NTSB preliminary report on the B-17 crash

A reporter sent me the NTSB preliminary report on the recent B-17 tragedy at Bradley in Connecticut. Here’s what I wrote back…

This more or less eliminates a popular speculation that the plane was mistakenly fueled with Jet A. Two people were there at the fueling. A good FBO will sample the fuel truck every morning for water, sediment, etc. This particular truck was thoroughly checked as part of the investigation.

The NTSB found no skimping on maintenance. The plane was within its annual inspection and had received progressive inspections at 25-hour intervals. Three of the engines were more or less fresh from overhaul (0 hours in January 2019; 268 hours of operation since then). Based on a quick search, I found that gently operated big radial engines in airline service after World War II were able to go 3,000+ hours between overhauls.

The plane should have been light. Out of a total fuel capacity of 1,700 gallons for a standard B-17, only 160 had been added that morning (to whatever was held in reserve from previous flying).

The report seems to eliminate another possibility, i.e., that the No. 4 engine wouldn’t feather properly, thus creating a huge amount of drag on one wing. (see ASA 2311 and ASA 529, both of wouldn’t have occurred if a prop could have been feathered; multi-engine planes are designed with the ability to twist the prop blades until they’re more or less at a knife edge to the wind, thus minimizing drag and workload for the remaining engine(s)).

The report hints at the No. 3 engine also being feathered. That would be bad. The plane isn’t designed to fly with two engines on one side and none on the other, though it probably would still be controllable at a low power setting consistent with approach and landing, especially at the “no flap” setting that they were using (flaps are essentially for landing on a short runway because they let the plane be flown slower and descend steeply at a slow speed, but they add drag and require extra power, so the pilots were being conservative in not extending them and relying on having a long runway for rolling out from a higher airspeed).

This doesn’t resolve any of the mystery, I don’t think. The failed engine was feathered, so the multi-engine plane should have been flyable just like the book says to fly it. The fuel was good 100LL. The pilots shouldn’t have needed more than a touch of power since they wanted to descend and had no flaps out.

The most surprising part of the report: “the airplane was about 300 ft agl on a midfield right downwind leg for runway 6.” Normal pattern altitude is 1000′ above ground level (AGL). “Right downwind” means they were going in the opposite direction of Runway 6 such that they’d have to turn right and right again to land. If this 300′ AGL altitude is correct, the plane was buzzing buildings on the SE corner of the airport (diagram) and better set up to land on Runway 33.

Still just as sad and still nearly as mysterious.


5 thoughts on “NTSB preliminary report on the B-17 crash

  1. If I understand the report, the plane would have been forced to make turns into the dead engines to make the runway. Could it be that the power of the working engines was kept way too low in order to maintain control? That could explain the loss of altitude. Perhaps it was impossible for the pilots to maintain both control and altitude.

  2. Perhaps one engine failed then they feathered the wrong engine resulting in two engines dead on the same side. Our fine host Phil has a history of doing this in just 2 engine planes. Imagine the complexity in a 4 engine plane.

  3. Pretty much how Louis Zamperini went down. #1 engine failed at 800 ft. They feathered the #2 engine & down they went.

  4. Comments from Matt Guthmiller on earlier article: Some of the old radial magnetos have issues with moisture penetration. Perhaps if they had to resolve that on the ground with #4 before takeoff, they assumed incorrectly that a right-side engine failure was #4 since they had just had issues with that, when in fact #3 failed instead? In that case they would have been flying along with #4 feathered, #3 failed but not feathered, and relatively low & slow while figuring it out.

    And I see the NTSB report also says the pilot reported to ATC a “rough mag” on #4.

  5. I come from a background in Mech Engineering and with a long time passion for high power radial aircraft engines and all technical aspects of aircraft. Have read hundreds of NTSB crash investigation reports and many dozens of articles and books dealing with aircraft accidents going back to the days of the great airships and flying the mail in the 1920’s. It has been interesting to discover, in all this study, that many accidents start with a mechanical failure of some nature, but in the end human error often becomes a factor more than any other.

    Minimum altitude agl, and minimum airspeed, were major factors in this accident. If what happened to precipitate this accident had happened at 5,000ft there would have been both speed and altitude to trade off….for time to think about the problems and the response. As I read this Nine O Nine was airborne for about 6 minutes. The radio call asking to return to the field was at least 2 mins after takeoff so the window for diagnosing the issue(s) and doing what one could to deal with them, was way less than 5 minutes! This is not much time to give one the opportunity to really think. Something has to be done very quickly. I write this NOT to second guess the pilots, none of us would ever choose to be in such a precarious position, but to simply point out that these general circumstances have indeed happened before, many times, in previous incidents, and on numerous occasions the problem engine, on multi engine aircraft, was misdiagnosed….and the wrong engine shut down! I am not trying to say that this is definitely what happened, just to point out that this kind of forced quick decision making has led to such an occurrence before, on numerous documented occasions……and is one possible explanation of events here.

    Think about this for a bit. At least one, and perhaps two, witnesses on the ground saw the aircraft fly overhead and said the problem engine was #3, which was described as “nearest the fuselage on the right wing”, so no question which engine the witness(es) thought was giving trouble. The crew had a very short time to deal with the issue. Having delayed the flight in order to dry out the mags on #4 could have easily led the crew, or any other one of us in that same position, to think immediately that it was #4 and feather that prop. If #4 had indeed been the issue, once the prop was feathered flying out on 3 engines, as lightly loaded as it was, should have happened. Circumstantial evidence at this point certainly makes this misdiagnosis a possibility, otherwise why did they crash if the problem engine was properly dealt with? I am sure that all the engines will be disassembled and studied by experts in the R-1820 engine and the final report will establish the condition of each of them at the time of the crash. Then we will likely know what happened to bring about such a tragic ending.

    One of the leaders of the NTSB is board member Jennifer Homendy, the blonde woman in the videos. She has a long background in “safety issues” in in mass transportation field going back to 2002. She has said that the goal is to not have anyone hurt ever, knowing full well that is not obtainable, but try to get as close as reasonably possible to that. I believe, from what I have seen of her, and NTSB reports in general, that she is a very competent lead investigator and will see that they get to the bottom of this and know what really happened. I think several things may come from this that will affect future flights. First of all I am adamant that we do all we can to keep these pieces of history flying as much as we possibly can. I would hate to see these historic aircraft turned into just museum pieces…and I don’t think that is going to happen. But I do think better procedures will come from this event.

    One of the things that is bad about having the pilot of the aircraft, flying passengers for hire, doing mechanical work on the aircraft himself, is that only one head is involved in evaluating the mechanical condition of the aircraft, in prelude to flight, a flight that is desperately needed in order to bring in revenue. Go to the Collings Foundation website on Facebook and you can read where they are almost begging for money to support the aircraft. I did a walk through of Nine O Nine, and the B-24, in Bangor, ME in about 2006 and would have accepted the risk and flown on either if I had had the money to spare at that moment. But in the scheme of things here, the NTSB is going to look at these flights as a commercial operation as money is exchanged for the chance to fly in the aircraft so the NTSB is going to look hard at what all of these organizations are doing….in a whole new way as they are responsible for oversight of public safety……and rightfully so.

Comments are closed.