, 

    Is there locking time difference between the hopping frequency range? customer did some experiment if the hopping frequency is less than 100MHz the locking time can support 7uS. but if hopping frequency range large than 100Mhz and locking time will increase to 37uS. 

    Is any solution can improve locking time with hopping frequency range large than 100MHz?

    below hopping frequency large than 100MHz

     

    and below hopping frequency less than 100Mhz.

    Thanks & Regards

    Eddie

  • 0 in reply to Eddie Chou
    TI__Guru** 107660 points

    Hi Eddie,

    Are the plots taken using InstCal?

    What is the target lock time?

  • 0 in reply to Noel Fung
    TI__Expert 6180 points

    Hi Noel, 

    Yes, customer use Instant Calibration（InstCal）mode for testing. 

    The target locking time is less than 10us and if the hopping range less than 100Mhz locking time is  around 7 ~ 8us but if hopping range large than 100MHz ( Ex. 5650Mhz ~ 5800MHz) locking time will increase to 32 ~ 37us.

    customer have questions below, 

    1. The locking time 7~30uS is within our spec. ? any measurement method can suggest for customer?

    2. for large range hopping, is there any register setting can help increase the locking time?

    Thanks & Regards

    Eddie

  • 0 in reply to Eddie Chou
    TI__Guru** 107660 points

    Hi Eddie,

    In general, 10µs is possible for any frequency change. You may need to scarify some phase noise for faster frequency switching. 

    C29 should be 470nF or less. C31 should be less than 10µF.

    Loop bandwidth should be wide enough to reduce PLL analog lock time. The bandwidth should be above 100kHz.

    Your test result (1st plot) does not look right to me, LD should remain HIGH before R0 programming is complete, just like the second plot. Did you check all the doubler buffer enable bits?

  • 0 in reply to Noel Fung
    TI__Expert 6180 points

    Hi Noel, 

    Customer use our EVM without change components value,  below are some experiment use Instant Calibration mode, 

    Step1：enable Instant Calibration mode, set output 5650 MHz，。

      • load register value as attached, " Step1_5650mhz.txt"

    Step2：change output frequency to 5850 MHz（+200 MHz hopping）。

      • LD pin change from  HIGH to LOW， PLL  unlock 。
      • load register as  "Step2_5650mhzto5850mhz.txt"

    Step3：Toggle FCAL and  LD come back to HIGH. 。

      • Measured LD from LOW → HIGH is around 30 µs。
      • the register setting is " Step3_5850mhzcal.txt"

    Questions from customer, 

    1:  Do we have experience use LMX2820 frequency hopping more than 100MHz but Lock time less than 10uS?

    Can we provide example setting change from 5650MHz to 5850MHz base on our EVM?

    2. from pllatinum sim simulation lock time is 30.3uS, but from customer experiment , they can sometime get 10uS locking time but unloack sometimes. 

    Is it related to parameter setting ( VCO_CAPCTRL、DACISET、VCO_SEL、INSTCAL_EN) ? 

    Can you help above questions?

    Thanks & Regards

    Eddie

    1. Fullscreen Step1_ 5650mhz.txt Download 
      R122	0x7A0000
R121	0x790000
R120	0x780000
R119	0x770000
R118	0x760000
R117	0x750000
R116	0x740000
R115	0x730000
R114	0x720000
R113	0x710000
R112	0x70FFFF
R111	0x6F0000
R110	0x6E001F
R109	0x6D0000
R108	0x6C0000
R107	0x6B0000
R106	0x6A0000
R105	0x69000A
R104	0x680014
R103	0x670014
R102	0x660028
R101	0x6503E8
R100	0x640533
R99	0x6319B9
R98	0x621C80
R97	0x610000
R96	0x6017F8
R95	0x5F0000
R94	0x5E0000
R93	0x5D1000
R92	0x5C0000
R91	0x5B0000
R90	0x5A0000
R89	0x590000
R88	0x5803FF
R87	0x57FF00
R86	0x560040
R85	0x550000
R84	0x540040
R83	0x530F00
R82	0x520000
R81	0x510000
R80	0x5001C0
R79	0x4F011E
R78	0x4E0001
R77	0x4D0608
R76	0x4C0000
R75	0x4B0000
R74	0x4A0000
R73	0x490000
R72	0x480000
R71	0x470000
R70	0x46000E
R69	0x450011
R68	0x440020
R67	0x431000
R66	0x42003F
R65	0x410000
R64	0x400080
R63	0x3FC350
R62	0x3E0000
R61	0x3D03E8
R60	0x3C01F4
R59	0x3B1388
R58	0x3A0000
R57	0x390001
R56	0x380001
R55	0x370002
R54	0x360000
R53	0x350000
R52	0x340000
R51	0x33203F
R50	0x320080
R49	0x310000
R48	0x304180
R47	0x2F0300
R46	0x2E0300
R45	0x2D0000
R44	0x2C4000
R43	0x2B00FA
R42	0x2A0000
R41	0x290000
R40	0x280000
R39	0x2703E8
R38	0x260000
R37	0x250500
R36	0x24001C
R35	0x233100
R34	0x220010
R33	0x210000
R32	0x201001
R31	0x1F0401
R30	0x1EB18C
R29	0x1D318C
R28	0x1C0639
R27	0x1B8001
R26	0x1A0DB0
R25	0x190624
R24	0x180E34
R23	0x171102
R22	0x16E2BF
R21	0x151C64
R20	0x14272C
R19	0x132120
R18	0x120000
R17	0x1115C0
R16	0x10171C
R15	0x0F2001
R14	0x0E3001
R13	0x0D0038
R12	0x0C0408
R11	0x0B0612
R10	0x0A0000
R9	0x090005
R8	0x08C802
R7	0x070000
R6	0x060A43
R5	0x050032
R4	0x044204
R3	0x030041
R2	0x0281F4
R1	0x0157A1
R0	0x006420
      Fullscreen Step2_ 5650mhzto5850mhz.txt Download 
      R122	0x7A0000
R121	0x790000
R120	0x780000
R119	0x770000
R118	0x760000
R117	0x750000
R116	0x740000
R115	0x730000
R114	0x720000
R113	0x710000
R112	0x70FFFF
R111	0x6F0000
R110	0x6E001F
R109	0x6D0000
R108	0x6C0000
R107	0x6B0000
R106	0x6A0000
R105	0x69000A
R104	0x680014
R103	0x670014
R102	0x660028
R101	0x6503E8
R100	0x640533
R99	0x6319B9
R98	0x621C80
R97	0x610000
R96	0x6017F8
R95	0x5F0000
R94	0x5E0000
R93	0x5D1000
R92	0x5C0000
R91	0x5B0000
R90	0x5A0000
R89	0x590000
R88	0x5803FF
R87	0x57FF00
R86	0x560040
R85	0x550000
R84	0x540040
R83	0x530F00
R82	0x520000
R81	0x510000
R80	0x5001C0
R79	0x4F011E
R78	0x4E0001
R77	0x4D0608
R76	0x4C0000
R75	0x4B0000
R74	0x4A0000
R73	0x490000
R72	0x480000
R71	0x470000
R70	0x46000E
R69	0x450011
R68	0x440020
R67	0x431000
R66	0x42003F
R65	0x410000
R64	0x400080
R63	0x3FC350
R62	0x3E0000
R61	0x3D03E8
R60	0x3C01F4
R59	0x3B1388
R58	0x3A0000
R57	0x390001
R56	0x380001
R55	0x370002
R54	0x360000
R53	0x350000
R52	0x340000
R51	0x33203F
R50	0x320080
R49	0x310000
R48	0x304180
R47	0x2F0300
R46	0x2E0300
R45	0x2D0000
R44	0x2C4000
R43	0x2B00FA
R42	0x2A0000
R41	0x290000
R40	0x280000
R39	0x2703E8
R38	0x260000
R37	0x250500
R36	0x24001D
R35	0x233100
R34	0x220010
R33	0x210000
R32	0x201001
R31	0x1F0401
R30	0x1EB18C
R29	0x1D318C
R28	0x1C0639
R27	0x1B8001
R26	0x1A0DB0
R25	0x190624
R24	0x180E34
R23	0x171102
R22	0x16E2BF
R21	0x151C64
R20	0x14272C
R19	0x132120
R18	0x120000
R17	0x1115C0
R16	0x10171C
R15	0x0F2001
R14	0x0E3001
R13	0x0D0038
R12	0x0C0408
R11	0x0B0612
R10	0x0A0000
R9	0x090005
R8	0x08C802
R7	0x070000
R6	0x060A43
R5	0x050032
R4	0x044204
R3	0x030041
R2	0x0281F4
R1	0x0157A1
R0	0x006420
      Fullscreen Step3_ 5850mhzcal.txt Download 
      R122	0x7A0000
R121	0x790000
R120	0x780000
R119	0x770000
R118	0x760000
R117	0x750000
R116	0x740000
R115	0x730000
R114	0x720000
R113	0x710000
R112	0x70FFFF
R111	0x6F0000
R110	0x6E001F
R109	0x6D0000
R108	0x6C0000
R107	0x6B0000
R106	0x6A0000
R105	0x69000A
R104	0x680014
R103	0x670014
R102	0x660028
R101	0x6503E8
R100	0x640533
R99	0x6319B9
R98	0x621C80
R97	0x610000
R96	0x6017F8
R95	0x5F0000
R94	0x5E0000
R93	0x5D1000
R92	0x5C0000
R91	0x5B0000
R90	0x5A0000
R89	0x590000
R88	0x5803FF
R87	0x57FF00
R86	0x560040
R85	0x550000
R84	0x540040
R83	0x530F00
R82	0x520000
R81	0x510000
R80	0x5001C0
R79	0x4F011E
R78	0x4E0001
R77	0x4D0608
R76	0x4C0000
R75	0x4B0000
R74	0x4A0000
R73	0x490000
R72	0x480000
R71	0x470000
R70	0x46000E
R69	0x450011
R68	0x440020
R67	0x431000
R66	0x42003F
R65	0x410000
R64	0x400080
R63	0x3FC350
R62	0x3E0000
R61	0x3D03E8
R60	0x3C01F4
R59	0x3B1388
R58	0x3A0000
R57	0x390001
R56	0x380001
R55	0x370002
R54	0x360000
R53	0x350000
R52	0x340000
R51	0x33203F
R50	0x320080
R49	0x310000
R48	0x304180
R47	0x2F0300
R46	0x2E0300
R45	0x2D0000
R44	0x2C4000
R43	0x2B00FA
R42	0x2A0000
R41	0x290000
R40	0x280000
R39	0x2703E8
R38	0x260000
R37	0x250500
R36	0x24001D
R35	0x233100
R34	0x220010
R33	0x210000
R32	0x201001
R31	0x1F0401
R30	0x1EB18C
R29	0x1D318C
R28	0x1C0639
R27	0x1B8001
R26	0x1A0DB0
R25	0x190624
R24	0x180E34
R23	0x171102
R22	0x16E2BF
R21	0x151C64
R20	0x14272C
R19	0x132120
R18	0x120000
R17	0x1115C0
R16	0x10171C
R15	0x0F2001
R14	0x0E3001
R13	0x0D0038
R12	0x0C0408
R11	0x0B0612
R10	0x0A0000
R9	0x090005
R8	0x08C802
R7	0x070000
R6	0x060A43
R5	0x050032
R4	0x044204
R3	0x030041
R2	0x0281F4
R1	0x0157A1
R0	0x006420
  • 0 in reply to Eddie Chou
    TI__Guru** 107660 points

    Hi Eddie,

    Here is my data, after finishing the one-time InstCal calibration, I change the frequency to 5850MHz, measure the lock time, then change the frequency to 5650MHz, measure the lock time. I use CSB pin to trigger test equipment.

    From 5850Mhz to 5650MHz:

    LD pin timing same as PNA. 

    From 5650MHz to 5850MHz:

    LD pin timing.

    To change the frequency, simply type the frequency you want in RFOUTA box. Output frequency will not change because double buffering is enabled. Then click the Calibrate VCO button (this will program R0) to change the frequency.