• State TI Thinks Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 2 answers
  • Subscribers 21 subscribers
  • Views 1277 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:

LMX2595: LMX2595 in ramp mode

LMX2595: Ramp mode in LMX2595

Varaprasad Rayudu
Prodigy 150 points
Part Number: LMX2595
Other Parts Discussed in Thread: LMX2820

Hello,

We are planning to use LMX2595 part# to generate the linear ramp sweep for frequency range of 14.15 GHz to 14.45 GHz with 80 MHz Ref and 160 MHz Fpd frequency.

1) Can we cover the 300MHz BW in ramp mode with out auto calibration enabled and which are the parameter will decide the ramp bandwidth and is there any relationship between loopbw & Fpd & ramp bandwidth.

2) If the Ramp mode is not supporting for full 300MHz, then can we send cal register setting values in fully assist mode after each 100MHz ramp BW

 if this possible, please suggest the method of performing calibration after each pre-defined ramp size.

 Also please confirm the tuning speed in the ramp mode to reach from start to end of the ramp frequencies and how to calculate the tuning speed.

3) As per the application requirements, tuning speed is very critical and need to cover the full ramp of 300MHz BW in 3-5usec with frequency step size of 100kHz.

Please suggest the suitable solution to meet the tuning speed of <3usec.

Regards,

RVP

  • 0
    TI__Guru** 107290 points

    Hi RVP,

    I am afraid LMX2595 cannot meet your sweep requirement. 

    (1) 300MHz linear ramp is not possible because the VCO tuning range without calibration is less than 300MHz.

    (2) Assume 100MHz linear ramp is robust, however, you have to finish a 100MHz ramp, exit ramp mode, run full-assist to "re-calibrate" VCO, enable ramp mode and start ramping again. All of these operations take time. Furthermore, you have to figure out how to manually control the timing of these operations.

    (3) 3-5µs ramp time for 300MHz sweep is not realistic. With freq step size of 100kHz and 3µs sweep time, that translates to 1GHz fpd, which is not possible for any PLL.

  • 0 in reply to Noel Fung
    Prodigy 150 points

    Dear Noel,

    1) What will be maximum linear ramp bandwidth possible with LMX2595 part# without calibration?

    2) What will the best possible sweep-time with fpd of 160MHz.(for example to cover 100MHz sweep)?

    3) Do we have any possibility of  knowing ramp completion by polling any of the register value, do run the full-assist mode to re-calibrate the VCO and during this do we see any non-linearity or frequency jump in ramp mode.

    4) Any improvement in sweep time by using LMX2820 in complete full-assist mode to load the registers with continuous SPI speed of 50MHz.

    5) What is the calculation between fpd and sweep time?

    Regards,

    RVP

  • 0 in reply to Varaprasad Rayudu
    TI__Guru** 107290 points

    Hi RVP,

    1. I would say around 100MHz should be doable.

    2. The ramp engine will increment frequency once every fpd cycle. Given your step size and ramp range, you can calculate the time taken to complete the ramp.

    3. No, there is no flag out for ramp completion.

    4. to program a register requires 24+1 SPI cycles. At the max. of 50MHz SPI clock, program one register takes 0.5µs. Your system requires at most 5µs sweep time. You could only make 10 register writes. Continuous SPI write to sweep frequency is also not feasible.

    5. answer in 2. above.