Engine Management

This article contains some very general tips on management of the larger Lycoming or Continental engines, particularly in high altitude flights, in normally aspirated (not turbocharged) aircraft.

It is generally assumed that the aircraft has an engine monitor, indicating the EGTs and CHT for at least one cylinder. This is the popular EDM700 which indicates all of them:

Also, smooth engine operation at peak EGT or LOP may not be possible unless the engine is fuel injected and even then it may need the injectors custom matched to the engine (e.g. GAMI).

Engine management becomes particularly relevant at higher altitudes. In the European IFR context, the minimum level acceptable to Eurocontrol tends to be around FL070 and practical weather avoidance strategies can push this to FL200. Most non-turbocharged IFR flight is therefore done with a wide open throttle, often either at a low RPM (for best range) or at maximum RPM (for the highest altitudes) and mixture control becomes very important.

There are two main aspects to engine management: fuel (mixture) management and thermal management. They are closely connected but are easier to describe separately:

 

Mixture Management

Any petrol/gasoline-fuelled piston engine delivers its best fuel efficiency at a specific air to fuel mixture which by mass is approximately 15:1. This is called stochiometric combustion, and it closely corresponds to the highest exhaust gas temperature.

So, why doesn't every aircraft engine set this ratio automatically and thus eliminate the mixture lever? There are several reasons...

One is undoubtedly tradition and extreme conservatism in the GA marketplace. This is far from misplaced, given the historically poor reliability of GA avionics. Having a DME fail is inconvenient but an engine failure would be much more serious. The traditional carburettor or fuel injection servo is a simple mechanical device which is highly reliable. Even the most recent attempts at FADEC (full authority digital engine control) on diesel engines have proved unreliable. So, we still have the magneto system with a fixed spark timing, when electronic ignition has been around for decades, is much simpler than full FADEC, would be an easy retrofit onto a magneto drive, but it does need a separate engine-driven power source (when your car battery goes flat, the engine stops). It can be done but the market will take a lot of convincing...

Aircraft engines need to be as light as possible, which is most cases has resulted in a rather minimalist air cooled design which is unable to deliver its full rated power when running stochiometric; the engine would simply overheat. So, when full power is required (usually during takeoff and climb), the over-rich mixture modifies the combustion temperature/pressure profile to produce less heat within the engine. The price is a big waste of fuel - around 30%.

A final and quite sensible reason is that an aircraft engine spends most of its time at cruise, which is a more or less constant power setting, so once the mixture is appropriately set, it can be left there. Contrary to popular belief, a 1950s avgas aircraft engine running at such a constant power is no less efficient than even the most modern petrol car engine. The big difference is that a car engine spends most of its time running at a low power but has to deliver a reasonable fuel economy and low emissions over a wide range of power outputs, and this is where most of the development of electronic engine controls has gone. An aircraft engine manually set up for stochiometric combustion, with its fixed ignition timing optimised for cruise, is running at the same parameters as if it was controlled by FADEC.

In traditional PPL training, the subject of "mixture" is largely avoided and most flight is done with the red mixture lever all the way forward, with various "if you touch this you will cook the engine" dark warnings from instructors. This approach delivers a cool combustion process, well suited to the erratic power changes typical in the training environment, but results in a huge waste of fuel, a correspondingly reduced aircraft range, a much reduced operating ceiling, and a higher risk of clogging up spark plugs. Fortunately, pilots wishing to move forward have found plenty of material on the internet, with a lot of it written by the famous engine management guru John Deakin.

The mixture setting is normally discussed in terms of where the exhaust gas temperature (EGT) is set to. The standard reference point is "peak EGT"; anything richer is called "rich of peak" (ROP) and anything leaner is called "lean of peak" (LOP).

For any petrol/gasoline-fuelled piston engine, the most fuel efficient operating point (stochiometric combustion) and the lowest specific fuel consumption (SFC) occurs around 25F LOP but, in practical flight, peak EGT is so close to stochiometric that the difference doesn't matter. The following diagram illustrates the basic idea:

The highest HP is obtained well rich of the more efficient settings.

The whole subject of operating ROP versus peak-EGT versus LOP has a long and complex history, with various religious positions having been adopted. The engine makers do authorise peak-EGT operation (generally only below 75% of max rated power) but have offered very little additional guidance. Lycoming have produced a notorious "Experts are everywhere" paper (local copy) which in essence suggests that lots of things are OK but due to a widespread (perceived) lack of engine instrumentation and pilot competence within the GA community they cannot recommend it... Nevertheless, great many GA pilots have embraced the more modern operating principles and there is no data suggesting that there is a problem with it. Historically, a favourite saying of many American pilots was "fuel is the cheapest thing you can put in your engine" but the recent massive fuel price increases appear to have ended this too. I have been flying peak-EGT or LOP since the engine was run-in and when it was opened up for the SB569 mandatory crankshaft change at the ~700-hour point, the engine was very clean and all parts were found to be within new limits, except the exhaust valves and even those were within overhaul limits.

Mixture management is quite simple. The phases of flight to consider are ground, takeoff, climb, cruise, and descent:

Ground and Taxi

Immediately after engine starts and stabilises, and also after landing, when the RPM is at idle, lean the mixture aggressively - almost to the point where the engine RPM starts to drop off. The lean combustion keeps the plugs clean and delivers the fastest engine warmup. Due to the very low power setting at idle there is no possibility of engine damage.

The idle/taxi RPM needs to be chosen with some care. The engine handbook recommendation may be say 1200 RPM but this can result in an excessive taxi speed, but lower values may not charge the battery, so some tradeoffs are inevitable according to how long one expects to be hanging around on the ground.

Fine wire (iridium) spark plugs suffer less from fouling than the conventional plugs, and seem to last almost for ever (mine are as good as new after ~ 1000 hours), but they are expensive and not immune to shorts caused by lead deposits. Invariably, the problematic plug is one of the lower ones which needs the lower cowling to be removed... However, to date, I have always succeeded in clearing a shorted plug using the traditional method of leaning to peak EGT at a reasonably high power setting.

Takeoff

The takeoff is simple: always at maximum available power. The fuel metering system (carburettor or fuel injection) is factory adjusted so that when at full throttle, max RPM (if a CS propeller is fitted), and full mixture, the engine is running very rich, with an EGT which is around 125F ROP. Now, at standard atmospheric conditions (1013mb, +15C), at sea level, the engine will deliver 100% of its rated power. However, the power output of a normally aspirated engine falls off immediately as one climbs.

It is less simple when departing from a high altitude airport! The takeoff is still done at the maximum available power but in this case it is not with the mixture full rich. The method is engine dependent but in general terms one leans the mixture until peak EGT is reached and that is the best power one is going to get. There are no high altitude airports in the UK but there are some elsewhere in Europe.

Note that due to the U.S. origin of nearly everything in this game, engine temperatures are expressed in degF.

Climb

Traditionally, the climb is also done with all three levers fully forward until top of climb. But there is a problem which manifests itself on higher altitude climbs:

As one climbs, the air pressure falls, and the fuel metering system detects the lower airflow and it reduces the fuel flow to maintain the mixture ratio and thus the engine operating point (in this case, 150F ROP). Unfortunately, it doesn't get it quite right. The air:fuel ratio is based on mass flow. It is trivial to measure mass flow of the fuel because at that stage it is still a liquid, and a liquid is virtually incompressible and thus its density is very nearly constant - even over wide temperature changes (avgas expands just 0.1% per degC). So any mechanism measuring liquid flow velocity or pressure in a fixed orifice will do, more or less. But mass flow of air is much more elusive; it is highly compressible and its mass flow cannot be directly measured with any simple mechanical device. There are modern electronic methods but you won't find one of these in a 1950s engine design... Both a carburettor and a fuel injection servo measure something halfway between mass flow and volume flow, and varies with altitude and temperature. Some fuel system designs incorporate altitude compensation but even they don't do the whole job, and introduce a nasty failure mode: a failure of the barometer diaphragm which stops the engine by flooding it with fuel.

The result is that as one climbs full-rich beyond say 7000ft, the engine runs progressively richer and doesn't run properly, and may eventually stop.

The traditional solution is to climb full-rich to a few thousand feet and then transition to a "cruise climb". There is nothing wrong with this, but it is not a comprehensive solution to high altitude climbs (say, 15000ft) because regular additional leaning is required even during the cruise climb.

A smarter way is the constant-EGT method. Very shortly after takeoff (within the first 1000ft or so) note the EGT of any particular cylinder. Obviously, this assumes one has an engine EGT instrument of some kind. Then, as the climb progresses, the EGT would naturally fall but one leans the mixture to restore and maintain the original EGT value. With a multicylinder monitor such as an EDM700, it doesn't matter which cylinder is used for this purpose but you have to stick to the same one all the way, and it is smart to use the one whose cylinder head temperature (CHT) is the hottest; on my engine it is #3, but #5 or #6 are more common as they are right in the back of the engine. Practically, one can do the mixture adjustment every 1000ft or so; it isn't really critical. The following pic (data downloaded from my EDM700) shows the general idea:

As can be seen above, the constant EGT also results in a nearly constant CHT which is a highly desirable situation because many engine installations have difficulty controlling the CHTs during climb - especially in the summer. Here, the CHT can be seen to actually drop slightly as the climb progresses, which is probably due to the reducing engine power and the cooling air getting colder.

The stepped nature of the EGTs during the above climb shows that the mixture was adjusted only periodically.

The constant-EGT climb technique enables a continuous climb from takeoff all the way up to the airplane's operating ceiling, with the periodic mixture adjustment being the only thing that needs attending to, and it is the most fuel efficient way to climb to any desired altitude.

I've done flight tests to measure the relative fuel efficiency of different climb methods. The primitive full-rich climb is definitely sub-optimal, as well as being of no use for higher altitudes. A very early transition to a cruise climb at peak EGT (or LOP) yields a fuel burn per distance travelled which is very close to the constant-EGT method but it is difficult - in the TB20, anyway - to keep the CHTs low enough while achieving a reasonable climb rate, so I don't use this method anymore. An LOP climb works with some engines, particularly turbocharged ones, in installations with very good cooling airflow.

Cruise

At top of climb, level off and wait till the target airpeed is reached, and then set the engine to the desired cruise operating point.

There are various options here on the power setting.

100% power: Taking the most primitive option first, most Lyco/Conti engines are rated at 100% power indefinitely so you could just burn along at whatever settings you ended the climb with - this will be the best power possible at that altitude. At low level, say 1000ft, this will obviously be all 3 levers fully forward; on my TB20 (IO-540-C4D5D) this yields about 165kt, with a ridiculous fuel flow rate of about 23USG/hr. This is OK if the airport is closing in 10 minutes, with a £1000 surcharge thereafter, but you might just make it...

60-65% power: It appears generally accepted that cruise flight at 60-65% of max rated power is optimal for engine longevity. This figure also appears in another well known Lycoming operating guide here (local copy) - pages 3, 10 and others. This power setting corresponds to approximately 23" MP and thus cannot even be achieved above about 8000ft with a normally aspirated engine.

There are now 2 cases to consider: a "best-economy" or "best-power" power setting, and there are variations on these.

Best Economy: As mentioned earlier, the best economy operating point (stochiometric combustion) occurs around 25F LOP although this point moves slightly according to the power setting. However, the efficiency curve is quite flat in that region and peak EGT achieves the same result, as close as is possible to measure (within 0.1 MPG). I fly practically all the time at peak EGT - except above FL170 or so when more power is required (see "best power" below). On low level flights where one has manually set the MP to say 23", it is easy with a bit of experience and an accurate digital fuel flowmeter (e.g. Shadin or JPI) to very quickly use the mixture lever to set a peak-EGT fuel flow, with little or no reference to temperature indications.

Setting up Best Economy without engine instruments: set the RPM and MP as desired, and slowly lean the mixture until there is approximately a 5kt drop in airspeed. This will be very close to the stochiometric mixture. It's not perfect but close enough, at the relatively low power settings used in cruise.

RPM: This produces very interesting results, with lower RPM delivering better MPG. 2200 RPM can improve the MPG by 10% over 2500 RPM. The reason for this is not well understood but appears to be a combination of: reduced engine friction losses; reduced pumping losses; and a time shift in the combustion pressure/time profile to a point where the "stuff" happens at a more favourable crankshaft/conrod angle. On all long flights, I fly at 2200 RPM and peak EGT; this yields fuel economy which is about 20% better than the flight manual! Beware: Many engines have "minimum RPM" limitations. This chart shows that my engine must not be run below 2300 RPM if the MP is above 27.2"; nobody knows why but it is rumoured to be a crankshaft stress limitation near the propeller flange. It could also be to prevent detonation. All things being equal, MP is a direct measure of torque. However, 27.2" is a very high power setting which, as the chart shows, cannot be achieved above 2500ft anyway... Another example is TCM SB07-8A (local copy) which sets a precautionary minimum cruise RPM of 2300 RPM, apparently due to previously poorly understood crankshaft counterweight behaviour.

"Deep LOP" operation (e.g. 50-100F LOP) has a religious following especially among Mooney owners but this is not borne out by flight test data which I collected carefully to exclude the unquestionably beneficial effect of flying a bit slower! I did a test flight on which the IAS was kept constant as this is a direct measure of thrust. The RPM was kept constant as this keeps propeller efficiency constant. The MP was varied to achieve the same IAS in all three cases. There was no measurable difference in MPG between peak EGT and any LOP setting. This result - easy to verify by anyone with a GPS-derived MPG readout - is unsuprising since power can come only from burning the fuel and any mixture leaner than stochiometric has to produce less power. However, LOP operation is - by definition - cooler than Peak EGT and is thus useful in some turbocharged engine installations which have problems keeping cool enough in high power cruise. Also, if LOP operation enables flight with a wide open throttle then engine pumping losses are reduced compared to the same fuel flow with a partially closed throttle. Deep LOP requires a well set up fuel injection system.

Best Power: For any petrol/gasoline-fuelled piston engine, the best power operating point occurs around 80F ROP. From my flight tests, the fuel flow is about 10% higher than at peak EGT. Given the price of avgas, this seems pointless but it has two uses: trying to beat the clock on high altitude flights (say FL100+, where 60-65% cannot be achieved anyway) where fuel is not an issue, and flight near the aircraft operating ceiling. Most airplanes will not reach their manufacturer-claimed operating ceiling unless the engine is set to the highest possible power regardless of efficiency, which means 80F ROP and maximum prop RPM. To reach my ceiling of 20000ft, I need to fly at best-power and max RPM (2575).

Optimal altitude: This is suprisingly non-critical. Flight tests suggest that so long as the engine is running with a wide open throttle (which, assuming 60-65% power, means above about 8000ft) and at a constant (preferably low) RPM, and is always set to peak-EGT or LOP, the MPG (i.e. the range) remains virtually constant. An even more interesting observation is that climbs do not significantly reduce one's range even though the speed is obviously lower; presumably this is because the potential energy built up during the climb is more or less recovered during the descent - provided that the aforementioned operating conditions remain true throughout the flight. If one has to use the best-power setting to climb very high, then some range is inevitably lost. Flight theory suggests that assuming a constant engine and propeller efficiency (which is probably reasonably valid, if always LOP, WOT, and at the same RPM) if you cruise at the same IAS at different altitudes, the no-wind range will remain the same. If, during the climb, or as a result of less available power when higher up, the IAS decreases towards the best range speed (which is theoretically equal to Vbg - the best glide speed - and is a lot slower than most pilots cruise at) then the range will increase, but the change is due to the lower IAS, not the higher altitude.

GAMI Injectors: For around $1000, GAMI will sell you a set of their injectors whose orifices (flow rates) have been selected to compensate for uneven airflows to different cylinders. According to Advanced Pilot Seminars the GAMI injectors also compensate for "covert" inter-cylinder fuel transfers which take place because the fuel is injected into each cylinder's inlet manifold rather than directly into each cylinder after the inlet valve has closed. Each injector is selected for a specific cylinder, for the specific engine. How much difference GAMIs make depends on how lucky one was with the standard injectors, but it appears that most engines can benefit from it. The general idea is that, with GAMIs, all cylinders run at the same operating point (e.g. peak EGT) at a given total engine fuel flow; this eliminates a major source of vibration and rough running which is caused by e.g. one cylinder running at peak EGT and another at 50F LOP. Similarly, if one cylinder is running 50F ROP and another is at 25F LOP, the former one is wasting a lot of fuel (any combustion rich of stochiometric implies unburnt fuel) which is why GAMIs improve fuel consumption. Before a purchase of the GAMI injectors, some flight data needs to be collected (basically, EGTs for each small increment in the engine total fuel flow, from all-ROP to all-LOP) and thus de facto the engine must be fitted with a multicylinder engine monitor to start with. Here are examples of pre- and post-GAMI data. GAMI have an FAA STC; the UK CAA issued various approvals a long time ago and these would now be grandfathered into EASA. There is no downside whatsoever to GAMI injectors... just make sure that if anybody takes them out, they put them back in the same holes!!

No Engine Instruments... this is a real problem as most of the methods mentioned here - particularly the constant-EGT method - cannot be done safely, or at all. In particular, there is a suspect region around 50F ROP (what Deakin famously calls the "red box") where detonation is most likely because that is where the time/pressure profile of the combustion encourages the highest CHT. This suggests that flying "slightly rich" with no instrumentation may not be a good idea at high power settings. However, tests done by GAMI suggest that detonation needs not only a high power setting (above 85%; not generally permitted at a mixture less than full-rich anyway) but also a grossly mismanaged CHT (around 500F).

Descent

Generally, there is not much to do because the engine is running at a low power setting and everything is thus that much less critical.

The MP naturally increases by 1" for every 1000ft so during a descent one sees a continuous increase in the engine power. For a fixed pitch trim setting, this causes the rate of descent to gradually decrease and the aircraft may even level off... eventually. However, if one is descending using an autopilot holding a preselected VS value (e.g. -500fpm), the increasing power will instead appear as an increasing airspeed. If this is an issue (e.g. to keep the speed below Vne, or below Va if in turbulence or in IMC where turbulence cannot be ruled out), it can be countered by gradually closing the throttle during the descent.

In theory, the mixture needs to be gradually enriched, in a mirror image of the gradual leaning done during the climb, and for the same reason. However, one is rarely short of power during a descent so this is really non-critical. A periodic enrichment is all that's necessary to prevent the engine going too far into LOP and consequent rough running.

This is almost outside the scope of "engine management" but do watch the engine power setting when descending with an autopilot down to a preset altitude. In most IFR touring airplanes one can descend at say -500fpm with the engine almost on idle, but when the autopilot captures the preset altitude and levels off, the result - given the absence of an auto-throttle - will be a pitch-up followed by a stall...

 

Thermal Management

Of the energy generated by combustion, around 44% goes out of the exhaust, 8% is lost via the oil cooler, and 12% is lost directly from the cylinders to the airflow (reference local copy). The rest is going usefully into the propeller. Looking at that "8% plus 12%", the airflow may appear to be less than important but actually it is vital because the cowling air intakes create a lot of drag and are thus normally designed to be as tight as is possible! Many airplanes suffer cooling problems as a result. The worst case scenario is any slow flight at high power e.g. during climb, or when flying very slowly with the gear and flaps down, or on the back of the curve. To put it in perspective, a 250HP engine during climb is trying to lose about 40kW to the airflow which is the maximum output of a large domestic boiler.

Ground: Following engine start, do not apply significant power until the oil temperature is in the green arc. This may mean a delay before taxi can commence if parked on grass, because getting the wheels out of the rut may require close to full power.

Climb: After takeoff, climb steeply (Vx) until obstacle clearance is assured, and at the earliest opportunity trim forward to get the airspeed well up; this also much improves forward visibility. At most European airports there is no problem doing this, assuming an IFR touring aircraft with a reasonable performance, but the departure (SID) vertical profile must always be studied as some do require a steep climb.

In a constant-EGT climb, the only item left to play with is the cooling airflow i.e. the airspeed, and this should be trimmed as soon as safe after takeoff to keep the hottest cylinder below 400F. Many pilots find this to be impossible to achieve if climbing at Vy, let alone Vx, so one normally trims to climb at a much higher airspeed. The rate of climb is barely lower at 120kt than at 95kt and the engine is much cooler. Cooling effectiveness improves dramatically for every few kt of airspeed. My TB20 has chronic problems keeping all the CHTs below 400F during climb and I have to quickly speed up to ~120kt which is not far below the economy cruise of 140kt IAS.

If there is a chronic problem maintaining CHTs below 400F, and the engine is in good condition, there are two possibilities. One is that the baffles around the engine are damaged and are allowing air to leak past, without going through the cylinder fins or the oil cooler. Even a small gap - 3mm - is bad enough. A great article on how to comprehensively repair the baffles is here. Another one is a mis-adjusted fuel servo; for each engine/airframe combination there is a specified band of full throttle fuel flow and it pays to have this adjusted (by a fuel servo specialist) towards the upper end of this range.

Cruise: There is rarely a problem keeping engine temperature within limits during cruise - unless there is something wrong with the engine or the airflow.

Descent: The key job here is the avoidance of shock cooling - rapid cooling resulting from a sudden engine power reduction. A huge amount has been written on this topic - including the above referenced article, and many of John Deakin's articles here mention it. Some pilots believe that shock cooling simply does not exist, but "something" is responsible for the cylinder cracks which are a common issue with larger engines, especially turbocharged ones operated at high power at high altitudes followed by long and steep descents. I am certain that shock cooling is a real hazard but only if the engine is hot enough to start with; data from one glider towing operator strongly suggests that there is a critical cylinder temperature below which a rapid power reduction doesn't do any harm. Flying schools also rarely get cylinder cracks despite heavy engine abuse but they tend to fly full-rich all the time. This all makes sense because the relevant grade of aluminium starts to lose strength rapidly above 350-400F. According to GAMI, shock cooling is unlikely to be an issue if the CHT is below about 380F.

The general solution to shock cooling is - obviously - to not do it i.e. fly such that the CHT doesn't fall too rapidly! The simplest way is to stick to gentle power reductions (of the order of 2" of MP or less, per minute) and gentle descents which don't generate a sudden massive airflow increase - all while remaining at peak EGT or LOP from the cruise. A smarter way, which is also excellent for fuel economy, is a gentle continuous descent all the way to destination; I once did this from FL180 down half the length of Croatia but this is possible only where ATC are pretty relaxed. If this cannot be achieved (e.g. ATC not authorising a descent until too late, not authorising a continuous descent, requesting an immediate descent, or in an urgent descent due to icing, etc) one needs to be more creative.

If the start of a steep descent can be anticipated, gradually reduce power while still in level flight, so that by the time one starts the descent and the airspeed and thus the cooling airflow substantially increase, the engine has already cooled to a temperature where it is not at risk; this is probably the only method usable in a checkride where the examiner will suddenly close the throttle (a popular alternative is to do it in his airplane!). An alternative method: prior to descent, select a deep-LOP mixture without reducing the MP (which cools the engine too) and once the descent commences, enrich to peak EGT which increases temperatures and helps to counter the extra cooling air.

If the start of a steep descent has not been anticipated, commence the descent without a power reduction, accept the higher speed, delay the power reduction until after the descent is established, and then reduce power gradually. For more extreme cases, drag is very useful: immediately the descent commences, extend the landing gear, flaps, air brakes, etc. but of course the limiting speeds for these devices must be observed.

Unfortunately, many pilots start a descent with a pitch-down (perhaps by setting e.g. -1000fpm on the autopilot) with a simultaneous large power reduction, creating a double whammy: cooling from increased air speed and cooling from reduced power. This can then be made worse with another standard training piece: a big mixture increase in the descent; this reduces the combustion temperature. Lycoming define (local copy) shock cooling as a CHT reduction rate exceeding 50F/minute and this can be done simply by suddenly advancing the mixture to full rich during level flight!

 

Last edited 27th December 2010

Nothing in this article takes precedence over anything written in any Flight Manual / Pilot Operating Handbook, or over anything else published anywhere else whatsoever. Use this information entirely at your risk.

Any feedback, reports of dead links, corrections or suggestions much appreciated:
Contact details

Aviation Home