Summary of essay by Javier Arango on Sopwith Camel handling characteristics

[Holtzauge]

Through aviation journalist Peter Garrison I have been fortunate enough to obtain a previously unpublished sixteen page essay by the late Javier Arango covering his impressions of the Sopwith Camel. Based on this essay and with Mr. Garrison’s permission I have added a new chapter to a paper I have written comparing the Sopwith Camel and Fokker Dr.1 turn performance which you can find here.

 

Would be interesting to hear other people’s impressions of how what is described in the essay compares to the in-game rendering of the Sopwith Camel. Some initial thoughts from me:

 

1. The sim behaviour of attaining roll angle with left rudder seems to be well aligned with IRL description.
2. Ailerons appear far too effective in game compared to what the essay say’s.
3. In the IRL description of the Camel’s adverse yaw this seems far more pronounced than what we see in-game.
4. While I think the tendency for the nose to rise when turning to the left and for the nose to drop when turning to the right due to gyroscopic forces is there I can’t say it matches the essay's description as this being “violent”.

 

Again, this is no attempt to snipe at Flying Circus which I think is a great game but simply an observation of some differences between in-game and IRL behavior.

 

Summary of essay:

 

This chapter is based on an unpublished essay by the late Javier Arango generously made available by his widow via aviation journalist Peter Garrison. The essay is a great read and provides a wealth of information about handling characteristics and comparisons to anecdotal evidence. However, due to the limited space, only parts can be included in this paper. Reading the essay, I found this passage from the conclusions especially moving:

 

”WWI airplanes can act as a Rosetta Stone, helping us understand the past by deciphering the meaning of their pilots’ words. Restoring the airplanes, working with them, and flying them also teach us that these airplanes were very much extensions of their pilots. The relationship between a pilot and his flying machine in WWI was still symbiotic. The capabilities of each airplane depended on the skill of its pilot, and what the pilot dared to do depended on the characteristics of his airplane.”

 

This observation is of course not only applicable to the Sopwith Camel and familiarity breeds confidence in any plane but it bears highlighting since it seems that while the Fokker Dr.1’s stall is benign, the Camel’s is not according to Reno Air Race and stunt pilot Chuck Wentworth who also flew Mr. Arango’s machines. Consequently, it’s not unlikely that the average Dr.1 pilot would feel more comfortable than a Camel pilot pushing the plane to the very edge of stall which is also what’s needed in order to attain the peak turn rates. Later in this chapter the effects of aileron, rudder and elevator deflection will be covered in more detail but on the subject on turns, below are two quotes from the essay that are especially interesting:

 

“Of the many manoeuvres we tested, the most revealing was the turn. In this essay I will discuss what we learned about the Camel’s turning behaviour. Concentrating in great depth on turning alone will best illuminate not only the controversies about the Camel but also our flight-test findings and how they relate to physical design features of the airplane.”

 

“Many of the Camel crashes occurred while it was turning. So we wanted to try to decipher what exactly made it simultaneously so lethal and so effective.”

 

Leaving turns aside for the moment, in the essay Mr. Arango makes a point of stressing the Camel’s agility and likens flying it to riding a bicycle and that it was exceptionally responsive and would feel like an extension of his body. It would turn just fine in both directions and the gyroscopic forces were not even noticeable. However, not all of his observations about the Camel are rosy:

 

“The Camel has no control harmony. The elevators are excessively light and responsive. The rudder is powerful but has no feel. The ailerons are very heavy and don’t seem to do much, except the opposite of what you want. The pilot needs to continually move all the controls all the time with varying degrees of force just to keep the airplane going straight. The Camel has no stability, so it keeps wandering about all over the sky.”

 

While the essay is mainly focused on the Camel’s handling characteristics, Mr. Arango several times points out the problem of terminology and that pilots of the past and present sometimes mean different things using the same words. As an example, he mentions the term “turn” and that for WW1 pilots this could mean both what we today would call a snap roll or a descending spiral depending on context. In addition, he points out that based on his experience, following these century old pilot’s instructions to the letter would in some cases produce “very, very unpleasant results” but at the same time tempering these observations by explaining how these statements from the past should be interpreted in a more modern context.

 

The Camel test flights were done with a data recording device activated and various manoeuvres designed to isolate and measure the effect of different control inputs. Interesting to note is that the flights were done with only a partly filled fuel tank but that even so the resulting balance was in the order of 31 % MAC which is very close to Mikael Carlson’s Fokker Dr.1 at 32% MAC meaning that both aircraft can on theoretical grounds be expected to be very lively in pitch. Extrapolating on this information it’s also easy to understand just how unstable in pitch a Camel could become if you topped off the tank!

 

Returning to the test results and beginning with the ailerons, when attempting to roll to the left with aileron input only, the nose swings violently to the right so that when a roll angle of 30 degrees to the left has been attained, the yaw angle to right is all of 30 degrees due to the excessive adverse yaw. In a modern plane the nose displacement tends to swing back but in this case the nose stayed where it was resulting in zero turn rate to the left! According to Mr. Arango, this unusual behaviour is most likely due to an unfortunate design decision to hinge the Camel’s aileron on the lower edge opening up a wide gap that resulting in what he called astonishingly powerful adverse yaw.

 

As opposed to the ineffective ailerons, these are Mr. Arango’s own word on the effects of applying left rudder: “The Camel responded immediately to the rudder input by correctly yawing its nose to the left. The wings quickly rolled left and automatically established a 30-degree bank. Surprisingly, the nose rose on its own, which is what a pilot usually needs to do by hand in order to make the airplane turn tightly without losing altitude. The Camel responded to a left rudder deflection by establishing a very nice turn to the left without further action by the pilot.”

 

However, it turns out that the results of applying right rudder are quite different: “Next, encouraged by the good results obtained by the application of left rudder, we tried a right-rudder input. I truly expected the Camel to do exactly what it had done before, except to the right. The Camel immediately responded to the rudder by yawing to the right, but then it acted strangely. The bank was only partial, never exceeding 20 degrees, and the nose dove down instead of rising.” Mr. Arango then goes on to explain that this leads to an uncomfortable flight condition with not much bank but with a great yaw angle and a nose down attitude.

 

As an outside observer, it’s here interesting to note that while the Camel yaws more rapidly to the right than to the left, just as claimed by contemporary authors, its roll rate was quicker to the left than the right contradicting some historical statements that the Camel has trouble rolling to the left.

 

Next an observation on the effects on pitch: ”The biggest asymmetry, however, is in pitch. Turning the left, the Camel’s nose came up; turning to the right, it went downward. Modern airplanes do not do this and certainly not at the violent rates the Camel displayed.”

 

Finally, the overall effects of rudder input are summed up: “It is most interesting to examine the consequences of a rudder input five seconds into the turn. Turning to the left, the airplane is adequately banked and turning in the intended direction, and its nose has descended from the initial pitch-up back to level flight. This is a very benign and controlled turn. But to the right, the airplane is turning more quickly, yet without adequate bank angle. The result is very unnerving. An airplane yawing quickly without enough bank angle is tending toward an uncontrolled and dangerous spin, and when its nose points downward it is in danger of entering a spiral in which its speed could quickly and dangerously increase.“

 

On the subject of gyroscopic forces, there are a couple of interesting passages in the essay that I think are best conveyed word for word:

 

“While discussing with him [Peter Garrison] my own experience of flying the Sopwith Camel, I noted that most of the accounts of WWI pilots seemed to me very exaggerated. I thought the Camel was a challenging aircraft but one that behaved mostly normally. When Peter asked me about the effects of gyroscopic precession due to the rotating engine—a contentious topic blamed for numerous accidents in the Camel¬—I again told him that it was barely perceptible and could be mostly disregarded by the pilot.”

 

“The Camel was thus not a defective aircraft that was susceptible to uncontrollable gyroscopic precession. In the most extreme manoeuvres it may have become uncontrollable, but so would most airplanes because of many other variables. The gyroscopic precession force in the Camel was not in itself the ultimate cause of such strange behaviour and of so many accidents.”

 

Another note of interest related to gyroscopic forces is that Mr. Arango again highlights the problem of terminology and correctly interpreting pilot statements from this time period since it seems that when they said torque, many were actually referring to precession effects and not to what we today would refer to as torque, e.g. the engine torque that turns the propeller.

 

Mr. Arango goes on to say that the high accident rates probably had less to do with the Camel being extremely difficult to fly but rather more to do with the pilots simply not being familiar enough with the aircraft and that inexperienced pilots may have become task-saturated and used inappropriate elevator and rudder combinations when attempting to correct the supposedly deadly right hand turns. Whatever the actual causes of the high accident losses were, Mr. Arango is quite clear in the essay that the Camel is perfectly controllable as long as you use the correct control inputs.

 

The trials also debunked the myth that the Camel’s turn to the right and left differed so much that a pilot would prefer to do a 270 degree turn to the right in order to make a 90 degree left hand turn: The flight data recorded showed that when making steep turn to the left and right using appropriate control input, it took exactly the same amount of time to complete a full 360 degree turn in either direction, namely 16 seconds. However, in the same passage, Mr. Arango observes that the Camel was unstable in the turn and that quite different control combinations were needed but that he as a pilot thoroughly familiar with the airplane corrected these instinctively and that the substantial difference in control input depending on turn direction as witnessed in the recorded flight data surprised him.

 

Summing up, it seems that many of the quirks attributed to the Sopwith Camel from legend do not stand up to a closer scientific scrutiny and while the Camel certainly was a difficult aircraft to master, probably to a large extent due to the concentration of weight close to the C.G and the very low level of stability in pitch due to a very rearward center of gravity, the thorough testing performed by Mr. Arango proves that it was nonetheless an aircraft that was perfectly controllable given that the pilots who flew them had sufficient training and experience to use the right control combinations.

 

Edited May 23, 2021 by Holtzauge

[Cynic_Al]

Regarding time taken to make a 360 degree turn, in-game I say that a realistic-looking 360 turn that does not lose altitude, in either direction takes no more than 10 seconds.  I wouldn't want to start a debate as to the reasons for the discrepancy.

[No.23_Starling]

Holzauge, this is excellent. Thank you for sharing. I didn’t think you needed to caveat the post by saying it wasn’t a criticism and that you love the sim. The performance of these aircraft is a nebulous and extremely tough topic, but the best approach is always as scientific or as first hand as possible. The question for me is the point of such posts in terms of the sim’s development. FC (thus far) is a port of RoF assets with tweaks (rolling back some FMs such as the Camel, and adjusting wing strength here and there). If there have been dev responses to such suggestions I’ve not seen evidence beyond the aforementioned wings and the incoming AU engine for the Dvii and Dva. Interesting read nonetheless. If you have anything similar on the N28 I’d love to read it.

[Canvas25]

Wonderful stuff. Thank you very much for sharing!

[Todt_Von_Oben]

Holtzauge, to help qualify your essay are you a certificated airplane pilot or instructor?   

[US103_Baer]

Awesome post @Holtzauge

Though it was like reading the first chapter of a book that teased the rest!

Does he go on to explain and detail further control inputs required to maintain or tighten turns in both directions?

[Holtzauge]

Thanks for the positive feedback guys!

 

Concerning the turn times, in my correspondence with Peter Garrison (He was involved in the trials with Javier Arango) he said he was not sure which of the Camels (Javier Arango had two) was used in that particular trial, the Gnome or what he called the Le Rhone powered one (Clerget?) and at what density altitude the test was done. In addition, he said that he suspected that Mr. Arango was not flying close to the stall for safety reasons which I think explains that it took 16 s. This was also why I removed my theory that the turn trial was done with the Gnome at the half throttle power setting from the paper linked in the OP.

 

About the caveat: Given I’m mostly pointing out things that differ IRL and in the sim in my posts I thought it would be a good idea to add this. However, given  I’m an engineer by profession we tend to focus on problems and ignore what works which may come out as negative even though it’s not meant to be. About posting more simulation results: Maybe, but I think the reception to the data I presented on the Nieuport 28 was lukewarm at best.

 

When it comes to the controls to maintain a turn, in the essay Mr. Arango pointed out that he was surprised himself that left rudder was needed in a right hand turn and that the correct procedure in the Camel was to lift the nose to the horizon with rudder, not elevator when going to the right: His theory was that many of the Camel accidents was due to inexperienced pilots using elevator, not rudder to raise the nose in right hand turns and therefore getting horrendous yaw angles and crashing due to this.

[ZachariasX]

Thank you for posting this @Holtzauge, very interesting!!

 

I think one aspect of the sim is that it doesn't really blanket out controls, such as the rudder in the Camel that in the real world adds litte to directional stability unless assisted by working the rudder. This excessive "keel effect" may dampen a lot on nose down in right turn or nose up in left turn. The gyro is probably rendered exact, but the aircraft is too stable on its own.

Edited May 24, 2021 by ZachariasX

[unreasonable]

That complements his lecture available on YT very well, thanks.  In terms of your comparison with the game Camel:

 

1) Agree - rudder not only initiates bank well, but left rudder needed in both turns in FC. 

2) I suspect this is true, but can it be quantified? Given how much faster banking with the rudder is than with the ailerons in FC anyway, this may not be such a great advantage.   

3) IIRC at some point fairly recently adverse yaw was reduced in the global GB/FC flight model: I am it sure used to be more noticeable.

4) My impression from the essay is that he is saying that the rise/drop of the nose in the real plane is not due to gyroscopic precession alone, but also something to do with the lack of control harmony. I suspect this is harder to model for FC: precession physics is much simpler than the control effects?

 

The first time I tried some turns in the FC Camel after a couple of years with no WW1 stick time, I managed to get into a RH spin pretty quickly with the nose dropping fairly quickly. Forgot you need left rudder to keep the nose level!

 

Just now, once in a turn of ~3.5 g I measured about ~7 seconds for a 360 turn, close to the sea but not necessarily absolutely flat, the same each way, but I found the RH turn easier to stabilise.  (1/2 fuel, full power).  We do not know how many gs it took to get the 15 seconds. Generally, with my stick on 1/1 curves on each axis, I find the Camel not very easy to steer - and hence aim - precisely, unless you are already flying fairly straight. 

 

So my guess is the FC Camel is not wildly out, but adverse yaw and aileron response might well be favourable in FC.   

 

  

 

 

[Holtzauge]

For sure about the keel effect @ZachariasX: Mikael Carlson said there was basically no directional stability in yaw on the Dr.1 and if you gave it rudder and yaw in one direction it would stay there until you used opposite rudder to actively bring it back. In the essay Javier Arango also mentioned a lack of directional stability and said he produced artificial feedback on the rudder by pushing with both legs to keep it centered and then reducing pressure on one side to turn, i.e. basically what Mikael say's about the Dr.1: No directional stability at all. Also, you may well be right about this aerodynamic trait being part of the problem in the right hand turns in the Camel: In-game we don't have to think about keeping the rudder centered while IRL it seems it was easy to miss this.

 

Maybe @Chill31 has some input on this as well? Although I guess his Dr.1 behaves much the same since I can't see that there are any significant differences between his and Mikael's Dr.1 on this point so I expect they would have the same (lack of) directional stability? Was also interesting to hear about the gyroscopic issues on the top of the loop and the need for full rudder (or close to it as I understood it): What Mikael told me was something similar and what he apparently does is to lean the loop slightly to one side in order to control in which direction he gets the departure if you get to slow on the top.

[Chill31]

I really want to see his data!  The Dr.I and Camel have similar attributes when compared from a broad view, ie. Engine power, Weight, Size, % MAC.  I find his general description of the Camels handling very similar to the Dr.I. 

 

I do have a feeling that he was not really working the Camel hard in his testing though.  In his description of the left turn where the nose rises, but settles in nicely...this must be a gentle turn.  Certainly nothing you would do in the heat of battle unless you just like providing target practice for some German pilot.  When I fly the Dr.I gently, it is relatively easy to handle, and I would go so far as to say that the difference between a rotary powered airplane and any other almost imperceptible.  

 

However, when I maneuver the plane as fighter aircraft, the gyroscopic forces play a huge role in what I have to do to fly it well and keep it lethal.  If I took a pilot with 50 hrs time and put him in a Dr.I, I would say "Turn Right if you are attacked! Roll right, tap the right rudder, and pull for all you are worth."  A left turn with a high turn rate requires a lot more coordination.  If you start the turn AND let the nose rise initially, you will NEVER get it back down after you start pulling the stick back.  So you end up in a tight turn, nose high, with decaying speed, which really means you are doing your best impression of a strafing target.  

 

I have been flying airplanes since I was a kid, so it is about like walking for me.  I may even be better at flying than walking...punchline here is this: that amount of experience lets me turn right 180 degrees (with the 80 Rhone) in about 4 seconds, to the left, 5 seconds.  If I take a fresh private pilot and put him in the plane, I bet a left turn takes him double the time to perform versus turning right.  

 

People who fly replicas today and do not work the plane hard can only give you a small glimpse of what it is like to really fly one.  You might get a broad picture, but the real character of the plane comes out when it is under a heavy demand.  

 

 

4 hours ago, Holtzauge said:

Maybe @Chill31 has some input on this as well? Although I guess his Dr.1 behaves much the same since I can't see that there are any significant differences between his and Mikael's Dr.1 on this point so I expect they would have the same (lack of) directional stability? Was also interesting to hear about the gyroscopic issues on the top of the loop and the need for full rudder (or close to it as I understood it): What Mikael told me was something similar and what he apparently does is to lean the loop slightly to one side in order to control in which direction he gets the departure if you get to slow on the top.

Yes, the Camel sounds like the Dr.I when it comes to yaw stability and control feel.  In the Dr.I the first 30% or so of rudder movement is effortless and only muscle memory and the seat of your pants will let you know that the rudder is not centered.  After that frist 30%, the rudder starts to have negative feedback, meaning it wants to add more rudder input on its own!  The center of pressure moves forward on an airfoil as the angle of attack increases. The aerodynamic balance on the Dr.I is large enough that the center of pressure can cause you to have to push opposite rudder to bring it back to neutral. If you simply take your feet off the rudder bar, it will remain displaced, possibly fully deflected, all by itself.

 

The 110 Rhone requires even more rudder use through the loop than the 80 did, which isn't all that surprising when you consider the additional size and mass.

[Holtzauge]

@Chill31: For sure, the 16 s turns indicate that the Camel was not taken to the limits of its capabilities which was also something Garrison pointed put: The turns Arango did in the trials were more designed to test and compare the aircraft’s behavior in left and right hand turns, not take it to the limit. Earlier on when I asked Garrison about the Camel stall he said that he would check with Chuck Wentworth who flew both Arango’s Camels and Dr.1 meaning he could compare them on this point and according to Wentworth, the Camel stall is not as benign as the Dr.1’s which could go some way to explain why Arango was careful not to push the envelope which also seems prudent given that the number of remaining Camels is limited and that it apparently can be a handful post-stall.

 

In addition, it’s unclear what engine was used in the Camel trials and if it was the Gnome run at the 50% setting (like there are video shots of Arango doing), then a turn time of 16 s a few 1000 ft up is what one would predict so summing up, while there are uncertainties about the exact test condition like engine used, altitude and power setting, I think Arango’s  trial is still a valuable contribution to understanding these planes even if it did not take the Camel to the limits.

 

Another thing to note about Arango’s turn trials is that he separates the sustained part when established in the turn which is 16 s in either direction and the turn entry: In the entry he describes the Camel’s nose swing and drop as “violent” and talks about yaw angles of 30 degrees which I think at least says something about how he did the entries anyway.

 

Concerning the Camel’s actual turn capabilities at the limits there are RAE reports talking of 8-10 s for a 360 degree turn in a Camel so doing a 180 degree instantaneous turn under 4 s in a Camel sounds perfectly possible as well.