Why? Surely if the phone is lying flat the g readings from the sensor should be (x, y, z) 0,0,1 so if they're 0.02,0.043,0.97 then the software knows it has to apply a -0.02,-0.043,+0.03 correction to all the data from then on? Isn't this what it does on a per-app basis when you tap the screen in the Level app to set it to zero?
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
iPhone 5s accelerometer/gyroscope/compass values incorrect
- Thread starter tharepairguy
- Start date
- Sort by reaction score
Why? Surely if the phone is lying flat the g readings from the sensor should be (x, y, z) 0,0,1 so if they're 0.02,0.043,0.97 then the software knows it has to apply a -0.02,-0.043,+0.03 correction to all the data from then on? Isn't this what it does on a per-app basis when you tap the screen in the Level app to set it to zero?
UncleGuido
macrumors member
There are two implied questions here. I'll try to address the second one first (and I think you already understand the answer). When you are running the iOS 7 Compass app, and using the inclinometer function, a tap on the screen does indeed zero out the reading. But this is not a calibration. You are just forcing the reading to zero, in whatever orientation the device is in at the time. It puts the app into a temporary "difference mode" which can be useful for comparing measurements at two different locations. Zero it on one surface, put it on the other surface, and the device will display the difference in their slopes. Tap the screen again, "difference mode" is cancelled, and the phone is again reading absolute orientation (which may be miscalibrated).
A true calibration would store data that corrects any offset errors (biases) in the hardware, so that the readouts are correct, relative to the true vertical defined by the force of gravity. Some third party apps (like Clinometer from PlainCode) provide a user-run calibration method that prompts the user to take measurements with the device placed in three orientations. This creates independent calibrations for each axis of the accelerometer chip.
It might be possible to calibrate all three axes by taking measurements with the phone lying on its back on a level surface, as in the method suggested by some of the Apple support reps. But for this to work, the sensing axes within the accelerometer chip would have to be perfectly orthogonal, and the offset errors of the axes would have to be correlated. Then, knowing the magnitude and direction of gravity, and the uncorrected readouts of the three accelerometer axes, you can do vector math to compute the angular error on each axis, and derive calibration corrections. If this is really the way it is supposed to work in iPhone 5s, I have no idea why it does not work. Maybe there really is a software bug.
I feel that the calibration method in Clinometer is likely to give better results. Ideally, Apple would provide a software update that adds a similar procedure within the Compass app, or perhaps within the Settings app. This procedure would store a system-wide calibration, correcting the accelerometer data for any app that used iOS services. And it would be optional. Those users who wanted max accuracy would run the calibration. Other users could ignore it.
Morac
macrumors 68030
iPhone 5s accelerometer/gyroscope/compass values incorrect
3 positions isn't enough, at least no for my iPhone 5S where on one landscape side the offset is 0 degrees and on the other it is -4 degrees. Calibrating on one landscape side wouldn't and doesn't correct the value on the other side.
A proper measurement would require 5 readings. That's what the Clinometer app does, which is take 2 readings from each side (excluding front facing down). That works great.
3 positions isn't enough, at least no for my iPhone 5S where on one landscape side the offset is 0 degrees and on the other it is -4 degrees. Calibrating on one landscape side wouldn't and doesn't correct the value on the other side.
A proper measurement would require 5 readings. That's what the Clinometer app does, which is take 2 readings from each side (excluding front facing down). That works great.
Last edited:
Technically they are not interlaced just looks like scanlines. Do a search for scanlines on here and you can read all about it.
Its a few days old, but here's Gizmodo's further take on the issue. I posted the blog they refer to earlier.
http://gizmodo.com/heres-why-the-iphone-5s-accelerometer-is-so-screwed-up-1445966306/@whitsongordon
Sadly, I do not believe the 5S fits with the Apple ethos of making the best devices. Like Antennagate, owners will only become aware of the sensor issue if they find the phone is behaving strangely. Once outside the 14 day new replacement, they may have to be content with a refurb phone, that may have the same issue until it is fixed properly.
We can only speculate whether Jobs would have tolerated this apparent design flaw pre release of the phone to manufacture. I suspect the answer is that he wouldn't.
http://gizmodo.com/heres-why-the-iphone-5s-accelerometer-is-so-screwed-up-1445966306/@whitsongordon
Sadly, I do not believe the 5S fits with the Apple ethos of making the best devices. Like Antennagate, owners will only become aware of the sensor issue if they find the phone is behaving strangely. Once outside the 14 day new replacement, they may have to be content with a refurb phone, that may have the same issue until it is fixed properly.
We can only speculate whether Jobs would have tolerated this apparent design flaw pre release of the phone to manufacture. I suspect the answer is that he wouldn't.
Has anyone recently exchanged their phone with Apple directly, and was the replacement any better?
S.T.O.P. it already! So tired of these Jobs comments!
Really?!
His antennagate ****-up is already in history books!
I did. It was quite a bit better but not perfect. See my previous posts for details.
waiting for my replacement. should come - according to UPS website - directly from China here to Berlin, Germany on next wednesday (after the new iPad event).
Then I hopefully see if there is a ios7 update which might fix my 1st iphone 5s (week 39) and I will have a brand new one from (presumably) week 42.
I assume because it was sent from there today. But of course I am not sure until I have it.
I keep you guys posted...
You can log into your apple account and get your serial number if your phone has already shipped, then you can will be able to find your production date now.
APPLE STORE Berlin
I tried (almost) all of them as well 1,5 weeks ago and they were all off 3-5°.
So I can corroborate that. 🙁
I was just wondering about the security guys behind their CCTV cameras who - once in a while - see a guy come in and "test" all their iphones on display for level.
I myself felt a little stupid and thought it was weird of me to do that. But in the end I wanted to know for sure....
----------
yay. thanks. just did not think of that.. 😀
UPDATE
I just checked that but I cannot find a serial number.
I only find Tracking number, order number and such...but no serial...
Last edited:
One of the biggest flaws of the 5s compass error happens when i drive my car using Apple or Google Maps with compass oriented map. If i drive on a straight road, it doesn't show vertical but 8 degrees turned to the right.
I find this very unpleasant.
I'm a French owner of iPhone 5s, which of course has the same 3 degrees inclinometer and 8 degrees compass error as yours.
I'm also developing a game app based on the inclinometer and it doesn't work well on my 5s.
I don't like the way Apple treats his customers on this case.
I find it strange that big medias don't tell about this big issue.
I find this very unpleasant.
I'm a French owner of iPhone 5s, which of course has the same 3 degrees inclinometer and 8 degrees compass error as yours.
I'm also developing a game app based on the inclinometer and it doesn't work well on my 5s.
I don't like the way Apple treats his customers on this case.
I find it strange that big medias don't tell about this big issue.
I wonder if anyone will ask Tim Cook some questions at the new ipad unveil next week?
I'm guessing the people there will have been specially selected to ask the 'right' questions!
Would be interesting to ask whether they have any intention of fixing the last product they released before back slapping each other about the new product!
I'm guessing the people there will have been specially selected to ask the 'right' questions!
Would be interesting to ask whether they have any intention of fixing the last product they released before back slapping each other about the new product!
that is right but I have not received a new invoice yet.
they always send that shortly before I receive the package per email.
and on the "old" invoice is - of course - the serial of the phone that I already have.
Your Link
interesting! your link did not show up until I quoted your post....
I was thinking that you forgot to write the actual link. But now it magically appeared.
Replacement iPhone 5s
well, I got the link thru Google search (which is stupid because by just logging in there was no profile whatsoever...) and it says it will be a
WEEK 42
Factory C3
let's see what happens next week...
MEMS device calibration
Here's some good information about calibration of MEMS devices (MEMS = micro-electromechanical system):
http://www.vectornav.com/support/library?id=86
For a product such as the iPhone I would think that the calibration equipment described would be used to calibrate the prototype units and define the assembly-line calibration procedures for the production run. Aircraft navigation units are tested individually, but I can't see that being possible on a high-speed assembly line.
Here's some good information about calibration of MEMS devices (MEMS = micro-electromechanical system):
http://www.vectornav.com/support/library?id=86
For a product such as the iPhone I would think that the calibration equipment described would be used to calibrate the prototype units and define the assembly-line calibration procedures for the production run. Aircraft navigation units are tested individually, but I can't see that being possible on a high-speed assembly line.
MuGeN PoWeR
macrumors 68040
Finally got my hands on a 5s, picked it up at the ATT store, the shipment came in yesterday. My gyro readings are -1, -1, 0, 0 when checked on all sides.
How do I find the build date?
How do I find the build date?