IWR1843AOP: Number of ADC samples and Bandwidth of LPF

Shoko Sakaguchi

Intellectual 560 points

Part Number: IWR1843AOP

Tool/software:

Hello,

I'm studying IWR1843AOP (It's working SDK 3.6).

1) If I change Number of ADC samples from 256 to 512, does the bandwidth of LPF change automatically?

Because Number of ADC samples is large, period of FFT will also be large, and as a result frequency resolution will be small.

2) If 1) is yes, is this LPF bandwidth applied not only to the SNR calculation but also to the noise power calculation?

3) If I use the following setting, radar cube size will be over shared L3 memory of IWR1843AOP.

What is expected to happen in this case?

Complex1x mode (this means 4 bytes per sample; 16 bits for the real part and 16 bits for the imaginary part)
Number of ADC samples: 512
Number of TX antennas: 3 (in TDM MIMO configuration)
Number of RX antennas: 4
Number of chirps per frame: 48

Radar_Cube_Size = 4 bytes * 3 * 4 * 512 * 48 = 1180KB

(The shared L3 memory of IWR1843AOP has a size of 1024 KB)

e2e.ti.com/.../3722508

Please advise.

Best regards,

Shoko

27 days ago

0 Saransh Gautam 26 days ago

TI__Prodigy 400 points

Hello!
We are looking into your query. Please allow us some time to respond.
Thanks.

0 Santosh Krishnan 26 days ago in reply to Saransh Gautam

TI__Mastermind 19450 points

Hello.

You probably will crash the demo or the DPUs won't configure properly since it won't have enough memory to allocate the L3 memory to. If it does, you may run into hard faults if you try writing or reading from memory that you don't have access to.

Sincerely,

Santosh

0 Shoko Sakaguchi 26 days ago in reply to Santosh Krishnan

Intellectual 560 points

Hello,

Thank you for your reply about question 3).

Could you please answer about question 1) and 2)?

Thanks,

Shoko

0 Santosh Krishnan 25 days ago in reply to Shoko Sakaguchi

TI__Mastermind 19450 points

Hello.

Apologies I did not see those questions. Let me look into them and provide an update by the end of the week.

Sincerely,

Santosh

0 Shoko Sakaguchi 21 days ago in reply to Santosh Krishnan

Intellectual 560 points

Hello,

Do you have any update about question 1) and 2)?

Please reply.

Regards,

Shoko

0 Shoko Sakaguchi 19 days ago in reply to Shoko Sakaguchi

Intellectual 560 points

Hello,

Do you have any update about question 1) and 2)?

Please reply.

Regards,

Shoko

0 Santosh Krishnan 18 days ago in reply to Shoko Sakaguchi

TI__Mastermind 19450 points

Hello Shoko.

Apologies for the delay; I am still looking into this. I will provide an update by the end of the day tomorrow. Thank you for your patience.

Sincerely,

Santosh

0 Shoko Sakaguchi 14 days ago in reply to Santosh Krishnan

Intellectual 560 points

Hello,

Do you have any update about question 1) and 2)?

Please reply ASAP.

Thanks,

Shoko

0 Santosh Krishnan 12 days ago in reply to Shoko Sakaguchi

TI__Mastermind 19450 points

Hello Shoko.

Shoko Sakaguchi said:
1) If I change Number of ADC samples from 256 to 512, does the bandwidth of LPF change automatically?

Because Number of ADC samples is large, period of FFT will also be large, and as a result frequency resolution will be small.

I believe this is done automatically. Looking through the ICD, I didn't see anything that would allow you to configure this value, so this should be the case but I will confirm this by tomorrow.

Shoko Sakaguchi said:
2) If 1) is yes, is this LPF bandwidth applied not only to the SNR calculation but also to the noise power calculation?

No; the SNR calculation is done by looking at the peak value from the range doppler heatmap and the cell noise computed as part of the CFAR calculation.

Sincerely,

Santosh

0 Shoko Sakaguchi 12 days ago in reply to Santosh Krishnan

Intellectual 560 points

Hello Santosh,

Thank you for your response and research.

> I believe this is done automatically. Looking through the ICD, I didn't see anything that would allow you to

> configure this value, so this should be the case but I will confirm this by tomorrow.

I'm waiting for your response. Thanks.

> No; the SNR calculation is done by looking at the peak value from the range doppler heatmap and the cell noise

> computed as part of the CFAR calculation.

Can I ask again?

Does the LPF bandwidth, which will automatically change depending on the number of ADC samples, apply to both the peak value of renge-doppler heat map and cell noise computed as part of the CFAR calculation?

Regards,

Shoko

+1 Santosh Krishnan 11 days ago in reply to Shoko Sakaguchi

TI__Mastermind 19450 points

Hello Shoko.

Shoko Sakaguchi said:
I'm waiting for your response. Thanks.

I have confirmed that this is correct. It is automatically set based on # of adc samples

Shoko Sakaguchi said:
Does the LPF bandwidth, which will automatically change depending on the number of ADC samples, apply to both the peak value of renge-doppler heat map and cell noise computed as part of the CFAR calculation?

It does not. It removes the higher frequencies from the ADC data but is not directly used in the SNR computation, and doesn't affect the SNR of the lower frequencies.

Sincerely,

Santosh

Sensors

Sensors forum

IWR1843AOP: Number of ADC samples and Bandwidth of LPF