You are not logged in.
Pages: 1
Hi,
I was trying to find a way to implement 40 MHz transmissions in the WARP 802.11 Reference Design for WARP v3. I have explored the forum seeking for related issues but unfortunately, I am still confused. I really appreciate the effort you put on answering such amount of topics, so excuse me beforehand if any of my questions were already answered.
Right now, I understand that the operative channels in the 802.11 Reference Design at 5 GHz are 36, 40, 44 and 48 20-MHz channels. That is, 40-MHz operation is not implemented yet.
As for 40 MHz operation in 802.11n, the following issues come to my mind. Let us assume a transmission in the 38 channel (center frequency 5.190 GHz and 40 MHz bandwidth).
- 1. Data packets would be transmitted in the 38 40-MHz channel. I assume this should be the easiest part since I already managed to transmit 40 MHz signals using WARPLab. However, I am not sure what modifications are needed for the 802.11 Reference Design.
- 2. In 802.11n, control packets like the RTS, CTS or ACK should be duplicated in the 20-MHz channels 36 and 40 when transmitting at channel 38. Would it be possible to implement?
- 3. Would that be possible to extend the operative 5 GHz band channels to ranging from 36 to 64, including the 4 40-MHz in between?
Thanks so much for your help.
Offline
The current 802.11 Reference Design for WARP v3 (v1.7.8) does not implement the standard 40MHz waveforms (known as HT40 (11n) and VHT40 (11ac)). The ref design does implement a non-standard 40MHz-bandwidth waveform realized by double-clocking the baseband and ADC/DAC. This 40MHz waveform is still 64 subcarriers vs 128 for the HT40/VHT40 waveforms.
As you noted, a PHY cannot be HT40-only, as control packets must be transmitted as 20MHz waveforms duplicated (NONHTDUP, we call it) in the upper and lower halves of the 40MHz channel.
Adding true HT40 and NONHTDUP support would require redesigning large parts of the Tx and Rx PHY models. This is a big project.
- 3. Would that be possible to extend the operative 5 GHz band channels to ranging from 36 to 64, including the 4 40-MHz in between?
The MAX2829 transceivers on WARP v3 support the full 5GHz band. You would need to modify the C code to tune the radio to other channels. The mapping of channel index to radio center frequency index is in w3_wlan_chan_to_rc_chan(). The radio center frequency index is an arbitrary index used by the radio_controller peripheral. The radio_controller supports many more frequency indexes than the default channel list.
Offline
Hi murphpo,
Thank you for the prompt response. I really appreciate that. I have one more question.
The current 802.11 Reference Design for WARP v3 (v1.7.8) does not implement the standard 40MHz waveforms (known as HT40 (11n) and VHT40 (11ac)). The ref design does implement a non-standard 40MHz-bandwidth waveform realized by double-clocking the baseband and ADC/DAC. This 40MHz waveform is still 64 subcarriers vs 128 for the HT40/VHT40 waveforms.
I understand that even though the standard 40MHz waveforms are not implemented in the ref design, it is doable. Essentially, I need to build a 128 subcarriers waveform. I assume that the FPGA core should be modified, i.e., it is not possible to do so through C code. Is that correct? If so, what should I modify at the TXer and RXer sides?
Adding true HT40 and NONHTDUP support would require redesigning large parts of the Tx and Rx PHY models. This is a big project.
Ok.
The MAX2829 transceivers on WARP v3 support the full 5GHz band. You would need to modify the C code to tune the radio to other channels. The mapping of channel index to radio center frequency index is in w3_wlan_chan_to_rc_chan(). The radio center frequency index is an arbitrary index used by the radio_controller peripheral. The radio_controller supports many more frequency indexes than the default channel list.
Great.
Thanks again.
Offline
I understand that even though the standard 40MHz waveforms are not implemented in the ref design, it is doable. Essentially, I need to build a 128 subcarriers waveform. I assume that the FPGA core should be modified, i.e., it is not possible to do so through C code. Is that correct? If so, what should I modify at the TXer and RXer sides?
Correct, you would need to make extensive modifications to the Tx and Rx FPGA cores (the wlan_phy_tx/rx System Generator models). This would require significant design effort, I cannot provide a guide via the forums. You would need to become very familiar with the existing PHY cores. Almost all subsystems would require modifications to support 128 subcarriers.
Offline
Hi murphpo, I will consider other options then.
Thanks again for your help.
Offline
Pages: 1