• State
  • Locked Locked
  • Replies 9 replies
  • Subscribers 10 subscribers
  • Views 240 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.

LMX8410L: Homodyne IQ output "hopping" when using the DIV2 Quadrature Generator

Raymond Harker
Prodigy 150 points
Part Number: LMX8410L
Other Parts Discussed in Thread: LMX2594

When setting the LO below 7500MHz, I'm seeing significant jumps in the IQ output between frequencies (custom board).  These jumps are random and will change frequencies on subsequent sweeps (as you can kind of see here).  If the jumps occurred throughout the entire bandwidth I'd start to suspect my ADC or backend processing, but the fact that it always settles out prior to the 7500MHz mark makes me start to suspect an issue with the DIV2 Quadrature generator.  I've used the TICS Pro tool to setup my registers and I'm copying it's suggested register write order on the frequency change, but with no luck.  I'm waiting until the VCO locks before I start my readings, so it shouldn't be an issue with an unstable output.  Increasing my dwell time on each frequency (3ms-12ms) doesn't change the output at all.  

Any suggestions on what I'm doing wrong?  I suspect that it's a simple register issue somewhere, but the documentation for this device is spotty at best.  

  • 0
    TI__Mastermind 28960 points

    Hello Raymond,

    It sounds like you are using the internal synthesizer for your LO generation. Are you sure the internal path is being selected, and you're not trying to source the LO externally? It's possible the VCO will still feed through the mux with no signal present on the LO pins, even at reduced amplitude; this would degrade the consistency of the divider output and could cause issues with the IQ phase precision.

    What is the y-axis on the graph above? I'm guessing green and blue correspond to I and Q outputs (or vice versa)?

    Can you upload the TCS file you're using to generate your settings?

    Regards,

    Derek Payne

  • 0 in reply to Derek Payne
    Prodigy 150 points

    0172.HexRegisterValues.txt 
    R127 (INIT1)	0x7F0003
R6 (INIT2)	0x060100
R127	0x7F0000
R126	0x7E8000
R125	0x7D0000
R124	0x7C0000
R123	0x7B0000
R122	0x7A0000
R121	0x790000
R120	0x780000
R118	0x760000
R117	0x7500C0
R116	0x740000
R115	0x730F00
R114	0x720000
R113	0x710000
R112	0x700000
R111	0x6F00B7
R110	0x6E0100
R109	0x6D0000
R108	0x6C0000
R107	0x6B0000
R106	0x6A0000
R103	0x672C00
R102	0x660000
R101	0x650002
R100	0x640002
R99	0x630000
R98	0x620000
R97	0x610000
R96	0x609218
R95	0x5F0000
R94	0x5E8080
R93	0x5D0000
R88	0x580000
R87	0x570000
R86	0x560000
R85	0x550000
R84	0x541900
R83	0x536A20
R82	0x520A23
R81	0x510310
R80	0x500009
R79	0x4F0040
R78	0x4E0064
R77	0x4D0000
R76	0x4C000C
R75	0x4B0800
R74	0x4A0000
R73	0x49003F
R72	0x480001
R71	0x470081
R70	0x46C350
R69	0x450000
R68	0x4403E8
R67	0x430000
R66	0x4201F4
R65	0x410000
R64	0x401388
R63	0x3F0000
R62	0x3E0022
R61	0x3D00A8
R60	0x3C03E8
R59	0x3B0001
R58	0x3A8001
R57	0x390000
R56	0x380000
R55	0x370000
R54	0x360000
R53	0x350000
R52	0x340820
R51	0x330080
R50	0x320000
R49	0x314180
R48	0x300300
R47	0x2F0300
R46	0x2E07FC
R44	0x2C1F22
R43	0x2B0000
R42	0x2A0000
R41	0x290000
R40	0x280000
R39	0x270001
R38	0x260000
R37	0x258204
R36	0x240050
R35	0x230004
R34	0x220000
R33	0x211E21
R32	0x200393
R31	0x1F43EC
R30	0x1E318C
R29	0x1D318C
R28	0x1C0488
R27	0x1B0002
R26	0x1A0DB0
R25	0x190624
R24	0x18071A
R23	0x17007C
R22	0x160001
R21	0x150401
R19	0x1327B7
R18	0x120064
R17	0x1100FA
R16	0x100080
R15	0x0F064F
R14	0x0E13F0
R13	0x0D4000
R12	0x0C5001
R11	0x0B0018
R10	0x0A10D8
R9	0x090604
R8	0x082000
R7	0x0700B2
R6	0x06C802
R5	0x0500C8
R4	0x041943
R3	0x030642
R2	0x020500
R1	0x010808
R0	0x00201C

    These are the register values on my device right after I set it to 4GHz, read directly from the LMX8410.

    I'm using the internal LO and also routing the VCO (or channel divider output) out through the LO port of the device.  I amplify the output, transmit it, and then route the received frequency back through the RF port to generate my baseband IQ values (which are displayed on the graph's Y axis in Volts).  I've set the internal termination of the LO port to High Z.  

  • 0 in reply to Raymond Harker
    TI__Genius 12550 points

    Hi Raymond,

    Derek, the subject expert is OoO and will be back next week. Can you be wait for respond by him?

    Regards,
    Ajeet Pal

  • 0
    Prodigy 150 points

    I played around with the data a bit more, it appears as though I am seeing a random 180 degree phase shift on the output data when using the DIV2 quadrature generator.  See below

    This is 10 different sweeps from 5500 to 6000MHz.  I lock the PLL, take 1024 samples, then increase the frequency and repeat the process.  Every 180 degree shift on the I data is accompanied by a 180 degree shift on the Q data (both have a sign change at the same time).  The output is clearly mirrored around a DC offset, but I still can't figure out what's causing the mirroring.  

    EDIT:

    It appears now that the 180 phase shift only occurs when I change VCOs.  I enabled VCO_CALSTART_CLOSE and the shifts are only occurring around VCO boundaries.  They are still random though.

  • 0 in reply to Raymond Harker
    TI__Genius 12550 points

    Derek can respond once he is back.

    Regards,

    Ajeet Pal

  • 0 in reply to Raymond Harker
    TI__Mastermind 28960 points

    Raymond,

    Are you ever synchronizing the internal synthesizer? I am wondering if the state of the DIV2 quadrature generator is effectively randomized by the VCO recalibration as it crosses the VCO boundary... it's been a few years since I drew the block diagram, but I believe I recall the SYNC feature can also reset the DIV2 quadrature generator, which should be able to eliminate any uncertainty in the LO phase even across VCO transitions.

    Regards,

    Derek Payne

  • 0 in reply to Derek Payne
    Prodigy 150 points

    I've implemented the SYNC function via the external pin but appear to be having the same issue.  I changed my MASH order to Integer to try out the Category 2 sync, but it didn't make a difference.  I'm pulsing the SYNC pin every time I change frequency, but the phase shift is still only occurring on the VCO boundaries which makes me think that the SYNC process isn't taking the DIV2 quadrature generator's output into account (or I'm not implementing SYNC properly, although I'm following the LMX2594 datasheet's directions and each pulse is triggering a VCO calibration).  

    On a possibly related note, I noticed that the N values in the TICS Pro software appear to be incorrect for this device when using the SYNC function.  The IncludeDivider factor is listed as 4, but I was unable to get a lock unless I used a divider of 2.  It seems as though the IncludeDivide factor already accounts for the DIV2 factor, so the software ends up factoring it in twice.  Maybe this is more evidence that the DIV2 Quadrature generator occurs separate from the SYNC?  

  • 0 in reply to Raymond Harker
    TI__Mastermind 28960 points

    Raymond,

    There's some bits in the register map R80 (11, 10, and 9) that I don't think are mentioned anywhere helpful, but which seem important for LO SYNC functionality. Do you have these enabled? I didn't see them in your original settings. The procedure (if I understand correctly) should set SYNC_PIN_IGNORE to 0 and all other phase sync bits to 1; it seems SYNC_PHASE_MIXLO bit explicitly handles synchronizing the LO path.

    I'm also asking internally to see if we have any demonstrated method to SYNC the DIV2 element.

    Regards,

    Derek Payne

  • 0 in reply to Derek Payne
    Prodigy 150 points

    For the SYNC implementation I was setting R80 to 0x0E09 and the SYNC_PIN_IGNORE to 0, as well as enabling the SYNC_PHASE_PLL bit in R0.  I wasn't able to poll my OSCin signal so I couldn't ensure the setup/hold times in fractional mode, but I put the device into Integer mode and still had the same problem.  If I understood the datasheet correctly, there are no setup/hold timing restrictions on the SYNC toggle when I'm in integer mode.