Signal Integrity and PCB CN0566 Rev. C

Hi DenisHW,

We're currently taking a look on the paper you sent and discussing it with the pcb engineer. For the mean time, can I ask what application y have in mind for the CN0566?

Regards,

Ivan

Hi Ivan. Thank you for you reply. We use the board to do some comparative radar cross section experiments for some objects. We use metal sphere as reference and compare with other objects response.

Thank you Denis!  Very interesting!  Do you have any layout suggestions to resolve the issue in Figure 8/9?  I guess just pull back that plane further?  

Hi Jon. Yes, as I can see the AGND shape on L2 can be just copy of that one from L3 where no breaks and holes are seen. But we should also check track widths on L1 top layer. As I can see from PCB stack-up, if we use L2 now as reference, L1-L2 height is 6.6 mil now instead of 10.5 mil before, and we need to adjust PCB track width to 0.35 mm to have 50 Ohm wave impedance. Let's also wait for PCB designer review. Thank you!

Hi DenisHW ,

We are currently investigating the issue you reported. May I ask if you have any actual test results for the 15–20 dB signal loss observed in your simulation?

Regards,

Ivan

Hello Ivan. Thank you for your reply. In our project I created radar equation and found that the measured radar range is shorter than it was calculated using radar equation. I found that the noise floor corresponds to theoretical level but the reflected signal is lower than expected. Then, to exclude the TX path, I used HB100 (3 of them) and this test can be done using examples for the phaser where we measure beam steering. Placing HB100 at the distance 1.5 m for example, the final dBFS is ~15 dB less than calculated. For calculation I used detailed signal path: TX power of HB 100, power density at RX antennas, RX antenna gain for 1 x 8 array, LNA ADL8107 gain, ADAR1000 gain, Mixer gain, Pluto RX gain and ADC voltage to code. Then we also visited accredited EMC lab, where still loss were confirmed. Also at EMC lab HB100 were measured and proved to have TX level very close to datasheet specification. After that, I decided to play on CN0566 PCB analysis and signal integrity simulation. I could make mistake somewhere, so I wait for your PCB engineer's feedback. We prepared some publication about our project but it is not approved yet internally.

Thank you!

Hi DenisHW ,

May ask for the board details of the EVAL-CN0566-RPIZ you are using? 

Also, can I ask the procedure you did in getting the actual test results? I want replicate the actual test on my end.

Thanks,

IVan

Hi Ivan. We use CN0566 board. For the problem I cncern, it can operate only in RX mode and example code like phaser_gui can be used as on the picture above. 

Place the HB100 TX module on a distance 1.5 m from a CN566 and then try to estimate the theoretical dBFS and compare with measured. We know TX power of HB100 as 13 dBm EIRP so we know the Poynting vector value at RX antenna. We can estimate the antenna gain for single patch antenna as 6.8dB (Balanis book). Then we sum the power from 32 channels. Then we have LNAs (22 dB), ADAR1000 gain, mixer LTC5548 (-12dB), Pluto (RX gain for example 1 as default setting). ADC of AD936 has 1.25 V reference (for peak to peak). HB100 at 1.5 m should give around 0..-4dBFS but we measure much less. I can try to create excel sheet with my calculation but please try to do this independently. Thank you!

Hi Ivan. We use CN0566 board. For the problem I cncern, it can operate only in RX mode and example code like phaser_gui can be used as on the picture above. 

Place the HB100 TX module on a distance 1.5 m from a CN566 and then try to estimate the theoretical dBFS and compare with measured. We know TX power of HB100 as 13 dBm EIRP so we know the Poynting vector value at RX antenna. We can estimate the antenna gain for single patch antenna as 6.8dB (Balanis book). Then we sum the power from 32 channels. Then we have LNAs (22 dB), ADAR1000 gain, mixer LTC5548 (-12dB), Pluto (RX gain for example 1 as default setting). ADC of AD936 has 1.25 V reference (for peak to peak). HB100 at 1.5 m should give around 0..-4dBFS but we measure much less. I can try to create excel sheet with my calculation but please try to do this independently. Thank you!