On Tue, 13 Jul 2010 13:51:46 -0700, Jim Lux
wrote:

Not any more.. recent phones (last several years) put the antenna at the
bottom to reduce the SAR number, since the top of the phone is next to
your head, and the bottom isn't. There's a really good explanation from
a guy who does, oddly, wireless device antenna design consulting..
http://www.antennasys.com/
specifically
http://www.antennasys.com/antennasys...-antennas.html

http://www.anandtech.com/show/3794/t...one-4-review/2 has some
information

Here's my wild guess as to what's happening, borrowed from my posting
in ba.internet and alt.cellular.attws. Apple announced that there's
going to be a press conference on Friday, where they'll announce
something. Note that I goofed on the location of the wi-fi/BT
antenna, but I'll leave the mistake here for now.


On Wed, 14 Jul 2010 13:05:27 -0700, AES wrote:

Is your assessment here that placing RF-lossy human flesh (with a high
water content) close enough to the phone antenna loads down (or maybe
just detunes?) the RF circuitry of the transmitter final stage to a
point that the active transmitter circuitry in the phone no longer
oscillates effectively? -- but just putting a dry and non-lossy
dielectric there doesn't, and also moving the flesh even a short
distance away reduces its loading on the circuitry enough that the
active circuitry still works OK?

Nope. My wild guess(tm), based upon what little I can extract from
the FCC ID page and from the iFixit autopsy, is that something drastic
is happening in receive. The -30dB (that's 1000 times drop) in signal
appears in receive, when you're NOT making a call and the xmitter is
inactive. Well, the xmitter is sending keep alive bursts every few
minutes, but nothing more. Methinks it's a receive problem, not xmit.

I'll go out on a limb and suggest that the receiver front end
(probably a GASFET or HEMT) might be a bit regenerative (border line
oscillatory). This gives it lots of gain, but only if nothing else
changes. In effect, the antenna becomes part of an oscillator
circuit, where the oscillatory conditions are partly provided by the
antenna Q (i.e. antenna efficiency).
http://en.wikipedia.org/wiki/Regenerative_circuit

Touch the antenna, and you reduce the antenna Q. The front end stops
acting regenerative (barely oscillating), and the sensitivity drops
like a rock. It's the only explanation I can conjure the will cause a
-30dB drop in received signal. It's been demonstrated (by me and
others) that strangling the antenna of other cell phones drops the rx
signal up to -10dBm (10 times drop). That's what I would normally
expect to see from touching a cell phone antenna. However, the -30dB
drop of the iPhone 4 requires some extra circuit design screwups.
Unfortunately, detecting the regeneration is going to require internal
probes, test fixturing, and plenty of expensive test equipment.

There's another related possibility. Most of the finger tests have
been with the finger bridging the gap. That means they're touching
BOTH the cellular and wi-fi/BT antennas. It's obviously the hot end
of the cellular antenna, but I can't tell if it's the hot end or
ground end of the wi-fi/BT antenna. If the hot end, another
possibility it that coupling to the 2nd antenna causes this antenna to
radiate into the case, thus causing the cellular front end to
oscillate. If the oscillations are bad enough, the signal level will
be high enough to induce blocking (overload) in the cellular front
end. This can be easily detected with a spectrum analyzer probe in
the vicinity. If touching/bridging the antennas causes a new signal
to appear on the SA, that's the problem.

It would be very interesting to know if touching ONLY the cellular
part of the antenna (the part that goes up the side of the phone), and
not the wi-fi/BT antenna (bottom of phone), causes the same drop in
signal. It would also be interesting to test the wi-fi signal
strength to see if it's effected by touching its antenna.

I have a few other guesses, but these are the best I can currently
conjure at this time.

[I'm an EE, but with no RF circuitry experience, much less any cellphone
experience. So, I can understand this happening with an active
transmitting circuit -- though I have somewhat more trouble seeing how
it would interfere in such a sensitive way with a passive receiving
antenna.]

Think regeneration and borderline oscillation.

The bad news is that if Apple stabilizes the rx front end, and
eliminates the oscillation or regeneration, my guess(tm) is that the
resultant phone is going to have a serious rx sensitivity problem. If
you look at the cellular antenna in the iPhone 3G and other internal
antenna cell phones, they're not very simple looking devices. Lots of
strange lumps, traces, pads, and oddities, all designed to optimize
performance in the 800/900 and 1800/1900 MHz bands. Getting them to
look like 50 ohms for minimum VSWR is also important. Lots of
articles in the IEEE Antenna and Prop proceedings on the topic of
cramming a dual band antenna in the smallest possible package.

On the other foot, the iPhone 4 antenna is crude. As far as I can
tell from here, it's a crude monopole (single wire) antenna. There's
no matching circuit for VSWR reduction, and no attempt to optimize
performance in the desired bands. I might have missed something in
the dissection, but all I can see is an end fed monopole, without a
counterpoise or underlying ground. Such an antenna is going to have
resonances at odd frequencies, and miserable antenna efficiency.

Disclaimer: I haven't found anyone willing to let me tear into their
iPhone 4 yet, so all the above is guesswork.

--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

On Wed, 14 Jul 2010 23:02:21 -0700, Jeff Liebermann
wrote:

However, the -30dB
drop of the iPhone 4 requires some extra circuit design screwups.
Unfortunately, detecting the regeneration is going to require internal
probes, test fixturing, and plenty of expensive test equipment.

Hi Jeff,

Regenerative designs inhabit the naked edge of
amplification/oscillation. You've already said as much. It has to
be, thus, a design feature and not a hack. Otherwise every third one
is going to break into oscillation and the FCC (or test lab) would
have certainly picked up on that. Further, if it is designed in, you
can observe that in the design. Special testing methods are too
elaborate to explain this. Regen is not a design feature left to the
vagaries of production variation in parasitic coupling.

If I'm wrong and Apple does use parasitics to encourage regenerative
feedback, then they may introduce a lanyard for hanging the phone
around the neck and call it the lavaliere option. No doubt they will
sell plenty of mirrors to see it, and someone will come out with an
image reversal app. Can a stylus be far behind? Whoops! It will
have to be chopstick for squeezing and expanding images.

73's
Richard Clark, KB7QHC

On Thu, 15 Jul 2010 00:37:22 -0700, Richard Clark
wrote:

On Wed, 14 Jul 2010 23:02:21 -0700, Jeff Liebermann
wrote:

However, the -30dB
drop of the iPhone 4 requires some extra circuit design screwups.
Unfortunately, detecting the regeneration is going to require internal
probes, test fixturing, and plenty of expensive test equipment.

I goofed slightly. The Anandtech article shows -24dB maximum signal
loss.

Regenerative designs inhabit the naked edge of
amplification/oscillation.

Not designs. I don't think it was intentional. Someone else
mentioned that Apple was so much into security on the iPhone 4, that
prototypes were tested in a similar case made to look like an iPhone
3G. That implies that there was little field testing in the new case.

You've already said as much. It has to
be, thus, a design feature and not a hack.

I beg to differ. Nobody wants oscillatory effects in the front end of
their receiver. They're unstable, vary with temperature, are
difficult to control in manufacturing, and will probably make the
device fail Part 15 incidental radiation test. Designing a front end
that's unconditionally stable with any antenna load does not yield the
optimum sensitivity. If you want the best performance, amps that are
unstable at specific loads will give better NF and gain.

Otherwise every third one
is going to break into oscillation and the FCC (or test lab) would
have certainly picked up on that.

Maybe. However, one would think that someone would have done a field
test with the final iPhone 4 and noticed that touching the antenna gap
produced a major signal drop. Also, FCC test are not preformed with
human hands wrapped around the phone due to the limited supply of
dismembered hands suitable for testing.

Further, if it is designed in, you
can observe that in the design. Special testing methods are too
elaborate to explain this. Regen is not a design feature left to the
vagaries of production variation in parasitic coupling.

That might partly explain why some users claim that there's no call
drop effect when they touch the antenna gap. Of course, users in
strong signal areas are not going to see a dropout, while weak signal
areas are going to be much worse.

Incidentally, when Consumer Reports ran a test with the rubber ring
thing, they slipped and didn't have the technician stand in front of
the test equipment this time.
http://blogs.consumerreports.org/electronics/2010/07/apple-iphone4-iphone-4-bumper-case-fixes-antenna-issue-problem-signal-loss-tested-verified-consumer-reports-labs-quick-fix.html
http://802.11junk.com/jeffl/crud/iPhone4-cmu200.jpg (normal)
http://802.11junk.com/jeffl/crud/iPhone4-cmu200-hand.jpg (with hand)
It's a Rhodes and Schwarz CMU200.
http://www2.rohde-schwarz.com/product/CMU200.html
I'm not familiar with the operation. The "Reported Power" appears to
be the signal generator output power (because it doesn't change when
the antenna is touched). However, I can't devine what the "Avg" or
"Peak Burst Power" means.

Also, note that the difference in (something) between the first photo
(+16.4dBm) and the hand on antenna measurement (-11.7dB) indicates a
-28dB drop in signal level. Ouch.

If I'm wrong and Apple does use parasitics to encourage regenerative
feedback, then they may introduce a lanyard for hanging the phone
around the neck and call it the lavaliere option. No doubt they will
sell plenty of mirrors to see it, and someone will come out with an
image reversal app. Can a stylus be far behind? Whoops! It will
have to be chopstick for squeezing and expanding images.

We'll find out after the press conference on Friday. My suggestion
was to ship a rubber glove with the phone. I also suggested gluing a
plastic thumb tack to warn against holding it improperly. Also, a low
signal warning tone. Others have suggested tape, rubber prophylactic
bump guards, duct tape, warning labels, rebates, bribes, and external
antennas:
http://gizmodo.com/5581253/there-fixed
Unfortunately, that will require running the FCC type certification
tests again.



--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

On Thu, 15 Jul 2010 09:37:38 -0700, Jeff Liebermann
wrote:

http://blogs.consumerreports.org/electronics/2010/07/apple-iphone4-iphone-4-bumper-case-fixes-antenna-issue-problem-signal-loss-tested-verified-consumer-reports-labs-quick-fix.html

Consumer Reports to Apple: "You've been served!"

The long and short of it is that Apple treats its product design like
bulimic runway models. It's up to you to see they get as fat as
necessary to live a productive life.

In other words, there's no room left inside the iPhone for an antenna
and it ruins the esthetics for it too bulge. So, if you don't want to
be viewed as part of the pocket-protector crowd, suck up and hold your
phone like Lady Gaga.

Support our troops at:
http://www.youtube.com/watch?v=-fMP0zZRIF4

73's
Richard Clark, KB7QHC

Jeff Liebermann wrote:


The bad news is that if Apple stabilizes the rx front end, and
eliminates the oscillation or regeneration, my guess(tm) is that the
resultant phone is going to have a serious rx sensitivity problem. If
you look at the cellular antenna in the iPhone 3G and other internal
antenna cell phones, they're not very simple looking devices. Lots of
strange lumps, traces, pads, and oddities, all designed to optimize
performance in the 800/900 and 1800/1900 MHz bands. Getting them to
look like 50 ohms for minimum VSWR is also important. Lots of
articles in the IEEE Antenna and Prop proceedings on the topic of
cramming a dual band antenna in the smallest possible package.

On the other foot, the iPhone 4 antenna is crude. As far as I can
tell from here, it's a crude monopole (single wire) antenna. There's
no matching circuit for VSWR reduction, and no attempt to optimize
performance in the desired bands. I might have missed something in
the dissection, but all I can see is an end fed monopole, without a
counterpoise or underlying ground. Such an antenna is going to have
resonances at odd frequencies, and miserable antenna efficiency.


I doubt they're using regeneration or some such similar scheme (too hard
to control for manufacturing variability)..

However,
the idea of self interference is interesting... (e.g. your finger
couples the WiFi to the cellular front end, which blocks)

It could even be as dumb as a dc bias shift thing (save a penny by
leaving out the capacitor, oops, that removed the DC block)


As far as match goes, in very compact systems, the idea of matching 50
ohms goes out the door unless one of your components happens to be 50
ohm impedance.. It's not like they're stringing together building blocks
with SMA connectors inside there. Lots of these RF circuits are
implemented with differential in and out these days, and the Z is
whatever it happens to be.

On Thu, 15 Jul 2010 08:43:04 -0700, Jim Lux
wrote:

I doubt they're using regeneration or some such similar scheme (too hard
to control for manufacturing variability)..

I'm not suggesting they did it intentionally. I'm suggesting that the
front end is unstable and that nobody noticed.

However,
the idea of self interference is interesting... (e.g. your finger
couples the WiFi to the cellular front end, which blocks)

I was looking at the FCC ID web site photos again. I can't even tell
which end of the antenna is hot or ground, much less where it's fed.
If the hot end of both antennas are facing each other, it's certainly
possible. However, if Apple did that, they wouldn't need two separate
pieces of metal for the two antennas. They could simply declare some
point along the frame as being the official ground point, and have the
cellular antenna go one direction, and the wi-fi/BT antenna go the
other. Because they didn't do that, my guess is that the hot end of
the cellular antenna is next to the ground end of the wi-fi/BT
antenna.

It could even be as dumb as a dc bias shift thing (save a penny by
leaving out the capacitor, oops, that removed the DC block)

Methinks not. The duct tape solution just reduces the effect, not
eliminates it. Assuming duct tape to be a good insulator, that would
also suggest that it's not a DC effect.

As far as match goes, in very compact systems, the idea of matching 50
ohms goes out the door unless one of your components happens to be 50
ohm impedance.. It's not like they're stringing together building blocks
with SMA connectors inside there. Lots of these RF circuits are
implemented with differential in and out these days, and the Z is
whatever it happens to be.

If you look at the inside photos of the BCG-E2380A on the FCC ID web
pile,
http://www.fcc.gov/oet/ea/fccid/
(don't forget to include the "-" as part of the product code), there's
a module labeled "Acoustic and Cellular Antenna Feed" on Pg 9 and 10.
https://fjallfoss.fcc.gov/eas/GetApplicationAttachment.html?id=1300576
There's a coax cable going across the module, which I assume is 50
ohms.
http://802.11junk.com/jeffl/iPhone4/iPhone4-accoustic-and-cellular-ant-feed.jpg
It doesn't appear to do any more than connect a tiny coax connector to
the RF module on one end, and a gold screw lug to the frame antenna on
the other. No matching in between. Due to the coax, someone must
think the antenna is 50 ohms at 800/900 and 1800/1900.

I've never seen a differential RF amp in a cell phone. If so, I would
expect either a balanced antenna feed, or possibly two antenna
connectors.

Disclaimer: I'm doing quite a bit more guessing than usual. Please
do not treat the above as authoritive.

--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

Jeff Liebermann wrote:

I've never seen a differential RF amp in a cell phone. If so, I would
expect either a balanced antenna feed, or possibly two antenna
connectors.


Things like vector mods (RF2422) have differential in/out, for instance.
Parts like the MAX2510 are similar.

It's true by the time you get to PA stage, that single ended seems to be
more common.