These are my findings from another forum I post on, feel free to pull them apart as you see fit.
Here are my findings from some experimentation, based on being at home last night and at work now. I live and work in Bristol, South West England. Work is a 10 minute walk from home, so I am probably using the same cell site at both locations. My phone is on 3's network.
Last night, when holding my iPhone, I experienced no issues with signal dropping. At work this morning, I did, while holding the phone in an identical manner. Now, my office is notorious for having a bad signal indoors, as we have tints on all the windows and bars across them for security. In other words, we are a pretty good Faraday cage. I held onto my phone in my office and saw the signal drop from full to none. I then walked outdoors, still holding the phone in the same way and the signal came back up to full. This happens when covering the outer edge of the iPhone in either a rubber band (similar to the Bumper) or paper from a post-it note.
So, my conclusions are this. Apple have pegged the signal strength meter on the iPhone very low (witness the lack of the engineering mode previously accessible by *3001#12345#*, which would tell us the actual received signal strength). There is no industry standard convention on phone signal strength displays correlating with actual received signal strength. Typically, phone signals vary from around -45 to -55dBm when stood right by a mast to -100dBm or worse when in poor coverage. Anything over -100dBm will be prone to dropped calls (I'm basing this on my experiences working with field engineers for Orange during my graduate training in the late 90's).
A manufacturer might decide that on their phone, anything better than -65dBm is full strength, -75dBm to -65dBm is 4 bars, -85dBm to -75dBm is 3 bars, etc, but it is purely arbitrary - there is no defined standard. What I suspect Apple have done is to peg their values much lower - perhaps they consider anything better than -85dBm to be full strength, whereas a Nokia or Sony Ericsson might show that as 3 bars of coverage.
So, my iPhone could be showing good coverage, even though it is only average in reality. I then pick up my iPhone in my office and my hand adds 20dB of attenuation (for example). Suddenly, the actual field strength drops from -85dBm to -105dBm and the signal meter shows it plummeting and I get no signal. Yet outside, I might have -65dBm signal, I pick up the phone (adding 20dB of attenuation) and the signal drops to -85dBm, still above Apple's thresholds for a full signal.
I'm not excusing Apple's design in any way, as there is a reason the original mobile phones used to have their antenna at the top of the phone, where it wouldn't get blocked by a hand. What I am suggesting is that Apple are disguising the iPhone's lesser ability to retain a good signal (in layman's terms, it's a bit deaf), by setting the signal strength indicators on the phone unusually low in real terms.
n.b. all figures and values used in this illustration are purely arbitrary, I don't have the means to measure actual signal strengths or indeed how much attenuation the human hand adds, but it does explain why some people are not seeing the problem at all (they already have a good signal and their hand does not cause it to drop below Apple's threshold for full bars). It also explains why some people still see the problem when using bumpers - it's the hand holding the phone that causes the problem, not bridging the gap between the different antennae on the iPhone (I tested this by insulating the gaps between the antennae, it made no difference).
