Filtered Flight Data, 47857-47868 [E8-18933]

Agencies

• Federal Aviation Administration
• DEPARTMENT OF TRANSPORTATION

[Federal Register Volume 73, Number 159 (Friday, August 15, 2008)]
[Proposed Rules]
[Pages 47857-47868]
From the Federal Register Online via the Government Printing Office [www.gpo.gov]
[FR Doc No: E8-18933]


=======================================================================
-----------------------------------------------------------------------

DEPARTMENT OF TRANSPORTATION

Federal Aviation Administration

14 CFR Parts 121, 125, and 135

[Docket No. FAA-2006-26135; Notice No. 08-08]
RIN 2120-AI79


Filtered Flight Data

AGENCY: Federal Aviation Administration (FAA), DOT.

ACTION: Supplemental notice of proposed rulemaking (SNPRM).

-----------------------------------------------------------------------

SUMMARY: The FAA is amending its proposal to prohibit the filtering of 
some original flight recorder sensor signals. Comments to the NPRM 
published in November 2006 and changes in available technology have 
caused us to reexamine our position on data filtering. We are now 
proposing that certain critical data parameters may be filtered if a 
certificate holder can show that the data can be accurately 
reconstructed. This proposed rule would improve the integrity and 
quality of the data recorded on digital flight data recorders while 
giving aircraft designers and operators more flexibility in system 
design and operation where allowable, including an option to filter 
data.

DATES: Send your comments on or before November 13, 2008.

ADDRESSES: You may send comments identified by Docket Number FAA-2006-
26135 using any of the following methods:
     Federal eRulemaking Portal: Go to https://
www.regulations.gov and follow the online instructions for sending your 
comments electronically.
     Mail: Send comments to Docket Operations, M-30; U.S. 
Department of Transportation, 1200 New Jersey Avenue, SE., Room W12-
140, West Building Ground Floor, Washington, DC 20590-0001.
     Hand Delivery or Courier: Bring comments to Docket 
Operations in Room W12-140 of the West Building Ground Floor at 1200 
New Jersey Avenue, SE., Washington, DC, between 9 a.m. and 5 p.m., 
Monday through Friday, except Federal holidays.
     Fax: Fax comments to Docket Operations at 202-493-2251. 
For more information on the rulemaking process, see the SUPPLEMENTARY 
INFORMATION section of this document.
    Privacy: We will post all comments we receive, without change, to 
https://www.regulations.gov, including any personal information you 
provide. Using the search function of our docket Web site, anyone can 
find and read the electronic form of all comments received into any of 
our dockets. This includes the name of the individual sending the 
comment (or signing the comment for an association, business, labor 
union, etc.). You may review DOT's complete Privacy Act Statement in 
the Federal Register published on April 11, 2000 (65 FR 19477-78).
    Docket: To read background documents or comments received, go to 
https://www.regulations.gov and follow the online instructions for 
accessing the docket. Or, go to the Docket Management Facility in Room 
W12-140 of the West Building Ground Floor at 1200 New Jersey Avenue, 
SE., Washington, DC, between 9 a.m. and 5 p.m., Monday through Friday, 
except Federal holidays.

FOR FURTHER INFORMATION CONTACT: For technical questions concerning 
this proposed rule contact Brian A. Verna, Avionics Systems Branch, 
Aircraft Certification Service, AIR-130, Federal Aviation 
Administration, 800 Independence Avenue, SW., Washington, DC 20591; 
telephone (202) 385-4643; fax (202) 385-4651; e-mail 
brian.verna@faa.gov. For legal questions concerning this proposed rule 
contact Karen L. Petronis, Senior Attorney for Regulations, Regulations 
Division, Office of the Chief Counsel, AGC-200, Federal Aviation 
Administration, 800 Independence Avenue, SW., Washington, DC 20591; 
telephone (202) 267-3073; fax 202-267-7971; e-mail 
karen.petronis@faa.gov.

SUPPLEMENTARY INFORMATION: Later in this preamble under the Additional 
Information section, we discuss how you can comment on this proposal 
and how we will handle your comments. This discussion includes related 
information about the docket, privacy, and the handling of proprietary 
or confidential business information. We also discuss how you can get a 
copy of this proposal and related rulemaking documents.

Authority for This Rulemaking

    The FAA's authority to issue aviation safety rules is found in 
Title 49 of the United States Code. Subtitle I, Section 106 describes 
the authority of the FAA Administrator. Subtitle VII, Aviation 
Programs, describes in more detail the scope of the agency's authority.
    This rulemaking is promulgated under the authority described in 
Subtitle VII, Part A, Subpart III, Section 44701. Under that section, 
the FAA is charged with prescribing regulations providing minimum 
standards for other practices, methods and procedures necessary for 
safety in air commerce. This regulation is within the scope of that 
authority since flight data recorders are the only means available to 
account for aircraft movement and flight crew actions critical to 
finding the probable cause of incidents or accidents, including data 
that could prevent future incidents or accidents.

I. Background

A. Statement of the Problem

    During several aircraft accident investigations, the National 
Transportation Safety Board (NTSB or Board) found that some flight data 
recorder systems were filtering flight recorder parameter signals 
before they were recorded. As a result, the recorded data did not 
accurately reflect the aircraft's performance or the movements

[[Page 47858]]

of the flight control systems before and during the accident or 
incident under investigation. Such signal filtering both hampered and 
delayed the investigations. Throughout the investigation of American 
Airlines Flight 587 (Flight 587), the NTSB expended significant 
resources and time trying to recreate the performance and movements of 
the flight controls of the accident aircraft.
    In November 2003, the NTSB issued three recommendations (NTSB 
Recommendations A-03-48/A-03-49/A-03-50, November 6, 2003) on digital 
flight data recorder (DFDR) recording requirements. The NTSB 
recommended that the FAA require all aircraft have a DFDR system 
installed ``capable of recording values that meet the accuracy 
requirements through the full dynamic range of each parameter at a 
frequency sufficient to determine a complete, accurate, and unambiguous 
time history of parameter activity, with emphasis on capturing each 
parameter's dynamic motion at the maximum rate possible, including 
reversals of direction at the maximum rate possible.''

B. Action by the FAA

    The FAA agreed with these recommendations. In 2006, we issued a 
notice that proposed a prohibition on filtering certain original flight 
data sensor signals. The 2006 NPRM contains a more complete discussion 
of the proposal and the events leading up to it (November 15, 2006, 71 
FR 66634).
    The comments we received on the proposed rule alerted us to several 
features of the proposed prohibition that could have significantly more 
impact than we would have predicted. The first was the proposed 
definition of filtering, which described it as a change to any original 
parameter signal for any reason other than the three specified in the 
rule. The comments indicated that the level of signal processing that 
is in use on newer flight data systems no longer coincides with more 
traditional concepts of filtering, and leaves in question whether 
current system designs would be considered filtering. A more detailed 
consideration of this issue is included in the discussion that follows.
    As we considered whether to change the proposed definition of 
filtering, we studied what is quickly becoming the standard in 
electronic signal processing. Our intent in the 2006 NPRM was to 
prohibit the processing of certain flight data sensor signals that 
would result in inaccurate data being preserved, as happened with the 
rudder movement data on Flight 587. That airplane crashed after takeoff 
from LaGuardia Airport in November 2001.
    The investigation following the crash of Flight 587 indicated that 
the issue was not that data were filtered, but that the actual rudder 
movement data could not be reconstructed once processed by installed 
filtering devices. While a prohibition like our 2006 proposal would 
solve the problem, current capabilities suggest that when properly 
processed and documented, data can be recovered from a system design 
that incorporates filtering. We are now proposing that data filtering 
be allowed if a certificate holder can demonstrate that the `filtered' 
recorded data meet the recording requirements of the regulations, and 
that the original sensor signal data can be accurately reconstructed 
using a documented, repeatable process.

C. Changes in This Supplemental Notice of Proposed Rulemaking

    We have moved the filtering regulation from proposed Sec.  
121.344(n) to its own section, Sec.  121.346 (or complementary sections 
in other parts), to avoid any confusion with the other flight recorder 
requirements. While today's document uses part 121 as a reference, we 
are proposing the same changes for parts 125, and 135.
    Section 121.346 begins with a definition of filtering that differs 
slightly from the one we proposed in 2006. The new proposed definition 
indicates that when a signal is formatted in any way to be DFDR 
compatible, it is not considered filtered. As discussed below, we 
disagreed with commenters that thought the definition was too specific 
to allow certain current practices in signal conditioning.
    We have changed our position from a strict filtering prohibition to 
one of conditional allowance that distinguishes between two groups of 
flight recorder parameters. The first contains those that are 
prohibited from being filtered unless a certificate holder can 
demonstrate that it has done the tests, analyses, and maintains the 
procedures necessary to reconstruct the original sensor signal values 
from the filtered recorded data. The FAA understands that such 
technology is available, and where it can be shown to be reliable, it 
will be allowed. The second includes those parameters whose signals may 
be filtered without further action.
    We have split the 88 parameters listed in Sec.  121.344(a) into two 
groups.
    The first group includes those parameters that may be filtered only 
if a certificate holder can demonstrate that the original sensor signal 
data (values) can be reconstructed using a valid, repeatable procedure. 
This group includes parameters such as flight control positions, flight 
control input forces and flight control surface positions that are 
sampled at higher rates. Filtering can and has been shown to result in 
incorrect information being recorded. This group also includes those 
parameters that are considered critical to accident investigation, 
including altitude, airspeed, heading, vertical and lateral 
acceleration, pitch attitude, roll attitude and engine thrust. 
Parameters that may not be filtered unless valid reconstruction is 
demonstrated are numbers 1-7, 9, 11-18, 26, 32, 42, 43, 68, 70, 77, and 
88.
    The second group includes all parameters that may be filtered 
without further showing by the certificate holder. We have found that 
filtering these parameters still results in acceptable recorded data as 
long as the accuracy and all other requirements of the applicable 
appendix are met. A significant number of parameters in this group are 
those that have only two states, on or off. These are known as 
discretes. We received several comments indicating that the information 
from these parameters is not affected by filtering. There are also 
discretes that may have three or four preprogrammed fixed values as 
their output, and are recorded as a binary numerical representation 
(BNR). Examples are traffic alert and collision avoidance system 
(TCAS), electronic flight instrument system (EFIS) display, and primary 
navigation system reference. Filtering these parameters would not 
result in any non-programmed value being recorded, nor would it affect 
the determination of the value at a given time. Other parameters 
included in the `may be filtered' group are secondary flight control 
surfaces and input positions, such as trailing and leading edge flaps 
and pitch trim surface position. Filtering does not appear to affect an 
investigator's ability to determine the position of this equipment. 
Parameters that may be filtered without further action are numbers 8, 
10, 19-25, 27-31, 33-41, 44-67, 69, 71-76, and 78-87.
    The option not to filter any or all parameters remains an 
acceptable means of compliance with the regulations. In all cases, the 
accuracy and all other requirements of Appendix M (or Appendix E or 
Appendix F) must continue to be met. The ability to reconstruct data 
does not forgive any appendix requirement for any parameter. For 
purposes of this SNPRM, when we refer to Appendix M in our response to 
a comment, this response also applies to operations covered by

[[Page 47859]]

Appendix E to part 125 or Appendix F to part 135 unless otherwise 
noted.
    The proposed time for compliance with this rule is four years after 
the effective date of the final rule. Within that four-year period, one 
of two things must happen.
    If a certificate holder elects not to filter any of the restricted 
parameters, it has four years to test its DFDR systems, verify that 
none of the restricted parameters are being filtered, or, if any 
restricted parameter is being filtered, modify that parameter.
    If a certificate holder chooses to filter a restricted parameter 
and show by test and analyses that the originating signal can be 
reconstructed, the procedures for reconstruction would have to be 
submitted to the FAA after the next heavy maintenance check of an 
airplane (beginning six months after the effective date of the final 
rule), but not later than two years after the effective date of the 
final rule. If a certificate holder has several of the same make, model 
and series airplane (group) with the same certificated DFDR system 
installed, the procedures need only be submitted once for the entire 
group of airplanes with identically installed systems. The compliance 
date for a group would be tied to the first airplane going in for a 
heavy maintenance check six months after the rule is final. Submission 
of the data to the FAA would be required no later than the time the 
first airplane of a group completes that heavy maintenance check. This 
compliance schedule is intended to allow the FAA to determine that the 
submitted reconstruction procedures are repeatable, but still allow 
time for other compliance action (within the four years) if 
repeatability is not accomplished. A certificate holder that is unable 
to show repeatability for any restricted parameter would be required to 
modify the parameter to eliminate filtering before the four-year 
compliance period ends.
    We have not included in the rule text a time limit for submission 
of the reconstruction procedures to the NTSB following an accident or 
occurrence that requires the NTSB be notified. We presume that the 
reconstruction data are included as part of the recorder and its data 
that are subject to Sec.  121.344(i) and the NTSB's authority under 49 
CFR 830. We invite comment on whether a specific, brief time for 
submission needs to be included separately in the rule for the 
reconstruction procedure data.

II. Discussion of Comments to the NPRM

A. Summary of Comments

    The FAA received 16 comments in response to the 2006 NPRM. AirTran 
Airways, L3 Communications, the National Transportation Safety Board 
and two individuals supported the intent of the 2006 NPRM, but offered 
detailed changes.
    The Air Transport Association (ATA), Airbus, Boeing, FedEx, 
Dassault Aviation, Northwest Airlines and the Regional Airline 
Association (RAA) opposed the proposed rule. They also recommended that 
any final rule should only apply to newly manufactured airplanes.
    The remaining commenters did not specifically state support or 
opposition, but suggested changes to the proposed rule.

B. Definition of a Filtered Flight Data Signal

    We proposed to define when a flight data recorder signal is 
filtered as occurring when an original sensor signal is changed in any 
way, other than changes necessary to:
    (1) Accomplish analog to digital conversion of the signal;
    (2) Reformat a digital signal into a DFDR-compatible format; or
    (3) Eliminate a high-frequency component of a signal that is 
outside the bandwidth of the sensor.
    Boeing stated that it uses methods other than those in the proposed 
definition to enhance the quality of original sensor signals by 
protecting them against aliasing and noise. These methods go beyond the 
three specific kinds of processing allowed under the proposed 
definition and would be prohibited. Boeing suggested that the proposed 
list of acceptable signal conditioning is incomplete and does not allow 
for technological advances. Boeing indicated its preference for a more 
general definition, and suggested the following:
    ``Filtered data is that for which the actual sensor signal is 
amended, altered, modified, or changed in any manner such that it 
cannot be readily retrieved using existing, easily understood 
instructions and commonly available tools and techniques.''
    Boeing's concern that its signal conditioning methods would be 
prohibited is unfounded. Anti-aliasing techniques and noise removal are 
each different methods of eliminating high-frequency components of a 
signal. High-frequency components are not necessarily outside the 
bandwidth of the sensor. The eliminated high-frequency component of a 
signal can be within the operational bandwidth of a sensor, but outside 
the operational bandwidth of a parameter. In this case, anti-aliasing 
and noise reduction techniques, by our definition, would be an 
acceptable form of filtering.
    Accordingly, we do not agree that a more general definition than 
that proposed would be appropriate. While Boeing noted that we used the 
more general definition in a June 2002 survey on filtering, that 
definition was not subject to the extensive analysis and review that we 
use when developing a definition for a regulation. When subjected to 
such analysis and review, the failings of the general definition become 
apparent. The general definition fails to address two common and 
acceptable changes--converting a signal from analog to digital and 
formatting a digital signal for DFDR compatibility.
    It was unclear from the comments whether there was some 
misunderstanding of the phrase ``operational bandwidth of the sensor'' 
in paragraph (3) of the definition. Comments are invited regarding the 
application of this phrase in the context of the filtering definition.

C. Appendix M Introductory Text

    When we amended the DFDR regulations in 1997 to increase the number 
of parameters required to be recorded, we created Appendix M to part 
121 (and comparable appendices in other parts). These appendices 
establish the upgraded technical requirements for each of the 88 
parameters. The introductory text to Appendix M currently states:
    ``The recorded values must meet the designated range, resolution, 
and accuracy requirements during dynamic and static conditions. All 
data recorded must be correlated in time to one second.''
    In the 2006 NPRM, we proposed language that would present a more 
detailed explanation of what we meant by the term ``dynamic 
condition.'' We proposed an additional sentence indicating that dynamic 
condition meant the parameter is experiencing change at the maximum 
rate available, including the maximum rate of reversal.
    In its comment, the NTSB asked that we change the term ``maximum 
rate available'' to ``maximum rate possible.''
    The NTSB did not provide any rationale for this suggested change, 
and we are unable to agree that the term ``maximum rate possible'' is 
an improvement. Since we consider the word ``possible'' to include 
states that are well beyond the normal operational range of equipment, 
the suggested

[[Page 47860]]

change appears inappropriate as a regulatory standard. The proposed 
introductory text for Appendix M is unchanged from the 2006 NPRM.

D. Parameters Covered by the Filtering Prohibition

    In drafting the 2006 NPRM, we reviewed each of the 88 parameters to 
determine which are too critical to allow any filtering beyond the 
proposed allowable signal conditioning. The parameters that met this 
critical performance requirement included flight control surface 
positions, flight control input positions and flight control input 
forces. We also determined that discrete signals should not be filtered 
since they, by definition, show that something is either on or off, and 
we knew of no need to filter these data. We sought comment on these 
choices.
    While the 2006 NPRM proposed a general prohibition on filtering any 
flight data signal other than those specifically excepted, some 
commenters referred to those signals for parameters covered by the 
general prohibition as being on a ``no filter'' list. While this is not 
technically correct, we are using the term ``no filter list'' when 
discussing these comments to prevent further confusion.
    Boeing asked that we remove discrete signals from the ``no filter'' 
list and provided several examples of how discrete signals are 
filtered. For example, Boeing noted that discrete signals may be 
filtered to eliminate false transitions that are typically encountered 
with mechanical devices. Boeing argued that the amount of signal 
conditioning that may be applied to a discrete signal should be 
determined by what is needed for the equipment to function as part of 
the overall aircraft design, not what is required for a DFDR recording. 
Boeing concluded that the DFDR should only be monitoring the status of 
functions interfaced with discrete signal inputs, and not dictating 
conditions that may prevent an improvement in signal integrity.
    Airbus recommended that we add 33 parameters to the list of those 
that may be filtered. Airbus stated that the 33 parameters are not 
related to flight controls or flight control inputs, and that there 
would be ``no critical loss of information'' even if the raw data were 
not recorded.
    The ATA questioned why we included the majority of the 46 
parameters covered by the filtering prohibition when they do not relate 
to primary flight controls. The ATA noted that only 6 of the 88 
recorded parameters (7 in airplanes manufactured after August 18, 2002) 
record inputs to, or surface positions of, primary flight controls. The 
ATA also argued that we should not include discrete signals because 
their inclusion could expose certificate holders to technical 
noncompliance when the prohibition serves no purpose. The ATA 
recommended that the prohibition on filtering be limited to parameters 
numbers 12 through 17, and 88.
    We note first that, in our review of the parameters for this SNPRM, 
we reanalyzed each parameter (as suggested by Airbus) in light of the 
information provided by several commenters. We based our original 
proposal on two presumptions. First, we believed that currently 
operating airplanes did not record discretes in a manner that would be 
considered filtering under the proposed definition. Second, we thought 
that by prohibiting filtering of discrete signals, future system and 
data bus/network designs would not change these signals in a way that 
might prevent recording information accurately. However, based on the 
comments, we are now allowing discrete signals to be filtered under the 
language proposed in this SNPRM.
    In contrast to the other commenters, the NTSB recommended we 
increase the number of parameters for which filtering is prohibited. 
The NTSB asked that we include time, pressure altitude, indicated 
airspeed, normal, lateral and longitudinal accelerations, heading, 
pitch, roll, radio altitude, angle of attack, brake pressure, yaw or 
sideslip angle, and hydraulic pressure. The NTSB characterized these 
parameters as critical, and noted that their signals could be distorted 
by filtering, and that filtering could introduce random timing biases 
that can change the timing of one parameter relative to another.
    We agree with the NTSB's characterization of these parameters as 
critical for the work of accident investigators. Accordingly, we have 
included them in the list of parameters that are restricted in this 
SNPRM.
    An individual who is an accident investigator also recommended that 
we include parameter numbers 9 (thrust/power on each engine) and 43 
(additional engine parameters) in the list of those parameters for 
which signals may not be filtered. After our review, we agree that 
these two parameters are critical to accident investigators and they 
are included in the ``no filter'' list.

E. Applicability

1. Existing and/or Newly Manufactured Aircraft
    In the 2006 NPRM, we proposed that the prohibition against 
filtering apply to both existing and newly manufactured aircraft. While 
Airbus, Dassault Aviation, Northwest Airlines and the RAA opposed the 
proposed rule, they also suggested that any final rule should only 
apply to newly manufactured airplanes.
    Airbus noted that covering only newly manufactured airplanes is 
less costly for all parties, since manufacturers would be able to 
combine new designs into other flight recorder system improvements. 
Airbus also noted that the filter conversion algorithms have been 
solved for in-service airplanes, concluding that the problem will not 
occur again, and thus there are no further benefits to be achieved by 
the proposed rule.
    The RAA viewed the safety concerns expressed in the 2006 NPRM as 
applicable to the design and certification process, and thus better 
suited to newly manufactured airplanes.
    As we stated in the 2006 NPRM, we considered the regulatory 
alternative of limiting the filtering prohibition to newly manufactured 
aircraft. While this approach is always less costly than a rule that 
affects the in-service fleet, it would also fail to address the large 
number of aircraft currently operating on which filtering is occurring. 
We are also concerned that failing to cover in-service aircraft could 
lead to future system modifications that result in more filtering.
    Experience with the Boeing 767 and the Airbus A300 has shown that 
filtering has caused problems during accident investigations. We 
disagree that the efforts expended during the Flight 587 accident 
investigation to reverse-filter the critical rudder data account for 
all the benefits that might accrue from a stricter limit on filtering. 
We also disagree that there was an acceptable outcome following the 
investigation. The NTSB has not released any formal opinion that the 
results for the Flight 587 data reconstruction were satisfactory, or 
that the processes involved in that data reconstruction were 
acceptable. We recognize that data reconstruction, when satisfactory 
from an accuracy standpoint and shown to be repeatable, is an 
acceptable alternative. That is why we are issuing this SNPRM. We do 
not believe the problems uncovered by the Flight 587 investigation have 
been solved. Allowing airplanes to remain in

[[Page 47861]]

the fleet while filtering critical data is not an acceptable 
alternative.
2. Airplanes Omitted in the NPRM
    In the 2006 NPRM, we inadvertently failed to include airplanes 
covered under Sec.  121.344(b)(1) and Appendix B to part 121, and Sec.  
125.226(b)(1) and Appendix D to part 125. We had no reason for 
excluding these airplanes. While we realize that they are older 
airplanes, many of the parameters included in the proposed filtering 
restriction are among the 18 required to be recorded by these 
airplanes. These airplanes are now included in this SNPRM. We 
anticipate that some of the equipment installed on these airplanes may 
not be sophisticated enough to allow filtering. However, we do not know 
the exact composition of this fleet, and have determined that it is 
appropriate for airplanes covered by Sec.  121.344(b)(1) or Sec.  
125.226(b)(1) to be subject to the proposed analysis to ensure that 
they are not filtering restricted parameters.
    We are not including the older airplanes operating under Sec.  
135.152(a) or (b) because of the complex nature of the applicability in 
those paragraphs. The applicability of the filtering regulations for 
newer part 135 airplanes remains as proposed in the 2006 NPRM.
3. Part 129 Airplanes
    Airbus stated that aircraft operating under part 129 (Foreign Air 
Carriers and Foreign Operators operating N-registered aircraft engaged 
in Common Carriage) may also be affected because Sec.  129.20 
references Sec.  121.344 for applicability. Airbus concluded that 
omitting these airplanes from the cost/benefit analysis will result in 
more significant costs.
    While the proposed rule would affect certain airplanes operating 
under part 129, the applicability of Sec.  129.20 is limited to those 
airplanes registered in the United States. We included these airplanes 
in the cost/benefit analysis for the 2006 NPRM and in the cost/benefit 
analysis for this SNPRM.
4. Part 91 Airplanes
    Both the NATA and the GAMA noted that the proposed regulation would 
have a significant cost and burden impact on the owners and operators 
of aircraft that are equipped with DFDRs as required under Sec.  
91.609. They each estimated that more than 5,500 aircraft would be 
affected, resulting in a $9,000,000 impact on the general aviation 
community for no measurable benefit.
    We did not propose any change to part 91 in the 2006 NPRM. The 
proposed regulations apply only to airplanes operated under parts 121, 
125 or 135. Part 91 certificate holders would not be affected by these 
regulations.

F. Compliance Time

    The proposed rule included a compliance time of four years for 
airplanes manufactured up to 18 months after the effective date of a 
final rule. Airbus, FedEx and Frontier Airlines asked for additional 
time to develop design solutions, manufacture parts, and perform 
modifications during scheduled heavy maintenance checks.
    Airbus stated that it would need 18-24 months to develop new system 
analog to digital converter (SDAC) or electronic instrument system 
(EIS) software, plus an additional 2-3 years to complete the retrofit 
of over 800 airplanes. Frontier Airlines, which operates 34 Airbus 
airplanes, asked for at least eight years because Airbus has not 
designed or communicated a proposed solution for EIS-1 equipped 
airplanes.
    FedEx proposed a compliance date of five years from the release of 
the new rule to coincide with the heavy maintenance check cycle for its 
A300/A310s (about 30 months).
    The 2006 NPRM proposed that aircraft manufactured 18 months after 
the effective date of the rule be compliant at manufacture. Boeing 
asked that we provide a compliance date rather than a compliance period 
associated with a date of manufacture because Boeing does not normally 
provide a date of manufacture. Boeing also suggested that 18 months is 
insufficient to assess airplane designs, coordinate with parts 
suppliers if changes are needed, and incorporate any changes into 
production.
    The compliance time proposed in this SNPRM is no longer tied to an 
airplane's date of manufacture. For those parameters that are 
restricted from filtering, a certificate holder has the choice of 
removing the filtering, or demonstrating by test and analysis that an 
original sensor signal value can be reconstructed from the filtered, 
recorded signal data. Verifying that there is no filtering occurring on 
the restricted parameters, or finding filtering and modifying the 
affected parameters is required within 4 years.
    If a certificate holder discovers signal filtering and chooses to 
show that the data can be reconstructed by test and analysis, the 
compliance period begins six months after the rule's effective date, 
when an aircraft goes in for its next scheduled heavy maintenance 
check. If an airplane in for a heavy maintenance check has a DFDR 
system identical to other airplanes in a certificate holder's fleet, 
compliance by test and analysis need only be submitted once for the 
entire group. This compliance schedule is intended to allow the FAA to 
determine that the submitted procedures are repeatable, but still allow 
time for other compliance action (within the four years) if 
repeatability cannot be accomplished.
    A certificate holder that is unable to show repeatability for any 
restricted parameter would be required to modify the parameter to 
eliminate signal filtering before the four year compliance period ends.
    We believe these compliance times provide ample opportunity for 
certificate holders to make choices about their equipment and conduct 
the necessary analyses during a regularly scheduled heavy maintenance 
visit, reducing potential impact on scheduled operations or additional 
out-of-service time. In addition, a four-year compliance time is 
consistent with FAA actions in previous flight recorder regulations and 
has been supported by the industry as an adequate time for retrofit and 
for introducing new system designs into aircraft being manufactured.

G. Cost/Benefit Analysis

1. Inspection Cost
    Boeing disagreed with the estimated cost of $75 per airplane to 
assess an FDR system for filtering. While the 2006 NPRM suggests that 
this assessment is the responsibility of the operators, Boeing stated 
that it expects the operators to request Boeing to research individual 
FDR systems installed on its airplanes. Boeing stated that it will take 
a significant number of work hours to assess each airplane because of 
the complexity of unique configurations with interfacing systems. Out-
of-production airplanes will also require more than one hour per 
airplane because of the research involved in determining the data frame 
and the service bulletins that have been incorporated over the years.
    Airbus agreed with Boeing that our estimated cost of $75 is too 
low.
    The commenters misunderstood the underlying factor in our estimate 
of this cost. In the 2006 NPRM, we averaged the cost of these 
assessments over the entire fleet of approximately 14,000 affected 
airplanes. We did not mean to imply that each individual airplane would 
require a unique evaluation or that such an evaluation would cost $75. 
Rather, a manufacturer would assess its airplane models and apply the 
analysis of each model to as many of its airplanes as have the same 
DFDR

[[Page 47862]]

system. We assumed that there was a high-level of similarity of DFDR 
systems within individual airplane models and a few analyses per model 
would suffice for most of the fleet. On that basis, our estimated total 
cost was about $1 million to determine whether filtering was occurring.
2. A300/A310 Retrofit Cost
    In the Initial Regulatory Evaluation for the 2006 NPRM, we stated 
that Airbus reported that it had already completed the engineering 
analyses to modify the A300/A310 series airplane. Consequently, there 
would be no incremental engineering costs for these models because 
those costs have already been expended.
    In its comment, Airbus stated that the engineering analysis to 
modify the A300/310 has not been completed. Therefore, we need to add 
these costs (about $750,000) to our cost analysis.
    Airbus also stated that it has not yet calculated the cost to 
modify the A320 family of airplanes. Frontier Airlines added that this 
lack of analysis is the reason it is unable to comment on the costs 
estimated for the proposed rulemaking.
    As we stated in the proposed rule, we do not consider the cost of 
compliance with Appendix M for Airbus 300/310 and the A320 family to be 
a cost of this rule. Even though the 1997 regulations do not 
specifically prohibit filtering, the Flight 587 investigation revealed 
that the airplanes' recorded data did not meet the accuracy performance 
requirements of Appendix M. The cost of compliance for Airbus airplanes 
is the cost of complying with Appendix M, which has been in effect 
since 1997. This would be true whether we had ever proposed a change 
regarding filtering.
    The added compliance option of this SNPRM does not change these 
circumstances. The recorded data for each parameter must meet all of 
the requirements of Appendix M, whether the data are recorded 
unfiltered or are filtered and can be reconstructed.
3. Risk Assessment
    The RAA considers the FAA's Safety Management System (SMS) to be 
applicable to all rulemaking activities. Given the low accident rate of 
commercial air carriers over the last 10 years, the RAA does not 
consider this rulemaking to be justified under the SMS. The RAA 
considered the likelihood that an accident in which the elimination of 
signal filtering will make a difference in the accident's investigation 
to be extremely remote. The RAA concluded that the FAA should conduct a 
risk assessment before issuing a final rule.
    The SMS initiative is still under development at the FAA, and it 
will cover much more than proposed regulatory changes. There is nothing 
to suggest that it will require every rulemaking be subject to a full 
risk analysis, and those that are will most likely be chosen on a case-
by-case basis. Regarding this rule, we already know that signal 
filtering has complicated accident investigations, and allowing 
filtering to continue is unacceptable. Further, until we get more 
information on the estimated costs of this rule and the option added in 
this SNPRM, we would not have sufficient data to support a risk 
analysis.
4. Benefits Not Demonstrated
    The ATA questioned the basis for the proposed rulemaking. The ATA 
stated that this rulemaking is not based on a demonstrated inability of 
existing FDRs to realize their intended function, which is to aid in 
the determination of the probable cause of an accident.
    The NTSB has repeatedly stated that filtered flight data have 
hampered its investigations. This issue is of such a concern to the 
NTSB that it has issued several recommendations to address it 
specifically. The NTSB has found that existing FDRs are unable to 
fulfill their intended function when filtered flight data are recorded. 
We agree with the NTSB's findings.
    The ATA also indicated that prior experience with filtered flight 
data does not represent a pressing aviation safety need. Other than the 
loss of Flight 587, prior filtering events occurred 14 to 15 years ago 
with no resulting injuries.
    We cannot discount the loss of Flight 587 (and its 260 passengers 
and crew) and the lessons we learned from investigation of that 
accident. Whenever an accident occurs, it is critical that the NTSB has 
accurate flight data to determine the probable cause of the accident. 
In addition, we need accurate data to determine whether we need to take 
action to prevent future accidents. We know that there are hundreds of 
airplanes operating with DFDR systems that filter critical flight data 
before it is recorded. We will not allow that situation to continue 
when the known result is data that cannot be accurately replicated when 
needed for accident investigation. The risk continues to exist. More 
importantly, such signal filtering has hidden the underlying problem 
that the recorded data are not in compliance with the accuracy 
requirements of Appendix M, which went into effect in 1997.
    The ATA also noted that Airbus has developed algorithm-based 
methods for reconstructing the filtered signals in its airplanes to 
establish actual system performance histories. These methods should be 
sufficient to prevent the potential for filtered data to impede any 
future investigation, minimizing the value of the proposed rule.
    The NTSB has never been satisfied with the results of the `reverse 
filtered' data from the Flight 587 accident investigation, nor the time 
it took to get it. The method used has never been subject to FAA review 
or approval, and without this proposed rule, the use of reconstructed 
data is not authorized. There is also no regulatory requirement to 
develop or maintain proper methodologies for data reconstruction. The 
Flight 587 accident demonstrated a deficiency in our own flight 
recorder rules that cannot be satisfactorily addressed by after-the-
fact efforts.

H. Harmonization

    Three commenters (Airbus, Boeing, Embraer) expressed the need for 
harmonization of FAA requirements with those of other civil aviation 
authorities. The commenters raised concerns that industry will be faced 
with conflicting requirements, high costs of compliance, and 
potentially complex system designs in an attempt to satisfy two 
different sets of regulations. They view harmonization as an 
opportunity to standardize flight recorder requirements.
    The FAA is working with the European Aviation Safety Agency (EASA), 
ICAO and other non-U.S. regulatory authorities to harmonize our 
regulations whenever possible. While we support harmonization, our 
efforts regarding filtered flight data are considerably ahead of other 
aviation authorities worldwide. Since the NTSB has already encountered 
problems, we are moving forward with our efforts to correct known 
deficiencies and forestall future problems that might be corrected by 
new designs. We remain open to assisting other aviation authorities 
that seek to benefit from our discoveries and experience with these 
regulatory efforts.

I. Propriety of Change to Part 121

    While AirTran supported the intent of the 2006 NPRM, it opposed 
using the operating rules of part 121 to impose technical requirements 
unique to a specific model of aircraft or unique to the design of an 
aircraft system. AirTran believes the proposed requirement to be 
outside the scope of part 121. AirTran noted that operators do not 
typically install or alter components that affect filtering of DFDR 
data, and data filtering

[[Page 47863]]

is a function of the DFDR system design and not the responsibility of 
the operators.
    While we understand AirTran's position, DFDR requirements are a 
part of the operating rules. The only effective way to implement 
changes to currently operating aircraft are through the operating 
rules. For the most part, our certification rules are not retroactive. 
Since the parameters that we are specifying as restricted exist only in 
parts 121, 125 and 135, a change to part 25 would be confusing while 
not changing the impact on certificate holders.
    AirTran also stated that it would like to see the flight recorder 
rules of part 121, 125, and 135 reviewed to redefine responsibility 
based on capability. AirTran suggested that information such as 
correlation of parameters at certification, filtering of parameters, 
and other data that are beyond the capabilities of most operators be 
placed in a special federal aviation regulation (SFAR). This could be 
accomplished by the manufacturers who have the expertise and data to 
provide the analysis. The results of this analysis should come to the 
operators in the form of FAA approved instructions for continued 
airworthiness to ensure continued compliance, similar to Sec.  
121.370(b) regarding fuel tanks.
    We appreciate AirTran's input. Such an approach would, as 
suggested, require considerable rulemaking and an assessment of the 
impact on manufacturers for equipment that may vary considerably 
between certificate holders. While we will keep AirTran's comments 
under advisement, we have determined that there is an immediate need to 
correct the current problems and prevent future ones. The comments to 
the proposed rule also indicate that aircraft manufacturers know their 
roles in compliance.

J. Conflicts With Other Proposed DFDR Regulations

    The ATA stated that the changes proposed in the 2006 NPRM are not 
coordinated with other flight recorder changes proposed separately. The 
ATA stated that the proposed filtering changes are redundant of the 
changes proposed in our NPRM entitled ``Revisions to Cockpit Voice 
Recorder and Digital Flight Data Recorder Regulations'' (70 FR 9752; 
February 28, 2005). The ATA noted that seven of the parameters that we 
proposed to have increased recording rates in the 2005 NPRM are also 
proposed to be restricted from filtering. The ATA stated that both 
changes are not ``needed to derive flight control histories or 
accomplish the intended purpose of FDRs.''
    The ATA also stated that the filtering proposal is not fully 
coordinated with the SNPRM entitled ``Revisions to Digital Flight Data 
Recorder Regulations for Boeing 737 Airplanes and for Part 125 
Operators'' (71 FR 52382; September 5, 2006). The ATA reiterated its 
disagreement with the proposed rule for 737s, and stated that if the 
FAA goes forward with both rules, the compliance times must be 
coordinated for 737s. The ATA is also concerned with the impact of 
filtering changes on older 737s that only record 18 parameters, and 
were specifically not required to upgrade equipment in the 1997 rule 
changes.
    We disagree in both instances. The proposed rule for Boeing 737 
airplanes would add three additional parameters that are not part of 
this rule, and that proposed rule does not speak to filtering. As to 
their coordination, both rules are proposals, with one addressing the 
addition of three parameters on one specific model of airplane, and 
this rule addressing the propriety of filtering certain parameters on 
all airplane models. The FAA will consider the filtering status of the 
proposed 737 parameters when a final rule is promulgated.
    Similarly, the 2005 NPRM addressed many topics, but filtering was 
not one of them. Both data sampling rates and filtering affect what 
gets recorded and can play major roles in data integrity and 
unrecoverable loss of data. But the two factors are separate and the 
proposed solutions are not inconsistent. The ATA states that both 
modifications are not necessary to achieve the ``intended purposes of 
FDRs'' but does not state what it sees as this purpose. History has 
already shown that filtered signals can raise serious questions in an 
accident investigation. It is also accepted that lower sampling rates 
provide much less reliable data. While the two characteristics of 
recorded data are related, they are not as inextricably intertwined as 
the comment suggests. The two proposals result from different problems 
and analyses, and reflect upgrades in technology that were not 
anticipated when the previous regulations were written.

K. Exemption Request

    The RAA asked that we add a provision that allows for deviations 
for the current fleet ``upon approval by the Administrator.'' The RAA 
was unconvinced that, given what the NTSB now knows about the 
characteristics of the Airbus FDR parameters, the Board would be unable 
to successfully and timely complete an investigation of a future 
accident on an Airbus airplane that filters data. The RAA does not 
support a retrofit program for any aircraft type until the 
characteristics of the filtered data are known to be sufficiently 
nonresponsive. The RAA's proposal would allow a certificate holder that 
has a deviation to avoid a costly retrofit if the deviation is 
considered minor and would not hinder a future NTSB accident 
investigation.
    We disagree with the RAA's premise that the NTSB would be able 
today to complete a timely and comprehensive accident investigation of 
an Airbus airplane that filters data. While significant work was put 
into data reconstruction during the Flight 587 investigation, the Board 
has never stated its satisfaction with the reconstruction. In addition, 
Airbus has never been required to formalize the analysis or results or 
maintain any data reconstruction technique. Nor can we support the 
inclusion of some unspecified deviation that would allow for an 
incalculable number of different flight recorder system changes that 
would only worsen the problem of data reconstruction from individual 
airplanes.

L. Relationship to the 1997 Regulations

    The ATA submitted a lengthy comment on what it considers the effect 
of the 1997 rule. While the ATA correctly concluded that the 1997 rule 
does not specifically prohibit filtering, it is not clear how the ATA 
connected this rule to the amount of data being recorded, or how it 
could provide ``more valuable raw data.''
    We are unable to respond to this because we do not understand the 
connection the ATA is trying to make between filtering and recording 
greater amounts of data. The FAA has never made any correlation between 
filtering and the amount of data being recorded, nor was it addressed 
in the 1997 rule.
    Airbus noted that the preamble language to the 1997 rule says ``the 
standards proposed are harmonized with the current JAR-OPS, which will 
be based on ED-55 standards.'' Airbus commented that the ED-55 
recommendation to correlate recorded data to flight deck display led it 
to record filtered rudder data.
    We believe this is Airbus's interpretation of its compliance with 
ED-55. Airbus appears to have concluded that the recorded rudder data 
must be the same ``smoothed'' data that are displayed in the cockpit, 
and that to be correlated, the data must be filtered. However, 
correlation does not require that identical data be recorded and 
displayed in the cockpit. Correlation

[[Page 47864]]

requires a mathematical relationship between the displayed and recorded 
data, not that recorded data be filtered.
    Although the 1997 rule language does not specifically prohibit 
filtering, in the case of Flight 587, the data recorded were not in 
compliance with the accuracy requirements of part 121 Appendix M. This 
failure to meet Appendix M, and not the fact that the data were 
filtered, has always been the compliance issue following the Flight 587 
investigation.

III. Regulatory Notice and Analysis

Paperwork Reduction Act

    This proposal contains the following new information collection 
requirements. As required by the Paperwork Reduction Act of 1995 (44 
U.S.C. 3507(d)), the FAA has submitted the information requirements 
associated with this proposal to the Office of Management and Budget 
for its review.
    Title: Filtered Flight Data.
    Summary: The FAA is amending its proposal to prohibit the filtering 
of some original flight recorder sensor signals. Comments to the notice 
of proposed rulemaking published on November 15, 2006, and changes in 
available technology have caused us to re-examine our position on data 
filtering. We are now proposing that certain critical data parameters 
may be filtered if a certificate holder can show that the data can be 
accurately reconstructed. This proposed rule would improve the accuracy 
and quality of the data recorded on digital flight data recorders while 
giving aircraft designers and operators more flexibility in system 
design and operation where allowable, including an option to filter 
data.
    Use of: This information collection supports the Department of 
Transportation's strategic goals on safety and security. If a 
certificate holder chooses to show that filtered data from certain 
critical data parameters can be accurately reconstructed, the 
certificate holder must maintain documentation of the procedure 
required to perform the reconstruction. This will allow the FAA and 
accident investigators to access accurate data promptly.
    Respondents (including number of): The likely respondents to this 
proposed information requirement are those certificate holders that use 
a flight data recorder system that filters the data from any of the 
critical data parameters.
    Frequency: The proposal would require operators to have a one-time 
engineering analysis performed (spread over two years) to evaluate 
whether the DFDR systems on their airplanes record certain flight data 
parameter signals after they have been filtered.
    Annual Burden Estimate: Since the one-time engineering analysis 
will be performed on individual type certified DFDR systems, we 
estimate that there would be 16 of these engineering analyses that 
would be completed and that each would require an average of 16 hours 
of engineering time to complete and submit to the FAA.
    The total burden is estimated as follows:

------------------------------------------------------------------------
                                                         Present value
    Documents required to show        One-time hours    discounted cost
  compliance with the Final Rule                           (in $2008)
------------------------------------------------------------------------
Report to FAA of an Engineering                   256            $28,160
 Analysis of whether Certain
 Flight Data Parameters Are Being
 Filtered before Being Recorded...
                                   -------------------------------------
    Total.........................                256             28,160
------------------------------------------------------------------------

    This rule will result in a minimal annual recordkeeping and 
reporting burden.
    The agency is soliciting comments to--
    (1) Evaluate whether the proposed information requirement is 
necessary for the proper performance of the functions of the agency, 
including whether the information will have practical utility;
    (2) Evaluate the accuracy of the agency's estimate of the burden;
    (3) Enhance the quality, utility, and clarity of the information to 
be collected; and
    (4) Minimize the burden of collecting information on those who are 
to respond, including by use of appropriate automated, electronic, 
mechanical, or other technological collection techniques or other forms 
of information technology.
    Individuals and organizations may send comments on the information 
collection requirement by November 13, 2008, and should direct them to 
the address listed in the Addresses section at the beginning of this 
preamble. Comments also should be submitted to the Office of 
Information and Regulatory Affairs via facsimile at (202) 395-6974.
    According to the 1995 amendments to the Paperwork Reduction Act (5 
CFR 1320.8(b)(2)(vi)), an agency may not collect or sponsor the 
collection of information, nor may it impose an information collection 
requirement unless it displays a currently valid OMB control number. 
The OMB control number for this information collection will be 
published in the Federal Register, after the Office of Management and 
Budget approves it.

International Compatibility

    In keeping with U.S. obligations under the Convention on 
International Civil Aviation, it is FAA policy to comply with 
International Civil Aviation Organization (ICAO) Standards and 
Recommended Practices to the maximum extent practicable. The FAA has 
determined that there are no ICAO Standards and Recommended Practices 
that correspond to these proposed regulations.

Regulatory Evaluation, Regulatory Flexibility Determination, 
International Trade Impact Assessment, and Unfunded Mandates Assessment

    Changes to Federal regulations must undergo several economic 
analyses. First, Executive Order 12866 directs that each Federal agency 
shall propose or adopt a regulation only upon a reasoned determination 
that the benefits of the intended regulation justify its costs. Second, 
the Regulatory Flexibility Act of 1980 (Pub. L. 96-354) requires 
agencies to analyze the economic impact of regulatory changes on small 
entities. Third, the Trade Agreements Act (Pub. L. 96-39) prohibits 
agencies from setting standards that create unnecessary obstacles to 
the foreign commerce of the United States. In developing U.S. 
standards, the Trade Act requires agencies to consider international 
standards and, where appropriate, that they be the basis of U.S. 
standards. Fourth, the Unfunded Mandates Reform Act of 1995 (Pub. L. 
104-4) requires agencies to prepare a written assessment of the costs, 
benefits, and other effects of proposed or final rules that include a 
Federal mandate likely to result in the expenditure by

[[Page 47865]]

State, local, or tribal governments, in the aggregate, or by the 
private sector, of $100 million or more annually (adjusted for 
inflation with base year of 1995). This portion of the preamble 
summarizes the FAA's analysis of the economic impacts of this proposed 
rule.
    Department of Transportation Order DOT 2100.5 prescribes policies 
and procedures for simplification, analysis, and review of regulations. 
If the expected cost impact is so minimal that a proposed or final rule 
does not warrant a full evaluation, this order permits that a statement 
to that effect and the basis for it to be included in the preamble if a 
full regulatory evaluation of the cost and benefits is not prepared. 
Such a determination has been made for this proposed rule. The 
reasoning for this determination follows:
    This SNPRM proposes to allow certain parameters to be filtered if 
the certificate holder can show that the data can be reconstructed. The 
expected economic impact would be a minimal cost with positive net 
benefits and, therefore, a regulatory evaluation was not prepared. To 
estimate the costs for any final rule, we request specific economic 
data in response to the questions in this section.
    The FAA has, therefore, determined that this proposed rule would 
not be a ``significant regulatory action'' as defined in section 3(f) 
of Executive Order 12866, and would not be ``significant'' as defined 
in DOT's Regulatory Policies and Procedures.

Aviation Industry Affected

    The proposed rule would apply to all aircraft operated under parts 
121, 125, and 135 and required to have a DFDR system. Aircraft operated 
under other FAR parts would not be affected.

Benefit and Cost Methodology

    The baseline for determining the proposed rule's benefits and costs 
is that each affected aircraft must be in compliance with appendix B or 
M of part 121, appendix D or E of part 125, or appendix F of part 135.
    The first step in assessing costs of this SNPRM is for each 
certificate holder to determine whether any of its aircraft are 
filtering DFDR data before it is recorded. For purposes of filtering, 
we have divided the 88 required parameters into two groups:
     Those parameters that may not be filtered unless the 
certificate holder can demonstrate that the original sensor signal data 
(values) can be reconstructed using a valid, repeatable procedure (1-7, 
9, 11-18, 26, 32, 42, 43, 68, 70, 77, and 88); and
     Those parameters that may be filtered without this 
demonstration as long as the accuracy requirements are met (8, 10, 19-
25, 27-31, 33-41, 44-67, 69, 71-76, and 78-87).
    A certificate holder must first determine whether any of the 
original sensor signal data from any of the parameters in the first 
group are filtered before being recorded. The cost of this 
determination is a cost of this proposed rule.
    If no parameters in the first group are being filtered, no further 
action is required, and the only cost is reporting this status to the 
FAA.
    If any of the parameters in the first group are being filtered, the 
certificate holder must determine whether the recorded data meet the 
accuracy requirements of the applicable appendix. If the requirements 
are met, no further action is necessary, and the only cost attributable 
to this rule is to report the status to the FAA.
    If the requirements of the applicable appendix are not being met, 
the allocation of costs will depend on the option chosen by the 
certificate holder to address the non-compliance. If a certificate 
holder chooses to remove the filtering in order to comply with Appendix 
M, the cost is attributable to compliance with the 1997 rule. If a 
certificate holder chooses to show by test and analysis that the data 
can be acceptably reconstructed, the cost is attributable to this 
proposed rule because it is a new option in this SNPRM. We are 
requesting estimates of the costs of these two options and their 
expected use.

Benefits

    The benefit of this proposed rule would be to make accurate flight 
data more quickly and unambiguously available to accident 
investigators. Inaccurate data have been shown to delay accident 
analysis, and could suggest unnecessary corrective action. These were 
experienced during the Flight 587 accident investigation.

Benefit Cost Analysis

    We believe that the SNPRM would provide a less costly means than 
the 2006 NPRM for certificate holders to address compliance with the 
applicable appendix. It will also provide accident investigators with 
the more accurate and less ambiguous data necessary to determine the 
causes of airplane accidents.
    To estimate the costs of a final rule, we need data on the costs of 
complying with this SNPRM. We are requesting the following information, 
including any supporting documentation:
    1. The aircraft makes, models and series in your fleet.
    2. The number of each aircraft make, model and series in your 
fleet.
    3. By aircraft make, model and series, how many type certificated 
DFDR systems are installed in your aircraft.
    4. For each installed DFDR system, by characteristics, determine 
the parameter numbers in the first group that are filtered before being 
recorded. For any filtered parameter, which do not meet the accuracy 
requirements of Appendix B or M (or Appendix D or E or Appendix F)?
    5. For the information requested in 4, do you have the 
technical capability to determine the characteristics yourself or is 
outside technical expertise required? How many engineering hours would 
be needed to determine whether the DFDR system is recording filtered 
flight data for the first group of parameters? What other costs would 
be required for this analysis?
    6. For each installed DFDR system that records filtered data for 
the first group of parameters (the data for which do not meet either 
Appendix B or M, Appendix D or E or Appendix F), how much would it cost 
to remove the filtering?
    7. For each installed DFDR system that records filtered data for 
the first group of parameters (the data for which do not meet either 
Appendix B or M, Appendix D or E or Appendix F), how much would it cost 
to develop and accomplish the tests and analyses required to comply 
with the reconstruction option proposed in this SNPRM?

Regulatory Flexibility Analysis

    The Regulatory Flexibility Act of 1980 (Pub. L. 96-354) (RFA) 
establishes ``as a principle of regulatory issuance that agencies shall 
endeavor, consistent with the objectives of the rule and of applicable 
statutes, to fit regulatory and informational requirements to the scale 
of the businesses, organizations, and governmental jurisdictions 
subject to regulation. To achieve this principle, agencies are required 
to solicit and consider flexible regulatory proposals and to explain 
the rationale for their actions to assure that such proposals are given 
serious consideration.'' The RFA covers a wide range of small entities, 
including small businesses, not-for-profit organizations, and small 
governmental jurisdictions.
    Agencies must perform a review to determine whether a rule will 
have a significant economic impact on a substantial number of small 
entities. If the agency determines that it will, the agency must 
prepare a regulatory

[[Page 47866]]

flexibility analysis as described in the RFA.
    However, if an agency determines that a rule is not expected to 
have a significant economic impact on a substantial number of small 
entities, section 605(b) of the RFA provides that the head of the 
agency may so certify and a regulatory flexibility analysis is not 
required. The certification must include a statement providing the 
factual basis for this determination, and the reasoning should be 
clear.
    While we have only limited information on the costs of this 
proposed rule, that information suggests that the rule would be 
unlikely to have a significant impact on a substantial number of small 
entities. We are specifically requesting information regarding the 
effect on small entities to assist in our final analysis.
    Therefore, the Acting Administrator certifies that this proposed 
rule would not have a significant economic impact on a substantial 
number of small entities.

International Trade Analysis

    The Trade Agreemen