Smart Meters and Stupid Governments – Time for Dedicated Spectrum

Are our governments really serious about Smart Metering, or are they just throwing money away as a political gesture?  Increasingly it looks as if it’s the latter.  Barack Obama just made a headline announcement that the U.S. Government is prepared to waste $3.4 billion putting smart meters into 13% of U.S. homes.  The reason for my cynicism is a lack of standards, particularly with respect to the choice of a wireless specification to link the meters with each other and to appliances around the home.  The current choices are not based on any understanding of technology, rather than lobbying by companies desperate for funding.  As a result, there’s a strong chance that these meters will not work.

I was at the Wireless Congress in Munich last week and listened to at least four different wireless standards explain why they’re each the best choice for smart meters.  Not one of them was really convincing.  Most had slick marketing presentations, but underneath, there are some very good technical reasons as to why NONE of the current pretenders are the correct one to choose if we really want smart energy to work. 

The critical problem is the choice of the 2.4GHz frequency band, which is where most of the contenders operate.  Ten years ago that portion of spectrum, known as an Industrial Scientific and Medical band (ISM) was virtually empty.   Microwave ovens used it, but only for a few minutes each day.  Wi-Fi, Bluetooth and ZigBee were all still dreams.  It was like a freeway built before cars arrived.  Today it is already congested and each new evolution of Bluetooth and Wi-Fi eat up even more of it.  In another ten years, which is before the Smart Meter rollouts will even have been completed, it is likely to be at a standstill. 

Smart Metering is an initiative that will cost billions of euros / dollars to install and which needs to continue to work for a lifespan of twenty or more years.  All of the prospective wireless technologies being considered for use in Smart Meters operate in open frequency bands that are likely to be heavily congested before the smart meter installation program is even started.  To use this spectrum for something as critical as smart metering is folly.

If Smart Metering is going to provide benefits, it deserves its own wireless spectrum and standard.  It’s not too late for regulators to set aside spectrum and for standards bodies to get together to produce an optimal standard.  If they don’t, we risk wasting trillions of dollars and failing to achieve any reduction in energy consumption.

There is no question that Smart Metering is a good thing.  According to EnOcean – a company in the business of making buildings more efficient, buildings are responsible for 40% of our energy consumption – more than is used for transport. 

Smart meters promise to help reduce that figure significantly.  To operate efficiently, it’s generally agreed that they need a wireless link to connect them to other utility meters within the house, as well as household appliances consuming energy, usage displays and controllers.  Unfortunately there seems to be little consideration of whether the current choices of wireless will work.

According to a speaker in Munich last week, Germany has made the decision to use wireless M-Bus for its smart meters.  They already use M-bus for wired metering and a spokesman made the comment that it was a small step to move to wireless M-Bus.  I hope he wasn’t standing at the edge of a cliff when he thought that strategy up, as his small step could have disastrous consequences.    

I’d like to offer a little history lesson to the Wireless M-Bus folk.  Wireless is different from wired.  It involves adding security, connection methodologies, and both topology and power management.  It took the 802.11 standards group around 13 years to go from their starting point of translating a wired 802.3 connection to the first reliable and interoperable wireless one in the  802.11b specification.  Wireless is not a small step.  If you think that it is and decide to ignore the evidence of history, you run the risk of making a catastrophic mistake that has major ramifications for the future costs of your country’s energy usage.

In terms of the eight or nine different wireless solutions out there are the moment, the best suited are probably ZigBee (because of its mesh capability) and Bluetooth low energy (because of its robustness). 

ZigBee currently leads all of the others in deployments, with the flagship “ZigBee City” network of 255,000 nodes in Goteburg.  (Although if we’re being honest, this isn’t actually using ZigBee.)  This aside, it’s a good example of what can be achieved by adding a mesh technology.  The mesh capabilities mean that the expected need for 3,000 repeaters was reduced to just 17.  (So if you’re in the market for some cheap 802.15.4 repeaters, head for Goteburg.)

The problem is that both ZigBee and Bluetooth low energy (and most of the other contenders) operate at 2.4GHz.  2.4GHz is not an ideal frequency for a fixed in-home application like this, even ignoring the issue of a congested spectrum.  For a start, it suffers from multi-path fading.  That’s not too bad if you’re a frequency hopper, like Bluetooth, but not so good if you’re not, like ZigBee.  For Wireless Standards that sit on fixed channels, it can be a challenge, as there’s no easy way for them to find the best channel to operate on.  Even if they support frequency agility, which is the ability to change their mind, changing channels can result in sleeping nodes getting lost, with battery life being wasted if they have to keep configuring the network.  It’s a particular problem for smart energy applications, as a device’s location is often fixed.  Unlike a mobile phone, you can’t move your gas meter by a couple of metres on a daily basis to find a better signal.

Working at a lower frequencies should help get over this issue.  They’ll also perform better, because they have a greater range.  But there’s very little spectrum available and it’s different in different countries, which doesn’t help with a global solution.  Plus, once again, it’s an open band, so anyone can bring out other products which could eat into the it and stop devices working.

There are other good reasons to work at lower frequencies.  The radios are simpler to design, they use less power and they have greater range.  Which all help with making meters cost-effective.  The only negative factor is that as the frequency decreases, the antenna for an effective range gets larger.  In practice, that means there’s a sweet spot for this application somewhere between about 300MHz and 600MHz.  That will give full house coverage at reasonable power with an antenna that can fit within consumer devices.

If the EU and US Governments are serious, and want to enable energy savings in a form that will continue working for the next twenty years, then they ought to set aside some spectrum for it to happen.  Choosing the range / antenna size sweet spot suggests that it should be between 300 MHz and 600 MHz.  The band should be dedicated to smart energy and have three or more channels to let the radios have the flexibility to cope with multipath issues.

As we turn off our analogue TV channels, regulators could easily dedicate some of this spectrum for smart energy.  It means they won’t be able to sell it off for broadcasting more sport or porn, but the world probably needs that rather less than it does reduced energy usage.

In the next twelve months we’ll start to see the results of some of the first deployments based on current technologies.  Some of the whispers I’ve heard suggest that they are less than impressive.  That suggests that the issues I’ve covered above are already rearing their heads.  If that’s the case, then what we have at the moment still needs to go through several generations of development to produce something close to useable. 

Doesn’t is make more sense for the industry to forget their egos and sit down to design something that is fit for purpose, rather than being cobbled together.  It is always difficult to free up wireless spectrum, because it is a limited and valuable resource.  But there is little more valuable in the medium and longer term than trying to control and contain our energy usage.

Why combine standards instead of just shifting frequencies?  Because no single one contains all of the best features.  Bluetooth low energy was designed from a clean sheet of paper to give the lowest possible power and benefits from hind-sight, improving on most of what has gone before.  ZigBee can offer the mesh that makes deployment simpler.  And both understand the importance of a qualification and conformance regime to enforce interoperability.

Some companies will argue that the choice must be made now.  In some cases that’s because they’re pinned their future on smart energy and need to sell their chips now if they’re going to survive.  That should not force a Government’s hand.   Smart Energy is a solution for the world’s future, not a short term bail-out for a few start-ups.

It will take time to develop a new standard and take it to market.  But the alternative is to iterate the current solutions for two to three years to the point where they just about work.  That leaves a real possibility that they will fall over in eight to ten years time as other devices moves into the same spectrum. At which point someone has to pay to upgrade the complete system.

So let me issue this challenge to the EU and US Governments.  If you really believe in Smart Energy and Energy conservation, forget being mugged by one of the current wireless technologies and set aside some suitable dedicated wireless spectrum for your smart energy applications.  Then bang the heads of the standards groups together to come up with a solution that combines the best parts of each to solve the problem.  Smart Energy is not a short term – it needs to work for successive generations.