5000iP SerHello Erik
Good to hear that you have used 1024 QAM over a 65 km path in Western Norway, using the 6.7 GHz frequency band.
We are also looking to deploy Higher capacity Microwave Link for long distance above 65 Km using higher order modulation like 1024 QAM.
Could you please share me some technical parametrs ( like TX power, antenna type, antenna dimension, antenna gain, Channel Bandwidth, Modulation,Rx level,Rx sensitivity, Fade margin,Channels) with the link budget calculation used. Which Company Microwave radio you are using for such link.
Hope you information would be very helpful to me and my companyies
Hi, indeed 1024QAM pose huge challenges in several areas. The peak to average ratio depend strongly on how the 1024QAM constellation is implemented. In the radio described in this case the peak to average for 1024QAM is really not much different to 256QAM. It does require Tx linearization and sufficient power backoff, but compared to 256QAM it’s really not very different. Long Haul microwave radios operate typically with 28-30, 40 or 56 MHz carrier bandwidth (varying with different frequency bands). With such bandwidth, it is indeed susceptible to multipath fading. The debate of OFDM vs single carrier, is not clear cut. OFDM offers both pros and cons in this area. OFDM is potentially offering better robustness to selective (multipath) fading but requires more signal processing and secondly the OFDM signal could offer better spectrum utilisation (due to the more “square” spectrum) but spectrum masks issued by the regulatory bodies (FCC, ETSI) are optimised to single carrier modulation.
1024 QAM presents a huge challenge to both tx and rx. On the ttx side, 1024 QAM has a much higher peak to average power ratio which means a big issue to the non-linear power amp. Would that be a compromise between the linearisation technique and the power backoff? On the RX side, it is susceptible to multiopath fading, which can imply a very expensive adaptive equaliser to compensate it. Given the maturity of technologies such as OFDM or SC FDE, should the system be considered with one of these advanced technologies?
Thanks for the interesting article. I'd like to know more about the additional complexity and cost needed to increase the throughput. It seems like you will definitely need more linear components, lower phase noise and maybe more DSP. Does the extra cost scale with the extra throughput? Or does it cost more than 1.25x to get 1.25x data throughput?
I guess it's important to understand that we are talking about carrier grade long haul (transport) links here. These links are dimensioned/engineered for very high availability, typically in the range of 99.995 or 99.999 % availability. This means less than half hour of outage per year! In order to achieve these figures the links have ample fading margin; typically in the range of 30 to 40 dB. Increasing modulation to 1024QAM reduces fading margin with 6dB compared to 256QAM. The effect on availability is just a factor of about 2 (doubling the outage), considering flat fading and interference free conditions. It has very limited effect on distance, as these links are not operating at the maximum possible distance anyway (due to needed fading margin). Interference can be an issue, again the sensitivity is increased by 6dB, so it needs to be factored into the availability calculations. Higher lever modulation does not change power consumption or other factors. So bottom line; as it uses adaptive modulation it offers 25% more capacity for 99.99% of the time. For the 0.01% of the time the links will scale down to lower modulation.
We are running a trial link over a 65 km path in Western Norway, using the 6.7 GHz frequency band. This has been running so far for about 2 months without even once changing to lower modulation. I hope this has helped clarify the technology.
It seems to me a waste of hardware upgrade. Performance requirements such as SNR for 1024-QAM is so high that it will appear to be defensless against channel fading, not to mention interferers. With adaptive modulation, it turns out to be running as 256-QAM at best as before. But it can be a good fit to wired channels such as cable, etc.
Blog Doing Math in FPGAs Tom Burke 3 comments For a recent project, I explored doing "real" (that is, non-integer) math on a Spartan 3 FPGA. FPGAs, by their nature, do integer math. That is, there's no floating-point ...