Backup gyros - which do you trust?

"Steve House" wrote:
I've been reading with interest the several threads where a number of
people
have strongly pointed out the advantages of a backup electric AI to
supplant
a vacum driven main AI. But I'm reminded of the saying "A man with a good
watch always knows what time it is. A man with two watches is never
sure."

This is a very interesting issue, to me. Reading the records of IMC
loss-of-control accidents is very unsettling to this single pilot IFR flyer
because of the cases where there *was* backup attitude instrumentation
available. Even when there wasn't, the pilots usually had at least the turn
coordinator to help keep the aircraft upright. It is too simple to chalk up
all these accidents simply to lack of proficiency. There is something else
going on - some human factors issue that has not been properly identified. I
suspect it may be related to task saturation. If so, instrument panel
clutter could be a contributing factor.

So I'm toodling along in IMC with no outside horizon reference and I see
my
two AIs don't agree with each other. How do I determine which to trust?
If
I had a third, I could go with a 2 of 3 voting strategy of course, but
with
only two, what do you do to decide which is operating properly and which
one
has faulted? Obviously I can look for consistency with other
instruments -
does my DG or Turn indicator show I'm turning, does the VSI show a climb
or
descent - but what would be the best strategy given the various ways
vacuum
or electric driven instruments can fail?

My strategy is to include a yoke-mounted GPS displaying a synthetic HSI in
my scan. This works wonderfully well in training, but I am not sure how well
I would do in a real situation where my AI suffered a gradual failure.
--
Dan
C172RG at BFM

There are several things you can add to help with the identification, In my
plane I have a low vacuum warning light (part of the precise flight backup)
mounted between the AI and DG. The AI is one of the sigmatec ones with a vacuum
flag, so that if vacuum is lost in the instrument but not in the system I still
know about it right away. These warnings cover identification of the more
common cause of loss of the AI. The other failure mode would be failure of the
gyro, in which case I don't believe you get the insidious gradual spin-down like
you do with loss of vacuum. I also fly with the GPS on the HSI page to offer
yet another source of redundancy.

Personally, I think the instrument scan typically taught relies too heavily on
the AI given its relatively low reliability. Unfortunately, the alternative is
a scan that works a bit more like a partial panel scan using the AI as
supporting, not primary. Such a scan is much harder to master and requires
considerable finesse to keep from chasing the needles. It is not one I would
expect to be able to teach someone just learning to fly by instruments.


--
--Ray Andraka, P.E.
President, the Andraka Consulting Group, Inc.
401/884-7930 Fax 401/884-7950
email
http://www.andraka.com

"They that give up essential liberty to obtain a little
temporary safety deserve neither liberty nor safety."
-Benjamin Franklin, 1759

Dan Luke wrote:

This is a very interesting issue, to me. Reading the records of IMC
loss-of-control accidents is very unsettling to this single pilot IFR flyer
because of the cases where there *was* backup attitude instrumentation
available. Even when there wasn't, the pilots usually had at least the turn
coordinator to help keep the aircraft upright. It is too simple to chalk up
all these accidents simply to lack of proficiency. There is something else
going on - some human factors issue that has not been properly identified.

Concur. I, too, don't think it's entirely lack of proficiency. I think
there are pilots who have training and proficiency, who, in the words
of my CFI, "ought to be able to do it", who don't. And clearly backup
AI is no panacea.

Obviously I can look for consistency with other instruments -
does my DG or Turn indicator show I'm turning, does the VSI show a climb
or descent - but what would be the best strategy given the various ways
vacuum or electric driven instruments can fail?

Having an inventive CFI who has little habits like mind-f***ing
me into doubting my AI while palming the TC fuse, I think the best
strategy is delimited above. *Instrument cross check is essential*

Rod Machado's "Instrument Pilot Survival Manual" delineates something I
haven't seen elsewhe

Turn triangle of agreement: AI, TC, compass
Pitch triangle of agreement: AI, VSI, alt static on/off

The point is to deliberately cross-check instruments
which depend upon independent power sources.

The problem (for me anyway) in training is that my compass is
mounted on the windshield bow and it's impossible to keep it
in my scan in VMC under the hood w/out extensive "cheating".

I also think Machado's under-utilizes ASI and hearing. I
think the reasoning is that there are three sources of ASI
failure and only two for VSI, one of which alt static
eliminates. But when forced to fly instruments without static
instruments, I found hearing was a fairly precise means of
pitch control (at constant power for a fixed-pitch prop)

Interested to see what others say: this topic should elicit a
lot of opinions.

Cheers,
Sydney

You have three sources of bank information in a typical panel: the attitude
indicator, the turn coordinator, and the heading indicator. If two agree and
the third does not, it is faulty. Add a fourth source and it makes
elimination that much easier.

Bob Gardner

"Steve House" wrote in message
...
I've been reading with interest the several threads where a number of
people
have strongly pointed out the advantages of a backup electric AI to
supplant
a vacum driven main AI. But I'm reminded of the saying "A man with a good
watch always knows what time it is. A man with two watches is never
sure."
So I'm toodling along in IMC with no outside horizon reference and I see
my
two AIs don't agree with each other. How do I determine which to trust?
If
I had a third, I could go with a 2 of 3 voting strategy of course, but
with
only two, what do you do to decide which is operating properly and which
one
has faulted? Obviously I can look for consistency with other
instruments -
does my DG or Turn indicator show I'm turning, does the VSI show a climb
or
descent - but what would be the best strategy given the various ways
vacuum
or electric driven instruments can fail?

Dan Luke wrote:

It is too simple to chalk up all these accidents simply to lack of
proficiency.

I was just reading the NTSB report of the King Air that crashed in
Colorado, attributed to spatial disorientation after a partial panel
failure. It seems representative of the problem. The facts are
chilling:

- IMC, alt. 23,200 ft.
- Two person cockpit.
- Experienced pilot - 5117 hours total, 2520 in type.
- Partial panel loss due to AC power failure.
- Failure immediately indicated by flags on affected instruments.
- Remaining instruments, powered by vacuum:
Left - airspeed, turn/slip,
Right - airspeed, turn/slip, altimeter, attitude.
- Aircraft began gently increasing turn within one minute of failure.
- Time between instrument loss and impact - one minute, 33 seconds
- Flight path consistent with graveyard spiral

http://www.ntsb.gov/publictn/2003/AAR0301.pdf

There is something else going on - some human factors issue that
has not been properly identified. I suspect it may be related to
task saturation. If so, instrument panel clutter could be a
contributing factor.

One comment in the report was that the pilot might have had a tendency
to focus on a single problem, and mot paid attention to other things.
He could have been trying to troubleshoot the electrical problem, and
not handed control over to the copilot, who would have had a better view
of the remaining functional instruments.

In any event, it is amazing how quickly the pilot lost control of the
aircraft, considering how this should have been fairly routine: If an AC
inverter had failed, then the changeover to the remaining inverter is
accomplished with a simple flip of a switch, and should have been almost
a reflexive action. The failure would have been immediately obvious, so
it wasn't one of those insidious failures that people don't notice at
first. An experienced IFR pilot should have been aware of the need to
maintain attitude and yet lost control almost immediately. In reading
the report, it seems like such an avoidable accident, yet...

The presence of two attitude indicators is especially valuable when they
disagree. That disagreement will direct your attention to the needle/ball
and basic flight instruments to help determine which one is correct. With a
single AI you could more easily follow a gyro error without noticing a
difference in the other basic instruments until it was too late.

--

Darrell R. Schmidt

B-58 Hustler History: http://members.cox.net/dschmidt1/


"Steve House" wrote in message
...
I've been reading with interest the several threads where a number of
people
have strongly pointed out the advantages of a backup electric AI to
supplant
a vacum driven main AI. But I'm reminded of the saying "A man with a good
watch always knows what time it is. A man with two watches is never
sure."
So I'm toodling along in IMC with no outside horizon reference and I see
my
two AIs don't agree with each other. How do I determine which to trust?
If
I had a third, I could go with a 2 of 3 voting strategy of course, but
with
only two, what do you do to decide which is operating properly and which
one
has faulted? Obviously I can look for consistency with other
instruments -
does my DG or Turn indicator show I'm turning, does the VSI show a climb
or
descent - but what would be the best strategy given the various ways
vacuum
or electric driven instruments can fail?

"Ray Andraka" wrote in message
...

Good reason not to depart in low IMC I guess. I agree, that the AI is a
weak
link and carries with it some risk. While nice to have, cost and panel
space
make the back up AI difficult in some aircraft.

I think it is a matter of priorities... in an airplane which is IFR
certified, it is hard to believe there is not some space or economic
compromise which could not be made if a pilot felt this were an important
enough issue. I know I will get diagreement on this as always, but I think
the electric AI comes first before an IFR GPS. Even C152s sometimes have
Garmin 430s/530s nowadays; an electric AI would make much more sense IMHO.

As an even better solution, Hal Sheevers of Sporty's has for quite some time
been lobbying the FAA to permit an electric AI to replace a turn
coordinator... it does not seem as if the issue is getting very far with the
FAA, but I do think that would be a very good compromise if we started to
see electric AIs installed in place of the turn coordinator on planes where
panel space is tight.


--
Richard Kaplan, CFII

www.flyimc.com

The problem isn't just finding a home for it, it is finding a home for it that
is in a spot where it will be in the pilot's primary scan. I don't think it
would do much good on the other side of the panel where it might only be
referenced once a minute or less. I also agree that an IFR GPS is not
necessarily a high priority. Frankly, that (the GPS) is a lot of money for very
little added capability. In the case of a GPS/Nav/Comm, it also introduces a
single point of failure for all of the electronic nav gear with no back-up other
than what might be in your flight bag.

It is a shame that many of the simulators do not fail the AI the in the gradual
way it fails for real. I think training with realistic failures is the best way
to be able to recognize the failure (some of the accident reports indicate even
with redundant AI's recognition is not guaranteed). Unfortunately, the failures
can't be realistically simulated in the aircraft (at least not without an
illegal mod to the vacuum system), so ground based simulators must fill in
there. I've hear that some of the high end simulators such as the Frasca's do a
realistic AI fail, but the PCATDs I've played with all just pop from working to
tipped over instantaneously. Any of the current breed of PCATD's do any
better? Mine is a really old version of Elite, from about 1995 or so.

Richard Kaplan wrote:

"Ray Andraka" wrote in message
...

Good reason not to depart in low IMC I guess. I agree, that the AI is a
weak
link and carries with it some risk. While nice to have, cost and panel
space
make the back up AI difficult in some aircraft.

I think it is a matter of priorities... in an airplane which is IFR
certified, it is hard to believe there is not some space or economic
compromise which could not be made if a pilot felt this were an important
enough issue. I know I will get diagreement on this as always, but I think
the electric AI comes first before an IFR GPS. Even C152s sometimes have
Garmin 430s/530s nowadays; an electric AI would make much more sense IMHO.

As an even better solution, Hal Sheevers of Sporty's has for quite some time
been lobbying the FAA to permit an electric AI to replace a turn
coordinator... it does not seem as if the issue is getting very far with the
FAA, but I do think that would be a very good compromise if we started to
see electric AIs installed in place of the turn coordinator on planes where
panel space is tight.

--
Richard Kaplan, CFII

www.flyimc.com

--
--Ray Andraka, P.E.
President, the Andraka Consulting Group, Inc.
401/884-7930 Fax 401/884-7950
email
http://www.andraka.com

"They that give up essential liberty to obtain a little
temporary safety deserve neither liberty nor safety."
-Benjamin Franklin, 1759

"Ray Andraka" wrote in message
...

The problem isn't just finding a home for it, it is finding a home for it
that
is in a spot where it will be in the pilot's primary scan. I don't think
it

Put it where the turn coordinator is located and the put the turn
coordinator off to the side somewhere.. the regs say you must have a turn
coordinator but do not say where the turn coordinator has to be on your
panel.


tipped over instantaneously. Any of the current breed of PCATD's do any
better? Mine is a really old version of Elite, from about 1995 or so.


The current Elite software allows a choice between instant or gradual AI
failure. They have a reasonable upgrade program for their software as well.


--
Richard Kaplan, CFII

www.flyimc.com

"Gary L. Drescher" writes:

In MS FS2002, the AI fails abruptly, in just a few seconds. But I've heard
that FS2004 (due late this month) has more-realistic gyro failures.

If you don't want to wait, FlightGear already has gradual gyro
failures. You can fail an individual gyro or an entire system
(i.e. vacuum or electrical):

http://www.flightgear.org/


All the best,


David

--
David Megginson, , http://www.megginson.com/