IWRL6432: chirpTimingCfg parameter and chirp idle times

Guido Gatti

Part Number: IWRL6432

Hello TI team,

I'm checking the configuration of the IWRL6432 and have some doubts about the chirp timing structure.

Starting from the Parameter Tuning and Customization Guide for the IWRL6432 Motion/Presence Detection Demo v2.0, in section 2.3.1, Figure 8 shows:

This picture shows a typical timing diagram for two chirp loops in the case of two Tx antennas.

It can be inferred that the sequence is as follows:

ramp time of the first chirp
idle time 2
ramp time of the second chirp
idle time 1
ramp time of the first chirp
idle time 2
ramp time of the second chirp
...

This is confirmed by section 4.1, which says:

At this point, I have two questions:

Q1: Is Idle Time 2 the chirpIdleTime parameter of the chirpTimingCfg command? I think so because section 3.1.2 states that chirpIdleTime "indicates the idle time in the chirp cycle before starting of the frequency ramp".

Q2: Does Idle Time 2 occur before each ramp, or only between two chirps of the same loop? The definition of chirpIdleTime ("before starting of the frequency ramp ") makes me think the former, but figure 8 above only shows Idle Time 1 before the first ramp.

At this point, whether chirpIdleTime is Idle Time 2 or Idle Time 1, a configuration parameter to retrieve the chirp repetition period is missing. In fact, all we have from the configuration commands are the ramp timing (chirpRampEndTime) and one of the two idle times.

Also at the burst level, the xWRL6432 Low Power Radar - Power Optimization Techniques application note (SWRA754) states in section 3:

which does not consider two idle contributions at the chirp level, but only one in the chirpIdleTime term.

So, here comes my third question:

Q3: Does the chirp timing have two idle terms as described in the IWRL6432 tuning and customization guide? And if so, how are they defined? And why are not considered in the interburstIdleTime computation?

Thanks,

Guido

11 months ago

0 Santosh Krishnan 11 months ago

TI__Genius 13690 points

Hello.

Guido Gatti said:
Q1: Is Idle Time 2 the chirpIdleTime parameter of the chirpTimingCfg command? I think so because section 3.1.2 states that chirpIdleTime "indicates the idle time in the chirp cycle before starting of the frequency ramp".

Q2: Does Idle Time 2 occur before each ramp, or only between two chirps of the same loop? The definition of chirpIdleTime ("before starting of the frequency ramp ") makes me think the former, but figure 8 above only shows Idle Time 1 before the first ramp.

You are correct with regards to both of your questions. The idle time referred do in the chirp timing command is regarding the idle time before the start of the frequency ramp, and this happens before every frequency ramp within the same chirp loop. The chirp parameters are defining characteristics for each chirp within a burst, so for every chirp in a burst within the frame, they all have the same characteristics.

Guido Gatti said:
At this point, whether chirpIdleTime is Idle Time 2 or Idle Time 1, a configuration parameter to retrieve the chirp repetition period is missing. In fact, all we have from the configuration commands are the ramp timing (chirpRampEndTime) and one of the two idle times.

I believe this parameter is in the frame configuration as framePeriodicity. Using this information and the other chirp timing information, you can calculate the remaining idle time.

Guido Gatti said:
Q3: Does the chirp timing have two idle terms as described in the IWRL6432 tuning and customization guide? And if so, how are they defined? And why are not considered in the interburstIdleTime computation?

It has two idle terms, but only 1 is explicitly defined. The idle time before the start of each frequency ramp is what is defined in the chirp timing command, while the other one is determined based on the values of all the other timing parameters. I believe the second idle time(idle time 1 in the diagram) is not considered in the interburst idle time computation because it happens after all the chirps have completed, which means all the bursts have completed and therefore has no impact on the inter burst idle time. However, let me confirm this for you and will have an update by tomorrow.

Sincerely,

Santosh

0 Santosh Krishnan 11 months ago in reply to Santosh Krishnan

TI__Genius 13690 points

Hello.

This is just to confirm the answer for Q3.

Sincerely,

Santosh

0 Guido Gatti 11 months ago in reply to Santosh Krishnan

Prodigy 20 points

Hello Santosh,

Thank you for your response. I will now summarize your answers and my conclusions, to avoid possible misunderstandings.

Starting again with Figure 8 of the IWRL6432's Tuning and Customization Guide, which I reproduce here for clarity:

Idle Time 2 can be set through the chirpIdleTime parameter of the chirpTimingCfg command

Idle Time 2 is applied before each frequency ramp of the burst/frame. As a consequence, referring to the image above, I expect to find it also before the second blue ramp, in correspondence with the purple circle

According to your answer, Idle Time 1 is not applied between chirps of the same burst, nor between different bursts, but only at the end of the frame (which makes it somehow related to the Interframe Idle)

I'd like your confirmation on the three points above. In particular, the third bullet is not quite clear to me. If Idle Time 1 is not between the end of the transmission of the second chirp of one cycle and the first chirp of the next cycle

Figure 8 seems to be very misleading, if not wrong (is the Chirp Repetition Period longer than a burst?)
The formula for the chirp repetition period for TX = 2: = (idle1 + idle2 + 2ramp)
would be wrong, because it only considers 2 ramp times, hence 2 chirps.
Also, all the examples of chirp configurations listed in Table 4 of the IWRL6432's Tuning and Customization Guide would be wrong as well, since all the combinations are consistent with the above formula: Chirp Repetition Time (Tr) = idle time for TX1 + idle time for TX2 + 2* Ramp end time
Nevertheless, it's not clear to me how idle time for TX1 and idle time for TX2 can be different in the red circled cases.

I hope you can help me with these doubts.

Thank you,

Guido

0 Santosh Krishnan 11 months ago in reply to Guido Gatti

TI__Genius 13690 points

Hello.

Thank you for your response. Let me look into your questions and provide a response by tomorrow.

Sincerely,

Santosh

0 Santosh Krishnan 11 months ago in reply to Santosh Krishnan

TI__Genius 13690 points

Hello.

Guido Gatti said:
Idle Time 2 can be set through the chirpIdleTime parameter of the chirpTimingCfg command

Idle Time 2 is applied before each frequency ramp of the burst/frame. As a consequence, referring to the image above, I expect to find it also before the second blue ramp, in correspondence with the purple circle

The idle time at the start of every burst is something that can be ignored as there is always some inherent latency before the start of chirping, but an easier way to think about this idle time is actually with respect to the burst periodicity. The burst periodicity will start at the start of the first chirp of the burst and will go up to but not include the start of the first chirp of the next burst. If you shift this a few microseconds to account for the idle time before the start of the first chirp, the burst periodicity will go from the idle time of the first chirp in the burst to the first idle time of the first chirp in the next burst.

Guido Gatti said:
According to your answer, Idle Time 1 is not applied between chirps of the same burst, nor between different bursts, but only at the end of the frame (which makes it somehow related to the Interframe Idle)

I would like to amend my statement from earlier. This time is applied between bursts, rather than frames as I indicated before, and is configured by configuring the burstPeriodicity parameter in the frameCfg.

I believe this should clear up your doubts regarding the chirp repetition time calculations.

Guido Gatti said:
Nevertheless, it's not clear to me how idle time for TX1 and idle time for TX2 can be different in the red circled cases.

There can be some latency at the start of the first chirp in each burst that can be accounting for this difference. But the main focus is that these idle times + the ramping time does not exceed the burst periodicity.

Sincerely,

Santosh

0 Guido Gatti 11 months ago in reply to Santosh Krishnan

Prodigy 20 points

Hello Santosh,

Santosh Krishnan said:
The idle time at the start of every burst is something that can be ignored as there is always some inherent latency before the start of chirping, but an easier way to think about this idle time is actually with respect to the burst periodicity. The burst periodicity will start at the start of the first chirp of the burst and will go up to but not include the start of the first chirp of the next burst. If you shift this a few microseconds to account for the idle time before the start of the first chirp, the burst periodicity will go from the idle time of the first chirp in the burst to the first idle time of the first chirp in the next burst.

Ok, this is clear now, thank you.

Unfortunately, I still have some doubts about the Chirp Repetition Time, however,.

My understanding is that with 2 Tx antennas, the Chirp Repetition Time defines the time between the transmission of the first chirp of a chirp loop and the transmission of the same chirp of the following loop. If this interpretation is correct, the Idle Time 1 would be within two chirp loops. This is also consistent with the meaning given to the Idle Times 1 and 2 in Section 2.3.1 of the IWR6432's Tuning and Customization Guide:

In your explanation, however, you say that the Idle Time 1 is applied between bursts, instead of chirp loops. In the case of a burst consisting of just one chirp loop, the two definitions are somewhat equivalent. But in the case of a burst consisting of more cycles, the meaning of Idle Time 1 is different.

In fact, let's assume a burst made by 2 chirp loops and using 2 Tx TDM antenna transmission. According to your definition, the timing of the burst would be as follows:

This definition would violate the formula given in Section 4.1: Tr = (Tidle1+Tidle2 + 2*Tramp). In fact, we would have Tr = 2*Tramp+2*Tidle2 in the case where Tr refers to a chirp loop, or Tr = 4*Tramp + 4*Tidle2+Tidle1 in the case where Tr refers to the burst.

Also, the idle times listed for Tx1 in Table 4 (see my previous post) for the 10-meter and 15-meter configurations would be incorrect.

At this point, I see only two possibilities:

The above formula and table are wrong
The idle Time 1 is a delay within a chirp loop (and not the burst), but its definition is not possible or left for future releases

Hope you can clarify this last topic.

Thank you,

Guido

0 Santosh Krishnan 11 months ago in reply to Guido Gatti

TI__Genius 13690 points

Hello.

I think where the confusion is rising from is your definition of chirp loop and burst. What you have labeled as a chirp loop is actually a burst, and then an X number of bursts make up a frame. In this example, you are creating two chirps per burst, 1 from each TX. The idle time 2 is defined for in between chirps, but after the second chirp, the chirping for that burst is complete. The remaining time from that point to the end of the burst period is the idleTime 1, which computed rather than configured as all the other parameters are configurable. Then the next burst will start, with the blue chirp, and end with the red chirp followed by idle time 1 again.

Sincerely,

Santosh

0 Guido Gatti 11 months ago in reply to Santosh Krishnan

Prodigy 20 points

Hello Santosh,

From your answer, it seems that a burst can consist of only one chirp per Tx antenna, which would be the same as a chirp loop.

However, this contradicts the IWR6432's Tuning and Customization Guide.

In fact, section 3.1.3 defines the frameCfg command, and its first parameter is numOfChirpsInBurst, described as the number of chirps in a burst, with a valid range of 1 to 65535 chirps.

The same section says:

which also shows that the number of chirps in a burst can be greater than the number of transmitting antennas. Referring to the example in my previous comment, I described a burst with numOfChirpsInBurst=4.

On the contrary, the situation you describe seems to be a special case of the motion/presence detection demo when configured in its Burst Mode.

Do you agree with this view?

And, if so, can you help me with my previous question?

Guido Gatti said:
At this point, I see only two possibilities:

The above formula and table are wrong

The idle Time 1 is a delay within a chirp loop (and not the burst), but its definition is not possible or left for future releases

Thanks,

Guido

0 Santosh Krishnan 11 months ago in reply to Guido Gatti

TI__Genius 13690 points

Hello Guido.

Guido Gatti said:
From your answer, it seems that a burst can consist of only one chirp per Tx antenna, which would be the same as a chirp loop.

I am referring to this example only when I say only one chirp per Tx antenna, as that is what is described in this example of TDM-MIMO. There definitely can be a case where a burst can have more than 2 chirps as you mentioned later in your post. The main idea is that after all the chirping the burst is complete, the idletime1 is executed until the start of the next burst, which is determined by the remaining time left in the burst period after the chirping is complete.

Guido Gatti said:
Burst Mode.

This is not a special mode. You can design each frame to have different number of bursts with different numbers of chirps in each burst, but there will always be bursts of chirps in each frame.

Guido Gatti said:
but its definition is not possible or left for future releases

This part is correct; you cannot define Idle Time 1 as it is dependent on the other parameters I listed in my earlier responses and is calculated accordingly.

Sincerely,

Santosh

+1 Guido Gatti 11 months ago in reply to Santosh Krishnan

Prodigy 20 points

Hello Santosh,

I see your point. Let me try to summarize.

Let's look at two examples, one with a burst of 2 chirps and one with a burst of 4 chirps.

Example 1

2 Tx antennas
TDM Transmission
numOfChirpsInBurst = 2

In this example, congruent with the Burst Mode of the motion/presence detection demo, only one chirp is sent per Tx antenna within a burst. Therefore, the chirp repetition period (i.e.: a chirp loop) is actually equal to the burst periodicity. After each first (blue) chirp there is an Idle Time 2 (defined by chirpIdleTime) and after each second (red) chirp there is an Idle Time 1.

Idle Time 1 can be calculated in two equivalent ways:

According to Formula 1:

we can calculate Idle Time 1 = Chirp Loop Time - (2*RampEndTime + Idle Time 2)
According to Formula 2:

we can calculate Idle Time 1 = InterburstIdleTime + Idle Time 2 = Burst Periodicity - 2*(RampEndTime + Idle Time 2) + Idle Time 2 = Chirp Loop Time - (2*RampEndTime + Idle Time 2)

Formula 1 and Formula 2 are congruent, so Idle Time 1 is clearly defined.

Example 2

2 Tx antennas
TDM Transmission
numOfChirpsInBurst = 4

In this example, congruent with the Normal Mode of the motion/presence detection demo, two chirps are sent per Tx antenna at each burst. In this case, the chirp repetition period is different from the burst periodicity. Each chirp is separated from the next by Idle Time 2 and, at the end of the burst, there is Idle Time 1.

Formula 2 can still be used to compute Idle Time 1:

Idle Time 1 = InterburstIdleTime + Idle Time 2 = Burst Periodicity - 4*(RampEndTime + Idle Time 2) + Idle Time 2 = Burst Periodicity - (4*RampEndTime + 3*Idle Time 2)

But Formula 1 doesn't fit because we have more than two chirps.

Can we conclude that the formula for the Chirp Repetition Period (Formula 1) given in section 4.1 of the IWR6432's Tuning and Customization Guide is valid only for the specific case of a burst consisting of only one chirp per Tx antenna?

Thank you,

Guido

0 Santosh Krishnan 10 months ago in reply to Guido Gatti

TI__Genius 13690 points

Hello Guido.

Formula 1 is specific to the example shown in that figure. The inter-burst idle time formula is more relevant to generic applications as it factors in the variability of certain parameters.

Sincerely,

Santosh

0 Guido Gatti 10 months ago in reply to Santosh Krishnan

Prodigy 20 points

Ok, the topic is clearer now.

Thank you,
Guido

Because of the holidays, TI E2E™ design support forum responses will be delayed from Dec. 25 through Jan. 2. Thank you for your patience.

Sensors

Sensors forum

IWRL6432: chirpTimingCfg parameter and chirp idle times