A few folks have asked me what Apple could to to fix the problems. As I wrote in earlier posts I think this is an antenna issue, that makes software fixes unlikely. It may be that the problem has a software component. For example, the signal strength bar display could be exaggerating the problem.

There may be a way to redesign the antenna to remove the problem without changing it much. In my opinion the dielectric properties of the hand are changing the capacitance between the metal bands on either side of the slot. That capacitance might not be a necessary part of the design. In that case it may be possible to solve the problem by making the plastic gap wider, that would reduce the capacitance in any scenario. If the capacitance is necessary though then things become more complex. It may be possible to move the sensitive area a little into the inside of the handset.

When the iPhone 4 antenna was announced I commented “This is a very difficult thing to do” in an article on the Wall Street Journal website. (I gave lots of other comments too, though these mostly weren’t used.)

Now the internet is abuzz with discussion of reception problems that have been found with the iPhone 4. Many are blaming the new antenna design. I don’t know for certain what the problem is, but I can make some informed speculation.

Customers who have bought the iPhone have noticed that when they touch a certain place then the reception gets much worse. That place is the small band of plastic in the out metal sidewall on the bottom left.

This video isn’t me, but it’s one of the better videos demonstrating the problem on the net.

A lots of folks on the internet have been asking if this problem is common to other phones. In a press release Apple said:

“Gripping any phone will result in some attenuation of its antenna performance with certain places being worse than others depending on the placement of the antennas. This is a fact of life for every wireless phone. If you ever experience this on your Phone 4, avoid gripping it in the lower left corner in a way that covers both sides of the black strip in the metal band, or simply use one of many available cases.”

It’s quite true that other phones suffer from similar problems. There are now several videos on the internet showing the signal strength report on various phones deteriorating when they are held. This has been shown using old Nokia phones, Blackberries and earlier iPhones. This is a well known problem. The important question though is: does the iPhone 4 suffer from it *more* than other handsets? In my opinion from the evidence available online the answer to that question is “Yes”.

The human hand contains a lot of water and other materials which absorb EM waves. It is also strongly dielectric, muscle has a dielectric constant of ~56. So, any capacitive effects are strengthened by the presence of the hand. Both of these have an effect on antenna performance. Suppose you cover over the antenna on your handset with your hand. Because of the absorbtion less radiation reaches the basestation. Also, the antenna contains distributed capacitance, this capacitance is increased by the dielectric properties of the hand, the antenna become more dielectrically loaded. The antenna has been tuned to work at the design frequencies with the normal capacitances. This disturbance causes “detuning” and the performance of the antenna reduces at frequencies bands it was designed for. Often there is a frequency shift and the antenna’s bands of good performance move up in frequency or down. People talk a lot about the absorption effect, but in practical cases the dielectric effect is often more important.

The magnitude of this problem was measured by G.F.Pedersen in several papers that comprise his PhD thesis. Pedersen got many people to hold a mobile phone like they would in a phone call and measured it’s performance as they held it, to find out the loss caused by the proximity of the head and hand. He collected the data for several types of antenna, including an internal PIFA antenna mounted at the top of the handset. Pedersen’s measurements show a great range of performance depending on how the specific characteristics of the user, such as how they hold the phone, the size of their head, etc. He found that for the PIFA antenna the average degradation was ~3dB. In the past I’ve found that modern handsets with internal PIFA antennas close to the top are a little worse than the one Pedersen used, but not much worse, I’d estimate the loss at 3 – 6dB. Those with PIFA antennas at the bottom are worse because because the user tends to cover over the antenna with their hand. I don’t have an estimate for the loss in that case.

The iPhone 4 doesn’t report the signal strength in dBs. Like most phones, it indicates it to the user by showing “strength bars”. At present nobody outside Apple knows how those strength bars work. The may include data other than strength too, such as signal to noise ratio. However, the various videos in the internet do show the signal strength dropping from 5-bars down to no-signal. Normally when a phone gives “full-bars” that means it’s receiving at least 10dB more than the minimum it can cope with. Often full-bars means that the phone is receiving 20dB more than the minimum. This indicates that the performance degradation for the iPhone 4 is especially bad. I can’t be sure of that, but that’s my view from the available evidence.

There are several ways that this problem could occur. Firstly, the loss introduced by placing a hand around the phone may be especially high. I don’t think this is likely because there have been demonstrations on the internet that show the problem is associated with a particular area on the phone. If the plastic band at the bottom left is touched then the performance deteriorates. The most likely explanation for this is associated with dielectric properties. If there is a lot of fringing capacitance between the two pieces of metal at either side of the gap then placing a dielectric object – such as a finger – across the slot will radically change the capacitance. That it turn will change detune the antenna and reduce the performance in some bands.

Some folks have suggested that the problem could be due to conductivity. They’ve suggested that the moisture on people’s fingers may cause a current path between the two pieces of metal. This is possible, if it did happen it would certainly cause big problems. Some folks on the internet have put a coin across the gap and showed that this causes performance degradation. However, there are some videos now showing the problem occurring with iPhones in rubber and plastic cases. These cases are unlikely to conduct well at GHz frequencies, so I think the dielectric explanation is more likely.

Some people aren’t reporting any problems with their iPhone. That’s not really so surprising since lots of people live in areas of high signal-strength (see my post on networks). Also, if dielectric detuning is the problem then it will not affect all bands and channels equally. It may be that some bands are entirely unaffected. So, we should expect only a subset of users to be affected.

Starting with the iPhone 4 Apple have started offering a cover, called a “bumper”, for the iPhone. It’s interesting that the experiments with covers and outer cases have been inconsistent. Sometimes the cover has solved the problem, sometimes it hasn’t, sometimes it’s helped a bit. This could be because the various covers are different, some dielectrically load the critical parts of the antenna, and some don’t. Perhaps some allows the users hand close enough that it causes dielectric loading but others keep the hand further away.

Lastly, an answer to something that has puzzled a few people – why does the signal strength display change so slowly? That’s because these sort of displays use averaging. They average the signal strength measurements over the last few seconds. So, even if the signal strength changes quickly the display doesn’t. This is a feature to make the display more user friends, all handsets do it.

New types of antenna are being invented all the time. Existing types of antenna are constantly being adapted to new roles. But, in most cases these efforts aren’t successful, most new antennas are outperformed by existing designs. The existing well-known types of antenna have their own niches. When designers attempt to apply them outside their normal niches in other niches where other antenna types predominate this usually fails. Even if the antenna works in its new application its normally found that the type normally used in that niche can outperform it. These stories are the tails of the failure of many research programmes and start-up companies.

The new antenna in the iPhone 4 is surprising and impressive because it’s designers have overcome these barriers. Most new handsets use one of the common types of antennas, the monopole, the vertical helix (or stubby) antenna or the PIFA. The antenna design is almost always new for every new handset, but it’s based on an existing archetype. When a new type of antenna being developed it will normally be released first in an obscure low-volume product.

In the iPhone 4 Apple have imported a new type of antenna into the handset niche. They’ve done this on their headline product, not just some minor experimental product that could fail without financial consequences. Steve Jobs mentioned the antenna in his keynote speech, he said that the stainless steel rim around the edge of the phone forms the antenna. He put up a slide showing it:

This slide doesn’t make it clear what type of antenna we’re dealing with here because it doesn’t show the feeds or the ground-plane. I think I’ve figured it out though by looking through some of Apple’s patent submissions. Apple have patented a lot of antenna designs, but two patents in particular looks very like the antenna in the iPhone 4 slides and teardowns on the internet. The first is US patent application #20100123632 and the second a granted US patent #7551142 and these describe types of slot antenna.

Like most patents it doesn’t describe the relevant ideas very clearly. So, I’ll explain it in my own way…. A normal slot antenna is a narrow rectangular hole cut in a large conductive surface.

The slot antenna is closely related to the dipole, it is the “complement” through the Babinet-Booker principle of the dipole. There’s an analysis of slot antennas from this perspective in Kraus & Marhefka. These simple slot antennas are used in antenna arrays for radars and various other types of electrically large antennas. But, they aren’t generally used for electrically small antennas. (Some folks may know of “slotted-PIFAs” these are PIFA antennas with slots in them but they aren’t slot antennas their design is very different). Slot antennas can be dielectrically loaded by filling the slot with a dielectric material such as plastic or ceramic.

The patent application suggests that Apple have designed a slot antenna that runs around the edge of the handset. The introduction says:

The antenna may be a slot antenna having a dielectric slot opening. The slot opening may have a shape such as a U shape or an L shape in which elongated regions of the slot run parallel to the edges of the portable electronic device. The portable electronic device may have a housing with conductive sidewalls. The conductive sidewalls may help define the shape of the slot. Antenna feed arrangements may be used to feed the slot antenna in a way that excites harmonic frequencies and that supports multiband operation while being shielded from proximity effects.

In the middle of phone there is a circuit board, battery, and display. All of these can be connected together, so that at RF frequencies they behave quite like a large block of conductive metal. This is common practice in handsets, as far as I can tell this is what’s being done in the iPhone 4. Around that there is a metal sidewall. The patent suggest that using a configuration like this slot antennas can formed from the gap between the metal sidewall and the internal conductive block.

The diagram above shows too slots both feed slightly off centre. One runs around the top and left side of the handset and the other around the rest. There are several ways of feeding such an antenna and the patent describes some of them. The simplest is a feed between the sidewall and main chassis driving a voltage between them. In the diagram above I’ve put on feeds and places where the slots end, these are just an illustration of the principle. I don’t know for sure if Apple have used this patent, if they have I don’t know where the feeds are. The patent explains that non-conducting front and back covers can be used to conceal and protect the antennas and internal electronics. Since these are covers of a dielectric material, that means there will be some dielectric load, and the antennas will be dielectrically loaded slots.

There’s lots more to talk about here. I haven’t covered much of the ideas in patent #7551142, I’ll do that in a future post. If I have time I’ll compare this to more conventional types of antennas like PIFAs. Also, I’ll comment on the problems recently uncovered with iPhone reception very soon.

In the WSJ article on the Apple iPhone antenna Ian Sherr mentions problems with dropped calls. I often here about dropped calls when I’m talking to Americans about cellular technology, but I hardly ever here this from Europeans. There are good reasons for this. Before discussing the iPhone antenna I think it’s useful to explain the differences between cellular networks across the world.

Suppose you work for a fast-food chain selling hamburgers or pizzas, you are a manager and you have to decide where to locate your next restaurant. You could put it in a small rural town. In that case there would only be a few other restaurants competing with it. The land to put the restaurant on would be cheap. But, on the other hand, there wouldn’t be many customers. Alternatively, you could put it in a big city. In that city there would be lots of competing restaurants and the rent would be high. But there would be lots of potential customers, if you could out-compete the other fast food restaurants then the potential for profit is very high. In both cases the restaurant itself would be quite similar, though you many need a few more seats, staff, friers and parking spaces if the restaurant were in the big city.

Cellphone network companies are in a similar situation. A cellular network in a sparsely populated area may have little or no competition. It’s relatively cheap to buy the land to put the cell towers on, but the number of potential customers is low. In a big city there is lots of competition and the land is expensive. However, the number of potential customers is very high. This is the logic that drives much of the cellular network business.

When cellphone networks first became practical they were applied first in the most densely populated cities in the richest countries. This was purely for the economic reasons I mention. The big cities allowed early cell networks to cover a large customer-base with few masts. Then, in time the cellular networks spread into more sparsely populated places and poorer places. These forces led to two distinct technical issue “coverage” and “capacity”. What concerned the network in big cities was capacity. They had lots of customers and lots of data to move. They had to make sure that the networks could absorb more customers without bandwidth limitations being reached. The people dealing with more sparsely populated areas had other concerns, they wanted to build a network to cover as wide an area as possible as cheaply as possible. They needed longer range base-stations and cheaper ones.

So, we could draw a capacity/coverage map of the world, and it would approximately follow population density. For example, in places like central Tokyo capacity is the concern, in the American countryside coverage is the concern. Lots of places fall somewhere in-between.

These problem have driven many specific technical solutions. For example, there are base-station antennas that are specially designed to deal with the situation where the network must cover a long road, but need not cover the adjoining countryside. A special type of small cell, the microcell, has been created to deal with areas where capacity is important. Many features of WCDMA are targeted at providing high capacity.

Whole countries can be looked at in a similar way, and this is where history comes in. In some countries the population is centred on large densely populated cities. In Japan, Korea, China and Spain it’s quite normal for people to live in apartments buildings in large cities. In some of these places the local culture and the law encourages this. In Britain there are strict planning regulations and taxes on transport fuel to prevent the growth of suburbs. In the US and Canada, for example, the culture is different people generally prefer to live in suburbs and the planning laws are set up to encourage use of the car rather than discourage it. This leads to different problems for network operators. In the US and Canada there are large areas of moderately populated surburbia, with some areas of high population and large areas of low population. European and Asian countries have areas of low population too, but the cities and suburbs are more highly populated.

This has direct effects on the technological problems. In areas where coverage is important the performance of the handset is more important. Below I’ve drawn a rough diagram of the differences between cellular networks, the scale is the same here…

In order to tackle the capacity problem in the densely populated Spanish city the network planners have used a large number of small cells. In the American suburb the planners have used a few large cells to serve their more spread-out customer base. In the diagram of the US case the cell tower is in the centre. It has four antennas on and four sets of base-station electronics producing four cells which I’ve marked in red, yellow, green and blue.

This has a strong effect on the cellphone market. In the US case a handset at the far side of the red cell is a long way from the base-station. So, the signal from the base-station is weak, and the signal received by the base-station from the handset is weak. If the receiver, transmitter or antenna on the handset is poor then it will drop calls on the edges of the cells. This isn’t the case in the Spanish city situation. There the cells are very close together and a cellphone would always be close to one of them. So, the quality of the RF electronics on the handset is less critical. Indirectly, coverage problems make the performance of the RF electronics (the antenna, low-noise amplifier, power amplifier, etc) more important. This is why Americans so often complain about dropped calls and bad coverage.

Cellular network operators worry about this because customers often blame the network for dropped calls. So, one way or another they encourage customers to use handsets with good RF electronics and encourage manufacturers to make them. A common method used by US carriers is to offer “good” handsets in their stores and to subsidise buying those phones with a contract. European cellular networks look at things differently, they don’t have to worry so much about coverage and dropped calls. The cellular networks in more densely populated places also offer special subsidies to the handset manufacturers but they’re often for different purposes. In that case the cellphone maker is encouraged to make a phone where the user has easy access to features that earn the network operator high profits. All network operators are “open to negotiation” on this point. If they think a new feature or a new phone will generate high revenue per user then they may be willing to sacrifice RF performance and put up with a few more people complaining about their network.

I was quoted very briefly on the Wall Street Journal website today. The subject was the new antenna in the Apple iPhone 4. Tomorrow I’ll write some more about this interesting antenna.

With a problem like blocking the engineer has to come up with an equation to fit the details of the problem. A good approach is to start off with a general equation and work from there. We can work out what the equation should look like by considering the parameters we need.

To do this I’ll start at the radio being blocked, the victim, and work backwards. At some particular power the performance of the victim receiver will fall due to blocking. Let’s call that X.

X = Blocking power in dB at the victim receiver which will affect the system performance.

How much power gets to the victim receiver though? Going from the victim receiver there will be a cable or feeding network, then an antenna.

L1 = Loss of circuits and interconnects between victim receiver and the antenna it is connected to.

Instead of going into great detail about the antennas here I’ll just define an aggregate figure…

C = Coupling between the antenna of the victim receiver and the antenna of the blocking transmitter. This is a coupling between the two antenna ports.

Then we have a loss corresponding to L2 at the other side…

L2 = Loss of circuits and interconnects between blocking transmitter and the antenna it is connected to.

And finally…

P = Power output of the blocking transmitter.

These parameters lead directly to the equation we need:
P – L2 – C – L1 = Y

If Y > X then there will be system performance degradation due to blocking.

The mismatch loss associated with the transmitter and receiver can be put into L1 and L2. All of the parameters I mention vary across frequency, the frequencies that are relevant here are those that the blocking transmitter produces.

In my opinion the architects of LTE have made some mistakes, this isn’t uncommon, many new protocols have had problems in their early days. The system they have created is susceptible to front-end blocking in particular bands.

Consider the band layout:
* LTE band 7 will operate from 2500 MHz to 2690MHz. This is split into an uplink band from 2500MHz to 2570 MHz and a downlink band from 2620 MHz to 2690 MHz.
* LTE band 40 will operate from 2300 MHz to 2400 MHz.

In between these bands is the 2.4GHz ISM band, which runs from ~2.4GHz to ~2.5GHz, the exact frequencies of edges of the band depend on the country. Many systems use this ISM band, WiFi and Bluetooth use it. More obscures systems such as video transmission systems and proprietary wireless keyboard and mice use it.

The diagram below shows the situation:

Yesterday I described the front-end blocking problem in general terms. You don’t have to be an RF genius to understand the problem the LTE architects have created here. The front-end receivers for LTE band 7, LTE band 40, WLAN, Bluetooth are fairly wideband, they will each respond to signals from all of the bands I’ve shown in the diagram.

GSM and WCDMA are “one band at a time” radios. When a GSM radio is using the 900MHz band it doesn’t use the 1800MHz band, or any other band, at the same time. The baseband controllers switches the radio between the various bands, sometimes it does this so fast that it can receive from the 1800MHz band while in call on a 900MHz band. This is achieved by time-division though, the radio never uses both bands at the same time. LTE radios are the same, that means that if a device is transmitting on LTE band 7 then it will not be receiving on LTE band 40. There is little chance of blocking problems within the LTE system.

However, the other radio systems involved, such as Bluetooth and WLAN, are controlled by completely separate baseband controllers. Let’s suppose there is a cellphone that has LTE band 7 and Bluetooth. When the Bluetooth module transmits then that signal will be transmitted from the Bluetooth antenna and received by the LTE antenna from there it will then enter the LTE receiver. Let’s suppose the base-station is transmitting to that same LTE receiver at the time too. As I described in my earlier post on blocking the front-end is unlikely to be able to cope with both of these signals. This will probably result in the LTE information being destroyed. The same applies in the opposite direction, the LTE transmissions can block nearby WLAN and Bluetooth receivers, especially if both are integrated into the same device.

In future posts I’ll discuss what can be done about this.

If you’re an RF engineer and you haven’t heard of front-end blocking then you’re lucky. I suspect you will hear about it soon, during the last few years I’ve done a lot of work on this problem. The diagram below shows the basic problem. A radio receiver only needs to be sensitive to the band of frequencies that it must receive, there is no need for the radio to have any sensitivity to other frequencies. However, that doesn’t mean that a receiver will reject incoming signals that are out-of-band. Most designs of receiver front-end are sensitive to a much broader range of frequencies than the band in question.

Suppose that a signal is being transmitted in the frequency range marked in green on the picture above. That signal will be captured by the receiver front-end along with the wanted signal. Then the receiver may cause the two signals to interact garbling the information from the wanted signal. There are several different ways that this can happen in different parts of the receiver circuit, I may write about those another day.

In this scenario the interfering signal isn’t in-band, it’s out-of-band, but it still causes a problem. For the sensitive receivers used in modern digital communications systems the power of the interfering signal doesn’t have to be high to cause reduced performance or complete link failure.

There are six general approaches to mitigating this problem:

• Move the receiving antenna away from sources of interference
• Break the path between the source of interference and the receiving antenna
• Make the receiving antenna more narrowband
• Add in a band-limiting filter between the antenna and the receiver
• Make the receiver less sensitive to out-of-band signals
• Change the receiver design so that the wanted signal and out-of-band signals don’t interact or interact less

I’ve used most of these strategies, in the future I’ll write about them in more detail.

.

My first job in Antenna design was designing internal antennas for cellphones. The cellphone market is enormous and every handset needs at least one antenna, and often two. It’s important for many of the people involved to understand the problems, so this is a very basic overview.

Due to the differences between handsets nearly every new model requires a new bespoke antenna design. The volume allotted for the antenna is different in each new handset, and it’s a different shape. The circuit board and the metal parts of the handset provide a groundplane for the antenna. This groundplane generally changes in size and shape with every new model. The groundplane characteristics affect the antenna significantly. An antenna built for a particular groundplane will not work well on a groundplane with a different size and shape. The plastics surrounding the antenna, and other nearby components also affect the antenna’s parameters.

Because of these issues it is virtually impossible to reuse designs directly. It can sometimes be done with external antennas such as whip antennas and stubby antennas, but not with the internal antennas commonly used today. It can sometimes be done with Bluetooth and Wi-Fi antennas, but not often with the main antenna for the cellular service.

So, Antenna Designers use a different sort of reuse. They gather knowledge about certain types of antennas, such as PIFAs and sub-types of them, they then learn how those sub-types behave and what factors affect their performance. An Antenna Designer learns how to “tune up” a few types of antenna. This allows him or her to quickly respond when given a particular product to work on.

To make this process as efficient as possible the major antenna companies have invested in rapid prototyping equipment, rapid antenna measurement equipment and other infrastructure.

RF engineering and Antenna engineering are often treated as Voodoo, as a “Black Art” akin to witchcraft. But, it is undoubtably scientific.  There are dozens of books and hundreds of journal articles on every aspect of it.  Every month there is a new issue of the IEEE journal “Transactions on Antennas & Propagation” and a new issue of “Microwave Theory and Techniques”.  If you liked you could accumulate thousands of pages of papers on the subject every month.

But, despite that information the subject continues to be veiled in a fog.  Professional Antenna Designers especially continue to use old tricks that aren’t documented anywhere.  This is because there is a gulf between engineering practice and research, to an engineer working on a specific problem little of the literature is useful and understanding what is useful is very time consuming.  So, in practice engineers fall back on their own knowledge and on experimentation because doing that is more effective.

I aim to promote my services as a consultant in Antennas and RF through this blog, but I won’t be doing that through posting ads for myself.  I will do that by bringing together the challenges of these branches of engineering and looking at them from various different perspectives. I hope doing so will help demystify the subject.  I specialise in antenna design for portable devices and notebook computers, I’ve designed and specified antennas for these products for many years and I’ll discuss that a lot.

I look forward to comments on these posts from other Antenna & RF engineers.