My Fairphone Experience

This is the English summary of what I wrote in German about the Fairphone in my blog.
Please use the collective English article if you would like to add comments in English.

Please let me make this absolutely clear: these are the findings of my tests. I do not intend to make the phone look bad. I am a very early adopter, my order number was 478, and I am still happy with the outcome. For me this is a massive step up from my earlier phone (Huawei Ascend Y200), and it solves everything I had to critize on my old one. As yet I have not run into any real show stopper with the Fairphone, and I do not intend to return it.
There are however weaknesses, which may or may not be addressed, if I find them, I write them down. They can only be addressed if they are known. And I do not just write them down, I do contact Fairphones support, once I think I have done all I could to find the cause.

Table of contents

Unboxing

At first of course there is the unboxing, which is mainly a collection of pictures, so you will have to live with the fact that I don't translate that.

Video

Then there is a video that I took of a firework on New Year's Eve. If I understand correctly, the quality of the video is the maximum possible. It's 720p, no additional effects, "Video quality" = "Fine". These 155MB come directly from the phone, no editing or recoding.

Pictures

Initial photos

Clicking on the image gives you the original file.

IMG_20140101_142214_r

IMG_20140101_142228_r

I took the following picture several times with differing resolutions (8,5,3,2 and 1.5MP), a click downloads a ZIP-package with all of them:

IMG_20140101_143229_r

Test pictures about the reddish center

In a quick test, Heise mentioned that the camera of the fairphone takes pictures with red tinting in the center. I did not notice this in my photos at first, but now that they menioned it, I can see this in the pictures above. Here is now a bit of testing.
I took a number of pictures of grey gravel in sunlight, with differing white balance and scene modes, to see how strong this effect was. Below are the histograms (first: upper right corner, second: middle) of this picture:

IMG_20140103_130205

Upper right corner:
HistogrammRechtsOben

Middle:
HistogrammMitte

This is an overview of the gravel-pictures I took. On click you can download a package with all of them (~46MB). I went through the list of white balance and scene-modes for these:

RedCenterPicturesOverview

I also took a picture of the gravel with the front camera (included in the package above, last one), and this one is interesting in more than one way. I will take some more pictures with this one...
It does show a similar red tinting in the center, which is quite puzzling, as it is a different camera. My first thoughts were that the codec could have something to do with this. But that is of course nonsense. Problems with a codec would show in a different way. So, probably neither camera nor codec (BTW: in Engineer mode you can take RAW pictures, these have the changing tinting as well), the only component left, that I can think of, is the lense. Is that the same for both cameras? But: shouldn't the lense in these cameras be an integral part of the component?
The other thing strange with this picture is, that the resolution is 2560x1920 (~5MP), although it is supposed to be taken with a 1.3MP camera...

FrontCameraRedCenter

FM Radio

I like the fact that the Fairphone comes with a built-in FM-Radio, as I am part of the audience of some talk radio stations. However, there are some things I have to critize.
The phone requires a headphone to be plugged in as an antenna, which is rather standard. I am not sure how the connection as an antenna is set up, but it is not overly stable. If you wiggle the plug, you quickly run into problems with the reception.
Reception is nothing to be proud of in the first place. Compared to my MP3/Radioplayer and particularly to the stationary radio, the Fairphone gives up on a weak signal much too early.
And even with top reception, there are some interesting effects. One is a delay of over a second between my other devices and the Fairphone. I have no idea what could cause this, but it is astonishing. The other thing is a rather high noise-floor, which is independent of reception or station and peaks at around 2kHz (varies by a few Hz). It is both audible and is part of a recording (recording is done to OGG-format, BTW).

FP01RadioNoiseSpectrum

Here an example, amplified to the max and in a loop:
Download

And there is a bug in the radio. As you have to insert a headphone, you can manually choose whether the speaker is on or not. When the speaker is switched on, the headphone stays connected. Apparently the amplifier gets turned up when the speaker is selected. With the effect that the headphone output is massively increased.

GPS

The GPS overall is quite acceptable. I recorded some tracks with OSMTracker. Time to fix is fine, the longest I could see (cold start, no support by EPO or A-GPS), was a little over 30 seconds. Usual times range between 7 and 15 seconds before the GPS is ready to be used for navigation.

EPO support

EPO (Extended Position Orbit) is supposed to help in finding a first fix, by supplying pre-calculated predictions for the satellite orbit. It's a Mediatek thing, so you have to download data from a Mediatek server.

FP01AboutEPO

I had some problems with that, when I clicked on "Download", the percentage indicator stayed at 0% whenever I tried. Sometimes it would give an error message, sometimes just stop.

FP01EpoScreen

As you can see, I finally managed it. It was not the same problem as described in the XDA-Forum, but this gave me the idea to look in /data/misc. There was a file called "EPOTMP" or so, with the current date. When I removed it, the download worked fine.

Recorded tracks

In this image you can see: FairphoneGPSTest1WalkingBackground

Another image, a little more options tested this time: FairphoneGPSTest2WalkingBackground

Power Consumption

First off, a warning: Lithium-Ion batteries can be dangerous, if not handled correctly. The follwing is not a guide on how to do things, this is what René Schmitz and I did. If you want to do the same, do it on your own risk. We do believe we know what we are doing, but we will not be responsible if you try to copy our actions, insane as they may be.

Here is a short video on how we built the test suite:

In Heises short test, they said that the energy in the battery was enough for 7.1 hours of video watching (according to the calculator, that means an average current of 281 mA), or 8.8 hours of surfing using WiFi (227 mA).
But for me this is quite meaningless information, most of the time my phone will just lie dormant, if I actually use it to watch something on Youtube, this is way of an exception. So what I want to know is: what is the baseline? How much does the phone take just waiting for something to happen?

Unfortunately the Fairphone is my only mobile phone, so I use it for various things. Especially now that it's new to me, I test a lot of things. Therefore there is not one longer stretch of time where the phone is used the way it will be once I'm used to it. So, I cannot just watch when the battery runs out under "normal" conditions. So, we need measurements, and if only to satisfy the curious nerd in us.
Unfortunately I did not have enough time to do any serious tests, as we had to spend some time on the setup first. In order to get some measurement-equipment between the battery and the phone, we had to have a mock battery. This is what it looks like when installed:
MockBatteryInPlace

This is where the original battery goes:
BatteryHolder

And this, finally, is the Fairphone booting. The battery is removed (and not burning!) and some measuring equipment installed:
PowerTestingFairphoneBoots

As I said, there wasn't enough time for any substantial measurements, but we now have a test-suite. The first check was whether the data reported by the engineer-mode matched what we got. Without any load except basic operation (2 SIM cards, display on), the engineer mode reported something between 115 to 120 mA, which matched the 118mA we measured with the external equipment.
So, a first result is: Fairphone with the display on (lowest brightness setting) and nothing much else to do takes ~118 mA, which means it will last 16.9 hours in this mode. All calculations based on the idea that the battery actually can deliver exactly 2000 mAh.
We'll keep you posted.