A professionally designed and built panel and gyro installation had a subtle bug:  the gyro cage knob released *most* of the way to its stop, coming to rest on the panel...  which caused latent problems with gyro bearings.  A glance at your panel will tell you if you do/not have a potential issue.

Pic 1:  Gyro with cage knob.  You can see that the knob rests against the gray instrument panel.  Note that the altimeter, to the right, does not have this problem because the panel has been cut out to accommodate the altimeter knob.

Pic 2:  Gyro without the cage knob.  The notch cutout for the cage knob is just big enough to accommodate the post, but not the knob.

This bug was subtle enough to pass functional check at Lancair Avionics and to allow the gyro to function fine for over 200 hours.  During this time the cage function worked fine, and when the knob was released the gyro tracked attitude perfectly.  Around 215 hours the gyro started drifting.  Mid-Continent Instruments found some bearings binding and overhauled the unit for free.  Both the loaner that they provided in the interim, and the overhauled original gyro, when installed, would cage normally but when released, would gradually go to a max pitch up attitude.  Cage/uncage looked normal and only after every other possibility was exhausted did we think to install without the cage knob, allowing the cage post to release fully, and the problem went away.  An easy hog-out with a dremel solved the problem permanently.

I suspect this could happen with any gyro that has a cage knob; this particular unit is the 2-inch electric Mid-Continent Instruments 4200-10.  It has an advertised MTBF of 2500 hours, a 1-year warranty, Lancair recommended it, and all the buzz on it is positive, although at $4K new / $2K overhauled it ain't cheap.  I toured the fabrication and repair shops and found high quality standards, thorough testing, and experienced people who were forthright in their answers to just about any questions.  They were mystified by this problem until we finally figured it out.  Their response to my initial email was a phone call back within 2 hours and they fixed it for free because they thought it should have lasted longer.  And when it was revealed to be an installation issue, they still didn't charge me anything.

Final side note:  Due to the recent dialog on LML about lightning and electrical systems, I asked about susceptibiity to lightning.  The 4200-10 has discrete components (resistors and caps) to power LEDs for lighting, and the rotor motor.  Turns out they have installed features on the 4200s (and, I assume, their other electric instruments) to block high current pulses, and they pointed me to design and test specs, which I will go track down.  So hopefully it is not as susceptible as integrated circuits or even unprotected discrete circuit boards.  

Charley Brown
Legacy #299  230 hours

--
For archives and unsub http://mail.lancaironline.net:81/lists/lml/List.html