On 07/11/2019 03:53 PM, lc wrote:
Thanks for all the response.
I should make my question more specific instead of using raw adc data
which is causing confusion.
The data I want to dump is (name based on b200.v)
rx_codec_d[11:0] which is directly from ADC
or
tx_data0[31:20]
tx_data0[15:4]
which are processed after b200_io
But I don't want FPGA to further touch the data.
From Ian's suggestion, seems I need to
- configure the USRP so that sample_rate=master_clock_rate
- use a tuning policy that doesn’t imply the internal CORDIC
processor in the FPGA to perform any portion of the tuning.
So my question is do I need to hack UHD in order to achieve item 2?
Or current matlab API could do this configuration.
Thanks
Chao
On 07/11/2019 03:53 PM, lc wrote:
> Thanks for all the response.
> I should make my question more specific instead of using raw adc data
> which is causing confusion.
> The data I want to dump is (name based on b200.v)
> rx_codec_d[11:0] which is directly from ADC
> or
> tx_data0[31:20]
> tx_data0[15:4]
> which are processed after b200_io
>
> But I don't want FPGA to further touch the data.
> From Ian's suggestion, seems I need to
> 1. configure the USRP so that sample_rate=master_clock_rate
> 2. use a tuning policy that doesn’t imply the internal CORDIC
> processor in the FPGA to perform any portion of the tuning.
>
> So my question is do I need to hack UHD in order to achieve item 2?
> Or current matlab API could do this configuration.
>
> Thanks
> Chao
>
>
Item 2 above can be achieved with the normal tuning API in UHD.
https://files.ettus.com/manual/classuhd_1_1usrp_1_1multi__usrp.html#a9b61448f392466e20572fdcb042e8ec6
and
https://files.ettus.com/manual/structuhd_1_1tune__request__t.html