• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Subscribers 29 subscribers
  • Views 883 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.

IWR6843ISK-ODS: Understanding of theoretical "chirp time" vs timing parameters within CFG file

Victoria Rudakova
Intellectual 770 points
Part Number: IWR6843ISK-ODS

Hello,

After going through the introduction videos for the mmWave and radar technology, I learned that chirp time (T_chirp) plays an important role for velocity calculations. For example, to calculate the velocity resolution:

V_res = lambda / (2 * N * T_chirp)

Looking at the CFG config file, we have different timing parameters like: idle time, ADC start time and ramp end time that seem to define the aforementioned T_chirp. 

When looking at the theoretical level of how the velocities are calculated, there is no idle time or other hardware-specific parameters to consider, so  I am trying to figure out which of the CFG parameters define the theoretical T_chirp and how they affect the velocities calculation:

1. Would it be correct to assume the chirp time is the time when the ADC sampling is performed, e.g.: T_chirp = <ramp end time>? Or is it the summation of <idle time> + <ramp end time>? 

2. How does idle time reflect onto theoretical formula for calculations of velocities? Does it play any role?

3. What about radar equation which is directly proportional to "T_meas" (measurement time) - is that also <T_chirp * N> and does it include idle time? - e.g., in order to improve SNR we could increase idle time? Or it only includes ramp end time and we have to work with that value directly?

4. Could you also help to understand the term "active chirp time" defined in mmWave SDK user guide: "active chirp time should be <= 50% of frame period"? Active means time during sampling? Is it correct to assume the frame time is given by T_frame = N_chirp * T_chirp - where T_chirp also includes idle time? 

  • Former Member
    0 Former Member

    Hello Victoria,

    I would recommend you refer to the calculateChirpParams.m file in the C:\ti\mmwave_industrial_toolbox_4_2_0\labs\area_scanner_orig\68xx_area_scanner\gui

    It will show you how we parse the profileCfg, chirp, and frame cfg parameters to calculate active chirp time, velocity etc.

    Amanda

  • 0 in reply to Former Member
    Intellectual 770 points

    Hello Amanda,

    Thank you for pointing out to that document, it helps with some of my questions (1, 2). 

    However, it is still not clear to me:

    * Q3 - about the radar equation and what does "T_measurement" mean. This is not covered in the script. From what I have seen in the script the chirp time means the summation of ramp end time and idle time. Would it still hold here? Does it mean one can improve the SNR by merely increasing idle time?

    * Q4 - about the "active chirp time": while the script provides the formula for its calculation, there is also a comment (Line 81):

    "this doesn't account for the low power mode where if idleTime is > 10usec the power amplifiers are turned off"

    It would be great to get some clarifications on what happens if the idleTime is much larger than 10usec and how it is reflected on the formula for the "activeFrameTime_msec"? Would the same formula still work?

  • Former Member
    0 Former Member in reply to Victoria Rudakova

    Q3 - Could you clarify what you mean by T_measurement or where you are sourcing this parameter from? But no increasing idle time will not help SNR -  if you look at the chirp diagram nothing is occurring during the idle time. Idle time is so the device can settle and if needed complete the digital processing. To increase SNR you generally want to increase the overall illumination (ex: more simultaneous transmits) or reduce the noise (ex more chirps)

    Q4 - Active chirp time is what we use to calculate the duty cycle for thermal considerations. If the idle time is >10us then analog components are considered duty cycled. It has no bearing on your calculation for max velocity. However if you want to accurately calculate duty cycle you should use:

    if((P.profileCfg.idleTime(i)<10)

    calc_P(i).activeFrameTime_msec = numChirps*(P.profileCfg.idleTime(i)+P.profileCfg.rampEndTime(i))*numLoops*usec2msec; numTx

    else

    calc_P(i).activeFrameTime_msec = numChirps*(P.profileCfg.rampEndTime(i))*numLoops*usec2msec; 

    end

    Amanda

  • 0 in reply to Former Member
    Intellectual 770 points

    Thanks for further clarifications, Amanda.

    Re: Q3 I refer to the document https://www.ti.com/lit/an/swra553a/swra553a.pdf "Programming Chirp Parameters in TI Radar Devices" - p. 3 has formula (2) that describes Range max based on SNR. There, a certain T_r notation is used to describe "Chirp time". There was the same formula in other document and it was noted as T_meas - measurement time and I was confused about it  since it was not clear what it meant; given the chirp time being depended on both idle and ramp time from the Matlab script, I wrongly assumed we can increase SNR based on increasing of either idle or ramp time. 

    Could you explain what the "Chirp time" means in the context of that SNR formula?

    >  To increase SNR you generally want to increase the overall illumination (ex: more simultaneous transmits) or reduce the noise (ex more chirps) 

    Could you please clarify what does "simultaneous transmits" mean? Like using more transmit antennas?

    "more chirps" - means to increase the number of loops within one frame?

  • Former Member
    0 Former Member in reply to Victoria Rudakova

    You should think of T_meas or T_c as the ADC sampling time (which is a subset of the ramp end time. Usually you have some delay in the adc start time for the integrity of the data and they you may or may not have some chirping after you are done sampling. But this excess ramp has no benefit since you aren't sampling) - you can refer to the .m file previously for the definition there.

    You need the ramp end time to be more than the adc start time + the adc sampling time. Whatever else is leftover is considered excess ramp time.

    You can can change the transmit power (backoff) per transmitter which would reduce SNR or you can enable multiple transmitters at once to increase overall transmit power. And yes more chirp loops or  defining more chirps would overall give you more total chirps per frame. This is usually bounded by timing or memory availability though.

    Amanda

  • 0 in reply to Former Member
    Intellectual 770 points

    Thanks a lot, that was helpful.