Let the games begin!

They've been at the game for years and so far have not been able to do so. (MacBook etc has very poor performance wrt wifi antenna)

Do it many times, and you'll probably see that it is the infrastructure's fault--that the call fails equally on the cheapest phone or any other phone.

AT&T infrastructure sucks.

You mean the Apple ads that came out in JANUARY?

nonsense. how conveniently you forget that htc infringed not only on apple intellectual property, but microsoft too.

that's not a controlled test in the least. there are way too many variables to be able to say it's *only* the antenna.

they have plenty.

WHAT?

NAVAS has a phone that ISN'T from MOTOROLA???

What happened, Navas? Moto didn't come through with the swag? Budget cuts, you suppose?

Hint: I do some antenna design in my spare time:

A log periodic antenna will not fit inside an iPhone. It's also rather directional, which is not what's needed in a cell phone. With small antennas, the efficiency is what usually goes down the drain first. Directionality goes next, as small antennas like to radiate in a spherical pattern, which mangles the SAR test. Most cellular antennas are loaded monopoles, PIFA, fractal, or just plain squashed into a pretzel. If I'm lazy, I just input the target specs into the modeling program, fire off the optimizer, and leave for a short vacation. When I return, I have a weird looking, best effort antenna, which usually works amazingly well.

Incidentally, I have built LPDA antennas on e=10 ceramic substrates, which will reduce the size by about a factor of 4. However, for

800-1900MHz, it's still to big for inside a cell phone.

Close. That's a conical dipole. Inverting the cones results in a bi-conical, which is an excellent broadband dipole. However, it has the same problem as the LPDA. It won't fit.

That's a discone. It's half a bi-conical, with a capacitive hat. It's one of the worst antennas ever sold, were it not for the advantage of having a huge bandwidth. For a base station, if you don't mind having all of your RF going well above the horizon, it's fine. That's why you see it at airports.

Hmmm.... 5 different antennas. Here's some exploded views of the iPhone that might help you identify which of the 4 antennas is in the iPhone 4:

Hint: The stainless frame is not a patch antenna because it's no suspended above a ground plane. It's also not a loop because it's broken into 2 seperate antennas (one for BT and Wi-Fi, the other for cellular). There's no meandering line or slot.

However, the iPhone 4 antenna might be a PIFA antenna, which is the "inverted F" in the picture. I can't tell from the photos or the FCC page. When I get my hands on an iPhone, or better photos, I can make a determination. I wouldn't be surprised if the frame antenna is just a piece of wire with a messy matching circuit.

Yep. UWB antennas are difficult but possible. However interesting, the iPhone 4 is not a UWB device and only needs to operate on 800,

900, 1800, and 1900 MHz.

That's a horn antenna. I'm trying to visualize how one would use a horn antenna with a cell phone. Certainly they're broadband, but are more suitable for illuminating a dish antenna, than cramming into a cell phone. Same problem as the others... too big and too directional.

It is difficult in a cell phone. If you don't care about size, shape, directionality, SAR, gain, and price, it's easy.

Hint... do some searching in the IEEE Antennas and Propagation Journal:

Type "cellular antenna" into the search box near the top of the page. You won't get the contents, but the abstract should be sufficient for this exercise. 3800 articles.

That's true for cell sites. I'm somewhat involved locally in the process. However, I don't think that politics had anything to do with the design of the iPhone 4 antenna.

No kidding. Since the composite AT&T map is now 9 months old: (2.6MB) I decided it was time to throw together an updated version. But first, I decided to see what it claimed about local coverage. I know where the local AT&T sites are located, and which way the panels are aimed, so that's easy for me. I was aghast. Not only does it claim coverage where I know none exists, but the islands of strong signals, which should correspond to the cell sites, is all wrong. At first I thought it might be a mapping offset (like conversion from NAD27 to WGS84), but it's almost random. I don't know how it's generated, but it's wrong.

I also noted that AT&T changed the colors on their "data" map, so that

3G is dark blue, while EDGE/GPRS, and roaming are slightly lighter blue with some black cross hatching. It's very difficult for me to separate these due to the (intentional) choice of indistinguishable colors.

Incidentally, I blundered across this site, which has links and screen grabs of many cellular maps:

If AT&T is bad, Verizon is the same, only in a different way.

Although Verizon map color differentiation is much improved over AT&T, you can't see any gradations unless you zoom down to the city level. There's no way to get a regional picture of coverage. From the region view, it looks like Verizon covers the entire area, which is at least locally, is grossly incorrect. Zooming in shows little detail. There are only 3 grades of service (full, marginal, none), which is useless.

Extra credit for Verizon still showing analog coverage on their map.

AT&T: We know, but we'll make it difficult for you to tell. Verizon: We know, but we're not telling.

Personal experience, and some info from Verizon dealers, employees and contractors. I don't want to reveal my sources. Note that dropped calls is their number one technical complaint (as of about 2 years ago).

To be fair, I don't know for sure if they're really intentionally dropping calls. I do know that I can't maintain a call for more than about 10 minutes at 5PM no matter how strong a signal. I was literally under the VZW cell site antenna when I was dropped. At other times of the day, that doesn't happen. I don't make too many calls at 5PM over 10 minutes, so my experience is limited. However, I deal with other local users that have had similar experiences. I really doubt that Verizon or any other carrier will admit that they're doing this but I do consider it a good way to free up channels when the system gets busy.

Cell breathing could easily cause the same effect. As the number of users increases, the base line noise level goes up. However, the way the cell site defends itself is to refuse to accept new connections until the noise level decreases to tolerable levels.

See graph on Page 5. In theory, by blocking new calls, the noise level will never exceed some pre-defined level. Were sites located with some overlap, and the adjacent site has available channels, breathing should not create a dead zone.

More:

"Again, its rarely the cell phone that is the cause of dropped calls, it is the wireless network including its towers and switches."

No, I didn't. I couldn't find anything of the sort with various searches. Got a link? All I could find is that Obama is "backing" the release, whatever that means:

Obama said the government would collaborate with the Federal Communications Commission (FCC) to make available a total of 500MHz of federal and non-federal spectrum over the next ten years, suitable for both mobile and fixed wireless broadband use, to be licensed and made available for shared access by commercial and Government users.

Note the "federal and non-federal spectrum". That's about as vague as any politician could make it. I would guess that includes all spectrum.

Never mind that any change in allocation will require ITC/WRC approval before it can be reallocated. At best, maybe 5 years for minor changes.

The last and only release of federal spectrum was the shared release of 3650-3700Mhz in 2007, a paltry 50MHz.

Unfortunately, it's useless within 50 miles of major metro areas thanks to the need to protect Sprint satellite uplinks. Where the FCC is going to find 500MHz of spectrum of uncontested spectrum needed for mobile and fixed wireless (actually for auction), is anyone's guess.

Gee whiz I doubt that. Considering I am thousands of KM away from them living in the land of Oz I think it more likely it is to do with the design of the phone.

And with laying down it doesn't change

David - who wonders where people get such strange ideas from...

John, I live in a rural area where there is only one provider, Telstra. So it is the phone

David

1. Increasingly out of date.
2. Not many people care, and the number continues to dwindle.

That's probably true of all major carriers (for different reasons).

Apocrophal story issues aside, while it may seem that being right under a tower would give you the least liklihood of being dropped, it actually tends to be just the opposite, since that location would tend to have the least good coverage from neighboring cells, and is probably outside the gain lobes of sector antennas.

I think it extremely unlikely: 1. The story would inevitably get out from disaffected employees, resulting in a firestorm. 2. Customers are more tolerant of connection problems than of dropped calls (c.f. above). 3. Call blocking is the standard method for managing load.

CDMA2000 issues are load and available capacity, signal strength, and call blocking (not dropping) policy. Channels are more a TDMA (GSM) issue.

What makes that an uathoritative article? The bit about switches getting overloaded and dropping calls is quite unlikely according to my friends at Motorola. Looks to me more like folk wisdom from D-AMPS days than current technology. Doesn't even mention the problem of CDMA cell breathing, a significant problem that I still think the most likely cause of dropped calls on CDMA2000 networks.

Good presentation on cell breathing:

A good deal of it is apparently expected to come from broadcasters who would voluntarily agree in return for compensation from the auction.

It can be shown easily that many different phones will show changing signal strength in many places.

It can be shown easily that many different phones will show changing signal strength in many places.

Sidekick was similar, except that data was lost after a reboot.

Although it's not as easy a reboot to lose your data in the event of an Exchange failure, if the Exchange server comes back up with a blank mailbox (using the same credentials) you'll lose all your data on the device. Even worse, should you attempt to remove the account from your device or change it to a new server, you'll again lose all your data.

The model of treating the server as authoritative is a good one in general (SyncML is an example of the disaster that happens without an authoritative owner) but it relies on the server administrators having proper backup procedures in place. This isn't too difficult (and is several orders of magnitude easier to implement than having ever user attempt to backup their own data)

The Sidekick data was recovered. The problem was incompetant administration by Microsoft, not the technology.

Not if it's backed up properly.

SyncML works quite well.

Yep.

There was backup. There was also fault tolerance. The problem was incompetant administration by Microsoft, not the Danger technology.

It makes sense to maintain your own backup, just as it makes sense to replicate your backup in the cloud.

actually they decided to upgrade the system without having an up to date backup in place and had to reconstruct it to recover most (not all) of the data.

and yes, it was incompetence by microsoft.

depends on what the data is. some things *can't* be in the cloud.