> >
>

# Help researching 1961 Electra crash

Accidents and Close Calls Discussion on accidents, close calls, and other unplanned aviation events, so we can learn from them, and be better pilots ourselves.

# Help researching 1961 Electra crash

28th Dec 2015, 03:43

Join Date: Nov 2015
Location: Seattle area
Posts: 212
We don't know what they knew. They're dead, and we can't interview them. We can only guess.

I set aside the root cause in the first few paragraphs, and explicitly said I don't take issue with that.

I take issue with a particular set of descriptions. They are not germane to the cause of the crash, but they are relevant if you care what the true story was. The errors in the CAB and ALPA reports are egregious and obvious. "Slapdash" is an appropriate term.

I don't accept that they are beyond question merely because the writers are probably now dead.
28th Dec 2015, 22:15

Join Date: Nov 2015
Location: Seattle area
Posts: 212
There is an interesting diagram in the CAB report showing the crash site in relation to the runway....it appears the diameter of the turn the plane made was approx 5000ft. Someone with more expertise should be able to deduce average bank angle from this information.
Ok, this is fascinating. I've been running some numbers.

There are probably formulas that would handle all this with the ability to tweak the bank angle change rate over time, but I'm a spreadsheet guy so I did it with a spreadsheet.

I divided the flight into one-second chunks, one per line. I start the plane headed 143 at zero bank, and for each second I increment the bank by an amount. Given the bank angle, I compute the radius of the turn and the number of degrees through which it turns in that second. The degrees turned gets added to the previous line heading. I run the numbers until the heading is 270, which was the heading at impact.

I also compute the distance by which the plane moves laterally from the runway axis for each second, which is the sine of the difference of the current and initial heading, convert from knots to feet per second. I accumulate each with the prior second.

The target is to end up with a lateral movement of 5000 feet when the heading is 270. It turns out a bank that increases by 1.1 degrees per second ends on heading 270-ish and 5000 feet from the runway axis at a final bank angle of 45 degrees.

I'm going to play with the numbers some more, and see what nonlinear changes to the bank angle give us.
28th Dec 2015, 23:02

Join Date: Nov 2015
Location: Seattle area
Posts: 212
I'm not saying this proves anything, ok?

This computation is highly sensitive to the way the incident starts out. If the plane starts to bank and then they get some control on it, you end up with a lower bank at the end.

So I started with the bank increasing by three degrees per second for four seconds. Then I increase it by two degrees per second for five seconds. Then I increase it by one degree per second for ten seconds. We're pretending the bank caught them by surprise and they got some control over it.

After 36 seconds they end up at 270 degrees 5000 feet from the runway at a bank of 33 degrees.

It isn't proof, it isn't even evidence, but what a totally charming result.

Last edited by BRDuBois; 29th Dec 2015 at 00:08. Reason: Miscounted my lines
29th Dec 2015, 08:42

Join Date: Jan 2009
Location: Oxford, UK
Posts: 1,548
Hello DuBois -

Back when your father was pilot of the Electra, I was a housewife with 4 kids, and for the first time had to get a babysitter and fly to Boston from Philadelphia on Eastern Airlines. It was an Electra, the Eastern service was called the Shuttle.

My first time on an airliner (had a ride in a light plane in l944). So I felt relatively safe, with the pilots in proper uniforms, the cabin staff very professional. The engines started, we taxied hither and yon, and then before entering the runway, the pilots were revving the engines, then doing it again, then doing it again, and again, and they certainly made some peculiar noises.
People in the cabin looking at each other nervously. Presently after ten minutes of this, the captain came on the radio and said "You may have noticed one of the engines sounds a bit rough, so we are going back to the gate and get a different airplane.!"

And with that, we all trooped off, and over to the other gate...no further excitement at all. My only recollection of an Electra flight.

Do you fly yourself? I have 3,000 hours in gliders and light aircraft, but in my day women did not often become airline pilots. We were supposed to stay at home and raise a family.

But I have read a few reports, when I knew the incident described, and as many above have said, witnesses do not always understand what happened.
Those who survived and were responsible for what happened tend to be economical with the truth! It can be too painful to admit responsibility when people are hurt.
29th Dec 2015, 13:59

Join Date: Nov 2015
Location: Seattle area
Posts: 212
I did a little sailplane flying when I was a teenager, but had to stop so I could save money for college. Loved it, though.

The scale in the CAB site map shows the impact was about 4100 feet off the runway axis, and the CAB report says it was 3800 feet off the runway. So I'll run some more numbers through my spreadsheet. It looks like the bank can never be higher than about 42 degrees to hit that window.
30th Dec 2015, 02:10

Join Date: Nov 2015
Location: Seattle area
Posts: 212
The iterations were too much like work, so I wrote a routine to do the heavy lifting. One problem is I had to add the speed calcs, which changes the turn radius continuously. Now I have a routine that calcs the flight path second by second, increments the speed until it hits 160 knots, and gives the turn diameter, time to impact, heading, all that stuff. The calc terminates when the heading goes past 270 degrees, as was reported.

V2 for this plane was about 120 knots, at a weight of 93,000 lbs. To be conservative, I'll arbitrarily use 125 knots as takeoff speed. So whatever flight time the bank rate allows us, I tweak the acceleration until it hits 160 knots before impact.

The first lesson is, you just can't realistically get to a 90 degree bank. The faster you increment the bank angle, the tighter the turn gets, and the less time you have to add more bank increments. Incrementing the turn by 10 degrees per second gets you to a 90 degree bank after about 8.5 seconds under a thousand feet off the runway axis. Any linear bank increase less than 10 per second never gets you there at all.

You can get to a 70 degree bank using a linear increment of three degrees of bank per second, but it's at less than 2500 feet from the runway. The lower the increment the farther it will get from the runway, but the farther you get from the 90 degrees reported. There doesn't appear to be any linear bank increase that gets you anywhere close to the target.

The flight time is estimated by Lockheed to be about 45 seconds. The CAB said it was two minutes from clearance to impact. We need a result that's somewhere on that order of magnitude. My computation starts from when the bank commenced, not when the plane took off, so there's some fudge factor.

We get a pretty good result if we increment the bank by 1.5 degrees per second, assuming an acceleration of 1.5 knots per second. With a steadily increasing bank, we end up with an angle of 51 degrees right at 4000 feet after 34 seconds. Problem is, we have physical evidence that the bank angle was about 30-ish degrees when it hit the railroad tracks.

Using a nonlinear bank angle, as mentioned above, we end up with a final angle of 33 degrees 4000 feet away after 27 seconds. That's the right ballpark for the angle and timing. This is based on all the bank incrementing happening in the first 8 seconds, and then the crew manages to stop the bank from increasing but are unable to immediately reduce it.

The intriguing aspect to this, of course, is that the crew must have been having an effect. If they were able to stop the bank from increasing, then they weren't without control.

Last edited by BRDuBois; 30th Dec 2015 at 10:33. Reason: corrected a number. too many numbers.
2nd Jan 2016, 10:34

Join Date: Nov 2015
Location: Seattle area
Posts: 212
I've uploaded a new version which incorporates the discussion of turn radius. It has a map showing the path of a 90 degree bank turn and others. Experimenting with the turns leads to a fairly small range of possibilities that can get the plane to the observed impact point, and the possible bank angles range from about 33 to about 41 degrees. The bank increments cannot be linear to get the plane to the observed impact point, which means something the crew was doing was having an effect. Very interesting.

The new link is http://we.tl/B9288LJnjV and this is now the link that clicking the background will take you to.
13th Jan 2016, 15:06

Join Date: Nov 2015
Location: Seattle area
Posts: 212
I took down the latest version after doing some more thinking. My turn computations were too simplistic, and they don't have the explanatory value I first thought they did.

So I'm still stuck with an intractable problem. Everything from the right wing hit on through the sequence looks like a belly landing. I can't show good evidence that the high-angle bank was wrong, but it's incompatible with the rest of the sequence. I can show that a low-angle bank is reasonable, but can't give any hard evidence for it. If the plane came down with a high-angle bank, I have no clue how it could have transitioned to the belly landing.

More work to do.
19th Jan 2016, 14:52

Join Date: Nov 2015
Location: Seattle area
Posts: 212
After removing the turn radius computations, but leaving in the discussion about how the ALPA came to the wrong conclusions, I put it back up.

http://we.tl/InGXbDKkkX

I'm working on a draft that starts at the final wreckage site and works backwards. We'll see if that one works better.

The problem with the chronological approach is that it's vital to understand the bank angle, but there's not any direct evidence until the right wing impact. People seem to come to a mental halt before that, because I'm disagreeing with the official reports on what is (at that point of the crash) thin evidence. Maybe if I start with the most obvious error I'll be able to carry the reader along better.
27th Jan 2016, 12:13

Join Date: Nov 2015
Location: Seattle area
Posts: 212
Uploaded a new version that reworks the turn and bank discussion. My first take at it was too simplistic.

This version is the kiss of death for the high-bank hypothesis.

Part of the new discussion uses the CAB map to determine what turn radii they were representing, and shows that their map conflicts with the witness narrative. The map constrains the maximum bank angle at impact to about 47 degrees. I also noticed a hint that at least some witnesses said the bank reduced, but the report writers chose to ignore them.

Another part takes my original computations and adds a little enhancement so I can insert the emergency during the normal expected turn. A series of calc runs shows how the flight would look and where it would end up. Turns out a final bank in the low 30's works out fine, and lower is possible.

The most interesting development in this approach is that it provides clear evidence that the bank was being moderated at the time of impact. The crew had leveled the plane from a maximum bank around 45-ish degrees to one of 35-ish, and were flattening the bank at a rate somewhere near one degree per second. In other words, they were very close to pulling out of this.

Take a look, let me know what you think.

http://we.tl/1qiw9nIJlw
5th Mar 2016, 14:53

Join Date: Nov 2015
Location: Seattle area
Posts: 212
From Help Researching 1961 Chicago Electra Crash — Tech Ops Forum | Airliners.net

An interesting suggestion

Although there was no aileron control, the decreasing bank angle, the radius of the turn and the relatively shallow descent could be explained by application of left rudder, probably to the stops. Application of full left rudder would have shallowed the bank angle and decreased the turning radius thus fitting your hypothesis. Still having elevator control and throttle control the crew just might have been able to and quite possibly did set the aircraft up for a semi controlled belly landing. Had they had another hundred more feet of altitude or if not for the railway embankment they might have made it. It would still have been an ugly impact but possibly a semi successful one with some survivors.

Not being a pilot, I don't have a feel for this, but it makes sense.

when the depth of the failings of the NW maintenance staff during the replacement of the aileron boost unit on N137US became apparent all eyes and all hands rushed in that direction and the impact sequence part of the investigation was left hanging. At that time everything about the Electra was under major scrutiny and a severe cloud of suspicion hung over the aircraft. Coupled with the serious failures of the NW MSP maintenance and inspection staff uncovered during the investigation I can easily envision a scenario where all eyes and attention were diverted in that direction due to the potential for further safety concerns. The downside to that shift being an overlooking of a possibly brilliant and heroic bit of flying on the part of your father and his crew.

We were intensely aware of the focus on Electra safety in general, in addition to the work being done to figure out why my dad's plane crashed. So this certainly explains why the impact sequence was ignored when the cause was discovered. However, having an explanation does not excuse the slapdash treatment of that part of the investigation.
6th Mar 2016, 21:33

Join Date: Aug 2006
Location: Timbukthree
Posts: 5,891
Buffalo Airways in Yellowknife, NT, has Electra pilots, as does AirSpray in AB.
7th Mar 2016, 00:23

Join Date: Nov 2015
Location: Seattle area
Posts: 212
Originally Posted by megan
It wasn't slapdash treatment, it just wasn't pertinent to the accident. You see the same in today's accident reports. Investigators are under the pump to find the root cause and to get the message out. Wasting time on non pertinent issues delays that process.
I have no quarrel with their finding of the root cause. But their work on the impact sequence does not magically become rigorous and proper just because they got the root cause right. The work on the impact sequence was still slapdash and it's fair to call it that.

Today based on a tip from a reader I went to the Chicago Tribune archives. On the back of the Monday edition there was an artist's conception of the impact sequence. It shows the main body sliding backwards to the final site. The paper didn't make this up; it had to come from the CAB.

As an old newspaperman I can tell you that the morning edition is put to bed about 11pm. In other words, on the day of the crash the CAB had ALREADY decided that the plane slid backwards, and they never wavered from that. This goes beyond unconscionable, which is what I called it in my analysis. To put out any affirmative statement on Day One is unheard of.

At the time the CAB put their stake in the ground on the backward slide, they had absolutely no idea what the root cause was.
7th Mar 2016, 01:30

Join Date: Nov 2015
Location: Seattle area
Posts: 212
Originally Posted by megan
And the proof is?
Evidence beyond a reasonable doubt. How many airplane crashes involve a plane sliding backwards? Ever heard of one outside of this crash? What are the odds that a reporter would make up such a story, entirely on his own? What are the odds that the CAB would match such an unlikely story with exactly the same one?

There is no need or incentive for a newspaper to posit such specifics. There's enough meat in this incident without a backward slide. Read the archives; it's dripping with pathos and tuggable heart strings.

The only way this was in the paper is because the reporter was fed it. This part, at least, I know how it works.
7th Mar 2016, 18:08

Join Date: Dec 2013
Location: Norfolk
Age: 62
Posts: 1
Having thoroughly read through the reports and reviewed the photographs, I find nothing inconsistent with the aircraft being banked at an angle of 90 to 100 degrees at the point of impact with the railway embankment.

The primary purpose of the accident report was to establish the cause and given the significant disruption and lack of flight recorder data, the investigators carried out a thorough investigation given the techniques and resources available at the time.

Given that the aircraft was at low altitude and climbing when the aileron cable detached, the description of the aircraft entering a continuously increasing bank and turn, before stalling into the ground is entirely consistant with the photographs of the scene. There is evidence that at least one control column was commanding a full left turn, presumably a full left rudder input was also commanded and the crew appear to have been attempting to control the turn with engine power, but simply didn't have the altitude to recover the situation.

I don't find anything in the official reports that suggest the crew acted inappropriately at any stage and there are definite indications that they fought to control the aircraft attitude all the way to the ground. The final radio transmission gave some pretty vital clues to the crash investigators and that is unusual in modern accidents where the emphasis is to concentrate on flying the aircraft. Perhaps one advantage of having more than two crew members in the cockpit?

The accident wasn't ever going to be survivable and I can't agree that the aircraft bank angle did anything other than continuously increase once the control cable separated.

Had the control cable separated in a high level cruise, it is possible that some measure of control over the aircraft attitude might have been regained, but I suspect that a spiral dive would probably have been the final outcome.

Given the high speed of the impact but only a modest low nose attitude, the fuselage would be expected to tumble and break up as it slid along the ground striking trees and other low level obstructions.

As far as I can tell, the crew did everything right, they just ran out of time and altitude. The official reports were as thorough as they could have been for the time. There doesn't appear to me to be any attempt to cover things up or hide any failings. It is clear that short cuts were being taken with maintenance operations, but no more than with any other similar organisations at that time. A lot of pilots and engineers had cut their teeth during World War 2 when the emphasis was in getting the job done, not getting all the boxes ticked.

Different times and different attitudes.

Sorry for your loss, but you should be proud that your dad kept trying to fly the aircraft right to the end and he was doing all the right stuff that he should to try and recover given the circumstances.
7th Mar 2016, 19:02

Join Date: Nov 2015
Location: Seattle area
Posts: 212
Originally Posted by G0ULI
I can't agree that the aircraft bank angle did anything other than continuously increase once the control cable separated.
It's entirely reasonable to say the bank angle in might continuously increase given the mechanical failure. But if it had continuously increased in this case, the impact would have been southeast of where it actually hit. The path would be a tightening spiral with more movement parallel to the runway and less orthogonal. For the plane to hit where it did, it required more movement orthogonal to the runway than parallel, and that means the bank must have moderated.

Given the high speed of the impact but only a modest low nose attitude, the fuselage would be expected to tumble and break up as it slid along the ground striking trees and other low level obstructions.
But it didn't tumble. The right wing was obliterated just inboard of the number four engine, but the left wing's full (or very nearly full) length was present at the final impact site. The empennage was virtually undamaged, and only the forward fuselage broke off.

There doesn't appear to me to be any attempt to cover things up or hide any failings.
I never suggested any ill-intent, merely carelessness. The breakup sequence just didn't matter much.

The recurring problem in this puzzle is figuring out how much we have to discard for the rest of it to make sense. Clearly not all the witnesses agree, so some of them must be in error. The question is to decide which ones.

I've been reading the Chicago Tribune stories. There is (embedded in some purple prose) a clear description of three powerful bounces, as opposed to a cartwheel. I find this fascinating because the pictures seem to show three bounces as well, and that is the scenario I present in my document.

That nugget, plus the evidence that the CAB had laid out their impact sequence on the day of the crash, is going to take a while to process before I publish a new version.

ETA: Another interesting bit in the Trib is a picture of a wing fragment lying on the west pair of train tracks. It's a small fragment no more than two feet across. But the large object lying on the east rails in my document is more like seven feet across. This adds weight to the idea that it's the number four engine lying on the east tracks in my picture, and that's a lock on the shallow bank argument.

Last edited by BRDuBois; 7th Mar 2016 at 19:21.
7th Mar 2016, 21:10

Join Date: Dec 2013
Location: Norfolk
Age: 62
Posts: 1
I personally think the object lying in the track is part of the wing tip and aileron structure, rather than an engine. The metal has been twisted making it appear more three dimensional than the originally flat wing structure.

Given that it was impossible to apply any corrective aileron input due to the control cable being disconnected, the only means of moderating the bank angle would be rudder and differential engine power.

The rudder alone may have been partially successful in moderating the bank angle initially. As the bank angle continued to increase, more and more engine power would have been needed to maintain lift. Once the bank angle passed 30 degrees, differential reduction of engine power to correct the bank would not have been possible. The aircraft would have immediately stalled and fallen from the sky.

The official reports are quite firm in their estimates of bank angle when the aircraft initially struck the high tension wires and the railway embankment. The wing spar alone would have been sufficient to displace the railway tracks. Railway tracks are designed to support heavy vertical loads. The sleepers linking the tracks are designed to maintain constant spacing between the rails, not to resist horizontal displacement. A certain amount of horizontal displacement occurs naturally with constant expansion and contraction of the rails due to seasonal temperature changes. So long as the relative spacing between the tracks remains constant, trains can use the line despite some quite severe looking kinks.

So I believe the debris on the track is a twisted part of the outer wing structure and the number four engine separated after passing over the embankment.

The rear fuselage and tail section are almost completely inverted where they came to rest. Whether they rolled, tumbled or spun to arrive at this position isn't really an accurate indication of aircraft attitude at the moment of initial impact with the ground.

I appreciate that you have put a great deal of time into your research and you have made a good case that the aircraft might have been at a shallower bank angle based on your interpretation of the photographs of the scene and ground witness marks. I simply can't agree with your well presented conclusions.

Unfortunately the original investigators notes and measurements are no longer available, but almost certainly the severed high tension cable lengths would have been measured to give an estimate (and rough confirmation) of the angle at which they were severed. Detailed measurements would have been taken to order replacement cables to splice in repairs no matter what. The artificial horizon had witness marks indicating a bank angle between 90 and 100 degrees. Just two bits of physical evidence in addition to what the investigators on the scene where able to establish, such as the impact point on the rails in relation to the high tension cable breaks.

Clearly the crew had some initial measure of control over the bank angle and roll rate, probably using the rudder, but eye witness accounts and physical evidence indicates that the aircraft progessively rolled further and further until contact with the ground. There is simply no practical way that the crew could have levelled the wings or reduced the bank angle at impact given complete lack of aileron control.
8th Mar 2016, 12:35

Join Date: Nov 2015
Location: Seattle area
Posts: 212
Originally Posted by G0ULI
The rear fuselage and tail section are almost completely inverted where they came to rest. Whether they rolled, tumbled or spun to arrive at this position isn't really an accurate indication of aircraft attitude at the moment of initial impact with the ground.
First of all, thanks for your thoughts. This is what I came here for.

The key evidence at the final wreckage is the position of the wings. They are inverted with the leading edge pointing back along the path. So either the plane slid inverted, which the remaining rudder rules out, or it arrived upright. That, plus the ditch, tells us a lot about that final site and how the plane got there.

I didn't mean to say (in my document) that the final position was simply maintaining a position it had been in at the moment of initial impact. Many things happened in between. My point was that to arrive upright at the final site with as much energy as it had means that the plane cannot have been dissipating that energy in a cartwheel. It's the final site's energy, as much as position, that rules out the tumble.

Unfortunately the original investigators notes and measurements are no longer available, but almost certainly the severed high tension cable lengths would have been measured to give an estimate (and rough confirmation) of the angle at which they were severed. Detailed measurements would have been taken to order replacement cables to splice in repairs no matter what.
I'm following various paths to try to locate more documentation. Something may turn up.

A previous post mentioned that the cables would be measured, but there's no need to do so from the power company's point of view. If they trusted the remainder of the snapped span, they could splice new line on one of the ground-level pieces, pull it to the opposite tower by pulley, and splice it in up there. If they didn't trust the snapped span, they would pull new line up to one tower using a pulley, over and on to the next, pull taut and splice it in.

None of that demands that they know exactly where the breaks were, and determining the break point would slow them down. I suspect they would measure only if asked by the investigating authority. I'm confident that such measurements would be routine today, but have no reason to think it would have been routine then.

I also don't have a description of the power lines. The measurement would be useful only if the lines had significant vertical separation. So in the absence of any explicit statement, I don't see any reason to say that some hypothetical measurement should be presumed to exist and therefore admitted as evidence. We just don't know.
8th Mar 2016, 15:13

Join Date: Dec 2013
Location: Norfolk
Age: 62
Posts: 1
The power lines are quoted as being two sets of three wires carrying 38,000 volts. These lines would be erected at a minimum distance from the ground and away from any horizontal obstructions to avoid the risk of arc over. The power lines would have complied with electrical standards imposed at the time and are unlikely to have been "over engineered".

So it should be possible to establish from the old standards, how thick the cables were, the typical breaking strain, at what height the cables were strung above the ground and the horizontal separation from the railway tracks.

This information may help to indicate the vertical path of the aircraft as it sliced through the power lines and the wing tip contacted the railway embankment. It should at least verify the estimates of a ten degree nose down flight path.

Standards for modern high voltage power lines are probably not too dissimilar to those as the time of the accident since the same laws of physics still apply to power handling and flash over voltages.

For example, in the UK, the minimum height above ground would be in the region of twenty feet. The cable thickness would be at least two inches, or more, depending on the current carrying capability.
8th Mar 2016, 15:26

Join Date: Nov 2015
Location: Seattle area
Posts: 212
Ok, thanks for seeing the power lines. The Chicago Trib says the power poles were 60 feet high and 300 feet east of the railroad embankment. This agrees very well with the map images in my document, using the scale on the left of the image.

That means the descent was very gradual, which agrees with witness reports. It also means anything steeper than 63% must have been in a stall and is not a feasible angle given the horizontal distance it had to cover to the tracks.

Once again we have to decide how much must be discarded for the remainder to make sense. The artificial horizon was treated with skepticism in the official reports. If we rule out the witness statements of the vertical bank, everything else fits.

ETA: The RR embankment was said to be the highest landform in the area. That means the power lines would have been somewhat less than 60 feet above the tracks, possibly considerably less.

Last edited by BRDuBois; 8th Mar 2016 at 15:37.