Why are 3G handsets and data cards so awful?
Answer: because the handset makers are deliberately making it impossible to show how bad they are, by making the tests meaningless.
Its easy to understand why a 2-year-old toddler might throw mud onto the breakfast plate before trying to eat it.
Its rather harder to understand why a large, successful corporation would sabotage its own products—but thats what is happening in 3G handsets.
Heres the problem, according to the people who test the things—Spirent: “The big phone companies have watered down conformance tests to make them easier to pass, and have optimized their products to pass conformance tests.”
The result, they say, is exactly what you would expect: the products pass the tests, but fail in the field.
To understand in detail why this has happened, you probably need to know more about how CDMA and Wideband CDMA technology works than I can honestly claim to know myself.
But I can give you the gist: with WCDMA, the number of users you can have connected to a single cell is not fixed, the way it is with GSM.
Instead, the more users you have, the more the background noise rises. To get around this, you need to boost the power of the signals, both from the terminal, and from the base station. Easy!
And then your customers all complain, because their batteries die before they can get home to recharge them.
Its starting to look, ironically, as if the problem will be tackled in the United States before it is treated seriously in Europe.
We all expected WCDMA to take off in the European Community first; the plan indeed was to have it as the de facto standard by now, and to be moving into Phase 2, with HSDPA (high-speed download packet access)—but in fact, this isnt what is happening.
Spirent is a very good oracle on the future of this technology, because it makes the test gear; sales of test gear predict sales of the products which are tested.
Thus it was, four years ago, that I was able to stand up on a podium and, using my most confidently authoritative voice, predict that we would see no massive 3G rollout before 2006. It was obvious; Spirent was not selling the equipment needed for testing the things.
States Take the Initiative
And what Spirent reports this year is that HSDPA is rolling out first in the States, with Europe six months to a year behind, and much slower.
The reasons are many, but one of the most striking is what Spirent is calling the “Verizon effect”—the adoption of rigorous testing of handsets by the operator, rather than the conformance authorities.
The situation is pretty grim if you believe Spirent (and I have learned to do so, the hard way!)—and to quote Nigel Wright, director of applications engineering: “We have seen of the order of 6-7 dB difference from best versus worst behavior in fringe reception.”
Exactly what does this mean? According to Wright, it means that one phone will be picking up a signal 6 dB stronger than another, just when it causes the most difficulty. That, pretty much, is the difference between “a good signal” and “no service.”
The problem is basically down to the fact that (in Europe, anyway) the authorities picked the 2.1GHz spectrum for WCDMA. Its an awful choice, because at that frequency, a weak signal pretty much stops dead when it hits a building wall.
The reasons for choosing this wavelength were political, not technical.
A far better choice would have been around 800/900MHz or even down around 450MHz—both spectra which are now coming to be available, but were not going to be free in 2002, when politics said 3G should be going live.
Around those levels, in-building penetration is excellent.
At 2.1GHz, it is negligible unless youre really close to a really strong base station signal. And crucially, as long as there arent many other 3G users in the cell: when numbers go up, the background noise goes up. And performance goes down.
What 3G users are finding is that they can get a signal and make a call, but only if they stand still.
Out in the street, it works quite well, of course; but we all know how easy it isnt to make a call in busy traffic.
In any large city, youll see people wincing and ducking into doorways so they can hear what the caller is trying to say.
If theyre 3G users, youll see them suddenly taking the phone away from their ear, staring angrily at the display, and shutting it down—because from four bars in the street, the display is now showing “no signal.”
What Verizon has been doing, according to Spirent, is devising tests for terminals that show up fringe-reception failures.
And, taking their lead from the way Verizon customers are responding to this, European operators too are now saying: “It does not matter if your smart new phone meets conformance tests. It has to pass our own standards, or we wont touch it.”
And not before time, either: public disillusionment with 3G is far greater than you would guess from the statistics.
My prediction: 3G phone charges are going to have to drop, enormously, in the next six months, as user numbers go up. And (as Ill try to explain next week) the beneficiary will be mobile data users.
Contributing columnist Guy Kewney has been irritating the complacent in high tech since 1974. Previously with PC Mag UK and ZDNet UK, Guy helped found InfoWorld, Personal Computer World, MicroScope, PC Dealer, AFAICS Research and NewsWireless. And he only commits one blog—forgiveable, surely?