DFC90 System Robustness

There has been some concern expressed over failure conditions and their affect on the Avidyne DFC90 and DFC100 autopilots.  The following information is designed to address these concerns.  For the purposes of this discussion, the DFC100 offers an added level of redundancy and safety simply  because the vast majority of installations are in dual-ADAHRS equipped aircraft.  In those configurations, the DFC100 autopilot is capable of taking input from either display or ADAHRS and this therefore, makes it less susceptible to any degraded capability.  Unless specifically identified otherwise, all statements made about the DFC90 can be applied to a DFC100 when accompanied by only a single ADAHRS.

As the attached slides indicate, there are only two conditions that result in a non-functional DFC90 autopilot. Specifically, they are if the sole AHRS is unavailable (indicated by Red-Xs over the attitude display on the PFD) or if the PFD has no power whatsoever (thereby rendering the AHRS to be unavailable).  

After discussion with many Avidyne customers over the years, it has become very clear to us here at Avidyne that the understanding of what constitutes a failed PFD is commonly misunderstood.   Many customers have associated a display issue (distorted or unreadable display, intermittent or non-functional buttons or knobs, etc) as a problem that would have rendered the autopilot unusable.   This is not the case. The same misunderstanding exists over the loss of air data.  Simply put, the loss of the display or air data, or almost any other failure mode does NOT render the DFC90 inoperative.

In order to provide some additional reference points in this discussion, every PFD return for 2009 was individually analyzed to determine what, if any, impact it would have had on a DFC90.   In parallel to that effort, we contacted the manufacturers of the turn coordinators (TC) that are on Cirrus aircraft and serve as the primary input to the S-Tec 55X system.   The TC manufacturers were kind enough to supply their reliability numbers to us for comparison purposes. We in turn were able to integrate that data with our own reliability data for equipment found on Cirrus aircraft and produce the proportional failure rates data that is presented in the attached charts.   Avidyne masked the vertical axis scale in order to honor our agreement with the TC manufacturers but the resulting proportional data is very telling.

The rate at which displays experience some form of failure and the rate at which TCs experience some form of failure are nearly identical.   And, they are both approximately 10 times higher than the rate at which the Avidyne solid-state AHRS experiences any kind of failure that would impact DFC90/100 autopilot operations.   This conclusion comes from empirical data from all of 2009 and not some type of estimate and is presented on the provided pareto chart.

The second part of the attached slides address each column of the pareto chart and attempt to provide the following for each failure mode:

• A brief description  of how to recognize the specific failure condition  
• A comparison of the  affect of that failure condition on both the S-Tec 55X and the Avidyne  DFC90  
• An identification  of what functionality is lost under the specific failure  condition  
• Avidyne recommended  pilot actions with respect to autopilot operations in the case of that  specific failure condition  
• A brief description  of the functionality that remains under the specific failure  condition
  

It is interesting to note that in most cases, the remaining autopilot functionality for an Avidyne DFC90/100 autopilot is higher than the remaining functionality of a S-Tec 55X under the same conditions.

< Click here to download Failure modes slides in PDF format >

Edited by BigDaddyJake - 30 Jan 2010 at 2:56pm

In reference to Brian's question (and thanks to him), there is now a pointer to this forum on COPA at http://www.cirruspilots.org/forums/p/116035/489994.aspx#489994

BTW, disregard the "Avidyne Confidential" tags on the bottom of the powerpoint slides.

Thanks for the questions.

The lack of labeling of the y-axis was done to protect the agreement we made with the turn coordinator manufacturers.   That being said, there is no variability between the y-axis incremental markings.  In other words, the proportionality between the bars is a 1-to-1.  If one bar is one tick mark high and another bar is 5 tick marks high, the 2nd bar has a 5-time higher failure rate.

I'll take a look at improving the labeling of each bar but the slides that are accessible via the green link try to add more description and explanation for each bar on the pareto.

Question - What control does either the DFC90 or the DFC100 have/exert over my rudder? Is yaw a consideration in climb? ... in turns?