CONNECT GOES LTE
Download article here
When our yearly network test covered LTE in 2013, it was a premiere. Now connect even measures smartphones
in 4G networks, being the first European magazine to do so. The equipment is worth a fortune.
Last January we were sure of a catastrophy. The numbers were so terribly wrong that there was no doubt
about a major bug. More than one colleague cut me a look that said „Come on, give us the all-clear!“
Because if this was to continue, there would be no measurements for connect 3/14.
A connect magazine without smartphone tests? Unthinkable! What had happened?
LTE measurement setup
The LTE license auctioning in April 2010 in Germany had made it clear: As soon as LTE phone sales and grid expansion reached a critical mass, we needed to be able to measure LTE. Easier said than done! Surely, WEKA MEDIA PUBLISHING owns three wickedly expensive Universal Radio Communication Tester CMU200 by the German noble brand Rohde & Schwarz. These devices can simulate a cellular radio station, but only in the GSM and UMTS network, respectively. Since you cannot upgrade the Rohde & Schwarz to LTE, it was time for something new. We started searching for a suitable LTE measurement setup in the beginning of 2012; two years later, this setup would cause the aforementioned panic in the newsroom.
800 Mhz instead of 900
On this journey, our first discovery was: The long waves of LTE at 800 Megahertz make it easier for the network operators to cover large areas, but they’re too much for the absorber hall of our laboratory at only 3 x 3,5 x 3,5 meters. The lowest frequency this hall was suitable for was GSM at 900 MHz. Part of this setup were absorbing wedges made of foam; their job was to prevent outer metal walls reflexions from falsifying the direct emission of test aerial and smartphone. Also, the distance between the two radio partners was only good for precise measurements as minimal as 900 MHz. Since both preconditions weren’t easily to be changed – we couldn’t up the ceiling height – the only solution was to enlarge the room virtually through extremely flat absorbers. This conversion was reflected in the books with a whopping 63 000 euros.
For measurement precision it was also necessary to lengthen the distance between antenna and device under test – which we achieved with a super flat cross-polarized test aerial. It is a big part of the accessory that at 30 000 euros cost as much as a brand new Volkswagen Golf GTI. Still, we could rebuild the absorber hall so we didn’t have to find a new one.
But a perfect absorber hall is nothing without the right measuring device. Although „measuring device“ is plain understatement – with a single device you couldn’t even determine the smartphone’s aerial sensitivity TIS (Total Isotropic Sensitivity).
Having sounded out the market for mobile communications measurement technique we set our eyes on the wideband radio communication tester CMW500 by Rohde & Schwarz. The CMU200’s LTE-ready successor is lovingly being referred to as „Callbox“ by many testing engineers; every established smartphone manufacturer uses it. You can get the basic CMW500 for a good 40 000 euros, roughly as much as a Mercedes E-Class is, to stay in the picture. The callbox works as a base station and can therefore establish contact with smartphones running R&S SIM cards. It can also switch to special measuring modes and read the data provided. To boot, the CMW500 can measure by itself, but when it’s about accuracy it employs the extremely precise spectrum analyser Rohde & Schwarz FSP7.
Since the basic CMW500 did not meet the high connect requirements, we settled for the MIMO option which is dimensioned for two LTE antennas each on the transmitter’s and the receiver’s side. This option is about 13 000 euros – you can buy a Volkswagen Polo for that kind of money.
Running out of time
Another massive challenge for our measurement setup renovation were the many LTE wave bands. Wireless carriers are happy with this range because they can use up to three bands: Rural areas with less customers are served best with 800 MHz while highly frequented places like metropolitan train stations profit from small 2,6 GHz cells. Everywhere else you can operate with a combination of both or sometimes with 1800 MHz.
Which means that we must be able to measure all three bands as to not put people at a disadvantage – either in cities, on motorways or in the country. The consequence of this: Instead of measuring the three bands GSM 900, GSM 1800 and UMTS 2100 like we used to we now have to care for six. Trying to wrest another absorber hall plus a second CMW500 from the publishing director was hopeless from the start. But there are more phones coming in every day, and as the utilization was high even before, we had to look for new ways.
The many measuring points surrounding the smartphone when it is fixed on the artificial head used to be a hassle because each and everyone of these points had to be approached precisely with a positioner, vertically and horizontally, before measuring could start. To cover every measurement angle we used to rely on a turning device that could spin the artificial head itself around two axises.
The new Conical Cut Positioner NCD by Maturo (45 000 euros) solves part of this time consuming problem: It spins the artificial head only vertically. The test aerial sits on a long boom on the same level as the smartphone can be turned around the phone in a large circle to change the other angle level. A precise high performance motor provides quick changes of position. To supply this motor with enough power we placed a three phase current in the absorber hall. A laser target device projects a hairline cross onto the artificial head as to place the phone precisely at the intersection of the two turning axises.
To speed up the measurements even more we needed to update the software and purchase a sequencer option. This enabled us to measure a variety of frequency bands automatically and in rapid succession.
As the LTE measurement setup evaluation was completed in March 2013, it called for an investment of 300 000 euros. Such a sum required close scrutiny and approval at the highest level at WEKA MEDIA PUBLISHING. Purchasing less expensive technology was discussed, outsourcing the measurements was considered. But finally the renovation got the nod.
The renovation starts
As soon as that we had to find two weeks time between producing issues of connect magazine; we started at the end of November 2013. Shortly before Christmas the new system was in place, and connect engineers were trained to operate the measuring equipment and the software.
Come January, we firstly put established UMTS phones to the test – and the measurement reports scared us to death. The new sensitivity figures differed radically from the ones we had gathered previously. We changed the measurement bandwidth, the measurement channels, the measured subcarrier – nothing did much of anything. Was there a system error?
After quite a lot of experiments, redemption came: In order to speed up the measurements a new technique to determine the bit error rate in relation to the smartphone’s signal level was used. Here, experts talk about a combination of measurements – Receiver Signal Strength (RSS) over RSS Patern Measurement, RSS Pattern Linearization and Sensitivity Search.
This method’s big advantage: The important but time consuming bit error rate measuring has to be conducted at one measuring point only. All other measuring points – there are hundreds – can be calculated from this result. That saves plenty of time, but it requires a connection between Callbox and smartphone that we previously had not used. After having chosen this option, old and new figures were in unison – and LTE measurements could finally begin.
For starters we gave the current smartphone stars a going-over. And we won an important insight: When the wireless coverage is good, an LTE phone will transmit the maximum data rate – 100 Mbit/s or 150 Mbit/s at LTE 1800 or LTE 2600. But let the signal level drop only about 1,5 dB below the necessary threshold, and the bit error rate will climb from almost 0 to 100 per cent; the signal turns useless.
When signal quality goes bad, LTE is able to at least switch from the 64-QAM modulation scheme to the more robust 16-QAM or to QPSK. This way, data is transmitted, if only at a quarter of the speed possible. Therefore you should look out for a smartphone’s wireless quality. connect shows this quality – it is the only magazine in Europe to do so.
The measuring tower
connect has bought plenty of measuring equipment in order to be able to determine a smartphone’s wireless qualities.
Since December 2013, connect has been conducting its wireless measurements with the over the air tower Rohde & Schwarz (R&S) TS 8991. At the top of the head-high rack there is a Maturo NCD controller: it can place the turntable – with the smartphone fixed at the artificial head – precisely and quickly in any position. The same goes for the long boom with the antenna representing the cellular radio station. Below sits the R&S SMT03, the measuring system’s standard signal generator. One storey down we find the spectrum analyzer R&S FSP7, serving for all high-precision performance tests. Next up is the R&S Vector Signal Generator SMBV100A – for GPS measurements, it can simulate as much as eight satellites simultaneously. It is followed by the LTE measurement’s centrepiece, the wideband radio communication tester R&S CMW500; it supports many other modern communication standards, too. Finally, there’s a tried and tested R&S CMU200 for GSM and UMTS measurements. The R&S switch unit OSP120 takes care of connecting the correct ins and outs during measuring. The R&S AMS32 software manages the measurements; the going rate for our version is about 40 000 euros.
CONCLUSION Bernd Theiss, connect editor
300 000 euros for a maximum 11 decibel difference between two smartphones – that sounds a bit exaggerated, doesn’t it? Not so fast. Consider this: Our LTE measurements show that in the country you can be three times as far away from a cell tower with a Nokia Lumia 925 than with a LG G2 before the data rate starts to sag. That shows how important repeatable LTE measurements are. Especially since up to now we have only measured the figurehead smartphones by top manufacturers – cheaper models will presumably show even bigger differences. Only measurements expose certain weak points a phone has. connect is proud to provide you with this information first handedly.