DARPA LIFT: I Want My $6.5 Million for Solving Your Non-Problem — Here Is the First-Principles Aerodynamics
DARPA LIFT: I Want My $6.5 Million for Solving Your Non-Problem — Here Is the First-Principles Aerodynamics
For context:
I entered the DARPA LIFT Challenge after telling its organizer, Phillip Smith, that I believed 20:1 payload-to-weight was easily achievable.
I later told him in writing to bring a WWII Jeep and an M1097A2 Humvee as competition payloads — over 40:1 and 100:1 respectively for a 55 lb aircraft.
I am posting the underlying first-principles argument publicly.
If the physics is wrong, please identify exactly where it is wrong.
Mr. Smith,
When I entered the DARPA LIFT Challenge, I told you plainly that I intended to walk away with all six prizes and the full $6.5 million.
That wasn't bravado.
It was because I believed the Challenge was built around a remarkably simple conceptual error:
CONVENTIONAL DRONE PERFORMANCE HAD BEEN MISTAKEN FOR A LIMITATION IMPOSED BY PHYSICS.
The error is easy to expose.
Take a hypothetical aerodynamic surface operating at:
L/D = 1,000,000 : 1
If it produces one unit of drag, it produces one million units of lift:
Drag = 1 unit
Lift = 1,000,000 units
Nothing mysterious is happening, and nothing violates conservation of energy.
The aerodynamic polar defines the lift-to-drag force relationship at that operating point.
A PROPELLER IS ALSO A WING.
Its blades operate from aerodynamic polars.
The motor supplies torque against the aerodynamic resistance of those blades while their aerodynamic force is resolved principally into useful thrust.
Now use that thrust to drive an aircraft wing, or use a propeller to tip-drive another rotor.
You have created successive aerodynamic force transformations.
The analogy is a compound lever: force ratios can be stacked.
In simplified force terms:
System force ratio = R1 x R2 x R3 x ... x Rn
For example, four hypothetical aerodynamic stages each providing an effective 100:1 force ratio give:
100 x 100 x 100 x 100 = 100,000,000
Every real aerodynamic stage must, of course, have a physically valid combination of radius, RPM, chord, pitch, Reynolds number, velocity triangle and airfoil polar.
And the entire system must obey conservation of energy.
Those are engineering constraints.
They do not change the underlying force relationships.
AND HERE IS THE CRITICAL POINT:
We can keep adding aerodynamic working area to target essentially any desired finite payload-to-aircraft-weight ratio until actual structural, geometric, Reynolds-number, mass, or mission constraints establish the practical limit.
The lift equation is elementary:
Lift = 1/2 x rho x V^2 x S x CL
where:
rho = air density
V = aerodynamic velocity
S = aerodynamic working area
CL = coefficient of lift
At the same air density, velocity and CL, lift increases directly with aerodynamic working area S.
If additional lift is required while maintaining the desired aerodynamic operating point:
INCREASE S.
ADD AREA.
There is no law of aerodynamics requiring a drone designer to hold lifting area fixed and compensate for increasing payload by continually increasing power density, aerodynamic loading, battery mass and aircraft mass.
That is a DESIGN CONVENTION.
It is not physics.
My LIFT designs exploited exactly this principle at different scales.
When more lift was required, I didn't accept the performance envelope of existing drones as an immutable constraint.
I ADDED AERODYNAMIC WORKING AREA.
That is why I entered the Challenge.
I did not believe I had discovered a way around physics.
I believed the Challenge had confused the performance of conventional drone architectures with the limits imposed by physics.
That is confirmation bias.
I recognized this before entering, which is exactly why I confidently told you that I expected to win all six prizes.
I then diverted approximately nine months of engineering effort and substantial commercial opportunity to solving the problem DARPA put in front of me.
So let me be equally direct now:
I WANT MY $6.5 MILLION FOR SOLVING YOUR NON-PROBLEM.
If my first-principles analysis is wrong, show me where.
Show where the aerodynamic polar fails.
Show why successive physically valid aerodynamic force transformations cannot be engineered.
Or identify the physical law that prohibits increasing aerodynamic working area until the required payload-to-weight ratio is achieved.
If that cannot be demonstrated, then America deserves an answer to a much simpler question:
WHY WERE WE GIVING AWAY $6.5 MILLION TO SOLVE A SUPPOSED DRONE-PERFORMANCE BARRIER THAT WAS SUBSTANTIALLY THE PRODUCT OF A CONCEPTUAL ERROR AND CONFIRMATION BIAS?
I am quite willing to have that discussion publicly.
John L. Bass
Senior Engineer
KISS Engineering
------------------------------------------------------------
P.S. — I WAS FEELING PARTICULARLY CONFIDENT ON MY BIRTHDAY
The following was sent directly to Phillip Smith on April 19, before the present dispute:
Re: Contact Us Via DARPA.mil
Hi Phillip,
When I started the conversation below I suggested 20:1 was easily possible, which I'm certain you felt was beyond belief at the time.
Hopefully you have other designs by now that are crossing well above 10:1 so that doesn't feel entirely impossible at this point.
Would you be willing to provide a WWII Jeep and a M1097A2 Humvee as the payloads at the competition since the stack of barbell weights is getting a bit unwieldy at these points? Yes that's a bit over 40:1 and 100:1 for a 55lb drone. Makes for much better photo ops for your competition.
John
That wasn't a prediction I made after the fact.
I had already told the Challenge organizer, in writing, to:
BRING ME A HUMVEE.