The American telecom operator Sprint is keen on mentioning Massive MIMO in the marketing of its 5G network deployments, as I wrote about a year ago. You can find their new video below and it gives new insights into the deployment strategy of their new 64-antenna BSs. Initially, the base station will be divided between LTE and 5G operation. According to CTO Dr. John Saw, the left half of the array will be used for LTE and the right half for 5G. This will lead to a 3 dB loss in SNR and also a reduced multiplexing capability, but I suppose that Sprint is only doing this temporarily until the number of 5G users is sufficiently large to motivate a 5G-only base station. Another thing that one can infer from the video is that the LTE/5G splitting is software-defined so physical changes to the base station hardware are not needed to change it.
Thanks for the new post.
I’m not 100% sure, but I just want to mention the point.
The RF signals in the left side (4G) and the right side (5G) should be synchronized to be able to do beamforming for the future massive MIMO. I don’t think that all the RFs are synchronized in the current system. Am I right…?
Then, this scaling can not easily be done with software defined radio, there should be more works (like clock synchronization).
I don’t have any other details than those provided in the video, but I would guess that the 4G and 5G parts will operate on different frequencies so there is no need for synchronization between them. A more “beautiful” solution would be to implement the base station with carrier aggregation of one 4G carrier and another 5G carrier. Then the entire array could be used for both systems.
Thanks 🙂
Thanks Professor Dr. Emil Björnson for this useful information. Any feedback on the Massive MIMO area, the upcoming Fifth Generation of Mobile Communication Networks (5G) as well as any practical engineering applications in these and other areas, are always welcome. On this post regarding the deployment by the Sprint Company of a Base Station equipped with 64 antenna elements, do you have any additional information on whether or not hybrid-analog digital beamforming was implemented in the design? Thanks in advance.
Thank you!
I don’t have any more information from Sprint, but I would be surprised if hybrid beamforming was used. There are already several 64-antenna LTE products with full digital beamforming on the market. One example is this product from Ericsson that is designed for the 2.5 or 3.5 GHz bands: http://ma-mimo.ellintech.se/2018/08/27/a-look-at-an-lte-tdd-massive-mimo-product/
Hybrid beamforming is a way to reach faster to the market with products at mmWave frequencies, but eventually digital beamforming will take over also at these frequencies. I have already seen examples of full digital beamforming implementations in the 28 GHz band.
Thanks Professor for that relevant information, which clarifies some doubts regarding the deployment of hybrid analog-beamforming in future Massive MIMO implementations. Taking into consideration the cost of deploying a Massive MIMO BS with a large number of antenna elements, it would imply using analog-digital beamforming as an alternative to face this problem, which becomes even sensitive at millimeter-waves frequencies, since the size, power consumption and other factors also take place. However, this approach (analog-digital beamforming) would induce other critical challenges in practical engineering applications of Massive MIMO, which make it preferable to use fully digital beamforming. By the way, I read the recommended article in your last comment and it is worth mentioning that, as you Professor pointed out in that article, this separate design between the baseband processing and the AIR 6468 unit offers a very flexible alternative. On this latter, and also I guess this question can be extended to the Massive MIMO product addressed in this post, do you have any information on how the frequency resources are allocated by these existing products? I guess a way of implementing this (just an idea) would be by grouping the users in such a way that the number of users per group be less than 64 (taking into consideration that there are deployed 64 antenna elements in the LTE-TDD Massive MIMO BS) and allocating to each group a 20 MHz (or even more taking into consideration carrier aggregation) channel bandwidth, since TDD Massive MIMO exploit the entire allocated bandwidth. Thanks in advance, Professor.
You can spatially multiplex 64 users with the 64-antenna array, but it is probably better so serve a smaller number of users at the time. I would guess that 5-10 users is what will give you the highest throughput when using a planar array of this size. Fortunately, it will take years before you have more users than that requesting data at the same millisecond.
Thanks for your valuable remarks, Prof. Björnson. I’m looking forward to more interesting posts.
Thanks Professor Dr. Emil Björnson for sharing.
Another scenario could be to align LTE frame structure with 5G to operate both on the same band. LTE and 5G inter-working is the initial approach towards 5G.
Please correct me if my understanding is not correct.