I didn't make it clear enough, and I think you quoted the wrong bit 
It was this bit
The GYRO chips based on MEMS technology nevertheless deliver information about their psition in Space, just like a mechanical or (lazter) Laser Gyro.
which is wrong. Neither MEMs nor any other type of gyro gives you a position in space. They are all just accelerometers.
The iPhone surely has enough sensors to be able to both sense turn rate, roll rate and pitch rate at the same time (although I'll readily accept the particular sensors they use might not be good enough for doing it in flight!)
Yes I agree, exactly. The sensor in the Iphone can, with suitable processing of the acceleration data, return pitch, roll and yaw. But it has a lot of drift and needs GPS input to be anywhere near usable. But the AHRS systems (G1000 etc) have the same problem. They all need a slow background erection process. Garmin solve it with a GPS input, Aspen do it with airdata (airspeed, heading, etc). But all can be confused if you e.g. invert the plane, and then power cycle the gyro. So long as the trajectory is a straight line, I think it will erect the right way up, wings level, no yaw.
That was not my topic, and i know that I just mentioned that the iphone has accelerometer chips, a gyro chip and a compass chip. Also i have never heard or read "accelerometer" for a gyroscope.
The only way to get a position using inertial navigation (measurement of acceleration on the 3 axes) is by starting off at a known position (which is why e.g. the gates where airliners park at a big airport have the lat/long on a big sign, though I am sure that stuff is in the FMS database anyway) and then integrating the accelerations.
Isn't that what all gyros, mechanical, lase or MEMS,do? Of course you start from a knowm position - but form there the gyro shows every movement away from the axis of the mechanical gyro or away from the vibrating plane (Mems).
So this discussion is about definitions or about iPhone apps for AI?
I have tried the one you have also, I think it neber really worked well.
But it has a lot of drift and needs GPS input to be anywhere near usable.
I never understood why Garmin needed GPS to erect the gyro. With a 9D sensor (3D gyroscope, 3D accelerometer, and 3D magnetometer), you don't need anything else to give you pitch, roll and yaw. And it does erect itself in (on average) straight and level (i.e. unaccelerated) flight. After all, a mechanical AI doesn't need GPS or airspeed either!
Yes the MEMS gyroscopes have a huge bias, so you need an online algorithm for estimating their bias. That is the reason why you usually have to couple MEMS gyroscopes harder to the gravity vector (compared to a good mechanical gyro), therefore the typical AI errors will be somewhat more pronounced than on good mechanical gyros.
The other problem is the magnetometer, it also has a huge bias, so it needs to be calibrated (usually rotated around all axes), and there are fairly huge magnetic fields in a typical cockpit, and they can change fairly much if you move the device a couple of cm, so you'd need to calibrate the magnetometer in the position it is actually used to get good results (otherwise it can be 30° off)
I never understood why Garmin needed GPS to erect the gyro. With a 9D sensor (3D gyroscope, 3D accelerometer, and 3D magnetometer), you don't need anything else to give you pitch, roll and yaw. And it does erect itself in (on average) straight and level (i.e. unaccelerated) flight.
I don't know the details but there are some complicated certification requirements involved before a MEMS gyro can be used for IFR.
Honeywell had a lot of problems with the KFD840 in this department. It was dropped early in 2013, though reportedly not for that reason.
It's obviously true that the pendulous vanes used to erect a normal vacuum AI are going to erect it to "perfectly level" if you turn on the airflow during a coordinated 30 degree bank - but that's a bit like saying coffee should be banned because it's full of illegal toxins and it would not be possible to market coffee if it was introduced today
Do any aircraft electronics use a magnetometer in the gyro? Normally it's just a fluxgate in the wingtip (etc) doing the heading and then you have a directional gyro to stabilise that heading.
It's obviously true that the pendulous vanes used to erect a normal vacuum AI are going to erect it to "perfectly level" if you turn on the airflow during a coordinated 30 degree bank
Why should that be a problem? It's still unaccelerated on average. So the gyro should still erect correctly given some time.
If your flight isn't unaccelerated on average, you'll have serious other problems to consider 8-)
but that's a bit like saying coffee should be banned because it's full of illegal toxins and it would not be possible to market coffee if it was introduced today
That's IMO a bad analogy
I was just thinking that a mechanical AI might not get certified if it was invented today
So the gyro should still erect correctly given some time.
Yes, eventually - as should a MEMS gyro, eventually.
My SG102 is a curious implementation. I can put that up at a 30 degree roll angle overnight and it still indicates 30 degrees in the morning. And it's exact. So clearly they have put in gravity sensing in there; not just acceleration. I have no idea what device(s) they are using and they won't say. But the SG102 is not certified for use as a primary pitch/roll reference - for that you still need external input... for whatever reason.
It would be interesting to find out why precisely the SG102 cannot get certified as primary.
(why can i edit some posts and others not?)
There is a 2 hour edit window.
so i better check for my typos sooner :-) embarrassing ...
Zuke here again across the pond in the desert of New Mexico. I just came back from a 2 hour flight. It was bumpy and turbulent. That makes for a good test for the pocket horizon pro app on my ipod along with the bad elf waas gps. I have it mounted just to the left of my vacuum sucking gyro. On a scale of one to ten with ten being perfect I would rate the performance of the PH app at 8.5. Overall. Upon arrival at at my home airport, I covered up my primary gyro and shot the ILS only using PH for reference. It performed well...easy approach only using backup.Had it been a dark night in IMC conditions with a total vacuum failure this app would of kept me alive. Though I have never flown behind dynons d1,I'll bet it isn't much better if at all.
Alexisvc....I could feel the frustration in your last words. Not fun to spend hours,days,weeks,and years trying to perfect a system that is lacking.....the apple gyro and slow drivers and gps.
I talked to an engineer at st microelectronics. They spit these marginal chips out by the hundreds and have no quality control. They install them in apple products and after passing minimal tests, they go to the consumer. What that means is probably no two chips will perform exactly the same.
Just thought you guys across the pond would like to know
