A Tale of Two Tails
When I joined Cessna as an undergraduate engineer in the fall of 1951, the Model 180 and Model 310 were paper airplanes – meaning preliminary design had been completed, detail design was well underway, and first flights were scheduled for somewhat under a year for the former, and just past the next year for the latter. I got to do some work on both airplanes during that time, but the action on the 180 had to be more immediate.
Even an inexperienced engineer like myself knew that the changes to make the 180 a high performance version of the established 170 were going to move the CG range forward and make three pointing more difficult, but incidentally provide a desirably better longitudinal stability within the allowable CG range than its predecessor had.
In flight test good longitudinal stability is shown by requiring increasing stick pull force to reduce speeds incrementally from a trimmed condition, and increased stick push force to achieve ever higher speeds from that same trimmed, zero force, speed.
The reason the CG on the 180 was going to be further forward than on the 170 was because a larger engine, constant speed propeller, and adjustable cooling air outlet (cowling) flaps were going to put extra weight in at the nose. Thus I did some analysis of an adjustable stabilizer that could increase the download at the tail for better three point landing capability, and that feature was incorporated in the design.
But flying the airplane showed that that wasn’t enough, so an extension to the elevator (of a few tried) was added on the prototype configuration to improve the ability to obtain good landings, which is of course more demanding at forward CGs. Static structural tests of that empennage had shown the elevator to be strong enough, but that deflections under load were quite large. But this was a strength test, not a deflection test, and it passed FAA (actually it was still the CAA back then) scrutiny.
So flight certification of the prototype 180 proceeded and also passed FAA scrutiny, but there was one odd result that bothered me a lot. The demonstrated longitudinal stability was generally as great as it was supposed to be, but at just the high speed end of the required cruise condition test, the forces tapered off and became essentially neutral, meaning it took only a forward movement of the stick, or wheel, but no additional force, to attain a somewhat higher dive speed. That did not disqualify the airplane, because it was over a very limited high speed range that it occurred, but I worried that actual force reversal might have resulted at a speed higher than then required to be tested, or an expansion of the CG range might have made the condition unacceptable to the FAA. And, naturally, I wanted to know why this force leveling “at the margin” at high speed happened.
But I couldn’t think of any aerodynamic, or aerodynamic configuration, reason for this condition to occur. So I hypothesized that it must be due to that flimsy elevator extension, which under the high dynamic pressure associated with velocities well above level flight cruise speed would bend a little, act like a trim tab, and reduce the stick force. But who would believe there could be structural flexure problems in the 150 mph range of speed on a stout general aviation model destined, among other things, to be a bush airplane?
Well, we could have addressed this with strategically placed strain gauges on the tail and an orderly arrangement of the resulting flight test data, but we were a cost conscious crew, so I quite cleverly conceived of a simple, conclusive and easy to conduct test that would prove my theory. I would take long thin pencil leads, which were readily available for the metal drafting pencils used back then, and carefully tape them on the surface perpendicular to the trailing edge all along the elevator. The stability of the cruise condition would be flight tested as before, and I just knew that at high speed the elevator would deflect, break the pencil leads, and upon landing I would remove the tape, inspect the leads and show everybody they were fractured and I was right.
I guess because it was my idea I was allowed to fly the test. But we decided to be conservative and at the same time show that the about to come on line production 180 configuration suffered the same problem as the prototype. So we took a production stabilizer off the line and installed it on the prototype airplane for flight testing the proof of both things.
I flew the test according to the regulations and was so excited about my soon to be realized analytical triumph that on landing I couldn’t abide the time taxiing back to the experimental hangar, so pulled off the runway, stopped the engine, set the brakes (I think) and ran back to the tail.
I carefully peeled off the tape and found – all the pencil leads perfectly straight, completely intact and ready to be used for the next drawing that came up.
Let’s say I was a bit disillusioned, and unsure how I could face everybody back at the hangar, so I quietly went inside and, as I should, plotted up the data, and it showed – no force leveling at even the highest speed I had flown, and it was a higher speed than that required for certification! I then remembered that I was pushing pretty hard at that speed, but I guess in all the excitement it hadn’t dawned on me that it was a departure from the expected.
And then I found out, and I think nobody in Flight Test knew, that the structures guys had gotten together with the detail designers and added a stiffener on the production elevators. Not to solve our non-problem, but just to reduce the deflections encountered on the static test of the prototype tail. I took all this as solid, convincing proof that my theory was, after all, absolutely correct.
We decided that I would do further testing with the production horizontal on the prototype, which included all the conditions for certification, and found that with the stiffened production tail the CG range could be greatly expanded and still meet all requirements – though I don’t think that placard change was made for some time, if ever. For good measure I extended the speed range tested in the cruise condition to over 200 mph IAS (from trim at about 150 mph) and found it took over forty pounds of pressure to keep the airplane diving that fast!
So the two tails of the title looked the same when installed on the 180 prototype, but performed very differently when tested, causing me some anxiety. But, my short lived trauma ended up letting it be determined that the longitudinal stability on the 180 was really as good as we had always expected it to be – and at any, or should I say every, speed.
Our Silence Spoke Volumes
As we worked with the 180 we ran up on another challenge.
One of the first things you check in the flight program for a new airplane is for the presence of carbon monoxide, chemically identified as CO, in the cabin, because it is a noxious gas that in enough concentration can cause dizziness, illness, feinting, and even death. And from the very first the 180 showed a concentration of CO that well exceeded the minuscule amounts allowed by the FAA, but not so much that flights couldn’t be conducted if proper precautions were taken.
Since the onset of carbon monoxide “poisoning” caused disorientation, a test of its presence was whether the affected person could legibly write his own signature. So the pilot and I would periodically write our name and show it to the other person to see if it was too squiggly. (In retrospect that was kind of silly, because if one of us was affected enough to write squiggly the other by then probably wouldn’t have been able to tell the difference anyway. But it made us more comfortable.)
And, as long as we were comfortable enough to continue, limiting the CO level was not expected to be a big problem. After all, engines are the source of engine exhaust, which always contain CO, so we were pretty certain where our CO was coming from – that big new 180 engine right in front of us. The typical solution for a single-engine configuration is to seal the firewall, which had already been done on the prototype, but obviously not well enough.
So, we had it done again, and expected that this time they would do it right. On checking it in flight, we found the new sealing hadn’t reduced the CO concentration even one molecule. We asked for an inspection, and related adding to the sealing – and the result was still the same, no lowering of the CO level. This was a puzzlement, since we thought the exhaust from that engine had to have been completely isolated by that last improvement. We did continue flying, and checking signature squiggliness, but with growing unease. But soon we had put enough hours on the plane to require a routine periodic weight-and-balance check on the aircraft, to be sure the ballasting and testing and things hadn’t somehow gotten those parameters out of whack.
Weight and balance checks were a ritual observed by the flight crew, the project engineer, weight-and-balance people and, of course, the Experimental ground crew. On the single-engine airplanes (all we had at that time), it was done by putting slings on the fuselage and lifting the airplane off the ground so huge scales could get the weight components and their moment arms, and by calculation locate the CG, and the above talented group was there to witness it.
As the airplane was lifted to be sure it was high enough to be free of the ground, the bottom of the fuselage came into view. And there, coming out of the cowl flap area, were two thin streams of exhaust residue, following the fuselage bottom to the rear of the airplane, then moving upward and disappearing into the cavity in which the adjustable stabilizer moved.
Our thoughts changed rapidly from weight and balance to noxious gases, and we all looked at one another, knowingly, but – I swear – nobody said a word. Nobody spoke because we all knew instinctively what had to be done. It was obvious that the pressure inside the cabin in flight was below atmospheric, not unusual, but there was a path for that exhaust originating in the engine compartment to come out of the cowling, flow back along the bottom of the fuselage, enter the fuselage at the lightly sealed cavity accommodating the adjustable stabilizer, then flow forward inside the fuselage, accumulating in the cabin and creating the problem we could not fathom the cause of.
Sealing a small bulkhead at the rear of the fuselage was easy compared to doing it on the firewall, and, on our first flight following that, the CO reading in the cabin was zero. We later found the same result, zero CO, with even a routine, production sealing of the firewall. We had followed common practices of design, highly and correctly suspected the source of the carbon monoxide measured, but never conceived how it could get into the airplane’s breathing space where the flight crew sat, which bothered us a lot. Until it revealed itself. And our silence spoke volumes.