Issue with Proximity Sensor during calls - Continued v2

Interesting. I haven't done anything with checking the pulsing and brightness of the sensor light, but I wonder why yours changed so much more dramatically than mine.

What week was yours made? Mine is week 37.

I have a feeling Apple may be making a number of small changes, especially if I find a trend in changes with week 46 or later phones that are like the only one I've gotten to test.

And they are not revealing them publicly. After all, I've tested enough to know that they made at least some change in the sensor for 4.2.1, yet I've never seen any official statements from Apple saying it was part of the upgrade. Could be a small hardware change(s) too. Not necessarily the choice of sensor either, though that could be part of it. They may have made a slight change in position or angle of the emitter or receiver to make it more effective. Or they could just have stepped up quality control.

Yet even if there are some hardware changes, we know they are also making software changes.

BTW, I've been trying to figure out why they would reduce power to begin with. You've said that even at full power, the drain on the battery would be infinitesimal. If so, then the only other reason I can see would be to keep from burning the unit out. Remember, even the older phones reduce the distance (well, at least my 3G does), but it is much harder and never gets nearly as low as the ip4 does. So it seems they want to reduce that power - but why?

Do you have any info on that?

Mine is from week 29.

Apple IPhone's 3G and 3GS models are build with a discrete PS sensor (build in the old fashion way, with an IR LED and a photodiode).
In the iPhone 4, they started to use a integrated PS/Ambient sensor, and this device obviously
needs a different working approach.

The 3GS discrete PS acts like an on/off switch. Once he detects that enough IR light is coming back to the photodiode, they trigger his output to signal the event. The trigger distance is thus fixed, and you only would see a very small (negligible) variations due to the electronics devices tolerances.

iPhone 4 is a different history. We know that Apple was changing the IR power in real time (easy to see with a digital camera) and it was changing the trigger distance (easy to see with your tests).

I always proposed that this behavior was flawed. In fact, this new sensor has the capabilities of work
like a discrete one and thus having a fixed trigging distance. And for my preliminary tests NOW the sensor is working exactly in that way. Also the IR power reduction is lower now.

In my modest opinion there's no need to reduct the power on the IR LED. If correctly programmed, the new sensor can act like the old discrete approach. Obviously, there's a lot of things that currently we don't know, like the exactly sensor used and if there's an imperative reason to play with the IR power.

If we remember the 3GS model, after the detection of an obstacle, the IR LED start to pulse slowly with the same power. But when you make a light blinks slower you are using less energy. My point is that from the beginning Apple is reducing the power once the PS is triggered. We know that the battery impact would be
negligible, but maybe are other implications that we can't see.

Also I must say that in those days the firmware development is very segmented, with the iOS code being writing in very smaller portions by different people. It not would be surprising to me that the guy who write the PS code tried to squeeze the battery power to a molecular level under pressure of management or only to add his personal touch.

my previous phone began working correctly some time after 4.1 release and definitely before 4.2 came out. (you can go back through my posts and verify this) that means i never updated it and it just magically started working.

4.2 does not do anything to fix the sensor. there was nothing in the release notes about a fix.

when i got my new iPhone 4 (4.2 pre-installed) it had PS issue for the first couple phone calls and then seemed to calibrate itself and has worked perfectly ever since.

i dont think the phones working now has anything to do with the release of 4.2. we had plenty of reports of people getting working phones before 4.2 was released.

this seems more like quality control has been stepped up and the sensor firmware has been corrected at the factory.

it seems to me that the Proximity Sensor on the iPhone 4 has a way to automatically calibrate itself. maybe this feature was simply broke and apple tried to increase output to compensate. now the firmware has been corrected and it is working again. thats the only way i can explain my new phones behavior and its the only way that seems to explain why some phones are still broke. If apple fixed the PS firmware via an update then how come a lot of people still have the issue and can only resolve through exchanging for new phone? the only way this makes sense is if apple cant flash the PS firmware via an update and the correction had to be made pre-manufacturing. that would make sense why the new phones dont have the issue and old ones still do.

Message was edited by: himoura

????

Whether they listed anything about the new update dealing with the sensor or not, the fact is that repeatable testing shows that EVERY phone I tested that failed test 6 under 4.1 passed it under 4.2.1. Something sure changed. We aren't talking about isolated cases which don't show a pattern.

my first phone changed before 4.2... how do you explain that?

and how do you explain all the people that exchanged before 4.2 and got working phones?

That wasn't the question. You said there was no update to the sensor in the latest version. Yet I've tested a number of phones, all of which failed the test #6 before the latest version and all pass it now.

Something had to have been changed.

its possible apple could have made changes and not listed them but things just aren't adding up at all...

even diego says that if 4.2 fixes it then how come some people still have the bug... very good question.

i could be completely wrong about the device calibrating itself who knows but i do know for sure that my first iPhone 4 got really bad on 4.1 and then a month or so later (way before 4.2 came out) magically started working 100%.

http://discussions.apple.com/message.jspa?messageID=12351515#12351515

people were exchanging pre-4.2 and getting working phones that passed the tests. are u guys gonna try to say that those phones weren't really working before but now in 4.2 all the phones work great? that's convenient

I don't think there is much doubt that they changed something. If it was just my phone, it could be one of those weird, isolated, occurrences - like when people upgraded to 4.0.1 and said it fixed their problem. Or people did a reset, etc, etc, etc.

But we had no pattern. It wasn't like taking 10 phones, doing the same thing, and all changed the same way. With this, I have a pattern. Not a big one, but 100% consistent.

If you look at those who passed the tests prior to 4.2.1, they also don't show a pattern. There were only a handful that said theirs work AND reported on the test. At least 2 of them ALWAYS worked - no change in distance from the start.

That's why I kept trying to get people to test, so a real pattern could be found, if any.

There are, in my opinion, two ways to show a meaningful pattern. One is to get LOTS of people to report. It isn't very scientific, but if the number is large enough, it works well enough. The other is to have one person checking multiple phones to see what happens. This eliminates the problem of just people who have problems tending to post.

If 10 people report something, is that typical? We don't know how representative they are. But if one person tests 10 phones and can report they all do the same thing, then we have at least a small trend.

That's why I'm hoping I can get to test more "recent" phones at Apple stores. Sure, the one from week 46 that I tested worked. But it is just one. Don't know why the other two from that week weren't set up to do calls (so I couldn't test them), but if I could even test just 3 and all worked, that would likely mean something was changed in the hardware too.

If anyone else can check at Apple stores - or have a number of friends with more recent phones - maybe they can do test #3 as well with them. That's the one that, even with 4.2.1, has problems with all I've tested except the week 46 phone. They might be better with 4.2.1, but the distance can still be reduced a lot.

we have had a pattern for months now.

people have been returning their phones prior to 4.2 and getting working models. thats a pattern...

also there have been lots of people to install 4.2 and say its still not fixed... that also is a pattern.

not really sure where you are going with this other then to say that if one person were to test 10 phones then somehow that is more legitimate then 10 people individually testing 10 phones.

The numbers of people getting working models is too small to tell us much.

It shows it MAY be improved in the hardware, but to really determine that we either need a lot more reports or someone who can test multiple phones to see if there is a real trend.

And "lots" of people have not said 4.2.1 doesn't solve their problem. At my last count, there were 5-6. Now, if one person tested 5-6 phones and all of them didn't work with the update, that would mean a lot. But random reports? Not so much. Those just hint at possibilities to be further researched.

Which is why I'll keep checking any phones I can get my hands on - to see if the isolated reports do lead to something more solid.

google 4.2 proximity sensor and troll the forums.

there are at least a few on every forum i have read that say it didnt fix their issue.

All I can say is that all of my 1-#4 iPhone 4's had PS issue on 4.0, 4.1 and even 4.2.1. Now my newly exchanged phone has yet to have one PS/face-dial issue. ZERO. I actually restored my new phone from backup so all of these theories to "restore as new" don't hold water either. I really believe about 99% that it's hardware related and only Apple can fix/resolve it and it will only truly be resolved for someone by getting >42 week build replacment with 4.2.1. 4.2.1 may have improved it some but not 100% and my #4 phone is proof of that. I really don't think there is much more need for testing. If you have an i4 with 4.2.1 and during a >5 minute call you face-dial and hold phone normally then there is only one option and that is to exchange.

I know there are a few - which is the point, a "few". Not enough to tell us much when from random sources. Only if someone tests "a few" and can reproduce it on all of them would it tell us much.

The problem is, we know there are at least a few older phones out there that never had a problem with my test, so when we hear about new ones working, it's hard to tell if they improved the hardware or if we are just seeing some random chance.

Thus I'm hoping a LOT more people will report in telling us they work and what week it is - especially if they don't have 4.2.1 on it yet since that update has helped a good number.

The testing now just changes focus somewhat - what date phones will consistently pass it (if any are really consistent). For that we either need LOTS of positive reports or reports from one or two who test multiple phones.

Something else I need to do and didn't think about it when at the store is to try the phone that passed my test on my shoulder. On my phone, unless I really make a special effort to press firmly to keep it against my ear, it will always come on. While I can live with that, it would be much better if it worked then too.

i agree 100%