These notes are being put up for
those who may be following the forthcoming spin trials of the HJT 36 Sitara. There
is some Physics/Math but it is High school stuff and helps in understanding the
conclusions given. No claims are being made as to accuracy of prediction though
some of the points are interesting and indisputable.
A quick review of the Programme
The HJT 36 programme is another
of those Defence related mysteries. Here was a programme which showed, briefly,
what standards we are capable of. Excellently managed (Take a bow, Yogesh
Kumar!) it went from sanction to first flight in 3 years which is about good as
it gets, only to hit a sandbank when the engine was changed. Talk of changing
horses in midstream! Mind you, the idea was not bad. The French, seeing an
opportunity, reportedly wanted higher prices for their Larzac and as usual
there was some weight increase/specifications change. The Saturn AL-55I was
chosen but things then began to unravel. The original design and project
management leaders retired practically en masse. There was no system of
retaining them. The engine was delayed by two years over the scheduled delivery
and when installed it reportedly would flame out during spins and the TBO was
only 250 hrs. To add to the above delays the prototypes suffered two accidents
– in one the canopy came off and the Aeroplane skidded off the runway during
takeoff. The other –a proper crash- was due to some engine related problem. If
that was not enough I believe somehow an ejection seat fired during final
assembly damaging and delaying the prototype. I think HAL made a classic
mistake of following the specifications and did not allow enough “let” in the
design. As the great Sydney Camm used to say “Follow the specifications too
exactly and you are a “goner” most of the time”.
The present problems on the HJT 36
My sources- no more than newspapers
and what go on in the www. - indicate the following “problems”:
i. Engine flame out during spins
ii. Poor engine TBO.
iii. Wing drop
iv. Spin trials not yet carried out
Engine flame out and TBO.
Very little actual details are known but it is surprising to hear of a TBO of 250 hours. This is what was
normal with Russian engines some 60 years ago. I cannot believe Saturn cannot do better than that today. They have been in this business for over 60 years.
Regarding flame out it is true turbofans are more sensitive to non uniformity of flow across the compressor
face. The Y intake duct will have the “in-spin” inlet getting some useful side flows but the ‘out-spin” side will be sucking in a lot of vortices shed by the slab sided front fuselage which is not a help. Both aircraft have a similar
height of forward fuselage but I wonder if the Hawk’s lower inlet position is not better for taking in some undisturbed air from the sink velocity flows. Another point you might want to take note of is that the HJT 36’s inlet is bang in line with the wing L.E. Somehow most people try to keep the inlet lip a bit forward by at least a foot or so of the l.E. -Kiran, Hawk, Galeb, Aviojet et al. What might be happening is that in a proper spin with the AOA hitting 50-60 degrees the intake may be full of
eddies being shed by the L.E. Is there anything in that I wonder? The Aero L 39 has an exactly similar relative
location but the Czech aircraft has a much more slender forward fuselage of more oval cross section so the yaw related eddy would be much less. Of course putting the inlet forward may require a new design for a boarding ladder for
the rear cockpit but c’est facile! Scarf the intake lip by 45º may be another idea.
A longer inlet duct always helps to smoothen out the flow. Going by the 3 view available on the net the designer
has an estimated 5130 mm between the air inlet lip and the engine exhaust. Take away 1210mm (?) for the engine leaves
3920 which is about six inlet diameters for the Y and flow stabilizing portion. It would mean that the engine has to be mounted fairly far back to have a reasonable amount of stabilization of the flow. All this, from so far away, of course
is conjecture but I know some simple solutions can be thought of.
Wing drop
Wing drop at low airspeed is not
really a problem unless it is vicious i.e. the whole wing drops suddenly and
without any warning. The idea is to delay the stall at the tip which should
stall only after the wing root has stalled so the pilot gets a “sink” rather
than a wing drop. There are several standard cures for wing drop at the stall.
i) Keeping the tip chord at a lower incidence
than at the root so that the wing root stalls first. Called ‘washout”, it is
the aero modeller’s equivalent of the “one eighth packing under the T.E.”
during building. In fact with straight unswept L.E.s there is a certain amount
of aerodynamic “washout” due to tip
vortices and viscosity and thus special “twist’ at the tip is often not needed
in light planes.
ii) A change in the tip aerofoil. The designer
will sacrifice maximizing lift to choose an aerofoil which is happier at low
Reynolds numbers. The Australian Victa had a sharply tapered wing and the
designer Henry Millicer had a NACA 23012 at the root and a NACA 4412 at the
tip. Our well beloved Gnat had the RAE 108(?) at the root but a deeply cambered
thin section (RAE ??) that spoke of circulation theory if you even looked at
it. That “conical cambered L.E.” only partially cured the problem because at
low speed i.e. on the approach, the Gnat would rock gently from side to side as
she came in over the fence.
iii) Automatic slots on the tips like the Me109. But
one wants to be careful with that because if un damped, one of the slots may
open a little earlier than the other and give one a bit of jerk or spoil one
aim in a combat turn.
Seeing how close the IJT 36 is to the HJT 16 it is a wonder that the design team did not use the
earlier aircraft’s flying surfaces and saved themselves a lot of time, bother
and expense. As a bonus they could have used the old undercarriage also. Using
existing sub assemblies happens often enough in Aeronautics. The Beaufort used
the Blenheim’s wing and the Supermarine Attacker jet used the propeller driven
Spiteful’s wing and the American High manoeuvre test prototype used bits from a
lot of other aeroplanes.
Stall
The stall is the
first step of the spin. As is well known the flow breaks away causing loss of
lift. Expanding a little on this what we can say is that at low velocities the
flow lacks the momentum (you can use that lovely onomatopoeic Hindustani word “dum”,
if you like!) to overcome skin friction and begins to break away near the
trailing edge. It is only on the blackboard that the flow over the wing is
‘homogeneous”. In reality it has micro variations so the breakaway point will
shift nearer and further away from the trailing edge causing the often
described judder near a stall. Once the speed i.e. momentum of the stream is
well below the requirement the breakaway will start close to the leading edge
and you have a full developed stall. I will mention here Dr. Winter’s
Zaunkoenig , a little 35 hp Zundapp engine parasol monoplane of the ‘40s which
had a wing so slotted and slatted that it was apparently impossible to stall
the entire wing at any one time and it could almost hover in a strong wing and
be flown controllably at as low a speed as 35 mph. Of course one must realize
that at those low energy and momentum levels the margins are paper thin. Many
young gentlemen wrote themselves off exploring the low speed characteristics of
“harmless” slow flyers such as the
Westland Lysander and the Fieseler Storch.
Spin
A spin is NOT a spiral drive. It
occurs when one tries to turn at low airspeeds such as in a dog fight or
turning off the downwind leg and manage to stall the aeroplane. The following
things happen.
i) The aeroplane stalls and because of the turn the “inner” (to the turn) wing is flying at a lower speed than the ‘outer” wing tip. It will stall first, lose lift and go down there by pushing up the local angle
of attack thus stalling it properly.
ii) The outer wing will go up thus reducing its local angle of attack. It is also moving slightly faster the stall speed and
so is generating some lift.
iii) The aeroplane now has its two wing halves, one
generating slight lift (remember that the mean airspeed is near the stall) and the
other no lift. Lifts and drags are related so the two panels have different
drags thus yawing the aeroplane.
The aeroplane is at now at a slow forward speed, slowly (relatively) sinking due to loss of lift and slowly
rolling and yawing due to the asymmetric and weak lift and drag. Mind you though the wing has stalled the tail plane and the fin are still ‘flying” i.e. it is not stalled. Thank God for small mercies! The empennage do not stall as they are usually of lower aspect ratio than the wing and the tip vortices “wash” over more of the span keeping local angle of attack down below the
stall. However there is the danger that if – as in low sectional density aircraft e.g. light planes, the forward speed will decay rapidly below the “flying speed” of the tailplane.
At this point we can make a simple mathematical model to give ourselves some figures and then use an
analogue so that we can compare the figures and draw some conclusions that is
not all hunch.
The spin model is simple:
The forward speed VH is
equated as the stall speed which is calculated using standard formulae. This speed will decay slightly with time as the engine is throttled to idle and because of the cranked up attitude at the stall, the wing and the fuselage acts as a large air brake. The sink speed VV is obtained approximately by calculating the flat plate area of the plan view of the aeroplane. We can
calculate the increase and the terminal sink speed VV. It is possible to calculate the rate of increase of the VV with time and in combination with behavior of VH get the instantaneous flow vector of the aeroplane.
The final AOA w.r.t. to the airstream in a spin is usually in the range of 40-700
The yawing and rolling rates can be approximated by knowing the aerofoil characteristics of the aircraft concerned. They are very slow about 0.25-1 turn/sec. and is ignored in this exercise.
So now we can visualize the aircraft as if trapped in a bubble of sluggish, disorderly air, from which it must break out if it is to recover. It is now easy to understand why spins in light aeroplanes can be very disconcerting. The aerodynamic forces are high relative to the inertia forces so they spin quickly and also snap out of the spin rapidly. My favourite story is about the Bucker Jungmann biplane which
could be spun rapidly but once the correct procedure was followed it would snap out of it so promptly that it often left the tyro pilot “shattered”.! In light planes the low sectional density means it will lose way quickly and so one could get into a flat spin from which recovery is more difficult. BTW there were aircraft which could not be “spun”. The Erco Ercoupe was one –achieving that by the very limited control movements giving rise to the dreadful pun “No(vices)can’t spin!” and there was the “unspinable” Australian CAC Winjeel. (Young Eagle) which was embarrassing because it was the basic trainer of the RAAF!
Recovery from the spin
Once the pilot realizes he is in a spin and if he is not doing show aerobatics he will first put anti-spin
rudder. Apart from subtle and interesting points such as “freeing more of the rudder area” the opposed rudder removes two disconcerting motions i.e. the roll and the yaw thus allowing the presumably frightened pilot to reorient him with
the world. He would then push the stick firmly forward to “unstall”’ the aeroplane and recover normally. Just as prototypes have refused to come out of spin - the prototype Fairey Swordfish was one such – despite the pilot’s best efforts - many aeroplane will also often come out of a spin if its controls are kept at neutral. The aerodynamic damping will slowly remove the yaw and the roll and the natural stability will re assert itself but prompt recovery action is recommended to reduce the considerable height loss.
Some physics and maths
One would need proprietary software and huge computers to model and predict spin behavior-and even then
get them wrong! There is a much cheaper way of being “wrong” and that is by using fundamentals of Physics and some arithmetic to calculate the relevant parameters. If we can then set up an “analogue” - an aircraft whose configuration and spinning characteristics are ‘known” we can make a reasonable prediction of the new aircraft’s spin behavior or some kind of a figure of
merit as a comparison. Here goes!
We need to know:
i) The pitch inertia PI. In this class of aircraft
it is approximated by 0.7 ML2. Why 0.7? Because the fuselage is taken to have
0.7 of the mass M and the wings plays no big role. The M is the clean half fuel
weight.
ii) The roll (RI) and yaw (YI) inertia. 0.7 ML2 +
0.3M. WS2. L is the length and WS the span.
iii) The tail volume TV which is the area of the tail plane and the moment arm of its centroid with the Cg. This will give the
comparative rate at which nose will be pushed down given the pitch inertia.
iv) The Fin volume FV. As above
v) This is the horizontal component of velocity. It will start as the stall velocity and decay with time. The time decay can be approximated by assuming a Cdo.
vi) Vv. The vertical velocity again can be calculated as indicated earlier.
vii) The tail angle θ is the angle that the leading edge of the tail plane makes with the wing trailing and leading edges. The tail plane must be clear of the wing wake to be able to unstall the aeroplane.
viii) The Fin coverage Fθ is the amount of the fin that is covered by the tail plane during the spin. The yaw inertia
approximation divided by the blanket factor will give the yaw correction time comparison.
ix) In all cases the height is taken as 2000 mts because the air density ISA is 1.00 and it is the minimum height you would want
to spin this class of aeroplane –when in service.
x) The fuselage side area moment which will indicate the damping available for opposing the yaw?
We will now select some aeroplanes which are in service and compare the factors to see where the IJT 36
stands. Our choices are as follows.
i. The HJT 16. The side by side sitting is an obvious disparity
ii. The Aero L29 which is remarkably similar configurationally.
iii. The Hawk because it is here even though the slight wing sweep adds to stability.
Finally I decided to compare with the Hawk because despite the slight swept wing it is one whose spin characteristics
we know. The Aero L 39 Alabatros was actually the better choice but I could not get all the details I needed to compare.
So the results:
Assuming that the
empennages are equally efficiently exposed to the resultant airflow,
then:
i. Compared to the Hawk the HJT 36 will lose less height during the spin at approximately 60% of the value of the Hawk’s and
pushing my luck the IJT 36 will stabilize in a spin in about 4 second compared to the Hawk’s 7 – both times should be treated as comparative rather than
actual.
ii. The rate of yaw will be slower at 60% to that of the Hawk due to greater damping.
iii. On application of opposite spin rudder the HJT 36 will recover 15% faster than the Hawk.
iv. On application of forward stick the HJT 36 will be in the correct attitude to recover by a time 30% less than the Hawk.
Whilst the figures are derived from a very simplified model they seem to be sensible. After all the HJT36 is a
smaller lighter aircraft than the Hawk. By the above approximation the figures are excellent because the young gentlemen can have a foretaste of milder jet spin of the IJT before going on to the harder stuff of the AJT.
Uh! Oh!
There is however one little fly
in the Ointment. It always struck me that the HJT 36 was “close coupled’ as
control line model flyers used to call the Peace Maker or Mercury Matador
designs. This close coupling CAN affects the airflow over the tail plane and
fin at the stall. If that is so the above results - where I assumed that fin and
tail plane effectiveness is same in the analogue as well as the specimen
aircraft - may not be obtained despite the potential to do so being definitely
there.
The desirable situation is that
when the aircraft stalls the tail plane is clear of the wing wake and the fin
is clear of the wake of the tail plane. It is then that the recovery actions
will be most effective. I calculated the position of the wing wake at the stall
and it appears that the HJT 36 stabilizer is deep within disturbed air shed by
the L.E and T.E. of the wing (tail angle 27 degrees minus stall angle 14 degrees equals 13 degrees on the wake whereas in the HJT 16 it is clear and in the Hawk it is 2 degrees in the wake). As a result, the controls of the HJT 16 will not be effective for the first 3-4 secs after the aircraft enters the spin during which the aeroplane may lose about 100 meters in height. This can critical in case of a spin at low altitude. With the Hawk or the Kiran, the tail moves down
and out of the wing wake. The sketch at the end of this note makes things
clear. Actually in the Hawk the tail stays within the wake at the root - giving
good stall warning - but because of the anhedral most of it is below the
disturbed wake and therefore effective. Wicked thought! Did BAC make the same
mistake as HAL and corrected it with the anhedral or it was a case of force
majeure - there was no place else to put the tail? Old Sir Sydney (Camm) would
never have allowed that even in his sleep!
The table below is interesting as
it compares the angle the Tail L.E. makes with the Wing’s L.E. and T.E. for
comparable types.
Type
|
Angle between Wing L.E. and Tail plane L.E
|
Do for Wing T.E.
|
Aero L39
|
9.5º
|
18º
|
Hawk
|
9.0º
|
16º
|
HJT 16
|
7.125º
|
13.4º
|
HJT 36
|
15º
|
27º
|
“Bobbery”? (Hobson Jobson- derived from Hindustani ‘Baap re’ and
indicating consternation!)
There is really no need to panic,
even if the above prognosis is true. A prototype crashing during spin trials is
yawningly boring. The HT-2 had spin related crash. The French Epsilon had a
spin related crash but, after Gallic shrugs no doubt, they simply sawed off the
old tail and fitted a completely new fin which was as effective as it was
elegant! BTW the location of the HJT 36 fin w.r.t. the tail plane, is pretty
good.
The above model of the wing wake
is true at the stall when the nominal angle of attack is circa 14 degrees. As
the aircraft begins to sink faster with time this wake will move
counterclockwise and the tail and fin will move into clear air after 3-4
seconds. The
total height required to recover from a spin may still be less because
inherently the HJT 36 should recover faster than the Hawk.
The real problem is human -psychological
and cultural. We have great faith in rigourous analysis and so when things go
wrong there is shock induced paralysis. Much time, I have seen, is spent in
fixing blame rather than fixing the problem! In the “feel” based approach,
which is sometimes, treated as a disability with sophistication, however, the
Designer is not only acutely conscious of possible failures but also he has
subconsciously prepared several “what if “ scenarios so the nettle is grasped
firmly and quickly.
Solution
What I have said is not a judgment
but rather a starting focus. What needs to be done now is perhaps to go down to
the flight Hangar and “talk” to the aeroplane, noting the subtleties of its
contour and using the rough figures of velocities to imagine how the air would
sluggishly flow around the nose, the intakes, the wings and from the bottom of
the rear fuselage upward towards the fin and tail. One would also imagine how a
strake or under fin would behave and on which longeron and frame the strake
could be riveted etc. This is to be
followed up using CFD ‘snapshot” assuming the proprietary spin software is not
available.
If major corrections are indeed
required then HAL Design team may have to accept to make a three four small
changes rather than try and cure by a one-step major change solution. The
former will be faster!
