Many operators talk loudly about their 5G network upgrades. Still, the new standard is not yet completely ready for our network test. However, umlaut conducted measurements for connect which reveal what the fifth generation of mobile communications ­can ­already offer today.

Wherever we look, we see 5G in the starting blocks. Our longtime network testing partner umlaut and we are also ready to ­evaluate the 5G network extensions with according measurements. The devices needed for this are available – as well as the necessary knowledge. However, 5G implementations stand only at their very beginning in Austria and Germany, ­whereas in Switzerland one operator – Sunrise – has already started a ­large-scale 5G rollout.

But the implementation of the still young mobile communications standard is accompanied by dependencies and problems. It is not a ­given that the young networks‘ functions are fully compatible with all new smartphones. And not ­every tuning approach in the networks is met by a mobile phone available on the market which supports this exact feature. Therefore, the early rollouts can not yet be fully evaluated. Consequently, we have excluded it from the setup of this year‘s network test.

As Vodafone and Telekom ­assiduously build 5G network cells and gather their first experiences with them, we were also keen on ­geting a first impression about what 5G is able to offer under the conditions of our network testing. But as the many tens of thousands of network cells in Germany only include a few hundred which are already supporting 5G, umlaut‘s drivetest cars would have only gathered a very small share of 5G measurements. Thus, for the moment we settled with measurements at a few fixed ­locations which already offer 5G. 

In Austria, the operator A1 Telekom is only allowed to use the frequency blocks which it purchased in the according auction starting from the end of this year. For our measurements, it activated a site with a test license. This allowed us to precisely evaluate which advantages 5G ­offers over LTE.

Measurements in Austria

So, Austria only ­allowed for a limited view, as 5G was not yet commercially available in all networks. As tje result of consultations with the operators, we had to consider further ­peculiarities on top of visiting the predetermined locations in Alpbach for A1 Telekom, in Linz for Three and in Klausen Leopoldsdorf for Magenta. Instead of umlaut‘s regular benchmark ­setup, we could only use specific smartphones with speedtest apps. For A1 we used the Samsung S10 5G, for Magenta the Huawei Mate 20X 5G and for Three the ZTE Axon 10 Pro 5G as their preferred devices.

In the measurements, 5G convinced with download data rates betwwen 500 and 1100 Mbps. On average, this is five times faster than the smartphones which we used for comparative 4G measurements – and also particularly faster than what can be commonly expected in the fixed-line network. Even considering that the resouces of an LTE network are distri­buted over more users, this is a signi­ficant speed advantage for 5G.

Measurements in Germany

In Germany, the network test ­experts of umlaut could already dig a little deeper. With a measurement car ­retrofitted with Samsung S10 5G, a drivetest team visited three suggested municipalities per operator. For ­Telekom they were located in Berlin, ­Cologne and Darmstadt, while Vodafone recommended places in Ratingen, Düsseldorf and Offenbach/ Main. Telefónica is obviously ­dedicated to optimize its LTE network and did not yet take part in the 5G comparison.

In the average of all throughput measurements, 5G showed more than three times the speed of LTE – again a remarkable performance boost. However, the advance shrank when we compared the maximum throughput rates. Here, the newest generation of mobile communications reached 695 Mbps – which is not quite as far ahead of LTE. The latter partially ows its good result of 401 Mbps to the very effective so-called carrier aggregation – the bundling of more than one frequency bands. This works well especially in not too crowded network cells.

An interesting observation is that the new mobile communication standard lags a little behind LTE when it comes to uploads. One of the reasons is that in the current implementation, non-standalone 5G ­always requires a 4G connection which controls the 5G data transfers. Currently, according to a Telekom specialist, only the downloads ­actually make use of 5G. Uploads are still executed via 4G, but with additional ­protocol overhead. When 5G will also support uploads – and even more with „stand-alone 5G“ which does not require a 4G link any more, the speeds of data uploads are definitely bound to change.

Measurements in Switzerland

In Switzerland which currently precedes Europe in terms of mobile communications, we went all the way. Simultaneously to our network test, we sent an additional drivetest car carrying Samsung Galaxy S10 5G on the the way. It drove the same routes as our official network test cars. In doing so, the distance between the 5G and the 4G measurement cars was always wide enough to make sure that one car would not affect the transmissions of the other. Thus, the measurements in the 5G model country were executed according to connect‘s network test standards. In this context, Swisscom claims to have commercially launched 5G in eight cities on April 17th, 2019. Sunrise specified that it ­already reached 152 cities with 5G by March 30th, 2019.

Of all the measurements, in the Sunrise network already 4.8 percent were completely conducted in the 5G network. For Swisscom the number was  substantially lower with a share of 0.43 percent. When taking into account those measurements that contain a partial share of 5G, the percentages came closer together with  7.1% over 6.5%. The lead of Sunrise is all the more surprising as this operator focuses its 5G implementation to smaller cities and even villages as a fixed-line substitution – while the network test has a larger share in the more densely populated urban areas in order to represent a sufficient share of the total population.   In any case, the determined data ­rates are remarkable, particularly regarding the expectable minimum download speeds (90 percent faster than). In this KPI, Sunrise shows a speed increase by a factor of 10. And even in the slowest it is still 10 percent faster than the maximum speed of ­many DSL lines in Germany. The minimum speed in the Swisscom network is also faster by a factor of 5, although it already marks the top at 4G within the test area of Germany, Austria and Switzerland. Sunrise also shows an increase in the upload data rates with a minimally reduced 5G reliability in comparison to LTE. In the Swisscom network, we observe the same slowdown as in Germany.

This is only an intermediate result of the 5G implementations in Switzerland which were recently somewhat slowed down due to the flare-up of protests by oponents of mobile communication (see box above).  But now it gains momentum again, especially as Swisscom plans to re-allocate parts of the LTE spec­trum for 5G on demand via Dynamic Spectrum Sharing (DSS) by the end of the year. Provided that DSS capable phones are soon available, this way a 5G coverage of all of Switzerland could become possible quickly. Sunrise also observes this technology, but currently upgrades its network the conventional way. At the time of writing, it was claimed to cover 331 municipalities – and counting.

And this is good, as umlaut and connect plan to re-visit Switzerland soon again in order to measure a complete 5G network.
As for Germany, we are keenly observing the development.

Generational Conflict

Each time a new mobile communications standard is introduced, this is also a wake-up call for the opponents of this technology. This is true again for the fifth generation, which stirs up the debate due to its new properties – a debate which was believed to be already resolved.

Among the arguments which oponents of mobile communication invoke against 5G, are the growing numbers of radio cells, the use of higher frequencies and the so called „beam forming“ – which means directing the radio signals straight onto the users.

However, critics often ignore that radio exposure is actually reduced with a growing number of network cells. The reason is that the intensity of a radio signal rapidly decreases with growing distances. In order to compen­sate for this, the transmitter has to ­tune up its power the farther it is positioned from the receiver. In order to save ­battery power and energy in the base station, the transmitter and the smartphone both reduce their transmission power for each connection as far as possible. The closer they are together, the less power they use for the transmission. Therefore, a tighter network of base stations actually leads to lower radio exposures – both for the user caused by his smartphone as well as for bystandards caused by the radio cells.

The higher frequencies of ­millimeter waves are currently beside the point, because the first 5G implementations start in the sub 6 GHz spectrum, in Germany at around 3.6 GHz. These frequencies are part of the well-tried range between 2.6 GHz for LTE and 5 GHz for WiFi. Concerning millimeter waves, research was intensified. But it is evident ­already that higher frequencies penetrate human tissue less deeply. Thus, for example the ­human brain is better shielded against millimeter waves.

It is understandable that ­beamforming – the focusing of signals to the recipient – can be viewed as a threat. But here again the output level is reduced so largely that only the necessary minimum reaches the receiver. The maximum radio exposition of a single user stays about the same. And bystanders are ­exposed at a much lower ­level ­reaching down to practically nothing, when they are situated ­outside the directed beam.

Even after almost 20 ­years in which more than 50 percent of the population in Europa have used mobile phones, the accu­rately recorded official cancer registers show no evidence of rising cancer figures due to ­mobile communications. This should give even the critics ­something to think about.