Accident Analysis - Single Point Failure

[rigpiggy]

Pilotdar that is under car/far 23, older stuff is grandfathered. Kind of like the jetstream, if you have a dual gen failure your standby runs off the ships battery. That is ok if it wasn't the battery that cuased the failure

[Blakey]

This was a very sad event and one that may not have been preventable given the lack of mechanical problems prior to IMC entry and the pilot's inability to control the aircraft without an AI. It is also VERY easy to sit here drinking my second coffee of the morning and expound on exactly what this pilot should have done. I have all the time in the world to decide on a course of action and a failure to make the correct decision will not cost me my life.

However, as a learning exercise, I would have to say that I am struck by the lack of clear communication here. This is not a new cause of course, nor one restricted to aviation, but some clearer, more focused communication may have helped.

The lack of a clear statement to the controller, telling them exactly what the problem was, has been noted but I wonder if the following may have assisted the pilot or assisted the controller in helping him:

1. Tell the controller that your IFR capability is limited. Why was the controller giving him descent clearance into cloud and heading changes in IMC? Because he did not know that these were not the best way to help this pilot.

2. Why is the controller asking him if he wants to divert to an airport with a 700' ceiling when his destination was 1600'? Because deverting to the closest airport is a good thing in most emergencies and he didn't understand that time was not the problem here but IFR control was.

3. We know the weather at his destination and at his alternate but what was the weather at altitude ahead of him and in his descent? Were there layers? Was it clear between layers? Were there any aircraft flying in that area who could tell you this information? Could he have been vectored in clear air, to a point where he could have just maintained a long, straight descent to a cloud-break on a long final? In the military, we will launch another ship to RV with the incident aircraft above cloud and "lead them down" to VMC or to a landing. Ever listened to "The Shepherd"? Was this possible? Who knows; I see no indication that any of this was even considered.

4. Did anyone have a smart phone with an AI app on it. In that location, they could probably have downloaded one while they transited North.

We can learn from this pilot's misfortune but, unfortunately, there are no new lessons here:

1. Ask for help. Ask clearly and declare an emergency if the situation warrants.

2. Divert to better conditions before that option disappears.

3. Use any and all aids to get yourself safely on the ground. This pilot was very little bother to anyone else, he didn't even want to declare an emergency.

[photofly]

4. Did anyone have a smart phone with an AI app on it. In that location, they could probably downloaded one while they transited North.

Just out of interest, I don't think this is a possibility: you actually have to have a specially designed ARS or AHRS to feed data to your phone to display. The hardware in the phone isn't sufficient, so having just an app won't work.

I have seen some AI-like apps, but as far as I know they all work like the ball (or a weight on a string) and so aren't any use for instrument flying.

[Blakey]

I couldn't say for sure; I have not tried it. Definitely not a viable flight instrument but, given only minor acceleration forces, it may help a pilot to keep their wings level.

https://www.youtube.com/watch?v=-M0sfZNfQcM

Costs $1.99.

[photofly]

The technical notes, such as they are, are here:
http://caffeinatedconsulting.com/aircraft_gyroscope
The author alludes to the difficulties, when he talks about drift removal.

I appreciate the sentiment, but a vacuum failure scenario and a real descent into IMC with three on board isn’t really the place to see how an experimental AI app works in the real world. So no, downloading that app en-route and trying it out in the circumstances of this flight isn’t a fab plan...

[Meatservo]

I'm pretty sure there is nothing in the retinue of solid-state parts in a smart-phone that will mimic the actions of a gyroscope. I think it's influenced by the same forces as your inner ear. I would be very surprised to find out that thing was any use at all, but I've been wrong before.

It was this exact scenario that finally caused me many years ago to lose my patience with the DHC-3 operator I worked for, who were very pushy about flying in poor visibility and ceilings. Like many young folk I tried to live up to their expectations until I had a flight very much like the one in the video. Obviously I had a happier outcome, but I was terrified and it changed my outlook. I remember noticing some time after the fact that one of the Otters (thankfully not the one I was in at the time) even had a vaccum-driven turn&bank indicator- driven from the same pump as the rest of the instruments!

Suffice to say I became significantly and permanently more hard-nosed about poor weather and VFR-only machines after that. The thing about one-mile visibility is that at any time, you can be a mile from ZERO visibility and not even see it coming.

[photofly]

I'm pretty sure there is nothing in the retinue of solid-state parts in a smart-phone that will mimic the actions of a gyroscope.

As a matter of fact, in a smartphone there is everything you need to mimic a rotating gyroscope. You have exactly the same kind of solid state MEMS gyro, accelerometer and magnetometer chip and GPS as is in things like the Garmin G5, and G1000. What you don’t have in a cellphone is an oven to stabilize the temperature, and hardware and firmware devoted to the task of maintaining an attitude solution accurate enough for flight. And the phone operating system doesn’t give you the access to the data you need. the tricky part Is keeping a stable reference over a duration of minutes without using the gravity vector. Aircraft are about the only systems where the gravity vector points other than down for periods of minutes at a time so it’s not a big disadvantage for most purposes.

I latterly used a single chip from Bosch that outputs a full attitude solution, connected to an Arduino, for my three axis position recorder; the chip had a built in micro controller doing the math, and it did a very good job. I’d previously used the arduino to develop my own AHRS, but I couldn’t temperature stabilize it and it had a tendency to topple when it got warm.

[digits_]

4. Did anyone have a smart phone with an AI app on it. In that location, they could probably downloaded one while they transited North.

Just out of interest, I don't think this is a possibility: you actually have to have a specially designed ARS or AHRS to feed data to your phone to display. The hardware in the phone isn't sufficient, so having just an app won't work.

I have seen some AI-like apps, but as far as I know they all work like the ball (or a weight on a string) and so aren't any use for instrument flying.

The older ones were simulated just based on GPS position, which was nice to see but completely useless. Interesting solution though to the early problems those programmers faced: not having accurate sensors, so getting the data from a completely different system.