Unusual Honeywell KLN94 aviation GPS failure - what could causethis?

I have seen that indication when the antenna did not have a clear view
of the sky.

The installed aircraft was an EADS/Socata TB-9 with gull wing doors.
Coming out of the pad, the pilot's door when up (as it was in the
summer) would obscure much of the sky and not allow the GPS to get a
position. Resolved by removing the blockage by repositioning the plane
or lowering the door. The antenna in this case was on the roof of the
cabin between the two gull-wing-doors.

Peter wrote:
http://www.peter2000.co.uk/aviation/kln94-failure/

This unit has been working perfectly for four years, and works fine
now.


Peter.
--
Return address is invalid to help stop junk mail.
E-mail replies to but remove the X and the Y.

Ooops. Only looked at pic number 1, not the ones showing the sat
reception. Only time I've had a slow position computation is when the
unit has been powered down for a LONG (weeks to months) time. Takes up
to 15 mins for it to "find itself" while remaining still.

An intermittent connection on the internal battery (there is a small
watch battery on the board inside - at least the KLN-89/B does) to keep
last known position and user waypoints stored. You didn't have any
error/hardware error messages on startup?

Dave

Dave S wrote:
I have seen that indication when the antenna did not have a clear view
of the sky.

The installed aircraft was an EADS/Socata TB-9 with gull wing doors.
Coming out of the pad, the pilot's door when up (as it was in the
summer) would obscure much of the sky and not allow the GPS to get a
position. Resolved by removing the blockage by repositioning the plane
or lowering the door. The antenna in this case was on the roof of the
cabin between the two gull-wing-doors.

Peter wrote:

http://www.peter2000.co.uk/aviation/kln94-failure/

This unit has been working perfectly for four years, and works fine
now.


Peter.
--
Return address is invalid to help stop junk mail.
E-mail replies to but remove the X and the Y.

How long were you in a static position on the ground? Less than 15 mins
or more than 15 mins? For whatever reason, if the box has "lost itself"
with regards to its last known position, or if the box is powered up
more than 50 or so miles away from its last known position, it can take
"a while" to find itself, and my understanding is that you need to be
still while this is happening. This process can be accelerated by
entering your current position into the box on one of the menu screens.

I've got two 89 series GPS's on my workbench, one of which will be used
for an upcoming experimental installation. My only references into this
matter are dealing with those two boxes, and the installation manual for
the 89 series.

You say the problem is fixed now, as in, not recurred? Do you have any
user defined waypoints? If there are/were, and are not now, this might
point to a problem in the memory (powered by that little battery)..
maybe there is an intermittent break in the power supply. You indicated
the box was opened within the past few months - could be pertinent in
troubleshooting this item.

Good luck,
Dave

Peter wrote:
There were no errors showing on the KLN94, other than the lack of a
fix.

I think this failure is unusual because - as far as I can tell - the
SV (space vehicle) signal indications are all normal, but the GPS
isn't producing a fix (STATE NAV D).

It is as if there was a disconnection between the GPS receiver and the
rest of the unit - all this is internal to the KLN94 so ARINC buses
don't come into it; the KLN94 doesn't have an ARINC bus anyway, and
the KMD550 MFD doesn't use ARINC, it gets the data via RS232.

It took an hour to get a fix, which is too long for any kind of slow
acquisition. The previous time it was on was about 2 days earlier; a 5
hour flight.

If the problem was just showing on the KMD550 then one could
understand a bad connection, but this is entirely internal to the
KLN94.

Satellite geometry can't be the issue because the two other GPSs (one
of which was using the same type of rooftop antenna as the KLN94) were
receiving fine.

The internal KLN94 battery should not be relevant once the power is
on. It was also changed a few months ago.

Presently I have no idea what this could be. I believe the KLN94
contains a separate GPS receiver module which emits an NMEA stream to
the rest of the unit. This stream is in fact accessible via a
connector pin on the back of the unit. A break in this stream would
account for this issue - but how does the main processor get the SV
reception data? I thought that stuff comes down the same NMEA data
stream, which is how (nearly) every GPS on the market is able to
display the constellation details.

I even changed the DME frequencies, while still on the ground, in case
it was a DME harmonic messing with the GPS signal (that's a well known
one). However, I may not have done that for long enough. The trouble
with that theory is that such interference should show up in the SV
signals too.

The selected VOR/DME would have been 113.4 MHz. I don't know if that
is a harmonic of the GPS signal.

Also the failure persisted for about an hour, by which time I would
have long been tracking a different VOR.



Dave S wrote


Ooops. Only looked at pic number 1, not the ones showing the sat
reception. Only time I've had a slow position computation is when the
unit has been powered down for a LONG (weeks to months) time. Takes up
to 15 mins for it to "find itself" while remaining still.

An intermittent connection on the internal battery (there is a small
watch battery on the board inside - at least the KLN-89/B does) to keep
last known position and user waypoints stored. You didn't have any
error/hardware error messages on startup?

Dave

Dave S wrote:

I have seen that indication when the antenna did not have a clear view
of the sky.

The installed aircraft was an EADS/Socata TB-9 with gull wing doors.
Coming out of the pad, the pilot's door when up (as it was in the
summer) would obscure much of the sky and not allow the GPS to get a
position. Resolved by removing the blockage by repositioning the plane
or lowering the door. The antenna in this case was on the roof of the
cabin between the two gull-wing-doors.

Peter wrote:


http://www.peter2000.co.uk/aviation/kln94-failure/

This unit has been working perfectly for four years, and works fine
now.


Peter.
--
Return address is invalid to help stop junk mail.
E-mail replies to but remove the X and the Y.

On 09/11/06 13:54, Peter wrote:
Dave S wrote

How long were you in a static position on the ground? Less than 15 mins
or more than 15 mins?

In a static position, maybe 5 mins. Then another 10 mins before
getting airborne.

The issue is that *if* the box lost it's position (for whatever reason),
that unless it's sitting still for up to 15 minutes, it may take it
quite a while to determine it's position (I presume this is due to
problems it has in determining it's position while it's moving).


For whatever reason, if the box has "lost itself"
with regards to its last known position, or if the box is powered up
more than 50 or so miles away from its last known position, it can take
"a while" to find itself, and my understanding is that you need to be
still while this is happening. This process can be accelerated by
entering your current position into the box on one of the menu screens.

The thing is that I have done 600hrs behind this box, and have never
had this before.

Once in 600 hours qualifies as an intermittent problem. What he was
suggesting is that if there was a battery glitch (while the unit was
turned off), then it may have lost its *memory* of its position,
necessitating the longer start-up time.


Normally, you switch it on, and as reliable as clockwork it starts up
and is ready in a few minutes. Same during flight.

This assumes it remembers where it was when it was switched off. The
owner's manual talks about this issue (actually, this is true of all
GPS units, both aircraft and personal that I've ever used).


I've got two 89 series GPS's on my workbench, one of which will be used
for an upcoming experimental installation. My only references into this
matter are dealing with those two boxes, and the installation manual for
the 89 series.

You say the problem is fixed now, as in, not recurred?

That's correct.

Do you have any
user defined waypoints? If there are/were, and are not now, this might
point to a problem in the memory (powered by that little battery)..
maybe there is an intermittent break in the power supply. You indicated
the box was opened within the past few months - could be pertinent in
troubleshooting this item.

How would the loss of the battery power prevent the GPS getting a
satellite lock?

See above. It doesn't prevent it really ... but it can make it take
longer. Moving the GPS before it's done with this can make it take
even longer.


Unless somebody can find a clue in the data next to the individual
satellites, or the date+time+location, I suspect this is a rare bug in
the firmware.

That's possible regardless of what people find, I suspect.


Of course if it happens again then I have a duff unit...

okay.



--
Mark Hansen, PP-ASEL, Instrument Airplane
Cal Aggie Flying Farmers
Sacramento, CA

Peter wrote:


How would the loss of the battery power prevent the GPS getting a
satellite lock?


It doesn't.. It legthens the time required to update its initial
position. Pretty much every gps, the first time its turned on, takes a
LONG time to obtain enough information to determine its current
location. The GPS cant tell you where you are going, or where you ARE,
unless it KNOWS where its been (even on a second to second basis).
Moving the box while this is happening lengthens that process.

If its moved more than 50 miles from its last known position (while
powered down) or if the "non-volatile" memory is lost for whatever
reason, then this LONG initial startup process occurs while box "finds"
itself. Not all of the data required for this is stored internally, and
if I remember right, is broadcast concurrently with the various time
differential signals that are used to compute the location fix.

As long as the machine REMEMBERS where it was, and its first signals
indicate that its not moved, the boot-up and acquisition process is
quite rapid.

For the average end user this is an issue that is never observed,
because this sort of thing is seen by the installer during installation,
or by the maintenance guy during update/replacement. Rarely, like when
the internal battery fails (after many years), an error message is
generated by the box that specifies the failure.

If you have an intermittent condition that isn't present on power up, it
may not annunciate it.

Good luck,
Dave

Peter wrote:

There were no errors showing on the KLN94, other than the lack of a
fix.

I think this failure is unusual because - as far as I can tell - the
SV (space vehicle) signal indications are all normal, but the GPS
isn't producing a fix (STATE NAV D).

It is as if there was a disconnection between the GPS receiver and the
rest of the unit - all this is internal to the KLN94 so ARINC buses
don't come into it; the KLN94 doesn't have an ARINC bus anyway, and
the KMD550 MFD doesn't use ARINC, it gets the data via RS232.

It took an hour to get a fix, which is too long for any kind of slow
acquisition. The previous time it was on was about 2 days earlier; a 5
hour flight.

If the problem was just showing on the KMD550 then one could
understand a bad connection, but this is entirely internal to the
KLN94.

Satellite geometry can't be the issue because the two other GPSs (one
of which was using the same type of rooftop antenna as the KLN94) were
receiving fine.

The internal KLN94 battery should not be relevant once the power is
on. It was also changed a few months ago.

Presently I have no idea what this could be. I believe the KLN94
contains a separate GPS receiver module which emits an NMEA stream to
the rest of the unit. This stream is in fact accessible via a
connector pin on the back of the unit. A break in this stream would
account for this issue - but how does the main processor get the SV
reception data? I thought that stuff comes down the same NMEA data
stream, which is how (nearly) every GPS on the market is able to
display the constellation details.

I even changed the DME frequencies, while still on the ground, in case
it was a DME harmonic messing with the GPS signal (that's a well known
one). However, I may not have done that for long enough. The trouble
with that theory is that such interference should show up in the SV
signals too.

The selected VOR/DME would have been 113.4 MHz. I don't know if that
is a harmonic of the GPS signal.

Also the failure persisted for about an hour, by which time I would
have long been tracking a different VOR.



Dave S wrote


Ooops. Only looked at pic number 1, not the ones showing the sat
reception. Only time I've had a slow position computation is when the
unit has been powered down for a LONG (weeks to months) time. Takes up
to 15 mins for it to "find itself" while remaining still.

An intermittent connection on the internal battery (there is a small
watch battery on the board inside - at least the KLN-89/B does) to keep
last known position and user waypoints stored. You didn't have any
error/hardware error messages on startup?

Dave

Dave S wrote:

I have seen that indication when the antenna did not have a clear view
of the sky.

The installed aircraft was an EADS/Socata TB-9 with gull wing doors.
Coming out of the pad, the pilot's door when up (as it was in the
summer) would obscure much of the sky and not allow the GPS to get a
position. Resolved by removing the blockage by repositioning the plane
or lowering the door. The antenna in this case was on the roof of the
cabin between the two gull-wing-doors.

Peter wrote:


http://www.peter2000.co.uk/aviation/kln94-failure/

This unit has been working perfectly for four years, and works fine
now.


Peter.
--
Return address is invalid to help stop junk mail.
E-mail replies to but remove the X and the Y.
You said the battery was changed a few months ago, then you started
having the problem. I wonder if the "repair" was not done well.

--

Regards, Ross
C-172F 180HP
KSWI