Written 2000

The Romanian aircraft manufacturer Avioane Craiova SA was established in 1972 to develop and build aircraft for the Romanian Air Force. Its position near the town of Craiova is in the south west of the country. Their first aircraft, the IAR-93 YUROM, was a joint venture with the Yugoslavians. This was initially a single-seat close air support, ground attack and reconnaissance aircraft. It first flew on the 31 October 1974. The experience gained from this aircraft, and its development over the years, enabled Avioane Craiova to produce the first Romanian designed, Romanian built aircraft: the IAR-99. IAR stands for Industry Aeronautical Romania. The first IAR-99 was developed for the advanced trainer and ground attack roles and is now known as the 'Standard'. It first flew on the 21 December 1985. The aircraft was ideally suited to the Romanian Air Force requirements at the time. However, with the MiG 21 'LANCER' upgrade it quickly became obvious that a new '99' (military trainer) was required. Similarly, with the Company's new-found freedom other markets were now open to them. It was with this in mind that the IAR SOIM (Little Hawk) was conceived. The initial aim was to develop a new trainer to meet their home Air Force's Operational Requirements but it quickly became obvious that the aircraft would be suitable for the old Eastern and Western customers alike. It was further decided to make the aircraft suitable for the whole training environment (from basic to advanced training) as well as the Close Air Support (CAS) role.

As part of an AFM visit to the factory I performed a three-flight evaluation of the prototype SOIM aircraft. All the flying activities were performed at the Romanian MoD Flight Test Centre, which was based at Avioana Craiova. The flights included general handling, low level flying (70-100m above the ground), systems functioning and formation flying. However, Avioane Craiova wanted to emphasize that this was the development SOIM and that small changes to the aircraft were already planned from the initial flight testing results.


The decision to produce a new trainer was made in late 1996. The aerodynamics of the IAR-99 Standard were considered suitable enough for a Lead-In-Trainer (LIT)/CAS aircraft and as such is was clear that the new aircraft needed the majority of the development work spent on its cockpit to bring it up to the modern standards. A two-stage programme was developed to allow the first Certification process to produce the LIT version, followed by a second stage to certify the CAS aircraft. It was decided that the aircraft should be able to carry both Western and Eastern air-to-ground and air-to-air weapons, as well as other 'smart' pods for the various attack options and electronic counter measures (ECM).

The avionics upgrading process began in 1990 with two aircraft (Sn 708 and 709) being equipped with Honeywell avionics. The first flights of these aircraft were performed on the 9 August 1990 and the 22 August 1990 respectively. A further aircraft (712) was fitted with Collins equipment and made its first flight on 6 November 1991. However, it was not until the IAR 109 SWIFT that the avionics integration started to progress rapidly. This aircraft was meant as an 'all-through' trainer (basic and advanced) with modern avionics and weapon delivery systems. The SWIFT was a joint venture with Israeli Aircraft Industries (IAI) and included Bendix-King avionics as well as Israeli systems. The SWIFT (7003) made its debut flight on the 2 December 1993. It was with the knowledge from all of these development aircraft that the IAR SOIM (718) was produced leading to its first flight on the 22 May 1997. To date only one SOIM has been produced, with the next three aircraft due to come off the production line early next year.


The aircraft was designed to meet modern European standards (MIL-8785B), with the main aim of being simple and easy to maintain at an affordable cost. Its basic construction was of aluminium alloy with fibreglass wing tips, tail fin tip cover and some doors and covers. The aircraft was fitted with a Rolls-Royce Viper Mk 632-41M engine built under Licence by Turbomecanica based in Bucharest. This engine was chosen because it has been proven to be very reliable, easy to re-light and has an engine life of at least 4000 hours. It was also relatively cheap, which was important during the selection in order to keep the overall aircraft costs down. The engine produces 4,000 pounds of thrust at sea level (17.8 KN). The wing was designed as a straight taper, fitted with two internal structurally integrated fuel tanks per side. There were a further two fuel tanks fitted in the fuselage, giving a total capacity of 1,370 litres (2,425 pounds/1,102 kilograms) internally and with two external tanks 1,820 litres (3,225 pounds/1,466 kilograms). On the prototype aircraft refuelling was by gravity (similar to a car refuelling nozzle) rather than a pressure system as is normal on equivalent Western trainers. I was told that this was deliberate because many of the developing markets might prefer this capability. However, pressure refuelling (as fitted on the IAR-93) will be offered as an option. The hydraulic system powered the undercarriage, airbrakes, flaps, main wheel brakes and ailerons; the ailerons also had a mechanical backup. An anti-skid system was fitted to the wheel brakes. The prototype SOIM was fitted with a Martin Baker Mk-10L ejection seat but it is planned to offer the Romanian equivalent ejection seat, made by Aerofina, as a customer option. At present the seat ejects through the canopy but the testing of miniature detonating chord (to shatter the canopy first) has been completed and will be a further option.


The cockpit was very modern and had been put together well with some novel features. A Modular Multi Role Computer (MMRC), designed and produced by Elbit Systems of Israel, controlled the whole avionics suit. This computer acted as the heart of the system running through a MIL-STD-1553B Data Bus. The front cockpit was fitted with a Flight Vision Systems' Head Up Display (HUD), a Multi Function Colour Display (MFCD) on the left head down and a Multi Function (monochrome) Display (MFD) to the right head down (both Elbit). There was no HUD in the rear cockpit but the HUD symbology was reproduced on a dedicated MFD, the ASHM (Aft Station HUD Monitor), mounted in the upper centre position of the front panel.  If needed the symbology could be swapped on to the right Head Down Display (HDD). Beneath the HUD was an Up Front Control (UFC) panel used to interface with the avionics system, including changing the radio frequencies, transponder codes, navigation functions and so on. The navigation system was called a 'hybrid' system because it combined a relatively cheap Litton inertial platform with a GPS to gain accuracy. Other than the cost, the advantage of such a system is the 'alignment' time, which was just 30 seconds. The cockpit could be 'monitored' by a high-resolution colour camera system, which could film the HUD and one of the HDDs. (A useful pilot debriefing tool at the end of a sortie.) Mission planning could be performed back at the squadron with the mission details being loaded by a data transfer system in the rear cockpit.

Weapon Systems

The SOIM had five external stations for carrying ordnance and/or fuel tanks: two on either wing and one under the fuselage. The fuselage station could be fitted with a semi-conformal gun-pod, fitted with a fast firing (3,000 rounds per minute) twin-barrel, 23 millimetre cannon or a 'smart' pod (LASER/FLIR, ECM, Photo-reconnaissance). The four pylons under the wings were capable of carrying a large variety of both former Eastern and Western weapons (up to 660 pounds/300 kilograms per pylon), including bombs, rockets, air-to-air missiles and smart bombs and pods. However, at the time of the assessment only the testing of Eastern ordnance had been completed, with the Western weapon testing due to be completed by the end of 2000. The aircraft was fitted with an Elbit stores management system, which was controlled by the MMRC. Avioane Craiova made the pylon release units. The payload configuration could be displayed pictorially on the MFCD. The cockpit employed the Hands-On-Throttle-And-Stick (HOTAS) concept allowing most of the 'fighting' functions to be performed by switches on these two controls. But to aid weapon management the aircraft was fitted with a Display And Sighting Helmet (DASH), designed by Elbit. In essence the technology displayed weapon sighting and navigation information on to the visor of the helmet. A helmet tracker mounted on the canopy followed the head movement in order to display the correct information correlated with the outside world. This potentially was a very impressive system that was going through its development and certifying stages.

On the self-defence side the aircraft was fitted with an Elbit chaff and flares system (auto or manual dispensing) as well as an all-round radar warning receiver produced by Elisra, Israel. However, perhaps the most surprising use of the computing power on the aircraft was a Virtual Radar (VR). This was a computer generated radar display (via the MMRC) designed to give the trainee pilot some experience in radar techniques and switching. The VR had two modes: the first, was that up to four SOIMs could be 'joined up' via data link to produce a pseudo radar picture by the computer interpreting the other aircrafts' GPS co-ordinates and then displaying their relative positions on the radar screen. Giving the impression that they were being identified by radar. The trainee pilot could then manipulate standard radar controls in the cockpit to become familiar with the techniques. The second VR mode was even simpler in that the MMRC internally generated targets on the radar screen for the trainee to encounter. Although, only five such scenarios are planned it would give the pilot a degree of radar familiarity. Unfortunately, as there is only one SOIM at present the multiple aircraft mode of the VR could not be tested.

Emergency Training

Following the theme of 'if you have computing power you might as well use it' the aircraft was fitted with a Fault Simulation Panel (FSP). This enabled the rear cockpit pilot/instructor to 'inject' emergencies/faults into the front cockpit. For example, engine fire, generator failure, oxygen problems and so on (10 faults in all). In theory this was a good idea but procedurally its use would have to be carefully considered. For instance you would not want the student to close the engine down, while reacting to a simulated fire emergency! However, all of the engine switches could be easily guarded from the rear cockpit to prevent this from happening but it would need careful thought. The other aspect of the system that may need to be reconsidered is that at present if a fault is simulated there is no way that the front-seat pilot can confirm the fault validity without the aid of the rear occupant. This may be a problem if non-aircrew passengers are carried in the back seat. Some form of duplicate switch in the front cockpit to 'arm' the system seems sensible and might aid general system integrity.


The preparation for the flight was relatively easy with a good set of documentation in English. The flying helmet was of American design and the strapping-in procedure was standard, as it was a Martin Baker seat. I then spent an hour or so sitting in the front cockpit of the aircraft in the hangar with electrical power connected. The Chief Test Pilot, Cristian Muscalagiu, gave me a good briefing on the aircraft, including the starting procedure. Because the aircraft was fully powered it meant that I could 'play' with the systems before the first flight, which is the best pre-flight training you can have. By the end of the session, I felt at home with the systems. It also served to highlight just what a simple aircraft this was - ideal for a 'first-jet' trainer.

At the aircraft I followed Cristian on his walk-round with very little to mention other than the odd accumulator pressure to check in the wheel well. Once in the aircraft it was very comfortable and quite roomy. As I had performed all of the checks the previous day in the hangar the pre-start time was minimal. In fact other than switching the avionics and radios on to ask for start, and then off again, there were only five switches (and ensuring the main fuel cock was open) that needed to be selected before the start button was pressed. Then at 10 per cent rpm, and with an oil pressure indication, the HP fuel lever was moved forward to on. At this point light up occurred, with the start taking about 25 seconds to complete. The engine was not fitted with any form of digital fuel computer (FADEC) to control the engine, but even opening the HP fuel lever as quickly as possible I could not cause a hot start. Again, relatively cheap and simple. Once the start had finished it was then a case of putting all the switches forward and setting up the avionics. Firstly, the aircraft's position co-ordinates were checked (either on the HUD or one of the HDDs) before switching the navigation system to align. It was an absolute joy when the system really did take 30 seconds to be ready to go. The other avionics switching was relatively simple through the menu driven displays. On the first trip it took 7 minutes and 28 seconds after engine start to be ready to taxy, but by the third trip it was down to just over 4 minutes!

Taxying the aircraft was relatively simple using the differential braking system; ie, pushing the brake pedal on the side that you wanted to turn to get the nosewheel to castor in the desired direction. However, it was noticeable that the undercarriage oleos (suspension) were very spongy. This was not too much of a problem at the start of the sortie or even during the take-off. However, it was a major cause for concern on the landings, where the aircraft had a tendency to roll from side to side quite harshly during the landing roll-out, particularly under heavy braking - something I would not expect a student to have to cope with! I later found out that this was because the oleo struts incorporated only one damping cylinder instead of two, but the design work had already be done to fit the dual cylinder version. These would be standard on the production aircraft.


Pre-take-off checks were minimal and basically required a 'left-to-right' check of the switches and that take-off flap had been selected. Even the emergency briefing was easy because the ejector seat had a zero height, zero speed capability. Once lined up the technique was to apply full power against the brakes to check the engine parameters (101.5 per cent rpm, 700 degrees jet pipe temperature for my flights), before releasing the brakes. The acceleration was reasonable with the nosewheel being raised at 165 kph (89 knots), resulting in the aircraft lifting off the ground at 210 kph (114 knots) after 20.5 seconds. The undercarriage was selected up when clear of the ground; however, it was recommended not to raise the flaps until the speed and height had reached 250 kph (135 knots) and 100 m (333 ft) respectively. This was because there was a slight tendency to sink as the flaps came in. Once airborne the aircraft was stable with reasonably light aileron control forces. However, the pitch control forces were high and needed active trimming to keep them light and manageable. The stick-top trimmer (for pitch and aileron force reduction) was very powerful and quick acting, which helped the situation. The Company did inform me that they were planning to review and hopefully reduce the pitch manœuvring forces.

General Performance

The first check of performance was to time a level acceleration from 250 kph (135 knots) to 735 kph (400 knots). This was a particularly interesting test because the aircraft had a relatively small amount of engine thrust matched with a light take-off weight (approximately 9,700 pounds/4,410 kg for my flight). The acceleration took 1 minute 28 seconds, which was reasonable for a jet trainer. By comparison the Aerovodochody L159 took approximately 1 minute for the same test but it does pack a large engine producing over 55 per cent more thrust. The next test on my performance list was a timed climb from 1,600 m (5,250 ft) to 3,100 m (10,170 ft), at 350 kph (190 knots). This took 1 minute and 17 seconds, or to put it another way the aircraft was climbing at 3700 feet per minute at a mean altitude of 7700 feet; again, not sparkling but satisfactory for this class of aircraft, albeit in the clean configuration. As far as the aircraft's rate of roll was concerned this was approaching fighter standards at 102 degrees per second at 350 kph (190 knots) and 140 degrees per second at 550 kph (300 knots).

As discussed earlier the forces required to pitch the aircraft were high, needing approximately 35-40 pounds of force to pull 4 'g' (without trimming). Another area that was interesting on the aircraft was the directional stability, which was quite weak. This caused the prototype to 'snake' directionally (left to right twisting movement, around the 'y' axis) whenever the aircraft was disturbed, either using the aircraft's controls or by turbulence. As a trainer aircraft this would not be too detrimental because the techniques are the important learning points not necessarily accuracy. However, when it comes to weaponry a stable platform is vitally important. The Company informed me that ventral strakes (to be fitted under the rear fuselage) had already been designed and would be fitted as standard on production aircraft to eradicate the problem.

The gear and flap up stalling qualities of the aircraft were benign with a gentle nose drop at approximately 180 kph (97 knots); with gear and full flap down the stall occurred at 145 kph (79 knots) with a marked pitch nodding motion. I then tried a few aerobatics (at a height of 2,500 metres, 8,200 feet) to see how much height was required for each manœuvre and to experience the general flying qualities of the aircraft. The loop was performed at 500 kph (271 knots) and, at 4 g, took 1,050 m (3,450 ft) to complete. The 'Split S', which is a combat manœuvre involving rolling inverted at 250 kph (135 knots) and then completing the second half of a loop by pulling through to regain level flight, took 900 m (2,950 ft) to complete, again at 4 g. Unfortunately, due to air traffic height restrictions, only the aircraft's incipient spinning (recovering after one turn) characteristics could be assessed. Using either the technique of releasing the controls or the normal spin recovery technique (full opposite rudder followed by forward stick) the aircraft recovered quickly without cause for concern.

Emergency Landings

Simulated emergency landings were practised from various positions, heights and speeds. They resulted in good judgement exercises for the pilot but were not representative of a modern fighter aircraft. This was due to the low drag of the aircraft even when the gear and flaps were selected down. The general technique in the vicinity of an airfield was to select the airbrakes out (to simulate a damaged, shutdown engine) and then to put the gear down to increase the drag. The flaps had to be selected quite early on in the procedure to get the drag up to an acceptable level. This prevented the usual technique (like in the BAe Hawk) of staying high and selecting the flap down when you were absolutely certain of making the runway. Cristian informed me that they were considering changing the landing flap setting, which would increase maximum flap from 40 degrees down to 60 degrees down. This would allow the desired increase in drag when required and would mean, in a real emergency, that the pilot could conserve height for longer knowing that he could get down easily when a runway landing was assured.  This modification should make the aircraft's gliding qualities more representative of a modern fighter.


As with any fighter pilot the most interesting stuff is when you start operating the aircraft in its intended environment. For the advanced trainer and CAS role this means low level flight, formation and weaponeering. Unfortunately, no ordnance was available to be either carried on the aircraft or dropped, which restricted the assessment to the handling of a clean aircraft in the tactical environment and the symbology aspects of weapons release.

Low Level Flight

Navigationally the aircraft was well suited to flying at low altitudes because of good steering information, backed up with GPS accuracy. The guidance was displayed in the HUD as well as on the HDDs. The HDDs also gave good situational awareness by displaying a 'gods-eye' view of the route as well as marking restricted zones and other airspace information - not as good as a moving map but a lower cost alternative. The steering waypoints (places to fly to) could be either inputted manually into the navigation system or be prepared back in the Squadron and then loaded into the aircraft using the Data Transfer capability. Guidance around the route could be set to allow manual sequencing of the waypoints or this function could be set to automatic; in automatic mode the computer changed to the next steering point as each waypoint was over-flown.

The ride quality was typical of an aircraft with low wing loading (small mass supported by a large wing area), reacting to the turbulence and giving a bumpy ride. However, even over some quite big hills the ride was not uncomfortable at the suggested low-level speed of 450-500 kph (250-300 knots). This low-level speed was quite low for an advanced trainer or CAS aircraft, although the Company did inform me that low-level could be flown at speeds as high as 650 kph (350 knots).  This is closer to a fighter low-level speed of around 775 kph (420 knots). It was easy to maintain accurate height above the ground using the radio altimeter readout in the HUD. Lookout was generally good with only the forward canopy frame restricting the view. On this prototype aircraft the slightly reduced directional stability meant that the aircraft's nose tended to oscillate left and right by up to +/- one degree. This had little effect on the overall training value of the low-level flying.

Weapons Capability

The aircraft comes equipped with the majority of the modern weapon delivery modes. For dropping bombs the three main modes were CCIP, CCRP and Dive Toss. For CCIP (continuously computed impact point) the aircraft's computer is continuously looking at the aircraft's height, speed, weapon trajectory characteristics and other pertinent parameters to try to work out where a bomb would fall if released immediately. The symbology displayed on the HUD illustrated this position to the pilot by displaying a bomb 'fall' line and a circle with a point in the middle to define the impact point. This type of delivery is the simplest to use and requires the pilot to 'fly' the bomb fall line through the target and drop the bomb as the circle passes over it. However, CCIP bombing does need the pilot to be able to see the target. If the target is obscured but the location is known the GPS co-ordinates can be entered into the system. Now the pilot simply flies the bomb fall line through the predicted target position (marked by a box in the HUD) and 'commits' to dropping the bomb by pressing the pickle button. The computer then looks at all the same parameters used in CCIP mode but it does not release the weapon until it considers the target will be hit. This is called CCRP mode (computer calculated release point). A combination of the two is known as Dive Toss. Here the pilot has a square in the HUD that he can position over the target and press a button to designate the target under the square. Once the designation button had been pressed the square was ground stabilized over the spot. Small corrections can be made using a controller on the throttle. Once the pilot is happy that the target is accurately designated he commits in the same way as CCRP. Now he is able to continue the dive towards the target waiting for the bomb release or he can perform a pull-up manoeuvre to make the bomb come off earlier (by literally tossing the bomb at the target). This allows a quicker turn away from the target area, which may be heavily defended. All of these modes were well produced and for training exposure were excellent.

The air-to-ground gun mode was similar to the CCIP bombs mode in that a continuously computed impact point for the bullets was displayed in the HUD. All that was required was to accurately track the target with the sight and open fire. The aircraft was not fitted with a laser ranging capability, although it was a customer option. For air-to-air gunnery the same principle was used but a 'Snap Hotline' was added, that ran through the targeting circle to indicate the fall line of the bullets. This line varied its shape considerably as the SOIM was manoeuvred, indicating the dynamic nature of air-to-air gunnery. Trying to follow a second aircraft, which was trying to manoeuvre away, was very demanding and generally resulted in 'snap' shots rather than continuous tracking.

For air-to-air missile combat the main sighting capability would come from the helmet display system (DASH). This form of targeting is ideal, as it simply requires the pilot to look in the direction of the target and place the designating box over it. When the correct lock has been achieved the missile can be released. Unfortunately, the system was not ready for outside assessment but it is hoped that it will be included next time. Radar missile handling could only be simulated using the computer generated Virtual Radar mode. This was reasonable training exposure but had no combat significance.

Overall, the weapon training value of the aircraft was good but it was not possible to determine the accuracy level of the equipment without releasing some actual weapons. In the CAS role the small directional disturbances would effect absolute accuracy and make hitting the target more difficult. The addition of the new ventral stakes should eradicate the problem and will be the subject of the next article!


During the formation assessment the aircraft was relatively easy to fly (with active pitch trimming!) and the engine response was good. Up-and-away manœuvring by a second aircraft was easy to follow and would be a good introduction to formation flying for the ab initio pilot. During the formation take-off and recovery to land the aircraft trajectory variations, with the selection of the services (gear, flap, etc), were all manageable and the engine response was quick enough to cope.


Avioane Craiova SA emphasized that the aircraft flown during the assessment was a prototype and that it was still under development. As such, they had some modifications already planned to address the majority of the concerns raised during flight test. It is reasonable to say that at present the aircraft is not as sophisticated as the latest British Hawk aircraft or even the Czechoslovakian L159; however, it is closing the gap and, according to Avioane Craiova, they are not after taking customers away from these aircraft (at least not yet!). They believe that they have an aircraft that is good value for money and therefore might appeal more to the developing markets, where some nations cannot afford the top level of sophistication. From what I have seen they may be right. As an entry level Lead In Trainer the IAR SOIM is fitted with a sufficiently complex suit of avionics and weapon delivery systems to make it effective. As far as the CAS role is concerned the flying qualities will need some improvement (as planned) and some weapon system and accuracy testing will be needed to ensure that the 'green' writing in the HUD is producing the required results when real weapons are dropped. At the end of the day you get what you pay for but in this very competitive market you can get a lot more for your buck than you used to - this aircraft proves that.


Length  36.11 feet (11 metres)

Height  12.79 feet (3.87 metres)

Wing Span  32.31 feet (9.85 metres)

Empty Weight  7,055 pounds (3,207 kilograms)

Take-off Weight Clean  9,680 pounds (4,400 kilograms)

Maximum Take-off Weight  12,258 pounds (5,572 kilograms)

Maximum Load Factor  +7 to -3.6 g

Maximum Speed  510 knots (940 kph)

Maximum Level Speed  467 knots (860 kph)

Maximum Climb Rate At Sea Level  6,900 feet per minute (35 m/s)

Take-off Distance (Trainer)  1,500 feet (457 metres)

Take-off Distance (CAS)  3,150 feet (960 metres)

Landing Distance (Trainer)  1,805 feet (550 metres)

Landing Distance (CAS)  1,970 feet (600 metres)

Endurance (Without Tanks)  2 hours, 40 minutes

Max Range (With Tanks)  595 nm (1,100 km)

Typical Cruise Speed  Mach 0.6 @ 33,000 feet

Sustained Turn Rate  20 deg/ sec @ Mach 0.42, 3,000 feet

Service Ceiling  42,322 feet (12,903 metres)

Pylon Carrying Capability  550 pounds (250 kilograms) per pylon