iPhone 5s accelerometer/gyroscope/compass values incorrect

[Mr. Luigi]

Report sent to Apple via my Developer Account

Just wanted to encourage those of you who have not found satisfaction despite upgrading to iOS 7.0.3 to submit a feedback report (http://www.apple.com/feedback/) to Apple. I have a developer's account so I submitted it on the developer's bug report site. Although it is clear that iOS 7.0.3 has helped, it still is not enough (IMO). FYI, this is what I said…

I updated to iOS 7.0.3. All 3 of my iPhone 5S's still do not read correctly using known perfectly vertical and horizontal surfaces (I am a teaching physicist and have access to accurate lab equipment). Readings, on average, are 1º - 1.5º off for surface and horizontal. (Although the iPhone only offers readings to the units place I can interpolate and tell that they are on the verge of reading 2º off). The phones perform better in the vertical dimension. One is 1º off and two of the iPhone 5S's do read correctly. This is a definite improvement compared to iOS 7.0.2. But the results are still disappointing and affect gaming and astronomy/planetarium apps. Errors were in the range of 2º - 4º in iOS 7.0.2.

1º - 1.5º off on level over the length of an 8 foot 2x4 means one end is about 2 inches higher than the other. So, no home building with my iPhone…or leveling a moderate size picture…or use in the lab in a pinch. More importantly, it compromises their use with the astronomy apps I use in my teaching.

Using the apps, iHandy Carpenter or Clinometer, all three iPhones read SPOT ON. This leads me to believe that the calibration method used is all important. Both iHandy Carpenter and Clinometer can employ a 1-step calibration method that uses known perfectly horizontal and vertical surfaces that allow for their superior performance. Clinometer can also use a "2-step" calibration method that involves calibrating by taking readings of horizontal and vertical with the instrument in two orientations-180º different from each other. This removes the need for a perfectly horizontal and vertical surface.

I understand Apple's aversion to using a calibration method like that used for iHandy and Clinometer. Those are not "It Just Works" methodologies. But, it would be HUGE advantage to have the option to use an OS WIDE calibration method similar to the one used in those two apps for customers who desire, or need, greater accuracy.

The 2-step calibration method available in Clinometer is essentially the same used by digital levels. What's good for the goose, I think, would be good for the gander.

Using the direction of the gravitation field to calibrate the iPhone's level sensors is still inadequate in my experience. Perhaps the coding could be further tweaked to improve its functionality. If not, I would urge Apple to offer a Clinometer type calibration OPTION in Settings to offer OS wide support.

 

[cdmoore74]

Just wanted to encourage those of you who have not found satisfaction despite upgrading to iOS 7.0.3 to submit a feedback report (http://www.apple.com/feedback/) to Apple. I have a developer's account so I submitted it on the developer's bug report site. Although it is clear that iOS 7.0.3 has helped, it still is not enough (IMO). FYI, this is what I said…

I updated to iOS 7.0.3. All 3 of my iPhone 5S's still do not read correctly using known perfectly vertical and horizontal surfaces (I am a teaching physicist and have access to accurate lab equipment). Readings, on average, are 1º - 1.5º off for surface and horizontal. (Although the iPhone only offers readings to the units place I can interpolate and tell that they are on the verge of reading 2º off). The phones perform better in the vertical dimension. One is 1º off and two of the iPhone 5S's do read correctly. This is a definite improvement compared to iOS 7.0.2. But the results are still disappointing and affect gaming and astronomy/planetarium apps. Errors were in the range of 2º - 4º in iOS 7.0.2.

1º - 1.5º off on level over the length of an 8 foot 2x4 means one end is about 2 inches higher than the other. So, no home building with my iPhone…or leveling a moderate size picture…or use in the lab in a pinch. More importantly, it compromises their use with the astronomy apps I use in my teaching.

Using the apps, iHandy Carpenter or Clinometer, all three iPhones read SPOT ON. This leads me to believe that the calibration method used is all important. Both iHandy Carpenter and Clinometer can employ a 1-step calibration method that uses known perfectly horizontal and vertical surfaces that allow for their superior performance. Clinometer can also use a "2-step" calibration method that involves calibrating by taking readings of horizontal and vertical with the instrument in two orientations-180º different from each other. This removes the need for a perfectly horizontal and vertical surface.

I understand Apple's aversion to using a calibration method like that used for iHandy and Clinometer. Those are not "It Just Works" methodologies. But, it would be HUGE advantage to have the option to use an OS WIDE calibration method similar to the one used in those two apps for customers who desire, or need, greater accuracy.

The 2-step calibration method available in Clinometer is essentially the same used by digital levels. What's good for the goose, I think, would be good for the gander.

Using the direction of the gravitation field to calibrate the iPhone's level sensors is still inadequate in my experience. Perhaps the coding could be further tweaked to improve its functionality. If not, I would urge Apple to offer a Clinometer type calibration OPTION in Settings to offer OS wide support.

Don't take this the wrong way but the level of accuracy your looking for is unreal. These phones are used by every day people and I bet that they don't even understand "direction of the gravitation field". I will admit; I don't even know what your talking about. I would say that a vast majority of people did not even notice this issue in the first place and those that did found that 7.0.3 gave them "acceptable" results. And the rest of those customers have either swapped those devices out or are still looking for the level of accuracy you just talked about.
And I'm curious to see if you tested older iDevices with the same precision. I could see the iPhone 5s being less accurate than a iPhone 5 but I don't believe the iPhone 5 was perfect across the board as well. The parts used in these phones cost Apple about $200 dollars and the chip in question is under $10.
For me I'm noticing that 7.0.3 is more accurate when it comes to the level but I also noticed that my battery life is not as great when compared to 7.0.2. Not trying to be funny but did you notice the same decrease in battery life and are you able to explain it in the same level of detail? My point is that the hardware and software on these phones are not perfect. You fix one issue you may introduce another. Just wait until iPhone 6 is released. I know for sure that there will be another "-gate" issue. Apple is not perfect but they sure do tolerate returns better than any manufacture I'm aware of. "Level off by 1 Degree"; sure lets swap it!!!

 

[gCloud]

Two of my sides is still off by 2. I'm going to get it replaced when I get a chance.

Did u dfu restore?

 

[HEK]

Update 7.03 solved it. Reads zero when level and compass points north. Now can go back to building house. Have to fix that one slanting wall though. #

 

[Morac]

Did u dfu restore?

DFU restore doesn't magically fix hardware nor is it any better than a regular restore, which it plan to try today. Actually there's no reason to do a DFU restore unless your device is so hosed that a regular restore can't be done. The only thing DFU does is bypass the check for applying an older firmware.

 

[gCloud]

DFU restore doesn't magically fix hardware nor is it any better than a regular restore, which it plan to try today. Actually there's no reason to do a DFU restore unless your device is so hosed that a regular restore can't be done. The only thing DFU does is bypass the check for applying an older firmware.

Ok good luck

 

[rkahl]

Just wanted to encourage those of you who have not found satisfaction despite upgrading to iOS 7.0.3 to submit a feedback report (http://www.apple.com/feedback/) to Apple. I have a developer's account so I submitted it on the developer's bug report site. Although it is clear that iOS 7.0.3 has helped, it still is not enough (IMO). FYI, this is what I said…

I updated to iOS 7.0.3. All 3 of my iPhone 5S's still do not read correctly using known perfectly vertical and horizontal surfaces (I am a teaching physicist and have access to accurate lab equipment). Readings, on average, are 1º - 1.5º off for surface and horizontal. (Although the iPhone only offers readings to the units place I can interpolate and tell that they are on the verge of reading 2º off). The phones perform better in the vertical dimension. One is 1º off and two of the iPhone 5S's do read correctly. This is a definite improvement compared to iOS 7.0.2. But the results are still disappointing and affect gaming and astronomy/planetarium apps. Errors were in the range of 2º - 4º in iOS 7.0.2.

1º - 1.5º off on level over the length of an 8 foot 2x4 means one end is about 2 inches higher than the other. So, no home building with my iPhone…or leveling a moderate size picture…or use in the lab in a pinch. More importantly, it compromises their use with the astronomy apps I use in my teaching.

Using the apps, iHandy Carpenter or Clinometer, all three iPhones read SPOT ON. This leads me to believe that the calibration method used is all important. Both iHandy Carpenter and Clinometer can employ a 1-step calibration method that uses known perfectly horizontal and vertical surfaces that allow for their superior performance. Clinometer can also use a "2-step" calibration method that involves calibrating by taking readings of horizontal and vertical with the instrument in two orientations-180º different from each other. This removes the need for a perfectly horizontal and vertical surface.

I understand Apple's aversion to using a calibration method like that used for iHandy and Clinometer. Those are not "It Just Works" methodologies. But, it would be HUGE advantage to have the option to use an OS WIDE calibration method similar to the one used in those two apps for customers who desire, or need, greater accuracy.

The 2-step calibration method available in Clinometer is essentially the same used by digital levels. What's good for the goose, I think, would be good for the gander.

Using the direction of the gravitation field to calibrate the iPhone's level sensors is still inadequate in my experience. Perhaps the coding could be further tweaked to improve its functionality. If not, I would urge Apple to offer a Clinometer type calibration OPTION in Settings to offer OS wide support.

I am absolutely certain that the iPhone 5S will never satisfy your needs.

 

[ScottW1]

I am absolutely certain that the iPhone 5S will never satisfy your needs.

He stated that the Clinometer app running on the 5S produces perfectly acceptable readings. So it isn't the hardware, and it is possible to produce software that reads/reports the hardware data correctly.

Why do you think the entirety of Apple's software engineering staff completely incapable of ever producing code that works as well as that which indie developers have produced in apps like Clinometer and iHandy?

That's all he is asking for -- and it really doesn't seem like it is over the top.

 

[WilliamG]

I think it's pretty funny people believing the DND fix is actually a fix.

Anyway, if your phone reads 0 to -1 degrees, that's pretty normal. Any more than that, at this point go and exchange your phone if it bothers you. It's that simple.

 

[Morac]

So no one else wastes their time, I did a restore through iTunes today and set up my 5S as a new phone. I skipped through all the setup questions and plugged the phone in a charger, turned on DND and let it sit for about 100 minutes. When I checked it later, the back and side were 1 off and the bottom was 2 off. Exactly what it was before I did all that.

In others words it made absolutely no difference.

 

[fiat]

He stated that the Clinometer app running on the 5S produces perfectly acceptable readings. So it isn't the hardware, and it is possible to produce software that reads/reports the hardware data correctly.

Why do you think the entirety of Apple's software engineering staff completely incapable of ever producing code that works as well as that which indie developers have produced in apps like Clinometer and iHandy?

That's all he is asking for -- and it really doesn't seem like it is over the top.

+1 here. I am completely with Mr. Luigi.

 

[califlorida]

I don't know if it's been mentioned nor do I ever remember seeing it before - There is now a "Compass Calibration" switch under "System Services" in the "Location Services" menu. Not sure if this is now what's "Fixing" the problem.

To me this still feels like it's masking the issue rather than fixing it. Unfortunate that so many people continue to act like this isn't a problem, or that we should just be happy with sub-standard tech. When something costs as much as an iPhone ALL features should work 100%. IMHO

 

[KurtWilde]

I don't know if it's been mentioned nor do I ever remember seeing it before - There is now a "Compass Calibration" switch under "System Services" in the "Location Services" menu. Not sure if this is now what's "Fixing" the problem.

To me this still feels like it's masking the issue rather than fixing it. Unfortunate that so many people continue to act like this isn't a problem, or that we should just be happy with sub-standard tech. When something costs as much as an iPhone ALL features should work 100%. IMHO

That options been there for a while, not new to current update.

 

[HeadphoneAddict]

Okay, so I posted previously that I returned my week 38 iPhone 5s on 10/17 before the AT&T 14 day return period was up, so that I could wait and see for sure if this was hardware or software issue. If I exceeded the AT&T 14 day period but then returned it during the 30 days I was given by Apple, I'd lose my AT&T subsidy on the replacement phone.

I picked up a new 5s at the Apple Store today, and it was reading -4 to -5 on it's side and flat, while my iPhone 5 and my son's 5 were level.

I installed 7.0.3 at the store and that fixed the problem. I was getting 0 to -1 flat on it's back, but it settled down to 0, and 0 on it's base or right side (the two side without buttons). This was the same as what both iPhone 5 showed.

 

[Mr. Luigi]

I am absolutely certain that the iPhone 5S will never satisfy your needs.

rkahi, Have we met before? You sound as if you know me better than I know myself.

1. I am happy that iOS 7.0.3 has solved the level problems for many people.

2. I was simply offering advice to those who have not found that degree (no pun intended) of satisfaction.

3. I don't profess to be an expert on Apple, but I suspect that their goal in a very competitive market is to satisfy the needs of as many customers as possible. I think they want information about the performance of their products so that they can make the best possible decisions as to how to allocate their resources. Submitting feedback reports is one way to do that. If things are fine from your perspective, then you have no need to send such reports.

4. I have had the good fortune to teach many very bright students in a long career. Five currently work for Apple as software engineers. I know that it is well within the capability of these former students to tackle the issue regarding the level sensors. They would be insulted if you told them that it was beyond their capability to solve this problem at a high level.

5. Although the iPhone 5S may only cost Apple about $200 to build, Apple charges, I believe, between $650 and $850 off contract for the phone in the US. In many countries you can only buy it off contract and for even more than I quote. For that price I would expect its OS wide level to perform adequately for the needs of 3rd party apps that depend on the level performing commendably. I have mentioned some of those apps I use in previous posts…Sky Safari and Sky Safari+, GoSkyWatch and GoSatWatch, Star Walk, Sun Surveyor, etc. My concerns do not really have much to do with the native level app at all. I use Clinometer and its ability to accurately render tilt to a tenth of a degree. My concern has always been on these other apps that depend on the information being fed to it by the iPhone's level sensor. I assure you, if I was a developer of one of these many apps that use the level sensors I would want them to perform adequately so that it did not compromise my app and its marketability.

6. I can buy a 6 inch Stanley bubble level at Home Depot for less than $10 and if it reads level to no more than one degree of accuracy I assure you folks who care about what they are building will return it.

7. The iPhone 5S is not your grandmother's flip phone. It is an incredibly sophisticated communication device and computer. We could have used it to send men to the moon back in the 60's. There is an incredible amount of processing power packed in that little guy. More processing power, I believe, than the computers used on board those Apollo missions. Reading level should be well within it's awesome capabilities.

8. IMO, the iPhone 5S should perform better than the 4S with regard to reporting level measurements. It is 2 years more advanced than the 4S and in technology, like dog years, that is a heck of a lot of time.

9. Which brings me to my last point (for now!). Although the number "One" conjures up images of small and insignificant it is actually quite a large amount of angular measure…especially for certain purposes. Just take your iPhone and tilt it on its side until it jumps from an angular measure to the next unit up. I think you might be surprised at the amount of tilt involved. That's a problem for some and they deserve to be concerned.

Just to set the record straight. I love my iPhone 5S. It is an awesome phone. Do I love one piece of its functionality…the sensors…no. But, overall I think it is the best smartphone on the market.

It makes me happy. That doesn't mean I should refrain from sending feedback reports to Apple.

Mr. Luigi

 

[GlenK]

Very good post Mr. Luigi. I agree with you completely on the 5S. I upgraded from the 4S to the 5S and the improvements are off the charts. Many on here complain about the smallest most insignificant things they can find. The negativity is unreal.

I love this phone and all my Apple products and appreciate their responsiveness when an issues do actually arise!!

 

[Morac]

I don't get how some people can say "it's only off by one, it's good enough".

Would you accept if the calculator was off by one? How about if all the phone numbers you dialed were off by one or maybe if the alarm clock was off by an hour?

 

[Gurashieruro]

I think it's pretty funny people believing the DND fix is actually a fix.

Anyway, if your phone reads 0 to -1 degrees, that's pretty normal. Any more than that, at this point go and exchange your phone if it bothers you. It's that simple.

Then why does it work for me then? After doing a full restore (delete everything), it shows it is 3 degrees off. After DND, it is spot on. Tell me this isn't so. Though I really doubt the 90-minute DND is really what fixes it and more towards leaving the phone to re-calibrate on a flat surface upon reset, it still works. Not saying that will work for others but eh.

 

[WilliamG]

Then why does it work for me then? After doing a full restore (delete everything), it shows it is 3 degrees off. After DND, it is spot on. Tell me this isn't so. Though I really doubt the 90-minute DND is really what fixes it and more towards leaving the phone to re-calibrate on a flat surface upon reset, it still works. Not saying that will work for others but eh.

Well I'm sorry. But I don't buy it. By that logic you should be able to restore and then see it be off by 3 degrees again until the "DND fix." I bet if you restore the phone you get the same readings as you have now.

----------

I don't get how some people can say "it's only off by one, it's good enough".

Would you accept if the calculator was off by one? How about if all the phone numbers you dialed were off by one or maybe if the alarm clock was off by an hour?

Faulty analogy is faulty.

 

[Morac]

Faulty analogy is faulty.

And stupid response is stupid.

 

[Mr. Luigi]

I'm not saying I understand why, but my wife's iPhone 5S…which had been off by 1 degree for days after updating to iOS 7.0.3 now is spot on after putting it in DND mode, plugged in, and letting it lay on a "level as level can be" surface for several hours.

WTF!

As for my iPhone 5S…no such luck. Still "a degree" off and you know how I feel about that.

But, as I have already stated, the iPhone rounds to the units place. So, even though my wife's iPhone and my iPhone were both showing 1 degree off, my iPhone could have been off by more than hers (My iPhone could actually be off by 1.49º and it would read 1º. Her iPhone could have been off by .51 degrees and it would read 1º). Perhaps the DND pixie dust magic trick didn't have enough "power" for my iPhone.

Again, these are just observations. I have no hypothesis.

 

[ScottW1]

Then why does it work for me then? After doing a full restore (delete everything), it shows it is 3 degrees off. After DND, it is spot on. Tell me this isn't so. Though I really doubt the 90-minute DND is really what fixes it and more towards leaving the phone to re-calibrate on a flat surface upon reset, it still works. Not saying that will work for others but eh.

Well I'm sorry. But I don't buy it.

LOL. "Facts, observations, or results that contradict my theory must be discarded."

I'm not saying I understand why, but my wife's iPhone 5S…which had been off by 1 degree for days after updating to iOS 7.0.3 now is spot on after putting it in DND mode, plugged in, and letting it lay on a "level as level can be" surface for several hours.

WTF! As for my iPhone 5S…no such luck. Still "a degree" off and you know how I feel about that.

WTF, indeed? Same experience here -- the "90 minute DND" thing did nothing for me, despite multiple attempts. I even poked some good-natured fun at "Gurashieruro" for maybe sending me on a Snipe hunt with the whole reset+DND thing. But I believe him when he says it worked for him and seriously appreciate the time he took to relay the information here, even though it didn't end up working on my device.

Sharing experiences -- what works, what doesn't -- is the one of the biggest benefits of this forum. Thanks to all who are helping with that.

 

[Mr. Luigi]

Sharing experiences -- what works, what doesn't -- is the one of the biggest benefits of this forum. Thanks to all who are helping with that.

Agreed! And I second that last part. Mr. L

 

[WilliamG]

And stupid response is stupid.

Clearly you don't understand. That's ok. If you want to believe your phone being 1 degree off is like an alarm clock being an hour off, go right ahead.

 

[Morac]

Clearly you don't understand. That's ok. If you want to believe your phone being 1 degree off is like an alarm clock being an hour off, go right ahead.

The only difference is YOU don't use apps that require the accelerometer to work correctly, but apparently do use the alarm. If you did use the accelerometer, then you'd realize there's no difference.