Hi,
I'm using the TrackAddict app on an iPhone6 for logging data only (video is switched off).
What is the best phone orientation to use for G force readings? Is laying the phone on its back with the top pointing in the direction of travel ok? Or should I still stand the phone up (in portrait mode) with it's back facing the direction of travel as if I was recording video?
I've been having a play around and can't figure it out.
Thanks in advance,
Binney
TrackAddict will automatically calibrate the accelerometer readings for most typical mounting positions, but that will usually only occur when it detects that the vehicle is stopped (which needs a good GPS signal), and this has to be before a session recording has started. So, you would first mount the phone where you want it, and just have the app running for a few seconds with a green GPS icon before you go to Record. In actual practice, you typically wouldn't need to wait any longer than until you get a green GPS icon.
There are two catches to that:
1) If the phone is upright, it has no way to detect any left/right yaw. So, it should be mounted straight, so that if the camera were on, it would be looking straight ahead with the vehicle's path of travel, and not at a left or right angle. Its back should also be to the front of the vehicle, so that the screen is facing the driver. The automatic calibration should account for pitch and roll, although it's ideal (but not required) if you mount it perfectly straight and level.
2) If the phone is laid down flat, then it needs to have its back facing down (ie so the screen is visible), and in a "portrait" orientation so that the top of the phone is the same direction as the front of the vehicle, as straight as possible. Although the center of a vehicle can be ideal for accelerometer readings, it is often a poor reception location for the phone's internal GPS (unless maybe if you have an open convertible top or roadster), so please keep that in mind if you are not using an external GPS.
If you were to start recording before mounting the phone, are not securely mounting the phone to the vehicle, or were attempting to set this up while the vehicle was moving, then the resulting G-force data probably wouldn't be very good. It's definitely important to secure the phone where you want it, with the vehicle stopped, before you tell it to start recording.
Thank you for the fast and detailed reply.
Just what I wanted to know!
Good info. The way my phone mount sits there is some amount of "yaw" present in it. I wish there was a feature to 'calibrate' the accelerometer or g-force info and always use that profile so you don't have to worry about having the phone mounted/car stopped before starting the logging. Some sort of yaw offset correction would be cool too, but not sure how feasible that one is.
Great app either way, I love how easy and straight forward it is to use.
Interesting...
1. Do Acceleration readings come from iPhone accelerometers or are they derived from GPS reports?
2. I assume the units are ft/sec^2 ?..... and is there a MAX
3. How are the axes oriented relative to the car? ...I assume "z" is up; if "x" is toward the front of the car, then "y" is out the drivers door?
TrackAddict uses the accelerometers in the phone. Calculating this data from the GPS results in substantially less precision, accuracy, and responsiveness. That is an option in our RaceRender desktop software, but it will use the real accelerometer readings by default when available.Originally Posted by my911
They are in gravitational units, G's; 1.0 G = 9.80665 m/s^2 = 32.1740 ft/s^22. I assume the units are ft/sec^2 ?..... and is there a MAX
I haven't had the opportunity to drive a car that could max out the iPhone's accelerometer hardware, but the practical maximum is expected to be between 2 and 3 G in most cases. There have been unconfirmed reports that some older iPhone models had an effective limit right around 2.3 G. I have not seen any updated data or theories for the more recent iPhone models, except for some mentions that the underlying hardware may have improved. On the Android side, it will vary depending on the device model, with some reportedly being good for 2.0 G and others for 4.0 G, although other factors may impact the effective maximum reading.
For reference, most production-based race cars on race tires tend to see around 1.5 G lateral in the corners... For many, that's just the peak that it spikes to; for the more exceptional, that's sustained with spikes up near 2.0.
X is longitudinal accel/decel (ie related to throttle/brake), Y is lateral (left/right), Z is up/down vertical3. How are the axes oriented relative to the car? ...I assume "z" is up; if "x" is toward the front of the car, then "y" is out the drivers door?
Weston
Thanks
Crisp info
Now I can go to work
1) Wow, very interesting and useful information. Thanks, Binney, for asking this question, and thanks, Weston, for such a thorough answer.
2) If accelerometer data is of no interest -- say, you're only concerned about Latitude, Longitude, Altitude, and Speed -- does iPhone orientation matter (as long as the phone isn't flopping around)?
Thanks,
Bill P.
The accelerometer data is primarily used for data logging and statistics, and should not impact the other readings. If you don't care about that accelerometer G-force data, the orientation and automatic calibration is not very important. The only other things that the accelerometers are currently used for is device orientation (for the on-screen display and video recording), and detecting vehicle movement when in Segment or Drag Race mode, although that should still work in most any orientation.
The phone will still want to be mounted sturdily and with a clear line-of-sight to the sky, for the sake of the GPS... We've found that if the GPS is free to move around, independent of the vehicle, it can really make a mess of the GPS data (which affects the map plot, speed reading, lap timing accuracy, etc). That tends to happen when motorcyclists or kart drivers place the phone into their pocket, or even if the phone is placed in a cup holder in a car... It may work for a quick test on the street, but things often go bad under race track conditions. A windshield mount is a good way to help avoid this in cars.
Weston: good to know (especially as a motorcyclist). Thanks for another clear and thorough explanation.
I have posted about this before and have tried a couple different mounts, but still no luck. Heavy/extended braking my G-force readings seem to get wonky and actually show that I am accelerating. The video below I am using a RAM mount with the phone oriented in 'landscape' with it in line with the car like the help files to direct you to etc. Very stable mount setup and I have now tried it on 2 devices-- same results (LG G2 and LG G4). The beginning of the video/front straight illustrates it well- when I let off the gas and ease into the brakes everything is good, but once I am into hard braking, the direction flips 180. Then as I go into the turn and slow my rate of 'decel', it jumps back the other direction.
Any ideas?
https://www.youtube.com/watch?v=idSIOb2LiI4
I've been reviewing past test data, and have found a few similar occurrences of this anomaly with the Android version of TrackAddict while in landscape orientation; the iOS version appears to be unaffected. I'm not yet sure if it's a hardware issue, something in the Android OS, or just our own bug somehow. I'm working on getting to the bottom of this, and hope to know more soon...
Sounds good, let me know if you need any more help with testing- I'd be glad to hook it up on my truck on the way into work or something
I was able to identify and resolve the problem this afternoon. I also found a small opportunity for improvement in some related code, which should boost the quality of accelerometer data in general. Those will in version 3.2.1, which I hope to release for Android in the next few days...
If it helps anyone I just purchased a Rokform "sport case" and they have an awesome windshield mount that's easily hefty enough for my iPhone 6S plus. They manufacture race parts for motorcycles and then decided to jump into mobile phone accessories. They're at Rokform.com if anyone wants to check them out.
Hi, I'm using v.3.3.3 on android and don't get any G-Readings when the phone lies on its back in portrait mode... Probably not 100% flat, but within about 10? nick. However in the logs I always get 0.00 for x- and y- acceleration.
Are you using the linear acceleration sensor from android or the accelerometer settings (andsosensors shows the difference quite nicely, the linear acceleration are probably what you want for g-readings and the accelerometer/gravitiy for detecting the device oriantation).
I'm happy to prove some debug input on these sensors. Not sure though if it's a problem of my phone..
BTW: if I use it in landscape mode camera facing front (to record) everything works perfectly!!
any idea/help appreciated, thanks!
I had some time to investigate that problem this weekend, and it appears to be a bug when calibrating the accelerometers for a flat device orientation. We're obtaining the raw sensor data on both Android and iOS, and then using our own calibration trigonometry to detect and account for both pitch and roll, but just had a simple bug that wasn't saving those calibration values in the special-case logic we use with a flat device orientation. As a result, it would either get stuck in calibration mode (reporting zeros), or it might even have bogus values in some cases.
I have developed a fix, will test on both Android and iOS in the morning, and then that will be released with our very next app update... The problem only affects when the device is laying flat. If it's upright in landscape or portrait mode (with the back of the device facing the front of the vehicle), it should work without issue.
sounds good, thanks! I'm happy to test and report if needed..
Reply With Quote