Mitigating the Problem with Tape/Gloves

I originally tested the iPhone 4 in a number of different positions in the hand, and in a bumper case, and generated the signal strength drops reported in our previous article.

Signal Attenuation Comparison in dB - Lower is Better
  Cupping Tightly Holding Naturally On an Open Palm Holding Naturally Inside Case
iPhone 4 24.6 19.8 9.2 7.2
iPhone 3GS 14.3 1.9 0.2 3.2
HTC Nexus One 17.7 10.7 6.7 7.7

After getting those numbers, my first thoughts were that two dominant effects were responsible for the iPhone 4 signal drop being measured. The first was detuning due to capacitance added by the hand making galvanic contact with the stainless steel, and possibly even coupling the two discrete antennas together. The second was simply attenuation due to our meatbag extremities (read: hands) being not perfectly transparent to RF at 850 MHz and 1.8 GHz. I made some Star Trek references that some of you caught about us being bags of mostly water - it’s true, and it’s something Apple has emphasized heavily in its letter - that all phones drop signal when your hand is in between the path to the base station antenna. The real question was how much of that 24 dBm drop was due to galvanic contact with your capacitive hands (detuning), and how much was due to your hands being mostly water, and so close to the radiative surface.

Almost immediately after the issue was identified, many took it upon themselves to apply adhesive insulative tape to the troublesome area. Others suggested testing with rubber gloves on to see how much the issue changed. I set out to test both.

Allow me to introduce you to my friend Kapton tape.

No, it isn’t a gold iPhone 4, nor have I dipped the stainless steel band in gold (you have no idea how many people have asked) - it’s the native amber color of the world’s most awesome tape. I managed to find a roll of 1-mil thick Kapton tape that is exactly the right width of the iPhone 4’s stainless steel band. It’s miraculous really how exact the match is, without any cutting or tweaking, it just fits.

What makes Kapton the most conclusive choice of tape here ever (and not your grandpa’s electrical tape or duct tape) is that it’s the industry standard for flexible printed circuits. In fact, it’s what’s used to insulate just about every flex PCB antenna around. The tape obviously has a huge impedance, so when I hold it, I’m insulated completely from the stainless steel band.

I wrapped the tape all the way around the phone - not just the lower left trouble corner - to guarantee complete insulation. Of course, it’s impossible to use the phone like a phone this way since you cover the dock connector, speakers, and microphones, but just for testing. The other thing the tape simulates is how the iPhone 4’s antenna would behave with a thick 1-mil (25.4 µm) coating. To test, I cupped the phone just like I did to cause the 24 dBm drop before.

I also took an ordinary natural latex glove (yes, really) and held the iPhone, this time without any Kapton tape wrapped around the phone. Nothing special here, just a laboratory glove and cupping the phone.

The results speak for themselves.

Signal Attenuation Comparison from Cupping Tightly in dB - Lower is Better
  Bare Phone 1 mil Kapton Tape Coating Applied Natual Latex Rubber Glove on Hand
iPhone 4 24.6 16.6 14.7
iPhone 3GS 14.3 N/A N/A
HTC Nexus One 17.7 N/A N/A

Instead of a 24.6 dB drop from cupping the phone tightly without a case, with bare skin, we see a 16.6 dB drop with tape all the way around, and a 14.7 dB drop wearing a rubber glove. Insulating the stainless steel completely from the hand completely results in 9 dB less of signal drop. The remaining 16 dB is then due to the hand being so close to the phone.

The takeaway is that the best coatings Apple could possibly apply would bring the drop down to 15 or 16 dB - in league with the Nexus One’s worst case drop, and almost in league with the iPhone 3GS worst case drop. It’s hard to argue that bringing the signal drop down to levels other phones have been selling with for a year now isn’t a problem solved type solution.

However, adding tape won’t completely eliminate the drop in received signal, nor does it mitigate the problem nearly as much as getting a case. In fact, if you’re really concerned about dropping signal on any phone, you should get a case anyways. It demonstrably reduces the signal attenuation added by having your hand so close to the radiative surface of the antenna.

Oxide on Stainless

I talked with a number of materials science wizards, and picked their brains about the possibility of applying a nonconductive coating to the iPhone 4’s stainless steel antenna bands.

The response I got back was that stainless steel is difficult to coat by very nature of it being “stainless.” The metal unsurprisingly develops a dull oxide which itself is a poor conductor, but forms a protective layer that resists tarnishing and corrosion. This same layer that makes the metal stainless makes it difficult to coat. Some grades of stainless are apparently much easier to coat than others, but nearly all grades would require abrasion or chemical etching, followed by vapor deposition of the coating.

It’s not impossible to coat though, and if rumors that new iPhones built in recent weeks are rolling out with coatings turn out to be true, it’s obviously being done. But coating the stainless steel bands is obviously something Apple had to have considered.

Update

I originally thought my Kapton tape was 5-mils thick, turns out it's 1-mil thick Kapton Polyimide, with adhesive for a total of about 2-mils of thickness. It's P-221 Permacel branded tape billed as the "ultimate" in electrical insulation.

Total Silent Recall?

Gizmodo reported (and iFixit followed up) on some users claiming that newer iPhone 4s had a different coating on their stainless steel band that mitigated the signal attenuation issue caused by tightly holding the phone. In theory, with the right coating, Apple could deliver the same sort of results we just showed using the Kapton tape. To date we haven’t been able to get our hands on one of these iPhone 4s with improved coating.

We found an iPhone 4 produced in week 28 of 2010 (digits 4 & 5 from the left of the iPhone 4’s serial number indicate production week) and took a multimeter to it. There was no measurable difference in resistance between it and our older iPhone 4s. In other words, the band was just as conductive. While this doesn’t rule out the possibility of Apple changing the manufacturing process on the phone, I wouldn’t waste time trying to hunt down a phone manufactured on a specific date just yet.

Proximity Sensor

Until two days ago neither one of us had experienced the proximity sensor issue with the iPhone 4. The proximity sensor on the iPhone detects if your face is close to the screen, like it would be during a phone call. If it does so, the iPhone turns off its screen to avoid any accidental input and save power. The proximity sensor issue manifests itself by the phone incorrectly assuming that you aren’t holding the phone up to your head and turning the screen back on. This happens in the middle of a call and often results in your cheek doing things on your phone without your knowledge.

Two days ago I was on a phone call when the proximity sensor all of the sudden decided that my face was no longer near the phone. My cheek then navigated into my contact list and tried to FaceTime with another contact while I was on the phone. I didn’t find out until the iPhone complained that a FaceTime connection couldn’t be established (due to the contact my cheek was trying to FaceTime with not having an iPhone).

I’m on the phone quite a bit and so far this was the first and only time the proximity sensor bug cropped up. We’re still looking into it but so far we can’t tell what the root cause is or if it’s helped by iOS 4.1.

Better at the Low End, Mixed Feelings Everywhere Else Final Words
POST A COMMENT

149 Comments

View All Comments

  • mfenn - Thursday, July 15, 2010 - link

    Brian,

    Can you update the article with the exact part # or a link to that Kapton in the appropriate width?
    Reply
  • Brian Klug - Thursday, July 15, 2010 - link

    Hey mfenn,

    Updated the article with a link and specifics about the tape. I also discovered that what I thought was 5 mil thick tape is the 1 mil thick Kapton - still an amazing insulator though.

    I didn't buy it online however, rather got it somewhere else locally. I don't know whether you can find that exact size online, but you can definitely cut it.

    -Brian
    Reply
  • Stokestack - Thursday, July 15, 2010 - link

    And maybe elaborate on why you think the new system is more accurate, considering that (in a digital system) a large range of signal strength does give you "full service".

    Shouldn't a phone show full bars for the entire range of signal strength that gives you totally effective performance? At the weak end is where you see performance variation, so that's where the bars should vary as well.

    What's the point of showing differing numbers of bars through a range that doesn't change the phone's performance?
    Reply
  • lunarx3dfx - Thursday, July 15, 2010 - link

    The difference, is that rf and digital are two totally different beasts. RF is like nothing you'll ever learn about. I have to deal with it a lot in the work that I do, and while a shift from -117 to -81 doesn't seem like that much, for every 3 dbm gain or loss, you are doubling or halving your signal strength (in wattage), but it still isn't that simple, for example a 10dbm signal will be expressed as 10 mW signal strength, a 20dbm signal will be 100 mW. You can't compare RF with anything else. It is it's own world. Reply
  • rushbc - Saturday, July 17, 2010 - link

    i understand about the RF energy being different than anything, it really is its own world (i know, i work for AT&T).

    but i think stokestack has a good point. if -94 dBm gives you great service and good speed for voice/data, then the end user shouldn't care if the display shows one bar or ten bars--their phone is working, transferring data and voice with no issues. so a large dynamic range of signal strength that shows the end user 5 bars does make sense. five bars simply indicates that you have enough signal to do what you need to do, and your device has enough signal to do what it needs to do.

    but since i do work for AT&T, and since i have talked to AT&T customers, i know that they DO care about the bars. i cannot tell you how many times people will tell me, "hey, i am only getting one bar on my cell phone at my office".
    and i say, "well, can you make and receive calls?"
    and they say, "Yes".
    then i say, "well, can you connect to the internet on your phone without any problems?"
    and they say, "Yes".
    then i say, "well, what is the problem, then, sir?"
    then they say, "but i am only getting one bar!" then i just smile to myself and try not to scream. it's almost like the famous old "who's on first" routine, but in real life.

    so i predict many users will be confused by the "new" bars. and some will be angry. but Apple is doing the right thing by displaying the "more correct" signal bars. honesty really is the best policy.

    Apple should really bite the bullet and completely ditch the whole bars thing.

    it would be another way for Apple to show great leadership, bravery, and innovation. lets just get rid of the silly bars, totally. we need Apple to just go with strict numerical dBm readout. sure people will be confused. sure people will complain. sure there will be a learning curve. but it would be a quick adjustment. and within two years every smartphone in the world would be doing the same thing with numerical readouts in dBm, not bars.
    it's called "standardization" (something completely lacking in signal strength indicators for wireless devices), and it would help the industry, and educate the consumers, and probably improve the wireless networks ability to respond to coverage needs and customer requests for network improvement. it would certainly help wireless consumers better understand how strong or weak their signal really is, and it could help people adjust their expectations accordingly.

    "Death to the Bars!"

    bring on the transparency of dBm numerical readouts on handset displays, i say.

    Disclaimer: I am an AT&T customer and an iPhone owner. I am also an AT&T employee. I do NOT receive any form of compensation for my posts. My posts reflect my own personal opinions and do not necessarily represent AT&T’s positions, strategies or opinions. I also do not possess any "insider knowledge" of future products or future releases. And if I did, I certainly wouldn't tell you! :)
    Reply
  • anandreader - Sunday, July 18, 2010 - link


    "I just smile to myself and try not to scream..."

    You're not listening to your customers. What they're telling you is that when the signal is weak, the voice quality is very poor. They may not be articulate enough to say that to you but that's the message that's being sent. You're just so fucking self-satisfied with your little syllogism that you're failing to hear the message. At&t's signal is crap in too many locations.

    Your attitude clearly illustrates why people detest At&t.
    Reply
  • thisma - Wednesday, July 21, 2010 - link

    Anandreader, I suspect that rushbc's interpretation is accurate.

    Occasionally a user will have a real problem but describe it using some idea which is quite remote from the actual problem. I find, however, a good match in my comparison between rushbc's statements and how I expect users do express themselves.

    My applicable knowledge, about how people communicate and what concerns users, has been developed during my several years of professionally teaching people how to use technology including, among other things, computers, phones, audio and video equipment, and all manner of peripherals.
    Reply
  • oPad - Friday, July 23, 2010 - link

    Wait a minute, it sounds like rushbc asked if there's anything wrong other than the bar strength display and they said "no",
    What's to articulate?
    Reply
  • thisma - Wednesday, July 21, 2010 - link

    Rushbc, I would love that! I fear however that the most prevalent perspective of users is that they don't care about precision of truth until they understand what that precision means. Even then a 5 bar display is so easy to interpret that a person doesn't even have to consciously check the indicator; if it's low you'll notice that out of the corner of your eye. My preference would be to have both the same way, perhaps, as you can optionally have the battery life indicator shown with the percentage of battery life next to it. Reply
  • bigboxes - Saturday, July 17, 2010 - link

    Sound waves act very similar. Reply

Log in

Don't have an account? Sign up now