Claybaugh 6-inch Rocket, Final Report

by Bill Claybaugh, RRS.ORG


EDITOR’S NOTE: This is a continuation of the reporting from the 10-16-2021 flight of the 6-inch rocket design, built and flown by RRS member, Bill Claybaugh.


This project is part of an effort to develop a two-stage sounding rocket capable of sending about 5 kg of usable payload to about 200 Km altitude.  This vehicle is intended to act as the upper stage of that two-stage rocket; it was—based on a systems analysis–sized for an eight second burn-time and about 1300 lbf thrust.

OVERVIEW

As flown the vehicle was 101.25” from nose tip to the fin trailing edges.  The Payload section was 40.125” in length and 6.170” in diameter; the booster was 61.125” in length and 6.00” in diameter.

Computer simulated rendering of the rocket

The vehicle had an aluminum nose tip, a filament wound fiberglass nose with a 5.5:1 Von Karmen profile, a filament wound cylindrical payload section, and an aluminum motor / airframe with aluminum fins.

VEHICLE DESCRIPTION

Inspection of the Forward Bulkhead showed it to be in good condition with no evidence of any gas leaks above the two O-rings.  The bottom of the Bulkhead showed some damage to the fiberglass heat shield from the ground impact of the rocket but showed plenty of

Pre-flight estimated motor performance was 1350 lbsf. of thrust with a burn-time of 8.35 seconds.  Burnout velocity was estimated at Mach 3.1 at 14,400 feet with a peak altitude estimated at about 71,000 feet. Total flight time was expected to be 143 seconds.  The booster had a streamer attached at the forward end to try and cancel horizontal velocity upon deployment at peak, thus limiting the range of the booster.

The payload also used a streamer for recovery, it was planned to separate from the booster near peak altitude using a pneumatic separation system that operated four pins which rigidly attached the payload to the rocket until pressure was released.

Final vehicle mass properties are shown below:

 Item DistanceWeightMomentC.G. 
   from  from 
   Nose Tip  Nose Tip 
        
 Nose Cone Assy. 24.005.55133.20 Measured
        
 Ballast 29.900.000.00 Estimated
        
 Instruments Assembly 33.008.97296.01 Measured
        
 Bulkhead & Sep. Sys.: 37.502.80105.00 Measured
 Bolts 38.630.031.28 Measured
        
 O/A Payload Length: 40.13    
        
    17.35535.4930.858 
        
 Bulkhead Retainer 43.041.1047.34 Measured
 Bolts & Nuts 43.410.3414.59 Measured
        
 Bulkhead Assy. 40.192.3092.43 Measured
        
 Tube 69.4313.05906.00 Measured
        
 Outer Liner 68.314.10280.08 Measured
        
 Fin Can 95.102.75261.52 Measured
        
 Nozzle 97.497.83763.31 Measured
 Bolts 97.190.2221.38 Measured
        
 Fins 96.495.55535.53 Measured
 Bolts 96.190.1816.93 Measured
        
 O/A Stage Length: 101.25    
        
    54.773474.5963.445 
        
 Propellant 68.3154.203702.54 Measured
        
    108.977177.1365.866 
        

FORWARD BULKHEAD

The forward bulkhead assembly consisted of the forward bulkhead with O-rings, fiberglass spacers, and a bulkhead retainer that incorporated the bottom portion of the separation system (four holes for the attachment pins and a 45-degree bevel to allow the payload to fall off the booster once the four pneumatically operated pins retracted).

Computer rendering of the forward bulkhead
Bulkhead retainer with separation system fittings

FINS

The Fins were attached to the motor tube via an internal “fin can” that served to provide the “meat” to allow four countersunk fasteners to hold each fin rigidly to the motor tube.  The internal Fin Can had a single O-ring at the top to seal between the phenolic propellant liner and the fin can as wall as two O-rings to seal between the fin can and the motor wall.

Computer rendering of the internal fin can

Note that the fins shown are the flight fins, post flight; with the exception of minor gouging the fins appear to be fully reusable.

Photo of fins post-flight
Computer rendering of one fin

NOZZLE

The nozzle consisted of an aluminum outer shell, a graphite insert, and a stainless steel nozzle extension with a plasma sprayed Zirconia overcoat on the inside diameter.

Photo of the nozzle
Photo of the nozzle from the other side

PROPELLANT LINER

The liner protecting the motor tube from the combustion gas was a phenolic tube with a 5.50 inch inside diameter.  The tube was originally slightly oversize for the motor tube’s 5.75” nominal inside diameter and was sanded as necessary to make it a tight slip fit into the motor tube.  It was then cut to a 48” overall length and fitted to the motor tube using a high temperature grease (550 degrees F).

Post-flight analysis shows that the liner had about 0.090” – 0.092” of the original 0.125” wall remaining in those areas exposed to hot gas throughout the burn; note that heating of the phenolic leads to expansion of the thickness of the liner, nonetheless, there was no evidence of hot gas having reached the motor tube wall.

PROPELLANT

The grain was cast in place using a dissolvable (polystyrene) mandrel that provided for four fins at the base of the motor and a simple cylindrical core at the upper end.  This grain design provided an approximately neutral thrust curve as the finocyl section regressed in burn area at a rate that very closely matched the progression of the cylindrical section of the grain.

Grain cross sections
Thrust and chamber pressure curves

The finocyl section at the base of the grain was 14.75” in length, the cylindrical section 31.25” in length for an overall 46” propellant grain length.

Because the grain design tools used for this project worked only in two dimensions, the 2.66 square inches of exposed grain surface at the top of the finocyl fins was not modeled in the simulation.  This represents 0.80% of the initial grain burn area and, accordingly, the actual performance was expected to be slightly regressive.

All grain design simulations were based on the 0.056 lbsm / cubic inch propellant density of the various static test motors; in the event, this grain came in at 0.059 lbsm / cubic inch due to changes in both the propellant mix and processing.  The effects of that higher density on flight performance will be addressed in the Analysis section.

AERODYNAMIC MODEL

Most dynamical simulations for this flight were conducted using RASAero II. The aerodynamic model estimated by that tool is shown below:

Aerodynamic model plot

Likewise, RASAero II provided estimates of Stability Margin over the flight profile:

Stability margin plot

A splash analysis was very graciously conducted by Chuck Rogers.  That analysis concluded that the initial launch conditions that minimized risk to the uninvolved public were a launch azimuth of 244 degrees and a launch tower angle of 87 degrees (that is, three degrees below vertical in a southwesterly direction).

PAYLOAD

The payload consisted of three subsystems: a pneumatic payload separation system, a main flight computer with integrated transmitter, and, a backup flight computer with onboard recording of flight engineering data.

PNEUMATIC SEPARATION SYSTEM

The separation system relied on four pins that rigidly locked the payload to the vehicle. The system was actuated by command from the main or backup flight computers, which command fired a nitrocellulose-based initiator that in turn drove a plunger through a burst disk.  Venting of the system allowed spring force on the four locking pins to draw them inward, thus allowing the payload to fall away from the booster.

The Separation System was o-ring sealed at all connections to assure it remained leak free under flight conditions. Initial testing showed the system could hold pressure (125 psia air) for 100 hours.  Pre-flight testing included a 50-hour leak down test followed by one minute on a shake table.  The unit was leak free and actuated on command after this final test.

The main flight data recorder and transmitter was a Multitronix Kate 2 System; backup flight data recording was provided by an Altus Metrum EasyMega.

Photo of the locking pin system
Photo of the pneumatic separation system

MAIN FLIGHT COMPUTER

The main flight computer was a Kate 2 Data Recorder and Transmitter from Multitronics, Inc.  This system used a 915 MHz ISM uplink and downlink with on-the-fly adjustable power output from 100 mw to 1 watt, it used Spread Spectrum Frequency Hopping and FSK Modulation with a 128-bit AES encryption.

The system fixes its GPS position every 200 msec and features unlimited GPS altitude reporting; the velocity lockout is at 1700 ft/sec.  A 50 g Axial Acceleremeter and 10 g pitch and yaw accelerometers record every 10 msec and report via telemetry every 100 msec.  A separate pyro board initiates payload separation and peak.

The transmitter link budget indicates a worse case net 26.5 dB at the receiver for this flight.

Link budget details for the flight computer transmitter

BACKUP FLIGHT COMPUTER

The backup flight computer was an Altus Metrum EasyMega with three axis data recording (acceleration and rates) and a barometric altitude estimator.  Separate batteries and switches powered the independent pyro initiation which was programed for one second after the accelerometer measured peak altitude.

FLIGHT SIMULATION MODEL

Simulation using RASAero II showed an estimated peak altitude of about 71,000 feet, a worse case total flight time of about 144 seconds (assuming no separation at peak), and a maximum worst-case range of about 75,000 feet.

Simulated trajectory plot

Baseline flight simulation (from RASAero II):

Baseline flight simulation

Launch Angle Vs. Range (from RASAero II):

Simulated launch angle vs range

Maximum Range Estimation (from RASAero II):

FLIGHT TEST RESULTS

Based on video analysis, ignition require 0.067 seconds from the rupturing of the burst diaphragm (a standard 1.5” rubber stopper previously tested to pass the nozzle at 40-50 psia) to first motion. From first motion, it required 0.35 seconds to clear the 24-foot tower at about 25 feet altitude and about 165 ft/sec.

Frame-by-Frame Video Analysis

(Red Indicates Clearing the Tower)

Estimated
CumulativeEstimated
FrameEstimateEstimateEstimatedEstimatedAverageInterval
NumberBurnFlightVerticalVerticalVerticalVertical
TimeTimeMotionVelocityAccelerationAcceleration
(ft.)(ft./sec.)(g’s)(g’s)
10.0170.0000.00
20.0330.0000.00
30.0500.0000.00
40.0670.0170.5060.00110.80110.80
50.0830.0331.0060.0054.90-1.00
60.1000.0502.0080.0048.6936.27
70.1170.0672.5075.0033.94-10.32
80.1330.0832.7566.0023.60-17.77
90.1500.1003.0060.0017.63-12.18
100.1670.1174.0068.5717.2514.97
110.1830.1335.0075.0016.4710.98
120.2000.1505.0066.6712.80-16.53
130.2170.1678.0096.0016.8953.66
140.2330.1838.5092.7314.71-7.10
150.2500.20010.00100.0014.5312.55
160.2670.21712.00110.7714.8819.07
170.2830.23314.00120.0014.9716.20
180.3000.25017.50140.0016.3936.27
190.3170.26719.50146.2516.0310.65
200.3330.28322.00155.2916.0215.85
210.3500.30025.00166.6716.2520.19
220.3670.31728.00176.8416.3417.96
230.3830.33331.00186.0016.3316.06
240.4000.35033.00188.5715.733.79
250.4170.36736.00196.3615.6313.52
260.4330.38340.00208.7015.9121.98
270.4500.40042.00210.0015.301.43
280.4670.41746.50223.2015.6423.60
290.4830.43354.00249.2316.8647.50
300.5000.45057.00253.3316.486.64
310.5170.46761.00261.4316.4014.08
320.5330.48364.50266.9016.159.19
330.5500.50069.00276.0016.1415.96
340.5670.51773.00282.5815.9911.26
350.5830.53376.50286.8815.707.00
360.6000.55085.00309.0916.4540.40
370.6170.56790.00317.6516.4114.94
380.6330.58394.00322.2916.167.64
390.6500.60097.50325.0015.824.06
400.6670.617102.50332.4315.7412.85
410.6830.633110.50348.9516.1129.77
420.7000.650115.00353.8515.918.13
430.7170.667119.50358.5015.707.67
440.7330.683123.50361.4615.434.52
450.7500.700128.00365.7115.236.92
460.7670.717132.50369.7715.026.55
470.7830.733141.00384.5515.2926.54
480.8000.750145.00386.6715.012.95
490.8170.767150.00391.3014.857.64
500.8330.783158.00403.4014.9921.55

Just after 0.50 seconds the vehicle began an unplanned turn to the Northeast.  This turn continued for 0.25 seconds before the vehicle resumed stable flight on the new azimuth and with a flight path angle of about 75 degrees.  After 0.80 seconds but before 1.0 seconds, the telemetry failed.  The cause of this failure is not yet established but appears to the manufacturer to have been a power outage; however, the battery was still connected to the main computer after recovery and the battery tested at an optimal 3.87 volts.

At about 1.0 seconds, the payload separation system appears to have been actuated by the backup flight computer; that computer is currently at the manufacture for data extraction to try and determine why it fired the initiators.

Based on video analysis, the vehicle appears to have coned twice following separation of the payload.  This coning could have been associated with the payload separation or with the deployment of the rocket’s streamer.  In either case, the vehicle resumed stable flight (as designed) without a nose cone. The payload assembly was located about 120 feet from the launch tower on the northeasterly azimuth.  The backup flight computer was still actively reporting (via “beeps”) it’s status but the main computer was not so doing.

Launch plus 0.50 seconds
Launch plus 0.75 seconds

The booster was located north and a little east of the launch site at a range of 14,300 feet.  Based on that range and the estimated motor performance a trajectory reconstruction suggests a maximum altitude of 21,200 feet, a burnout velocity of 1550 ft/sec and a terminal velocity of about 820 ft/sec with a total flight time of about 74.5 seconds.

The booster impact left an about 2.0-inch-deep depression in the hardpan before the hardware apparently fell on its side. Given an estimated terminal velocity of 820 ft/sec, this implies and average of 410 ft/sec to stop and thus that the vehicle came to rest in about 0.000407 seconds.  This in turn indicates an average deceleration of about 31,200 g’s on impact.

ANALYSIS – THE TURN TO THE NORTHEAST

The Turn to the Northeast

All testable reasons for the turn to the Northeast after 0.50 seconds have been ruled out: there was no hot gas leak nor any apparent disturbance to the thrust vector. The wind was from the Northwest and less than 5 mph, if it had caused the turn we would have expected the vehicle to turn toward the Northwest, not the Northeast. The temporary “hanging” of a part of the bellybands appears ruled out by the absence of any gap between the fins and the motor tube as well as by the absence of any damage to the fin leading edges.  Further, the bellybands all landed within fifty feet of the launch tower; given an estimated velocity of about 165 ft/sec at the top of the tower, this implies that each bellyband followed a nearly vertical trajectory following clearing the tower.

The remain hypothesis for the cause of this turn is that the vehicle ran into a “dust devil” that was not visible because it had not reached the ground.  Examination of the video using polarized glasses showed no evidence for such an event, but that is not conclusive as the sun angles may have been inappropriate for this technique.

ANALYSIS – TELEMETRY FAILURE

Telemetry failed after 0.80 seconds but before 1.0 seconds based on analysis by the manufacturer of data recorded by the receiver (data packets are sent every 0.2 seconds, one was received at about 0.80 seconds and no subsequent packets were recorded).  The cause of this failure is unclear: the manufacturer has initially concluded it was a power failure, however, the battery showed 3.8 volts at recovery and was still connected to the computer / transmitter; thus, a power failure would have to have been internal to the hardware. This failure might be associated with separation of the payload from the rocket which occurred around this time.  Transmitted data show that the main computer did not initiate the separation and had continuity to the initiator throughout the period during which data was transmitted.

ANALYSIS – PREMATURE PAYLOAD SEPARATION

The payload was recovered about 120 feet from the launcher on a Northeasterly heading.  Based on the location a trajectory reconstruction suggests separation may have occurred around 1.0 seconds into the flight at about 400 feet altitude.

Given the data indicating that the main computer did not command separation while it was operating and the observation, following recovery, that both initiators had been fired (firing of either initiator ignites the other), it appears that the backup computer may have initiated the separation.  That computer is currently at the manufacturer for repairs after which we hope to extract whatever data it may have recorded, including continuity data with respect to the initiator to which it was wired.

SUBSEQUENT FLIGHT

Following payload separation, the vehicle appears to have coned twice and then resumed stable flight on the new heading.  Upon recovery, the vehicle did not have its streamer attached and we assume it was lost to aerodynamic forces during the separation of the payload and subsequent coning; however, that streamer has not been recovered and so we cannot confirm when it came off the vehicle.

Per the trajectory estimate, it appears that even with a blunt front end, the vehicle may have reached around Mach 1.35 (1550 ft. / sec.) but that estimate is unconfirmed.

Note that the video measured velocity and acceleration up the launch tower was noticeably higher than the pre-flight estimate: pre-flight, velocity at the top of the tower was estimated at about 145 ft / sec while the measured velocity just after clearing the tower was about 165 ft / sec.  This difference may be due to the higher density of the propellant as compared to the pre-flight model; assuming that the ballistic characteristics of the propellant remained the same (very unlikely) modeling of the pre-flight propellant assumptions but using the higher density indicates it would produce about 5% higher thrust at about 8% higher chamber pressure due to the higher mass flow compared to the pre-flight modeled propellant.

Modelling of the vehicle performance using the actual range and these different propellant performance assumptions does not significantly change the estimated peak altitude or velocity: the somewhat greater energy of the flight propellant is spent on increased drag as velocity approaches Mach 1.35.

SUMMARY AND FUTURE WORK

The rocket motor appears to have performed as designed, albeit in off-design flight conditions.  In the absence of any explanation for the unplanned turn to the Northeast, no changes to the motor design are planned for the next flight vehicle other than the hard anodizing of the fins to help them survive future flights to still higher velocities.

The payload assembly appears to have been commanded off the rocket motor at about one second into the flight; the reason for this remains unclear at this writing. For future flights the internal payload structure will be made still more robust to prevent the internal structural failures that did occur upon impact of the payload; some of those structures will be rebuilt in stainless steel to help move the Cg forward (this was not an issue for this flight, but will be for eventual Mach 6 burnout velocities).

Further work is required on the base of the launch tower to significantly reduce the labor required to assemble and erect the tower.

The bellybands will be modified for greater strength and spring back by moving to 1095 spring steel instead of the 2024T-3 used for this flight; in addition, the guides will be lightened both to aid travel up the rail and to mitigate against any impact damage that might occur if they contact the vehicle during separation.

Claybaugh 6-inch Rocket, Post-Flight Inspection

by Bill Claybaugh, RRS.ORG


EDITOR’S NOTE: This is a continuation of the reporting from the 10-16-2021 flight of the 6-inch rocket design, built and flown by RRS member, Bill Claybaugh.


Post-Flight Motor Inspection

Recovery of the spent motor hardware allowed a detailed disassembly and inspection of the parts.  This revealed several useful observations:

Motor Tube

The recovered Motor Tube showed a dent just above the fins that was deep enough to have caused a pressure failure if it had been present while the motor was operating; we thus conclude that the dent occurred during or post impact.

Localized dent in the aluminum case, likely resulting from impact after burnout

Bulkhead

Inspection of the Forward Bulkhead showed it to be in good condition with no evidence of any gas leaks above the two O-rings.  The bottom of the Bulkhead showed some damage to the fiberglass heat shield from the ground impact of the rocket but showed plenty of fiberglass heat shield remaining after the about eight second burn.  The “nose” of the ignitor assembly remained in place in contrast to previous tests where this part had shattered upon ignition; the change to a steel “gun barrel” liner for the initiator appears to have resolved this issue.

The forward side of the bulkhead showing no leakage or damage.
Aft side of the bulkhead showing damage to fiberglass heatshield.

Fins

The four fins were intact and largely undamaged; they appear suitable for reuse in future flight vehicles.  Checking with a 0.002” feeler gauge showed there was no gap between the “nose” of any of the fins and the motor tube.  A further check using backlighting confirmed that there were no visible gaps between the fins and the motor tube at any location along the fin edges.

Nozzle

The graphite nozzle insert had broken free of its aluminum shell on impact; it was damaged at the exit end and is not suitable for reuse. The aluminum shell showed signs of erosion at the very top of the nozzle.  This area was covered by a ring-shaped fiberglass heat shield that was not present upon disassembly.  This suggests that the heat shield was fully consumed by hot gas erosion during motor operation; a thicker heat shield is evidently appropriate in future nozzles.

The titanium nozzle extension was undamaged and is suitable for reuse in future nozzles of the same design.

Nozzle was damaged in the impact.

Fin Can

The internal “Fin Can” showed some evidence of blow by of the O-ring that normally sits between the Fin Can and the phenolic liner at the base of the propellant grain.  No hot gas erosion was evident in the aluminum structure or in the O-ring, but soot was found on the downstream side of the O-ring.  If this O-ring were breeched, hot gas could—in principle—circulate between the liner and the motor wall; thus, this is a potentially significant issue.  Mitigating against circulation is the use of high temperature grease between the liner and the motor wall. There was no evidence of any soot or hot gas circulation along the interior of the motor wall. Likewise, there was no evidence of any hot gas leak between the fin can and the motor wall.  With minor refurbishment, the fin can does appear suitable for reuse excepting the potential change to two O-rings between the liner and the fin can.

Some “blow by” transient leakage past the seals was evident.
Opposite side of the fin can shows same pattern of the “blow by”.

Phenolic Liner

The propellant grain liner was partially consumed at the forward and bottom ends where the liner is exposed to hot gas for the full eight second duration of the burn.  There was no evidence of any hot gas contact with the motor tube wall and we thus conclude that the existing liner is of sufficient thickness to handle the current eight second burn time.

Conclusions

Based on this inspection it appears some minor redesign of the nozzle top heat shield is required.  It may likewise be prudent to replace the single O-ring used between the internal Fin Can and the phenolic liner with two O-rings.  The rest of the vehicle hardware appears to be in good shape and does not seem to require any design changes.

The lack of gap between the fins and the motor wall appears to rule out the possibility of part of the belly-band having become trapped on one of the fins and causing the unexplained turn to the Northeast.  The cause of that turn remains a mystery.


MTA Launch Event, 2021-10-16, First Update

by Bill Claybaugh, RRS.ORG


In a remarkable demonstration of persistence and luck, RRS President Osvaldo Tarditti was able to find the spent booster rocket. A few photos were captured of the recovered rocket.

Bill Claybaugh’s recovered spent booster casing brought back to the Mojave Test Area (MTA)
Closeup on the bulkhead shoved into the aluminum case of the booster from the impact.
The fins look great and the nozzle was recovered.

Based on the impact location, it was possible to reconstruct a possible flight trajectory by assuming the motor performed as designed and further assuming the front of the vehicle was a flat plate and that the mass did not include the mass of the payload.  We know from video, telemetry, and recovery of the payload that the payload separated from the booster about one second into the flight.

The recovery location on the map shows a northeast trajectory as confirmed by launch footage.

This analysis suggests a burnout velocity of about 1550 feet/second with a peak altitude of about 21,200 feet given the known range of about 14,300 feet. This analysis gives a flight time of about 74.5 seconds and an impact velocity of about 820 feet/second.

Given the observation that the vehicle stopped in about 2 inches (based on the depth of the depression in the hardpan) before falling on its side; we can estimate the impact deceleration.  Given an average velocity during impact of about 410 feet/second because the final velocity is zero and it took only 0.167 feet to come to rest, it follows that the impact occurred over 0.000407 seconds.  This, in turn, indicates an average deceleration of about 31,250 g’s.

The reason for the vehicle turning to the Northeast starting at about 0.20 seconds into the flight remains unclear. There is no evidence either in video or in images of the recovered hardware of any hot gas leak nor of any transient thrust vector anomaly.  The wind was less than 5 miles per hour and from the Northwest; if it had caused a turn, we would expect it to be toward the Northwest, not the Northeast as observed.  The only plausible theory at this time is that part of the belly-band became embedded between the nose of a fin and the rocket body causing the turn via differential drag and then fell away from the vehicle, causing the resumption of normal flight.  Once the recovered hardware is available for inspection, we will test each fin nose to see if a gap exists that might have caught the 0.020-inch thick belly-band.

The recovered payload segment was examined after it was found just north of the launch site.

It also remains unclear as to why the payload separated about 1 second after launch.  The recovered payload showed that both initiators had fired (by design, if one fires the other is ignited; thus, only one signal is required to fire both) but did not show any evidence of having been “swaged” or otherwise subject to being forced off the rocket by aerodynamic or other forces. Neither does the matching front end of the rocket show any evidence for the payload having been forced off. We thus conclude that one of the flight computers ordered the firing of the initiators.

The bellybands being fit checked in the launch rail.
Recovered bellybands have evidence of tearing from what is likely fin impact.

However, the main flight computer stopped working just after 0.80 seconds into the flight for an unknown reason after recovery it was still connected to its battery, which showed the expected 3.87 volts. Further, the limited data recovered from that computer shows that it did not initiate separation of the payload: the firing circuit shows continuity throughout the period that the computer was operating and separately records that no signal was sent by that computer.

Still image of the rocket just after launch making the unexpected hard turn.

This points to the backup flight computer.  That hardware is currently at the manufacture for repair, after which we hope to extract continuity data with regard to its firing status.  Hopefully, once that and other data is available from the backup computer we will be able to establish when it ordered the separation of the payload, and why.

Recovered payload with the main and backup computer.

A second update to this firing report is expected. The booster has been packaged up for a more detailed inspection.