Ookla Shows Impact of 5G vs 4G on Smartphone Battery Drain
The latest research from Ookla, which operates the popular broadband connection testing service Speedtest.net, has revealed that Smartphone users using 5G based mobile broadband networks suffer between 6% to 11% higher battery drain than for those using 4G.
The study was actually focused on identifying the difference between modern chipsets (SoC) on Android based Smartphones, which perhaps naturally found that Qualcomm’s latest Snapdragon 8 Gen 2 chipset wasn’t only more efficient than the previous Gen 1, but also recorded the lowest battery drain of all SoCs in their analysis (31% for users on 5G and 25% on 4G-LTE).
The same trend was also found right across the board, which shows that battery efficiency continues to improve on both 4G and 5G capable chipsets (partly reflecting the ever-shrinking nanometre scale of modern semiconductors – this helps to deliver greater efficiency).
Advertisement
However, it’s worth noting that Ookla’s methodology for this test was very basic, with the company measuring battery drain by identifying all devices recording 100% battery level during morning hours (6am-12pm), and then comparing that to their minimum battery level in the afternoon (12pm-6pm). But this overlooks key details about the device and user, such as what activities they were performing, display size, brightness and screen on-time etc.
Such details are important because otherwise the study is ignoring key aspects of context and performance. For example, the study suggests that 4G devices use less power, which is a logical assumption, but this overlooks the fact that 5G also downloads data several times faster too.
Likewise, the study doesn’t split the 5G dataset between Non-Standalone (NSA) and the latest Standalone (SA) networks, which is relevant because end-to-end 5G SA networks are known to be much more energy efficient than the initial round of transitional 5G deployments. A more scientific study would have been better able to dig into this sort of detail.
Advertisement
Mark is a professional technology writer, IT consultant and computer engineer from Dorset (England), he also founded ISPreview in 1999 and enjoys analysing the latest telecoms and broadband developments. Find me on X (Twitter), Mastodon, Facebook, BlueSky, Threads.net and Linkedin.
« Broadband ISP BT to Scrap UK Line Rental Saver Discount
Netomnia Covers 500,000 UK Premises with FTTP Broadband »
Latest UK ISP News
- FTTP (6046)
- BT (3644)
- Politics (2723)
- Business (2443)
- Openreach (2410)
- Building Digital UK (2333)
- Mobile Broadband (2153)
- FTTC (2086)
- Statistics (1909)
- 4G (1824)
- Virgin Media (1771)
- Ofcom Regulation (1588)
- Fibre Optic (1470)
- Wireless Internet (1464)
- 5G (1414)
Not splitting between 5G NSA and 5G SA makes the entire methodology suspect.
5G SA can run with just a 5G signal, and completely power down the 4G LTE parts of the modem while it’s on 5G SA. In contrast, 5G NSA carries most of the control signalling on the 4G LTE carrier, and you cannot disable the 4G LTE modem as a result; instead, the 5G NSA carriers are used as supplementary carriers in parallel to 4G LTE carriers, but only for user data.
So, it’d be surprising if any device is able to get better battery life using 5G NSA than using 4G LTE – there’s going to be some power overhead from maintaining two sets of carriers, and that’s got to be overcome by the higher speeds you can get from using the 5G NSA carriers allowing you to spend more time with both modems turned off.
In contrast, 5G SA only needs the 5G NR modem on, not the 4G LTE modem, and could save battery if the 5G NR modem uses less power to maintain network attachment than the 4G LTE modem does.
Totally, the fact its pretend 5g using 4g also means its having to receive both 4 and 5g signals, so it obviously will use more battery.
Strange, no iPhone SOCs? Apple still use Qualcomm modems (and it should stay that way), and that battery drop in range reflects my experience on 5G Auto. Personally the trade off is fine with me, and it’s good we have to choice.
Snapdragon 8.1 in my phone – 5G gets it so hot I can’t hold it after a while.. The phone charges in 17 mins and it gets hot then but 5G gets it very very hot.
I don’t think the Ookla/SpeedTest app can get the battery levels on iPhones/iOS due to how restrictive they are. Features like their coverage map are all from Android data, I wouldn’t be surprised if it was the same here.
Now you mention it, that would explain the omission
even more reason to avoid 5g. if i get a phone with 5G, the 5G will be disabled, that is if the provider allows it. Best thing is not to get a phone with 5G
If 5g is patchy around where you live, makes sense. The lower the signal the more the phone works too find a stronger signal. It’s better to have full signal to have the phone use minimum power when sat in standby on 5g.
@Michael V, 5G is very patchy where I am, so I have been told, but the networks deny it is, well some do.
I am not planning to get a new phone for a while, hopefully not for a few more years, hoping mine will last for a lot longer, already 3 years old, so another 4 or so would be nice.
Not buying a 5G phone 4 years from now would be like buying a 3G phone in 2020.
We’ll have 5G NR networks (proper 5G, not 4G with a 5G band attached like we have now) which use less power. Networks are already deploying 5G on low bands for extended coverage and even re-using 4G bands for it, so you’ll probably have worse coverage without 5G. 5G modems are also improving a lot: my 2nd gen flagship modem can’t compete in speed/heat/power with the current best one… all this in 3 years or so?
You seem to be very negative about everything new, Ad47uk. You don’t use 5G, don’t know how good the current coverage is or how it will be when you upgrade your phone, have not tested this in practice to see if you notice the (current) higher battery drain, etc, but you’ve already concluded that 5G is bad and that it needs to be disabled or not be supported by the device at all. Wouldn’t it be better to keep an open mind about these things?
Michael V says:
July 5, 2023 at 8:06 pm
If 5g is patchy around where you live, makes sense.
Like the entire country then…lol
Yeah iOS is left out as the operating system is restrictive for apps running in the background. Most Network mapping apps can’t run on them.
I don’t really see much of a difference between 5G & 4G on battery drain. But with 5G typically using the mid band around 3.8hz, it’s going to use more power. Seeing more 5G on low band at 700mhz will help.
This brings up the question on how much power 5g will use when ofcom auction off mmWave bands & networks deploy that. [26ghz & 34ghz bands]
Right now, I don’t think it would make sense to split 5G on SA & NSA. Vodafone have only just launched SA in a few cities in England. So is better to just class 5G-NR as NSA.
You expect 5G NSA to use more battery than 4G, since when you’re attached to a 5G NSA network, you are also maintaining attachment to a 4G network. With 5G SA, you can power off the LTE modem until you’re considering switching back to an LTE cell – just as a 4G phone powers off the UMTS and GSM modems until it’s contemplating a switch to a previous generation cell.
5G SA at least allows you to power off the LTE modem for the duration of your attachment to a 5G cell.
So results with 5G NSA and 5G SA mixed together are misleading – 5G NSA results are “power consumption when using both LTE and 5G NR modems simultaneously”, while 5G SA results are “power consumption when just using the 5G NR modem” and 4G results are “power consumption when just using the LTE modem”. And note that the data set they’re using is global, not just UK phones.
As a consequence of not separating out these three cases, the data on “5G” is not of use to anyone – if you do have 5G SA via Vodafone, you’ll see less power consumption than this test implies (but how much is unknown), while if the test included 5G SA networks (say in the USA, China, Hong Kong or elsewhere), anyone using 5G NSA will see higher power consumption than this test implies.
Completely irrelevant test methodology, so many unknown variables.
Excellent example of where correlation does not mean causation.
Maybe they were trying to prove something else. There is always a reason behind the results, them being the results they want to publish. As with any tech paper written these days its the money behind the research wanting a specific answer to a question proven.