The Supersonic dilemma

By Bjorn Fehrm

December 20, 2017, ©. Leeham Co: The Super Sonic Transport (SST) has a new spring. Aerion announced its new partner Lockheed Martin Friday and Boom got a new investor in Japan Airlines (JAL) the week before.

The design of a supersonic transport aircraft is exciting and difficult. Yet it isn’t the key challenge. The engine is.

The SST design challenge

How to design a supersonic aircraft was mastered in the 1950s. I’ve flown one of these designs, at the speeds the new SST’s aim for, the Mach 2 SAAB Draken.

The SAAB Draken was designed in the mid-1950s, like the Mirage III and Lockheed Starfighter. All capable of Mach 2.0 at high altitude.

The challenge the new SST projects wrestle with is not how to design the SST airframe. How was learned then. And it hasn’t changed.

Yes, Aerion introduces interesting supersonic laminar flow technology to make the SST fly farther on less fuel. This wasn’t available then and it’s new and exciting stuff. But it’s not the sauce that cracks the real SST nut.

The real problem all these projects wrestle with is the engine.

But the Mach 2.0 Draken had an engine, didn’t it? Yes, but not one that would be accepted in a modern SST. Too thirsty, too noisy.

The design of an SST engine

Since the days of the straight jet-driven Draken/Mirage/Starfighter, the bypass turbofan has solved the “too thirsty, too noisy” problem. The higher the bypass, the quieter and more economic the engine. Yet this is the problem. High bypass turbofans don’t work on an SST.

Why? We have learned the higher By-Pass-Ratio (BPR), the better?

We didn’t learn the full story. The mantra, “higher BPR” is better, only applies for subsonic airliners. It’s all in the fundamentals of how jet thrust is created. And it isn’t complicated.

We have several times presented the thrust of a turbofan as:

Thrust = Air mass flow * Air over-speed

The engine thrust increases when the air mass flow through the engine increases or the air’s speed out the back increases. We also learned; the propulsive efficiency of the engine increases when the over-speed (“specific thrust” in expert speak) is close to the free stream’s speed.

Consequently, today’s high BPR engines get a lot of air accelerated to a modest over-speed.

Now to a version of the formula that tells us a more detailed story (still not complicated):

The Air mass flow is the air entering the engine in the front, through the inlet. The Air over-speed is the difference between the air entering the inlet (freestream air) and the air going out the engine’s back. We can also write the formula as (read here to drill deeper):

Thrust = Air mass flow * Exit airspeed – Air mass flow * Inlet airspeed

The first term is called Gross Thrust and the second term Ram drag. The Thrust we now call Net Thrust, to distinguish from Gross thrust. So Net Thrust is Gross thrust minus the Ram drag.

The increasing Ram drag is our supersonic problem.

To understand it, we compare two engines: the engine powering the Boeing 737 MAX and the Aerion engine GE is working on.

An airliner engine versus an SST engine

The 737 MAX engine is the LEAP-1B, built by GE (core) and Safran (low-pressure part). Static thrust is between 22klbf and 29klbf, with a bypass ratio of 9. Overall Pressure Ratio is around 40, with the high-pressure compressor at 20. It’s a state of the art engine, with low noise and low fuel consumption.

The Aerion engine is projected as an 18klbf static thrust engine with a “medium BPR,” according to Friday’s press conference. It’s based on a GE commercial core which “has billions of hours of operation,” surrounded by a new, lower bypass, low-pressure system.

We will assume a BPR of 4 and an overall PR of 30.

Based on the information, we assume the core comes from the CFM56. It has a static thrust between 22klbf and 34klbf dependent on the version. Static Pressure Ratio for the most common version is around 27, with the high-pressure compressor at 10.

Its core design is from the 1970’s Boeing B1 bomber project, a 40-year old base design. We will come to why this core would be more suitable than the 40 years younger LEAP core.

Thrust versus Ram drag

We will simulate “flying” these engines at the M0.78 of the 737 MAX and then the M1.4 of the Aerion SST. We will use our engine simulation software, GasTurb, for the simulation.

As these are different size engines, we will focus on key ratios between the engines rather than absolute values. Our aim is to get a qualitative understanding of the SST problem.

Here the results:

  • At the typical cruise speed of M0.78 and altitude of 37,000ft of the 737 MAX, the LEAP has a Ram drag-to-net thrust ratio of 1.2. The Aerion engine cruising at the same speed and altitude has the ratio of 0.8. Its fuel consumption is 20% higher, as it has a lower core efficiency (lower PR) and higher specific thrust (over-speed) than the LEAP.
  • At the Aerion cruise speed of M1.4 and altitude of 45,000ft, the LEAP has a Ram drag-to-net thrust ratio of 3.7. The Aerion engine is now at 2.1, 55% lower than the LEAP. Now the fuel consumption reverses, as the Aerion engine works less hard to generate the net thrust. It has an 11% better fuel consumption (TSFC).

One could wonder if the Aerion engine would work for BOOM’s M2.2 aircraft. It wouldn’t. Ram drag-to-net thrust at M2.2 would be at 20 times and the TSFC goes through the roof. The faster you cruise, the less air you shall let pass your inlet. And the faster you should kick it out the back.

The SST dilemma

The dilemma is engine noise also scales with Specific thrust, but in the wrong direction. The lower the over-speed, the less noise generated by the air leaving the engine. Noise regulations coming 2020 requires an engine with a compatible Specific Thrust.

The SST engine design problem then is:

  1. Design an engine with the lowest BPR that fit the noise regulations.
  2. Check how fast this engine allows you to fly.

If you want to go faster, you need a variable cycle engine (one with high BPR at low speed and low BPR at high-speed).

The other problem

The BPR (or really Specific Thrust) dilemma is the principle problem. There is one more key problem (and then many smaller), needing a solution. An SST engine gets hot.

The air entering our LEAP at 35,000ft is at -54°C. When braking from M0.78, the air is heated to -28°C before it enters the compressor. At the end of our 40:1 PR compression, the air is around 520°C, below what the best compressor materials can handle. The subsequent higher temperatures in the combustor and turbines can be handled with cooling air tapped from the 520°C compressor.

If we fly the LEAP at the Aerion data, M1.4 and 45kft, the air entering the compressor is at +28°C, a 56°C rise. Now the end of the compressor is around 650°C. This is uncomfortably high for a three-hour cruise and the cooling air for the turbines and combustor is a bit hot.

To fix the high engine temperatures we need an engine with a lower compression ratio. This is where the lower compression core of the CFM56 fits.

For a Mach 2 engine, the entry temperature of the compressor air is 150°C. Now we will fry the engine with a compressor which has a PR higher than 15.


The cruise speed of the Aerion SST has gradually gone down over the years from M1.6 to M1.4. And for good reasons.

The Ram drag problem becomes more manageable and the engine runs cooler. The SST engine can be built on a modest compression ratio civil core, paired with a suitable low-pressure system.

If one wants to go faster, a suitable core is harder to find. The engine needs a lower bypass ratio at cruise, to not suffer under the created Ram drag, and a lower overall pressure ratio, to not get too hot from the high inlet temperature.

44 Comments on “The Supersonic dilemma

  1. Fantastic article Bjorn, thank you!

    Couple of things. GasTurb is a name with all the hallmarks of being chosen by a proper engineer. It is perfect:

    And if they’re still operating it, perhaps you should go for a ride in Thunder City’s EE Lightning in South Africa. Or perhaps wait for the project that’s running in the US restoring another Lightning to flight. I reckon you’d like it, and we’d all love to read your report on the experience, sort of a preview of what we could expect from a new SST.

    • Hi Matthew,

      would love to fly the EE Lightning, has two RR Avons of the type I had in the Draken. Should be fun!

      • Ah, it’d be just like old times!

        Though taking a look at the figures on Wikipedia, the Draken looks like it was pretty nippy. So perhaps the Lightning would be a bit to too slow (never thought I’d write that!)… And it looks like there’s more airworthy Drakens out there to choose from. Oh for a ride in one of those!

        • I always hate it when the USAF takes up the news reporters that have not a clue about aviation.

          I have waited 60+ years to take a ride in a Supersonic aircraft.

          Pretty much my bucket has a hole in it now.

          Bjron would be more than worthy.

          Lightening was one of my favorites (maybe the favorite) of that Era.

          Not just a hot rod, hot rod with 2 engines and stacked, great stuff, sorry that era is gone.

    • Hi John,

      you are OK with a straight inlet (like the Aerion one) up to about M1.6 (only about a 5% loss of air pressure, going to heat instead). At M2 or higher you absolutely need a variable inlet, otherwise, you have the heat and no pressure. A variable shock inlet like the Concorde’s is assumed in the calculations. If not, things are worse.

  2. Damn if that Aerion doesn’t look like a 104 “gone civie”! And doesn’t that make total sense in joint venturing on it with Lockheed? On a more engine specific note Bjorn, why not adapt and modernize the Concorde’s Olympuses? And, lastly, does the engine name get any better? Call it the “Grande Olympus”, or more prosaically, the “Olympus II”!

        • Hi Uwe,

          my guess, without knowing. In this phase, Aerion needed a manufacturing partner, probably one who was ready to invest as well (it costs a LOT to start a production, your initial losses on production are as high as the development costs. You have high losses before your learning curve has come down). Guess Airbus had less interest than LM in the production phase.

          • I knew the shape was bugging me.

            F-104 indeed.

            LM is doing some branching out, this would be part of it.

            How well it goes is ??????

        • Apparently the DoD are interested.

          [Which means commie Europeans out and get good ‘ol capitalist muricans in. ;-)]

  3. So when I understand this right, the problem is that there is no modern engine with a suitable low pressure ratio and the development cost for one would be prohibitive.

    • Hi Gundolf,

      Aerion has found its engine. At M1.4 the problems are solvable with available technology. I guess you mean Boom. At M2.2 the dilemma of take-off sound and efficient M2.2 operation is profound. Let’s see how it solves it.

  4. Sorry for my ignorance, but I need to know:
    a) A super-Duper SST, made for those who are rich, could afford to have 3 engines, like 2 for take off to Mach 0,95 ( a-la CFM Leap), and one for Supersonic (Mach 0,95 – 1.6)? Then, you could close the bigger Leap intakes. That could provide a very smooth Take off and landing, with an interesting low speed cruising. I did read some previous time that the Concorde was able to supersonic cruise (Mach 2 +) with just two engines, and that it was indeed very efficient at supersonic speeds.
    b) Isn’t the sonic boom the most difficult part of the equation? Usually, the engines have several parts moving at supersonic speeds (and many indeed move the air to supersonic speeds). So, in order to allow for a US overflight, the sonic boom should be addressed first?

    • Hi Ramiro,

      What I didn’t mention in the article is the nature of supersonic drag. If you have engine bodies who not gulp air they cause high supersonic wave drag. The two non-active engines would be a major drag source, even with the intakes closed. Supersonic aircraft shall be long and have a small cross-section. The Starfighter is the epitome supersonic aircraft.

      The non-allowed sonic boom over the US is a problem (not a rational ruling, just anti Concorde at the time. So much for playing fair). Europe says the boom cannot reach the ground, and Aerion is working on flying at M1.2 without this happening. More as it should be.

      • Bjorn:

        I hope I can put the anti sonic boom issue in some US perspective.

        It was the start of the environmental movement as well as the post Vietnam War era and the political environment (belief the government always lied)

        LA was the pits, I flew into that area in a C152, I was maybe 3 miles from the Airport before I saw it on an otherwise clear day (outside the bowl)

        Rivers severely polluted and getting worse, corporations doing what they wanted.

        In short it was a toxic environment (pun intended) for a sonic boom to be introduced when a lot of the country was not in the mood as well as the military and their (call them cowboys) who had a propensity to inflict sonic booms well outside their training area.

        I think it was a decision based on, its not ours, why should we fight it when we are getting (rightfully) pasted on this other stuff?

        My personal take is that the world benefited overall. The movement spread to Europe (I was shocked to see pictures friends took there and how polluted it was)

        And while its also personal, there is an anti well to do remnant of that in my soul that it benefits people with a lot of money vs the environment for a sane use of resources.

        You can flip that in regards to there would be working people making money to make the stuff the same as Executive jets (Yachts etc)

        While I grew up in that whiz bang age my views have changed and I question things a lot more.

        Now its more, we can do it, should we?

  5. I guess the second problem for SST is the business case. Are there enough routes where SST is possible and useful? It seems many countries have forbidden civil flights with supersonic speed over their territories. That was always a problem for the Concorde too. There are the routes over the Atlantic (from Western-Europe to the US-East-Coast). And if they find enough range maybe also over the Indian ocean (from the Arabian Peninsula (Dubai, Doha, etc.) to Australia). What else? Is it enough for a profitable production and operation of SST aircrafts?

    On top of that: In the past a flight was simply lost time. Saving flight time had a high value for some people. Today there is broadband internet onboard. For many people it’s possible to work normally during a flight.

    • Hi Guido,

      this is the $5bn question. I just addressed the technical problems before you get there.

    • What else? For example, Juneau, Vancouver, Seattle, Portland, San Francisco, LA, Phoenix, Las Vegas, Salt Lake, and/or Denver to Honolulu and/or Maui. All with transatlantic range profiles, with significant time savings. Most of these are with enough of a pool of affluents (remember, Denver and Salt Lake are major, interior West hubs) to pay for at least one round trip a day.

    • For me a flight is still lost time. I just can’t focus on anything really tricky, be it spreadsheets or contracts. I’ve read a study somewhere that this has to do with the lack of oxygen in our brain due to the relatively low air pressure on planes.

      So building planes that offer really comfortable “altitude” like 3.000 feet or so and also improved humidity should address that problem. Maybe here’s a new business idea…

      • It’s not the flight that’s the problem, it’s the airport. Just measure the time from entering airport to start of take off roll, similarly end of landing roll to departing airport environs. Add in any transfers and subtract long hauls and see where your time is wasted.

    • I think we have to look at a new SST as being something aimed at people who do want to go jolly quickly, just for the hell of it.

      So that’s impressively extravogant charter flights like Concorde used to do, maybe a few luxury commercial routes. If they could build an SST without a vast outlay, it might just work.

      There is kinda a precedent for this kind of thing. In some parts of the world there’s a boom in luxury train travel (e.g. Japan, there’s some in Europe too). Special trains have been built, and they’re booked up solid. That kind of hints that there is a pool of people out there who have money and want to travel in style. The sort of person who’d pays thousands of dollars for 1st class+ on an Emirates or Etihad A380.

  6. @bjorn : a couple questions:
    1: having read your cooling argument I have to wonder: the Leap generation engines are built specifically to handle more heat throughout the process (and I assume the GE9x and company even more heat as they have ~60:1 pressure ratios), whereas the CFM-56 core is designed for much lower temps. so don’t both cores have the exact same problem? or have they dramatically narrowed the temp margins on the newer gen cores?

    2: I remember reading (once upon a time) that the F-105 Thud had the most efficient supersonic inlet ever designed (and the eyeball test would agree as the inlets are remarkably small when compared to other SS aircraft of similar performance). I also know that the primary reason the B-1B was a M1.1 aircraft and not M2+ like the B-1A was the B-1B did away with the expensive and complex variable inlets. how much effect can careful intake design have on the ram drag over a simplistic variable ramp SS intake?

    • Hi bilbo,

      1. The new engines are more heat resistant in the combustor/turbine parts where they use smarter cooling with compressor air and advanced coatings/materials in combination. The SST problem is more the last stages of the compressor, where there is no higher pressure cooling air (the only one you have is the compressor air in the last stages and it’s the too hot air anyway and it hasn’t any higher feed pressure to use on itself!). Compressor T3 temp is more difficult than Combustor/Turbine T4x air (the Ts are the sections in the engine in engine speak).

      2. A lot on the efficiency side. A non-variable inlet loses about 35% of the pressure at M2.2. So you have the heat but lose pressure = efficiency. Link Ram drag is not affected much, it’s simply the aircraft speed times the gulped air mass. The lower inlet pressure means less air density = less mass but it’s not material. The pressure loss means the engine (Gross thrust part) has to work even harder to produce the net thrust = increased TSFC.

      • Thanks Bjorn.

        so, why don’t they use the same materials/coatings in the final compressor stages as they use in the HP turbine stage? I get that cooling air is a big part of it, but also they use more exotic alloys and coatings (and soon CMC)

        yes, cost is an issue, but from a technical perspective I would think that swapping in turbine grade materials in the last couple compressor stages in the Leap core would get some gains.

        • It’s not just the blades, we talk about the casings, rotors, stator vanes, the whole lot which cannot be cooled. It’s something new as would be a variable cycle engine. The rather small private venture SST projects are not the ideal ones to try this out for the first time. The Defence dep. has been working on variable cycle engines for some time but they are not ready for prime time, not even in high-risk military projects.

          • Thanks Bjorn.

            I’m surprised that for non-rotating components that some form of closed circuit liquid cooling or an interstage heat exchanger (a-la a low tech version of the Reaction Engines concept) has not been used. if you cooled all the compressor stators with fuel (to minimize the weight penalty) there would be a significant knock on effect on the TIT.

        • They are, the last stages of the compressor are getting powder metal disks like the HPT turbine disk. It is expensive but when Inco718 is not good enough you go to forged powder metals.

        • You probably can do all this, if you have a few billion in development dollars to throw around to make sure it all works right. Unfortunately there is probably enough technical and market risk in the SSBJ already. I can well understand why they might choose an engine core (CFM56) that was derived from a supersonic aircraft core in the first place (F101).

  7. Bjorn:

    Back on the technical side, a lot of past my ability to follow the math (never the greatest at it, ok with practical day to day use)

    Did I read it right that once up to cruise the SFC goes down for the SST?

    • Hi Transworld,

      no, the SFC goes up with increasing Ram drag. To produce the same Net thrust, the engine must produce more Gross thrust (which is made with burning fuel), so more speed = more Ram drag = higher SFC.

        • You are right in one sense though. There is a drag bump when passing the sound barrier up until ~M1.2, this is why the Concorde used the afterburner past the sound barrier during its climb.

          • With a really good supersonic inlet and a very high cruising altitude the Engine efficiency goes up with increasing speed. Drag increses with speed but as speed increases you can go higher into thinner air with less drag and thus often colder air enter the Engine helping the cycle . Still flying supersonic is gas guzzling but far off from flying hypersonic then fuel consumption rises exponentially.

  8. hello… is it possible (LAW or permission) to fly @ supersonic speeds within the US Country or other Countries ?

    the intent for a super sonic offer to the market will be only for flying at supersonic speeds only during the atlantic ocean ?

    • In the U.S., there’s an FAA regulation regarding not exceeding Mach 1, except in very limited circumstances (e.g. military, Space Shuttle, X planes, etc.)

  9. The CFM56-2 core engine came from the F-101 hence the CFM56 series core engine is pretty well suited for supersonic operations. GE can reduce the life limited parts life on those engines to approx 2500 cycles and maybe allowing operting stub life CFM56-7B cores in these engines for +500 cycles. The key is a new intake and LP system that can be modelled of the F414 latest development or the varibale bypass fighter engines in development. So for a business jet application with a few cycles /month it can be a pretty quick solution of a CFM56-7B+F414. For an airline pumping back and forth over the Atlantic in M1.4 day and night a bigger development program is required to get reasonable life and maintenance cost.

  10. You know, I think I’m beginning
    to understand how Elon thinks
    we should just do it all with
    Rockets, and a Large
    Economy Section.

  11. Interesting steep negative scarf on the tail vs wing nacelles. Would that be for noise reduction Bjorn?

    • I think it has more to do with increasing pressure recovery at M1.4 by having a two shock inlet.

  12. “An intricately-shaped variable geometry inlet uses shockwaves to compress oncoming Mach 2.2 air, efficiently slowing to the ideal subsonic speed for the engine. Digitally-controlled movable surfaces precisely position shock waves to achieve ideal compression at a wide range of speeds and flight conditions, while blow-in doors provide extra airflow for takeoff.”

Leave a Reply

Your email address will not be published. Required fields are marked *