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.

LMX2492EVM: Spur on ramp reset

Part Number: LMX2492EVM
Other Parts Discussed in Thread: LMX2492,

Hi fellow EEs,

I have been using the LMX2492 in its ramping (FMCW) mode, and programming it with TICS Pro(which proved to be an effective approach with relatively few problems apart from occasional crashes).

I used the ramping function with the following settings, and measured everything on the RFout output of my Rev.C LMX2492EVM :

  • 9600 MHz --> 9800 MHz frequency sweep range ;
  • 100 µs ramp duration.

The ramp loops on itself (previous settings are for ramp 0, thus RAMP0_NEXT is set to 0).
The end frequency of the generated pattern being different from the beginning frequency, RAMP0_RST is activated.

I successfully generate a ramp, however when measuring the power levels at a given frequency inside the sweep range (zero span view on SA), I also measured a spur on ramp reset.

When using a more performant SA with Pulse measurement features, I was able to get a time domain view of the ramp itself (not just its spectrum) :

As you can tell, on ramp reset, there is a significant parasitic pulse (see top left/bottom left-most).

However, this event and behavior between ramps is not seen in the LMX2492EVM's User's Guide (3.2.2, Fig.10).

Is there way to reduce or completely eliminate it? I have tried using a negative ramp segment to avoid using the RST function, to no avail.

Thanks in advance for your insight,

Regards,

Nicolas

  • Hi Nicolas,

    I think the problem is with the Vtune voltage being not enough to drive the VCO to 9800MHz. In the EVM, passive loop filter is used, as a result, the max. Vtune voltage from the PLL is around 3V. To get 9800MHz from the VCO, we need close to 4V Vtune voltage. We can modify the EVM to use active filter so that we can support higher Vtune voltage. 

  • Hi Noel,

    Thanks for the quick answer. I will try to stay in the PLL's charge pump output voltage range.

    However, in the EVM User's Guide (3.1.1, Table 1), there is a VCO frequency range specified, which I assumed the EVM was able to cover :

    What part of this range is the EVM effectively able to cover with a passive loop filter configuration ?

    Regards,

    Nicolas

  • I also want to add that experimentally, I was able to generate ramps with frequencies of up to 9.8 GHz, using the EVM with its factory passive filter configuration, without any problems.

  • Hi Nicolas,

    9.4 to 10.8GHz is the frequency range of the VCO, I should have make this clear in the user's guide, my bad.

    I reviewed the EVM schematic again, although the PLL is 3.3V operated but the charge pump supply is actually tied to 5V, That means, the output Vtune can be up to 4.5V. From the VCO datasheet, the EVM should be able to support up to 10GHz.

    in your last plot, were you also doing a sawtooth ramp?

    In your first post, if you sweep the frequency between 9600 and 9700MHz, will you see the same issue?

    Badly I don't have the EVM in the lab, otherwise, I could do a crosscheck.

  • Hi Noel,

    Thanks for clearing up the EVM's frequency range.

    In the last plot, I was indeed generating a sawtooth ramp.

    I changed the ramp frequency to operate between 9600 and 9700 MHz, as per your request. I got the following plot :

    I also experimented with the RFout/2 output, and got a nominal looking ramp (while still ramping from 9600 to 9800 MHz) :

    Having measured good looking ramps on the RFout/2 output, I decided to measure further ramps on this output.

    In my experiments, I used a ramp going from 9600 to 9700 MHz in 100 µs, and used another one to ramp back to 9600 MHz in 20 µs. The pattern's ending frequency being equal to the starting frequency, RST parameter was turned off.

    The spurs appeared again, both at the end of the pattern and between both ramps. Note that the power supply was set to a current limit of 400 mA during all measurements.

  • Hi Nicolas,

    My colleague has this board, I will check this out on Monday.

  • Hi Nicolas,

    Here is my test result, 9600 --> 9700MHz in 100us; 9700 --> 9600 in 20us. current consumption is 276mA including the on-board XO. 

    register file here. OSCin = 100MHz.

    4617.HexRegisterValues.txt
    Fullscreen
    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    R141 0x008D00
    R140 0x008C27
    R139 0x008B10
    R138 0x008A00
    R137 0x008900
    R136 0x008800
    R135 0x008700
    R134 0x008600
    R133 0x008527
    R132 0x008410
    R131 0x008300
    R130 0x008200
    R129 0x008100
    R128 0x008000
    R127 0x007F00
    R126 0x007E27
    R125 0x007D10
    R124 0x007C00
    R123 0x007B00
    XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

  • Hi Noel,

    I've loaded your exact config on the EVM, tried using another power supply (with a current limit of 1 A in order to give the component even more room), verified the 50 ohm terminator... I tried removing most variables out of the problem and got the following output :

    And even with all of this, I still see the spurs on the output, whenever the ramp changes direction. I'll experiment more with the component, and I'll update this post if I find a way to get rid of them. Thanks for your time and effort.

  • Hi Nicolas,

    Could you try this? 

    I added a 5us no-frequency-change ramp before ramping down.

    4657.HexRegisterValues.txt
    Fullscreen
    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    R141 0x008D00
    R140 0x008C27
    R139 0x008B10
    R138 0x008A00
    R137 0x008900
    R136 0x008800
    R135 0x008700
    R134 0x008600
    R133 0x008527
    R132 0x008410
    R131 0x008300
    R130 0x008200
    R129 0x008100
    R128 0x008000
    R127 0x007F00
    R126 0x007E27
    R125 0x007D10
    R124 0x007C00
    R123 0x007B00
    XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX