LTE has too many frequency bands – endangering global connectivity

LTE has too many frequency bands – endangering global connectivity

LTE is a triumph of standardisation at the silicon chip level. Never has the world been so converged on a common mobile technology future. But LTE has been much less successful at the Radio Frequency level. LTE has been specified in too many frequency bands. There are 30 different frequency bands for the FDD version of LTE and 14 for the TDD version. Just what problems are we storing up in the future by this alarming proliferation of 4G frequency bands?

No traveller has any fond memories of the analogue cellular age. Europe was such a hotch-potch of different cellular standards and frequency bands that a car equipped to work across Europe would be so full of different car phones that there would be no space left for any passengers. This traveller’s nightmare was over once GSM become the dominant technology across Europe and then most of the world. The number of GSM bands were limited to just 3. This comprised 900 and 1800 MHz in Europe/Asia/Africa and 1900 MHz in the USA. The cross-Atlantic GSM roaming breakthrough came with the Motorola Timeport in 1999 – the world’s first tri-band GSM mobile. Within a few years tri-band capability was being built into every GSM mobile. We have come to accept and value this universal connectivity. It makes our global economy more efficient. That could now be at risk in the new 4G era.

What are the chances that the world’s smartphone manufacturers will eventually finds ways to pack in all 44 frequency bands specified for LTE into every smartphone, or even half this number? Probably zero. With the current state of the art (discrete component RF architectures) every new band built-in requires a filter that adds loss. The losses are additive and degrades the smartphone performance.  Low end mass market smartphones manufactures will also go for the minimum as saving every cent counts. Then there are the temptations of mobile operators whose balance sheets are under stress from high customer churn. Why should they subsidise a 4G smartphone to also work on a different 4G frequency band where the only beneficiary is a competitor?

That is the concern but what is happening in practice? Below are 4 popular LTE smartphones and already 13 combinations of LTE frequency bands reflecting regional and operator difference:

Apple iPhone 5
Model A1428: 700b 850 1700 1900
Model A1429 (GSM): 850 1800 2100
Model A1429 (CDMA): 700c 850 1800 1900 2100
Nokia Lumia 920
RM820: 700 1700 2100
RM821: 800 900 1800 2100 2600
Blackberry Q10
SQN100-1: 700 850 1700 1900
SQN100-2: 700 1700
SQN100-3: 800 900 1800 2600
SQN100-4: 1900
HTC One
Europe Version 1: 1800 2600
Europe Version 2: 800 1800 2600
Sprint: 1900
ATT: 700 850 1900 2100
T-Mobile: 700 1700 2100

Nothing is impossible in today’s technology world but there is already clear evidence of regional specific 4G smartphones becoming the norm and we may see more examples of operator specific smartphones. The casualty will be the international traveller moving out of range of WiFi. It will become a matter of chance if their 4G smartphone connects to the local 4G network in a country they happen to visit due to incompatible frequency bands. Looking further into the future the “connected car” looks unlikely to be able to tap into the benefits of 4G networks if every time the car crosses a border it loses its connection. A 4G Internet of Things looks a complete non-starter until all this confusion of frequency bands gets resolved.

There are other implications where 4G is rolled out in different bands by different mobile operators in the same country. Incompatible frequency bands between operators is not new. It happened when 1800 MHz GSM (PCN) networks were first introduced into the UK but the situation was soon put right with the arrival of first the dual 900/1800 MHz band GSM mobiles. It was a driven by the GSM economies of scale rather than by operators or regulators. But we cannot just assume that smartphone manufacturers will automatically ride to the rescue.  The emergence of O2 from the recent Ofcom 4G spectrum auction without any 2.6 GHz spectrum and EE and H3G emerging with only  5 MHz of 800 MHz of spectrum (insufficient for a high capacity 4G network) has taken away a natural convergence of compatible UK smartphones.

If a polarisation of 4G smartphones occurs it will have consequences. Even though smartphones might start off on the right bands for the right network, users will switch networks or sell on their smartphones. This could lead to a growing disconnect between the installed base of 4G smartphones and the network provider a customers buys their SIM from. That will add stress on the less capable 3G networks and make it more difficult to re-farm 3G spectrum for the more efficient 4G technology. We can already see from the data above that the current Apple iPhone 5 does not work on any of the 4G spectrum recently auctioned by Ofcom (800MHz or 2.6 GHz) but just happens to work on the EE re-farmed spectrum at 1800 MHz.

An even greater worry for the industry is the impact of the excessive LTE band proliferation on carrier aggregation. Carrier aggregation allows LTE systems to switch customers between different frequency bands a mobile operator is using. It makes a significant contribution to LTE efficiency. But its use depends upon the right selection of bands in the smartphone design and whilst any particular combination is not a particular challege the totality is far more difficult. Even having many limited combinations eats into the scale economies of 4G devices with an adverse impact on their price.

The RF component industry is perhaps the most switched-on to this LTE band proliferation challenge. There are newer RF designs that use MMMB wideband technology (multimode multiband). But there is a significant performance loss as there is no opportunity to optimise for a specific frequency band. So if the LTE equivalent of the tri-band GSM mobile ever arrives it may come at a price of lower performing smartphones.  This can eat into a fully loaded LTE network capacity and force expensive additional investment to retrieve the lost network capacity. So perhaps the ideal goal is a compromise rather than a future smartphone working on all LTE bands and good compromises require the active engagement of all the players.

Radio spectrum management is a very long term game. In a blink of an eye on radio spectrum management time 5G will be upon us with demands for considerably wider radio channel widths. This will demand new frequency bands as trying to engineer such very wide radio channels in any existing bands already in use will be an operational nightmare. So more mobile bands will emerge for 5G to be accommodated in smartphones.

Are these concerns exaggerated? That depends. Somewhere in this mess is likely to be an optimal solution but there is no sign that the problems created by the proliferation of LTE frequency bands has reached the “to-do” list of any regulator or government or industry group. So the point of the blog is not that things will get this bad but they will if nothing is done about it.

The irony is that we may well be moving back to the future – where, in the new 4G era,   the only absolute guarantee of the most basic smartphone connectivity anywhere in the world will be GSM. So we had better hang onto our GSM networks that bit longer – just in case.

 

Comments are closed.