Power Consumption: 5G Basestations Are Hungry, Hungry Hippos

Dan Jones
3/6/2019

The increased power consumption of next-generation basestations may be one of the dirty little secrets of 5G, which might not be a secret much longer as operators roll out initial networks.

The power consumption of a 5G basestation is three times that of its 4G LTE predecessor, according to Zhengmao Li, EVP at China Mobile, at a presentation at Mobile World Congress last week, as noted by my Heavy Reading Colleague, Gabriel Brown:

5G basestations bump up the power requirements over 4G LTE, in part because of the massive antenna arrays (MIMO) used for the next generation tech. Earl Lum, president of EJL Wireless Research, says that MIMO increases the "power amplifiers" and "analog-to-digital paths" required, as well as overall digital circuitry in the units. For instance, typical 4G basestations now use 4 transmitter and 4 receiver (4T4R) elements, while 5G is expected to use 64T64R MIMO arrays.

"Sprint stuff is 64T64R," Lum notes. (See Sprint's Initial 5G Cities Will Get 4G Massive MIMO First.)

Lum expects this to increase power requirements at cellsites, "Now you'll need 10 kilowatts of power," he says, so that operators would likely want to have 15 kilowatts of power at hand. 4G cellsites use around 6 kilowatts for power, assuming a three-sector, 12 radio set-up.

This might cause some carriers to downgrade their MIMO requirements. South Korean operators are using 32T32R arrays for 5G, Lum says.

"There are lots of trade-offs, 64T64R makes sense for," Heavy Reading's Gabriel Brown notes. "Say Sprint with a US cell grid and the need for dual LTE-NR support, whereas somewhere like Korea or Europe 32T32R is a sweet spot."

Lum expects China Mobile to look to deploy 16T16R arrays in more rural areas, which will reduce costs but also speed and capacity.

Overall, power consumption should come down on 5G infrastructure over time. "Materials technology is improving," Brown says, while noting that "power-saving techniques" and AI will all come into play too over time.

While EJL Wireless Research's Lum says operators could hold off "maybe three years" in expectation of better power consumption results for 5G.

— Dan Jones, Mobile Editor, Light Reading

(9)  | 
Comment  | 
Print  | 
Threaded  |  Newest First  |  Oldest First        ADD A COMMENT
Gabriel Brown
Gabriel Brown
3/7/2019 | 4:09:39 AM
Hungry Hippos
If I understand correctly -- and I'm sure someone can set me straight if I'm wrong -- in the Sprint 64T64R deployment ithe radio/antenna unit is split between LTE 32T32R and NR 32T32R.

In sub 6GHz, 64T64R is needed where you have very high capacity demand and/or where you need vertical diversity -- e.g. where you have tall buildings. There is some debate about 16T16R -- some vendors are developing this product; others are not.

Also, note my addedendum to the tweet in the article 

https://twitter.com/Gabeuk/status/1100292194315980801

Some context to this tweet: these are important issues. However, the speaker was presenting at a vendor conference. It is entirely normal for senior operator execs to poke vendors to keep them on their toes, especially if they have also given praise. That's how I interpret this.
DanJones
DanJones
3/7/2019 | 9:13:35 AM
Re: Hungry Hippos
Gabe, the difference is that we're going from 2T2R, 4T4R, some 8T8R to 64T64Rb that's what's increasing the power requirements.
DanJones
DanJones
3/7/2019 | 9:13:35 AM
Re: Hungry Hippos
Gabe, the difference is that we're going from 2T2R, 4T4R, some 8T8R to 64T64Rb that's what's increasing the power requirements.
Gabriel Brown
Gabriel Brown
3/7/2019 | 11:12:52 AM
Re: Hungry Hippos
Umm, yes, that is indeed the point.
daveburstein
daveburstein
3/7/2019 | 5:30:28 AM
Great headline
Dan

And good data. But I think Earl is offbase in calling Massive MIMO 5G. It's been deploying on 4G since 2016 and isn't much changed. 
DanJones
DanJones
3/7/2019 | 7:31:32 AM
Re: Great headline
It has been used for 4G somewhat, it'll be commonplace for 5G.
bullschuck
bullschuck
3/7/2019 | 12:21:07 PM
The maths here look funny
Something here just isn't adding up.

So does a 5G node (as currently being deployed in mmwave in the US) use 3x as much power as a 4G RRH in the AWS/PCS band? I seriously doubt that. I don't think those 5G nodes are set up with 40 to 60 W amplifiers. But I would be happy for someone in the ecosystems to correct me if I'm wrong.

Does a mmwave 5G node use 3x the power of a 4G AWS/PCS small cell? Maybe. I wouldn't be surprised either way. But I doubt that those draw anything like 6 kw.

Now, if a carrier is deploying 5G in sub-6GHz bands, then OK, you need more juice to support all the amplifiers for all the antennas in the array. I get it. Lots of power. But then, do you need 3x the number of basestations? If you are deploying in mmwave, ok, sure, 3x is on the low side on the number of towers. But then you don't end up with the power issues, see above.

I get each of the issues but I don't see them coexisting. Either you deploy in mmwave and have to build a ton of nodes, but the power requirements are modest, or you are deploying in existing sub-6GHz spectrum and you have additional power requirements, but won't need additional towers.

There must be something I'm missing here. What am I missing?
Gabriel Brown
Gabriel Brown
3/7/2019 | 12:54:09 PM
Re: The maths here look funny
As per my comment down thread, I think it was partly this:

The speaker was presenting at a vendor conference. It is entirely normal for senior operator execs to poke vendors to keep them on their toes, especially if they have also given praise.

Fwiw, China Mobile is sub 6Ghz (150 MHz between 2515MHz and 2675 MHz and 100 MHz between 4800-4900).

And of course there are lots of variables deploying Massive MIMO on existing site grids -- say, if you design for uplink cell edge performance, you'll need more sites than downlink.
bullschuck
bullschuck
3/7/2019 | 1:41:12 PM
Re: The maths here look funny
Good point on the context. Clients should never own up in front of multiple vendors how they are poised to make a huge amount of profit. Throwing shade to control costs is a time-honored business tactic.