Samsung Galaxy Nexus Fault - "Ghost Volume Reduction"

[Hands0n]

Unless you've been living in a cave, or up a tree of late, it will not have escaped you that [arguably] the hottest, most anticipated new Android smartphone, the Samsung Galaxy Nexus, has arrived with a fairly significant fault.

The fault manifests as the smartphone automatically [ghostly] dialling down the volume to zero, all by itself. Stuck buttons? Dodgy volume rocker switch? Something much more sinister?

Unfortunately it seems to be the latter. But it doesn't affect everyone.

Over the past 24 hours the problem has been localised to be some interaction between the SGN and 2G 900Mhz networking. So thats Vodafone and O2 to you and me. Orange and T-Mobile 2G newtworking does not appear to cause any issue. Those two networks operate at 1800Mhz. Nor does any UK-supplied 3G that operates at 2100Mhz. It is purely, and demonstrably, 2G at 900Mhz.

The problem seems to manifest in one or both of two ways.

Firstly, by switching the SGN or having it fall back to 2G/900Mhz the "ghost" occurs.

Secondly, by placing a 2G/900Mhz handset next to the SGN the "ghost" occurs.

Clearly then, this is a radio frequency interference issue. It happens not only when the Android OS is running, but there is video on YouTube of the same condition happening with a 2G handset placed close to the SGN while it is in Bootloader mode. So this is most definitely a hardware issue.

Samsung remain completely tight lipped at this time. Not even a denial. Anyone with an SGN and some 2G/900Mhz networking can demonstrate it happening.

Clove have said it is a batch issue and that none of theirs are being affected. Well, they would say that. I'd want to see the YouTube tests repeated by Clove to demonstrate no effect. Otherwise, the house rule says that Clove are as impacted as anyone else.

Me? I've got an SGN and am quite happy with it. I will only use it on the Three network so no worries about 2G/900Mhz there then.

A shortie YouTube from me, taken this morning, for your entertainment http://youtu.be/LjwMzooNVCE - please do excuse the sniffing and coughing, I am heavily under the influence of Man Flu.

[Ben]

Woah, how weird is that!

What did you record the video on btw?

[miffed]

LOL ! Imagine if that had been the iPhone 4S , it would be headline news

Didn't realise you had one ! Not sure how I missed that ? How do you rate it compared to GS2 ? Suppose its a bit difficult to tell until ICS is here for the masses.

[DaveC]

I was able to replicate this last night using my Nexus One in 2G mode.
Being out today and on a good 3G signal the phone behaved well. Went into Asda and dropped to 2G and as soon as I tried to view an email the volume turned down and became erratic.

All I have had from O2 is the following reply to a tweet:-

"Samsung are investigating at the moment and we hope to have more information to share soon.Thanks."

I've got a week to get it back under the 14 day rule or else I have to take the 'not fit for purpose' route and all the messing that will involve.

The feeling out there is that this is hardware related, so any firmware / software fix is liable to be a botch.

[Ben]

It hadn't registered with me that Hands0n had one of these either for some reason. Looks really nice. Then the volume goes down all by itself. Lol!

DaveC - Mhmm yeah, I'd definitely be wanting to return within the 14 days if the problem is persisting. That'd get annoying fast.

[Hands0n]

Woah, how weird is that!

What did you record the video on btw?

I used my Samsung Galaxy S2 to record the video :)

[Hands0n]

Apologies gentlemen. I didn't announce that I was going to get one of these, mostly because I had no real intention of doing so. However, a "just used" one came my way at a price was was too appealing to say no to. And so I took the plunge.

The Samsung Google Nexus (SGN) satisfies my eager curiosity for Android 4.0 (Ice Cream Sandwich). I have Android 3.0 (Honeycomb) on a tablet and it is a work in progress at best. ICS is supposed to be the finished goods, the real McCoy, and it certainly does appear to be. Much more polished, finished off even. It is also annoyingly different to Gingerbread to be a very different OS to what has gone before. It does not "feel" as incremental as iOS 5 has been to iOS 4. That said, it is a very capable OS with lots of improvements and new features that I have yet to discover. I like a lot of it, but I also miss some of the "good stuff" that is in Cyanogen 7.1 (Android 3.x) such as the built-in Screen Shot capability. Even after all of this time Android 4.0 ICS doesn't have it, whereas the Apple boys have since day one. Stuff like that is totally annoying.

However, all of that is overshadowed by what is possibly a genuine hardware fault that Samsung will be unable to fix in code. It is to do with the 2G radio, possibly the data portion relating to GPRS/EDGE. Although the 2G handset that I used is disabled for data by my firm, so it sort of discounts the specific GPRS/EDGE angle, landing more firmly on just it being something to do with 900Mhz GSM, end of.

I have never seen anything like this on my Nexus One, that spent quite a bit of time in 2G mode courtesy of patchy Vodafone around the South East (not as bad as non-existent O2). Although I do recall the Nexus One did have an issue with its fixed soft buttons that were off centre, making it difficult and annoying to get the required action first time. I hated it for that alone, HTC never fixed it, so it was bound to be a hardware/design issue.

So, it does not bode well for Samsung and the SGN. I suspect either a recall and replacement costing Samsung millions, or a half-hearted attempt at patching over the issue in software, probably by weakening the radio and thus causing other issues. Samsung will be hoping that 2012 comes around quickly, the SGSIII arrives and everyone forgets the SGN. Except we won't.

[Ben]

I used my Samsung Galaxy S2 to record the video :)
Gosh. Ok. Was worried it might be an iPhone lol - a lot of jerking around, refocusing, and choppiness!

[DBMandrake]

If operating another phone on 900Mhz 2G next to it is enough to trigger the problem its definitely hardware, not something that could be fixed in software - take it back under the 14 day return period until the dust settles and a hardware revision is released.

Sounds like the input circuitry connected to the volume buttons is lacking proper RF suppression! As anyone who has placed a 900Mhz 2G phone next to an amplifier/TV etc knows, the TDM pulses of GSM get into anything that doesn't have proper RF suppression. It's very unlikely to be a manufacturing flaw, but rather a design flaw due to the handset simply not being adequately tested on 900Mhz 2G, as it is not used in many countries. It will take a (fairly minor) hardware revision to fix it.

[Wilt]

I don't understand how this issue could ever make it past testing...

Even if it is just a batch issue, I would expect Samsung to be taking random samples of batches in order to test build quality and things exactly like this.

[DaveC]

Samsung seems to keeping quiet. They are refusing requests for information from the technical press and from time to time deleting negative comments on their facebook page.
When people complain via email they get a standard reply and are told to phone up - so what's the use of email?

What seems strange is that some carriers, aware of the problem. are still selling the phone.

[Hands0n]

Gosh. Ok. Was worried it might be an iPhone lol - a lot of jerking around, refocusing, and choppiness!

LOL, no. That was as much me stifling my coughing fits. This Mab Floob is dreadful this time round, twice in a month! Normally I use my Panasonic HD camcorder mounted on a tripod. But couldn't find it in myself to go get it :)

[Hands0n]

Well finally, Google have spoken up to say that there is a fix on the way, as reported by Android Police here --> http://bit.ly/trEYSj

Samsung's response is, frankly, disgusting and reinforces my feelings of ambivalence to that particular company. They have absolutely no discernable quality of customer feedback and support. Something like this could have been defused by a simple message along the lines of "We have identified a fault in the first release and are working on a fix that we will push out OTA as soon as possible". That would have done it. There was no point trying to hide behind clean Facebook pages, the rest of the Internet cannot be silenced. YouTube is alive with videos, blog sites are awash with information on the fault.

The SGN is a distinctive and very nicely put together device. The OS is brand new, and has a great number of excellent features. And it looks damn good :)

But ... for some reason, I find it compelling to keep using my SGS2 - my original plan to eBay it as soon as I got the SGN has gone and stalled... Maybe when the bug is fixed, even though it won't affect me on Three!

[gorilla]

I like a lot of it, but I also miss some of the "good stuff" that is in Cyanogen 7.1 (Android 3.x) such as the built-in Screen Shot capability. Even after all of this time Android 4.0 ICS doesn't have it, whereas the Apple boys have since day one. Stuff like that is totally annoying.

Are you sure?


Sharing with screenshots

Users can now share what's on their screens more easily by taking screenshots. Hardware buttons let them snap a screenshot and store it locally. Afterward, they can view, edit, and share the screen shot in Gallery or a similar app.
http://developer.android.com/sdk/android-4.0-highlights.html

It can't be an easy function though if you haven't discovered it already lol.

I've been completely out of the loop for the last week, so congrats on the phone! I think ICS will (or could) be better on the SGII itself and am resisting testing early alpha releases.

[Wilt]

Hold the power and volume down button to take a screenshot.

[Hands0n]

Now I feel completely daft :)

Screenshots have indeed been implemented in ICS, just not how I expected them to be, coming from Cyanogen MOD 7.1. In that OS you press and hold the power button, a long menu comes up, select Screen Shot and you're done.

In ICS, however, you press and hold the Volume Down and Power buttons together for a moment. A copy of the screen is then saved in the Screenshots folder in Gallery.

Other stuff I like is the ability to group apps together on the screen in folders, rather like you can in iOS. I really save screen estate.

[miffed]

Samsung's response is, frankly, disgusting and reinforces my feelings of ambivalence to that particular company. They have absolutely no discernable quality of customer feedback and support. Something like this could have been defused by a simple message along the lines of "We have identified a fault in the first release and are working on a fix that we will push out OTA as soon as possible". That would have done it. There was no point trying to hide behind clean Facebook pages, the rest of the Internet cannot be silenced. YouTube is alive with videos, blog sites are awash with information on the fault.



I am not sticking up for Samsung here .... but with the Google branding of Nexus devices - would it be fair to say Google , rather that Samsung should be seen as the Vendor for this handset rather than Samsung ?

[DaveC]

I am not sticking up for Samsung here .... but with the Google branding of Nexus devices - would it be fair to say Google , rather that Samsung should be seen as the Vendor for this handset rather than Samsung ?

Well the back of the phone says Google and Samsung. I guess that may be part of the problem - each shifting the responsibility onto the other.

[DaveC]

Finally got the update this morning. As reported elsewhere, the build number has changed whilst kernel and baseband remain the same.
So far, so good but I need to move into a 2G area to be totally sure.

[Ben]

Looking forward to video demo take 2 from Hands0n! We need genuine shock and awe if it's fixed! :D

[DaveC]

No video I'm afraid, but I put the phone into bootloader and initiated a 2G call from my Nexus One net to the volume control and nothing happened.

[Ben]

Pretty cool they've managed to fix it in software. Hurrah! :)

[Hands0n]

Meanwhile ... I am still waiting for the OTA update. It is available to be run manually, but I'm digging my heels in this time. I want it OTA.

Either way, work or not, I will make the obligatory video for YouTube :)

[Hands0n]

Thats it, all done, bug fixed, Snoopy time :)

I received the OTA today, rather than force it manually as is available on the Interwebs. It was a small updated that patched some 30 files, according to the dialogue box showing progress of the update.

The OTA is rolling out in typical Google style. Seemingly random, people getting it at all different days of the week and times of the day. Does anyone actually know what Google's algorithm is for this. If Apple can do a global simultaneous update why can't Google? Bah!

So, I have made the video and posted it up here http://www.youtube.com/watch?v=1iAzVkPMPCU

NB: Some have commented on how extreme and unlikely the "test" is. But the point is to demonstrate how a 2G 900Mhz signal disrupts the Galaxy Nexus volume setting. It did happen when the handset was [itself] on 2G 900Mhz, but it was more reliably reproducible by using a donor 2G handset close up to the Galaxy Nexus. For the purposes of the video it was preferable to use that method. But I was able to witness the symptom occur by flipping the Galaxy Nexus to 2G only and waiting a short while.

[Ben]

These videos remind me a bit of Take Hart! Ahh, what a show!

Funny that some have commented as such, Hands0n... I didn't realise Samsung had 'fanbois'.

[DaveC]

HandsOn, did you try the test with the Nexus in Bootloader mode? Some of the original videos showed the bug affecting phones this way.
Have to say I tried it after the update and was unable to replicate the effect.

[Hands0n]

@DaveC - I had not tried it in Bootloader mode, although have seen the videos on it. There was much speculation that because Bootloader did not have the Android OS running that it had to be hardware. But that opinion did not take into account any actual radio adjustments that the new firmware would almost inevitably make. Some little "tuning" there could be all that was required.

I'll have a go this evening, but expect no worse results than yourself :)

[Hands0n]

The news is not so good in Bootloader mode. The fault is very much apparent --> http://www.youtube.com/watch?v=8ohOBHQm17c

This does strongly suggest that there is indeed a hardware fault that is being masked by code in the Android OS itself. That said, this is a non-fault in my opinion, now that it has been fixed (albeit in the OS). For the very few times that I will be in Bootloader mode, I am very, very unlikely to be rubbing a 2G/900Mhz handset up alongside the volume rocker buttons :D

[DaveC]

Funny how I could not get it to do that.
Just tried it again and still no response using a Nexus One in 2G only mode. - May get one of my old 2G handsets out.

Now everyone is worrying that the fix will use processor time and result in poor battery life in 2G areas. - You can't win

However well the phone works, I think in Europe and other 900MHZ countries at least, people will have niggling doubts and that may well affect sales.

[Hands0n]

Yea, I've had someone (from XDA) ask me to re-run that same test with a CPU usage monitor running. I don't think that will be at all scientific unless there were two Galaxy Nexus, side by side, one updated the other not, loaded with exactly the same apps, config'd the same, everything. I haven't that kit, nor the time really. But certainly, I could not perform a reasonably controlled test, so it would be somewhat pointless.

[DaveC]

Guy on this morning worrying like mad that this is going to cause his phone to run slower and drain his battery quicker simply because the processor may have to do more work.
They really do panic.
Have I noticed my phone running slower - No
Does my battery appear to drain quicker - No
Whatever you come up with in tests, its what is apparent in real usage that matters.
I bet if they dug deep enough with most smartphones on the market they would find something to complain about.

[Hands0n]

Yea, there is some real paranoia. One poster on my YouTube page has gone nuts against Samsung.

My own experience is that the GN has run for 12h 58m on battery and is at 69%. I've not used it much, but it has been on 3G only (Three, All in One plan). The top four consumers have been:
Screen 35%
Phone idle 25%
Android OS 22%
Mobile standby 11%
Everything else is 2% or less ...

I agree, whatever the measurements and the science, it is all down to real world usage that counts.

[DaveC]

Video of the CPU Test http://www.youtube.com/watch?v=h3T4my6Eyf8

[Hands0n]

A nice find. So CPU goes from 370MHz to 420MHz briefly. Not something to be at all concerned about in the least, I'd say.

Interestingly, the video of the CPU usage before the fix shows even higher usage http://www.youtube.com/watch?v=LlPN7XgXXsw up in the 700MHz range. So the OTA fix seems to be beneficial rather than negative.

The app used to monitor the smartphone in the video is SystemPanelLite Task Manager (free) available here --> http://bit.ly/tS8JEN There is a paid for version that has a whole load more functionality. But the all important activity monitors are in the Lite version.

[DBMandrake]

The news is not so good in Bootloader mode. The fault is very much apparent --> http://www.youtube.com/watch?v=8ohOBHQm17c

This does strongly suggest that there is indeed a hardware fault that is being masked by code in the Android OS itself. That said, this is a non-fault in my opinion, now that it has been fixed (albeit in the OS). For the very few times that I will be in Bootloader mode, I am very, very unlikely to be rubbing a 2G/900Mhz handset up alongside the volume rocker buttons :D

I hate to say I told everyone so ;) The fault will definitely be a lack of proper EMI design (electromagnetic interference) involving one or more of; poor shielding, insufficient RF filtering on the traces/cables going to the button, or poor PCB layout which makes one of the traces going to the button into a nice little antenna. In other words a hardware design flaw.

GSM pulses have a low duty cycle which means they are very short bursts with low average power but very high peak power, without the right shielding and filtering they wreak havok with all manner of devices. (How many iPod docks not designed for iPhone's go "dut-da-da-dut-da-da" when the phone is active on 2G ? Mine do.

What the software patch will do is look at the pattern of pulses inadvertently picked up by the button wiring - GSM pulses follow certain patterns that could be easily recognised which would be nearly impossible to simulate by rapidly pulsing the button. If one "clean" press of a certain minimum duration is received it will be registered as a genuine button press. If lots of very brief pulses are received that are too rapid to be a genuine button press, it will be ignored. That's what I would do if presented with the problem of "fixing" the issue in software. It looks like they didn't (or can't) do the bootloader though.

[Ben]

Fortunately I don't think this being a hardware issue is going to cause any lasting problems; indeed, looks like the software fix works rather well. Probably a lot better than the software fix for the iPhone 4's antenna problems did! I'm guessing Apple later modified the iPhone 4 to help a little, and suppose Samsung will tweak the Galaxy Nexus similarly.

Funny that some iPod docks aren't properly shielded for 900MHz GSM! #fail

[Hands0n]

Given that the OTA fix actually works without detriment to the Galaxy Nexus, anything else is academic. The persistence of the "fault" in the Bootloader is irrelevant inasmuch as that is not how the phone is ever likely to be used.

The thing is, with the complexity at hand in modern day equipment it is inevitable that design defects will occur. Who is to know how many similar hardware defects are resolved in firmware before the device is shipped? The same goes on in other industries. For example, only today Honda have recalled some 250,000 Gold Wing motorcycles for a rear brake defect. I can't think for how many years these machines have been on the roads with the defect, unknown to the owners.