NTSB report on Crash of Alaska Airlines Flight 261 1

[3DDave]

Report NTSB/AAR-02/01 is available free from the U.S. NTSB. The crash occurred January 31, 2000 and the report was adopted about two years later.

I suggest reading it. It contains a detailed failure analysis of a linear actuator, primarily from lack of lubricant, but also from a long chain of slightly bad decisions.

It answered a question I'd had concerning the mixing greases with potentially different bases leading to oil bleed-out. At least for the greases used, there was little incompatibility.

It skipped exploring what has been a particular irritant - that military specifications for grease cover performance, not ingredients. There are some specs where compliant greases have much different thickeners. Lithium soap is one, clay (dirt!) is another. In an investigation of a grease lubricant failure I contacted a grease-development chemist as well as the government agency responsible for the specification. The chemist pointed out the problem and the agency acknowledged it as flaw in the acquisition process.

What I learned at the time is that grease is not a lubricant - grease is a lubricant carrier. Its development in the late 1800s was to replace manual and drip oilers**. The manual version is seen in movies with steam locomotives as trainmen work around the train with oilers. Grease is a complex assembly that was developed to cause the oil to stick, and bleed oil out onto nearby surfaces. Don't confuse mechanical greases with cooking grease, which is really fats and oil.

The characteristic that causes problems is that once the oil has bled out, the thickener (sometimes clay) remains. In the actuator nut, the entire passage from the grease fitting was blocked with thickener, preventing grease from reaching the interface between the screw and nut. In many cases the amount of oil that can be delivered is limited by the size of the grease glob that can hang on, or the volume of the grease groove or boot. What exaggerates the problem is new grease can't entirely displace the dirt, further limiting the amount of oil delivered.

One feature that seems wrong is acceptance of a 32 microinch finish on a high load sliding contact surface without forced lubrication. Similar experience shows that any grease that might be in place is extruded in a short time from the center of the bearing area leaving the relatively rough steel surface to abrade and contact weld to the copper in the mating part. These welds are weak and tear loose, producing tiny flakes.

It should be possible to create top and bottom oil seals for the nut and put a drip feed reservoir to keep the nut filled. At the least this would give a chance to check the oil level, something impossible with grease, and debris could not block the lubricant path.

The nut would have radial grooves in the lands to allow oil to penetrate the full contact width at several locations on the contact path. I'd try for an 8 microinch or smoother finish on the screw. It should be possible to get 10,000 flight hours/ .001 inch.

**Perhaps you've heard of Mr. Zerk, as in Zerk fittings? He's an actual person who held patents on oil delivery systems and then developed those fittings and grease gun nozzles that attach to them. There were a large number of grease fittings and nozzles patented before the designs were became mostly reliable indicating a great interest in injecting grease.

 

[gearcutter]

The Discovery channel's 'Air Crash Investigation' episode called 'Cutting Corners' is what originally got me interested in this very sad story and the investigation that followed.

I agree with your comments about grease; how it is mostly made of thickener and actually carries a relatively small amount of lubricant.
I'm constantly amazed at how often it's used in splined joints and the resulting damage that can result, particularly if applied 'once only'.



Ron Volmershausen
Brunkerville Engineering
Newcastle Australia

http://www.aussieweb.com.au/email.aspx?id=1194181

 

[tbuelna]

3DDave-

I didn't take time to read thru the NTSB report, but my recollection of this event is that it was determined the root problem was excessive backlash had accumulated in the jackscrew/nut assy of the horiz stabilizer trim actuator. Increased backlash during service due to wear of the nut threads was considered normal, but there were periodic inspections required to determine if the backlash was within specified limits. And if not, the backlash needed to be adjusted. My understanding is that while backlash inspections were indeed conducted, there was an issue with the method used to check the backlash which result in incorrect measurements.

As for using grease to lubricate mechanical devices like flight critical actuation systems, while it can be shown to work well in theory, in practice it requires lots of specific detail design features to make it perform well. Unfortunately, many of the grease lube provisions I have seen on aircraft mechanical systems could have been designed much better.

Regards,
Terry

 

[3DDave]

Hi Ron and Terry,

Thanks for the comments. I've appreciated the quality of posts from you both.

Terry, In this case there was no adjustment to be made. Alaska Airlines just ran the jackscrew until the threads in the nut were worn almost completely away. In the photos what looks like little pieces of wire are wrapped around the screw, but they are all that remains of the thread in the nut.

Ron, about "Cutting Corners" - I'll keep an eye out for it. If you want a chilling read, the part of the report that describes the pilots attempts to save the flight and just ran out of altitude is dramatic.

As much as it appears that maintenance was not done right, it is also clear the design is such that getting an accurate assessment of the nut thread condition was difficult. I can imagine a way of decreasing the wear rate, but not to also give a direct measure of thread thickness without total tear down. Backlash measurement is only good if all the wear products are flushed from the joint.

 

[gearcutter]

Hi Dave,

Here's part 1 -

http://youtu.be/W_cJ6XzN3Cc

Ron Volmershausen
Brunkerville Engineering
Newcastle Australia

http://www.aussieweb.com.au/email.aspx?id=1194181

 

[tbuelna]

3DDave,

If you read the NTSB report you linked, the backlash limit noted in the Boeing service manual was 0.040", and any actuators having more backlash than this amount should be removed from service. It would be incorrect to claim that AA willfully chose to continue operating the actuator until the jackscrew nut threads "were worn almost completely away". The report noted that there were a few instances where similar actuators showing more than 0.040" backlash were removed from service.

According to the NTSB report, there was no fundamental issue with Boeing's design and certification of this actuator. In fact, the basic actuator design has performed quite well overall with the 717 and MD90/95 aircraft models. Unfortunately, this particular problem seems mostly to have been the result of a mis-interpretation of inspection procedures.

 

[CoryPad]

3DDave,

Thank you for initiating this discussion. I intend to read the report. I remember the head of the NTSB presented some preliminary findings at our local ASM International chapter meeting a few years ago and I recall images of the worn nut threads.

Your description of grease is not accurate in my opinion. Grease is a lubricant. While each formulation is different, typically greases are 80 % to 90 % oil, so the thickener is a minority constituent. Calling the clay additive "dirt" is a gross misrepresentation - clay additives used in greases (or plastics or coatings) are highly functional materials that are not contaminants as you alluded. Modern greases do not have intentional oil separation or bleeding to provide lubrication.

 

[3DDave]

Hi Terry,

I did read the report and did not say AA did it on purpose; but they did run it to failure.

While the NTSB/FAA may be happy with the design, I believe that items that wear at unexpectedly high rates indicate a fundamental flaw, either in design or expectation. Like the runup to the loss of Shuttle Challenger, where no o-ring erosion was expected, NASA had continued missions because unpredictable erosion became OK.

I don't recall reading in the report that the actuators were required to be disassembled and cleaned before the backlash check was done. I suspect the build-up of wear debris and grease thickener would mask the fact that the nut threads were wearing away too much. Since it isn't just backlash that is important, but thread thickness, I feel it is an oversight of all parties to not inspect the critical item.

 

[3DDave]

CoryPad,

I agree that when first installed the thickener is a valuable component, but after the oil is exhausted, the thickener is a difficult to remove contaminent. Considering grease as an engineered system of components, some of which leave the system over time, will better serve than to consider it a homogenous material.

Refering to the clay as "dirt" is to emphasize that that is what's left when the oil is gone. Since no one replenishes the oil that has left the grease, it's important to understand that an application of grease isn't going to displace dirt.

My source of information was a major Petro-company grease chemist who was very happy to discuss development and useage generally. I don't recall precisely how I found him, but I think I started with the QPLs for MIL-SPEC greases and either the QPL agency knew or it was listed in the contact info. I also got a good insight into the difficulties in bearing packaging and storage, as I was trying to find out why 'fresh' bearings had no lubricant. Turns out bearing makers, grease makers, and bearing packaging makers had all been working to resolve that problem, without universal success. We ended up requesting bearings without lubrication and one seal loose so that they could be greased just before install, eliminating any in-storage grease failure.