View Full Version : 'Dutch Roll'=Literally An Aircraft Version of Death Wobble

March 21st, 2014, 07:42 AM
AF: Crew Died When Tanker Did 'Dutch Roll'


The Spokesman-Review | Mar 14, 2014 | by Jim Camden

A KC-135R tanker flown by a Fairchild Air Force Base crew shook apart over Kyrgyzstan last May, an investigation shows. The tail section came off and the rest of the plane plummeted through the air until it exploded then crashed, killing the three on board.

Shortly after the tanker took off from Manas Air Force Base, the crew reported the plane was "waffling" and later that it was "bent."

The tanker was experiencing what pilots call "Dutch roll." The nose was moving left and right while the wings were alternating up and down. Unless it's brought under control, the phenomenon can exert too much stress on the structure of a plane.

That's what happened to the tanker. About 11 minutes after takeoff, the plane exploded in flight and crashed. Killed were Capt. Mark Tyler Voss, 27, the pilot; Capt. Victoria Pinckney, 27, the co-pilot; and Tech. Sgt. Herman "Tre" Mackey III, 30, the boom operator.

The Air Force released its findings on the May 3, 2013, accident Thursday morning in a briefing from Brig. Gen. Steven Arquiette, who led an investigation into the crash of the tanker that went by the call sign Shell 77.

The tanker, which had been in the Air Force inventory since it was delivered by Boeing in 1964, had been in Kyrgyzstan for less than 24 hours. Assigned to McConnell Air Force Base in Kansas, a different crew had flown it to Manas the previous day and reported it had a "right rolling motion," but the plane had no major maintenance problems in the previous 14 flights.

Voss, Pinckney and Mackey had been deployed from Fairchild to Kyrgyzstan about three weeks earlier and were the first crew to fly the plane after it arrived. http://ad.doubleclick.net/ddm/trackimp/N7384.4611.MILITARY.COM1/B8032641.107034634;dc_trk_aid=280091562;dc_trk_cid=57377200;ord=[timestamp]?
Shell 77 took off with reports of isolated thunderstorms. About two minutes into the flight, co-pilot Pinckney reported over the radio that the plane was "kind of waffling a lot." Later she said, "the jet's bent."

Data from the flight recorder shows the plane was going into a Dutch roll, with the movements getting more pronounced with each wag of the nose and dip of a wing.

"As it grows, it gets to be quite violent," Arquiette said.

The KC-135R has systems that are designed to help prevent or correct such rolls, but investigators who inspected pieces of the plane determined that at least one of them, a rudder lock lever, was significantly worn. It may not have operated properly and could have increased the problem of the roll.

Corrective action taken by the crew couldn't stop the roll and appears to have made it worse, based on data from the flight recorder. Eventually, the nose of the plane was swinging from 12 degrees right to 17 degrees left, with one wing lifting up then dropping as the other wing lifted. The stress caused the plane's tail section to fail and separate from the plane just ahead of the vertical stabilizer.

The tanker dove, nose down at 82 degrees, gaining speed. The stress forced the right wing off. Fuel from the wing tank spilled out and ignited. The plane exploded midair and crashed to the ground in pieces. Debris was scattered over an area more than 2 miles square in the foothills below the Himalaya Mountains.

Investigators know Shell 77's tail came off first because there was no fire damage on that section.

The investigation pointed to several problems with training and information provided to tanker crews on Dutch roll. Arquiette said when he was a young pilot in the 1980s, he flew the KC-135A, a version with less powerful engines. Crews practiced regularly with how to handle Dutch roll, which was a common occurrence.

"The R is much smoother," Arquiette said. "When its systems work, they work quite well."

When the engines were replaced, the revised KC-135Rs also had new systems installed to prevent Dutch roll. Crews train less on the problem, both in the air and in flight simulators. Investigators couldn't even re-create the conditions of Shell 77 in a simulator at McConnell. Those simulators are being updated to improve training.

As a result of the crash, KC-135R crews will get more training on how to handle Dutch roll. The procedural guidance in the flight manual on how to handle a Dutch roll, he said, is "cumbersome and disjointed." The manuals are being revised.

Maintenance is being increased and systems for controlling the rudders are being updated. Boeing is evaluating the rudder system, and flight data recorders are being updated.

The changes are important because the Air Force may continue flying the Eisenhower-era tankers through the next decade.

The remoteness of the crash site was among the reasons the investigation took almost twice as long as a standard accident investigation, Arquiette said. Briefing the crew's family members about the results was "the hardest thing I've ever done in my career," he added.

March 21st, 2014, 08:17 AM
Mmmmmm, dutch roll.

http://i46.photobucket.com/albums/f131/Waifer2112/DutchCrunch-Roll-Web1_zps2071ab9d.jpg (http://s46.photobucket.com/user/Waifer2112/media/DutchCrunch-Roll-Web1_zps2071ab9d.jpg.html)

March 21st, 2014, 08:18 AM
Hmmmmmm, dutch roll.

http://i46.photobucket.com/albums/f131/Waifer2112/DSCN13491_zps75b24a57.jpg (http://s46.photobucket.com/user/Waifer2112/media/DSCN13491_zps75b24a57.jpg.html)

March 21st, 2014, 08:22 AM

March 21st, 2014, 09:12 AM
so what happens - the swept wings push more length to the airflow on the side that swings forward, lifting that side and slowing it down, making that side fall back and makiing the other side do the same thing?
How would you stop that? small rudder and aileron in opposite directions? I don't think slowing down would help.

March 21st, 2014, 09:14 AM
With death wobble my jeep never explode into pieces and crashed in a ball of flames. That sounds so friggin terrifying.

Da Kine
March 21st, 2014, 09:22 AM
Great...now whenever I fly and hit turbulence, this is going to pop into my head...

March 21st, 2014, 09:23 AM
With death wobble my jeep never explode into pieces and crashed in a ball of flames.

It's a Jeep, I wouldn't write it off as not being a possibility.

March 21st, 2014, 09:24 AM
Great...now whenever I fly and hit turbulence, this is going to pop into my head...

lol.. was thinking the same thing

March 21st, 2014, 09:27 AM
I just looked it up. It works the way I thought, but to fix it, you just resist the roll. at the far end of the roll, as it starts to go back,roll away from the direction it's going - not enough tto move it, just to slow down how fast it comes back to (and through) level. I guess the motion's slow enough to react to. I was picturing like with death wobble - where you can't move fast enough to react to each cycle and have to do something static... slow down, or take a curve... or find a road with heavy crown. I came home once doing a slow serpentine.

March 21st, 2014, 09:37 AM
I came home once doing a slow serpentine.


March 21st, 2014, 09:52 AM
What happens as the yaw increases you actually loose lift on one wing causing the dip/roll.

March 21st, 2014, 09:54 AM
The crew of Shell 77


Da Kine
March 21st, 2014, 09:56 AM
On submarines we had something that could happen during a high speed turn called a snaproll. During a high speed turn the boat would roll toward the turn and and either up or down depending on which way your were turning. If it got too far, it could actually continue all the way over. Not sure if it ever happened in real life or just a theory, but we all took it very serious during high speed transits. Something about turning a nuclear reactor over was not something any of us wanted to see. The other issue was that our ballast tanks would be useless if were upside down as they had open grates on the bottom.

March 21st, 2014, 10:09 AM

Thanks for posting that.

March 21st, 2014, 10:17 AM
The crew of Shell 77



March 21st, 2014, 10:22 AM
Yaw dampeners. Get one. Beechcraft Bonanza v-tail owners swear by them. Pilots can aggravate the yall, and dutch roll like they can with pilot induced oscillation. Knowing how to use your rudder pedals helps. Trike pilots may not do that effectively. Some time in a Cub, Stinson, Citabria, or Bellanca Champ helps that.

March 21st, 2014, 10:25 AM
KC-135's do have a yaw damper system.

March 21st, 2014, 10:27 AM
Yaw dampeners.
They had one. It wasn't working, and because they had them, they weren't trained to deal with the problem without the dampers. The crash was caused by a failure in training.

March 21st, 2014, 10:31 AM
They had one. It wasn't working, and because they had them, they weren't trained to deal with the problem without the dampers. The crash was caused by a failure in training.

Yep. What do you do with death wobble on the Jeep? I've experienced it with mine. Pull back the power.

March 21st, 2014, 10:40 AM
Ok I read the full report the malfunction in the rudder control system was a slow oscillation. As the rudder moved back and forth plus the crew trying to counter the deflection the deflection got bigger. Now on the center pedestal is a safety covered switch for shutting off the damper system. What appears to have happened is when the crew shut off the system the rudder defaults to the center position. It was at full deflection and the rudder very quickly centered and the load literally tore the tail off. The crew reported to ground and it was confirmed they were tiring try to control the rudder (countering) a 3000 psi hydraulic actuator with manual input rudder. Now speculation on my part as yaw damper is part of the autopilot system is the yaw damper lost its electrical null. The system didn't know what zero was anymore and once it received an input it started searching for null and thus started searching or oscillating and it just get bigger. The yaw damper system is a separate system from the main autopilot and not controlled by the computer so there is no auto disconnect if this happen in that system.

March 21st, 2014, 10:47 AM
Here is what the system looks like for those interested.


March 21st, 2014, 11:04 AM
Thanks Oscar. Very interesting. If you pull the power back temporarily that should stop the dutch roll. I see the 250 KTS number, so if you reduced speed that should help. I've got the tail wagging in a Bonanza, but some pressure on the rudder pedal solved the dutch roll. This was in a 76 V-35B.

March 21st, 2014, 11:15 AM
Scary stuff.

RIP for those we lost. :usa:

March 21st, 2014, 01:43 PM
Yep. What do you do with death wobble on the Jeep? I've experienced it with mine. Pull back the power.

apply a lateral force if I want to maintain speed (road crown, serpentine), or slow down. I didn't figure slowing down would help with the aircraft.
How fast is that motion? I'd suppose in the little plane, 5 seconds, on the big one, 20 seconds?