View Single Post
  #1  
Old February 13th 04, 01:27 AM
me
external usenet poster
 
Posts: n/a
Default Strange compass behavior

Help, can someone help explain these compass symptoms? I know that the
compass should be about the most reliable piece of equipment in the plane,
but ours is acting strange.

We have a Precision PAI-700 vertical card compass mounted in a 1966 Cessna
182. We’ve swung the compass on the ground with the engine running at 1800
RPM and all electronics on. All readings are within 5 degrees; most have 3
degrees or less error.

In the air is another story. On a day with very light winds aloft we
checked the compass indication against the GPS (and the GPS against a chart)
and we had errors between plus 12 degrees and minus 19 degrees when in
un-accelerated straight and level flight, roughly on the four cardinal
directions as shown below.

GPS Compass Error
163 166 +3
272 253 -19
327 339 +12
067 062 -5

While on a long trip with the autopilot slaved to the GPS, we noticed that
the Compass does not swing smoothly in the air, (although it seems to on the
ground) but acts as though it has detents and moves in steps of more than 10
degrees. We took a reading on the GPS and the compass in un-accelerated
straight and level flight, then turned approximately 3 degrees to the right
and repeated to get the following set of readings:

GPS Compass Error
246 220 -26
249 230 -19
252 231 -21
255 232 -23
258 233 -25
261 243 -18
264 244 -20
267 254 -13

We also repeated the above measurements on a different heading, but with the
avionics master off and compared against the heading indicator.

HI Compass Error
135 135 0
130 125 -5
125 121 -4
120 120 0
115 109 -6
110 73 -37
105 60 -45
100 51 -49
95 48 -47

All of the above measurements seem very repeatable. The compass is
relatively new. At first glance it might seem that we should replace the
compass, but it swings fine on the ground, and seems to move freely on the
ground with no funny behavior. We just don’t want to replace the compass
and find we still have this problem (although that is probably our next
step).

Any ideas?

Thanks,
--Michael A.