The Car I Designed on Compost Gas, and Why the EV Market Just Found Its Way Back to It
A Systematic Innovation homework assignment, decades early
Herbert Roberts, P.E. | Inventor's Mind
When I was learning systematic innovation, one of my homework assignments was to design a car that would run on vegetable scraps.
I thought about composting — how a pile of decomposing waste gives off gas as it breaks down — and started sketching from there. My concept used a GM Fiero as the base, because it was light and the engine bay sat in the rear, out of the way of everything else. Instead of the Fiero's gas engine, I put in a small gas turbine, running on the gas captured from composting waste, spinning fast enough to turn a generator. The generator made electricity. The electricity drove a motor at each wheel, independently, with energy recovery built into the same motors on braking and deceleration.
No mechanical connection between the engine and the road at all. The turbine's only job was making electricity. The wheels' only job was using it.
Confession. I didn't design that car because I had some early insight into where the automotive industry was headed thirty-some years later. I designed it because the assignment forced a constraint — vegetable-scrap fuel — that made the obvious answer (a normal engine, a normal driveshaft) impossible, and the impossibility is what pushed the design somewhere useful. That's worth admitting plainly, because it means the wide-end move I stumbled into wasn't foresight. It was a homework assignment doing exactly what a good constraint is supposed to do: making the narrow-end answer unavailable so you're forced to look upstream of it.
The false issue, if you'd handed me that assignment and asked me to solve it the fast way, would have been "how do I get a normal engine to run on compost gas." That's a real engineering problem, and a hard one — compost gas is low-energy-density, inconsistent, nothing like refined fuel. The true issue, once the fuel constraint forced the question, was "does the engine actually need to be connected to the wheels at all, or can it just make electricity and let something else move the car." Once I stopped trying to make an unreliable fuel work in a mechanical drivetrain and started treating the engine as nothing but a generator, the fuel-quality problem got a lot smaller. The turbine didn't need to run smoothly enough to feel good through a driveshaft. It just needed to spin a generator well enough to keep a battery topped up.
The forensic correction.
Electric vehicle sales have slowed hard enough to draw a straight line through automaker balance sheets: Ford took a $19.5 billion charge on its EV plans, Stellantis booked $27 billion, GM absorbed $7.6 billion, and Honda cut its EV investment by $20 billion — all inside about a year. The narrow-end response to that kind of pressure is obvious, and on the surface reasonable: the market rejected the new architecture, so retreat to the old one. Go back to gas or diesel driving the wheels directly, the way it worked for a century before anyone tried anything else.
That's the trap, and it's the same trap the "how do I make compost gas run a normal engine" framing would have been, if I'd taken the easy way out on that homework assignment. It treats "EV" and "engine-to-wheels" as the only two shapes the problem can take, when the part of the EV architecture that actually struggled wasn't the drivetrain. It was the energy source. Range anxiety, charging time, and cold-weather battery performance are all problems with storing electricity, not with using electric motors to drive wheels. Electric motors are simpler, more efficient at converting energy into motion, and better at instant torque than a mechanical driveshaft has ever been. None of that stopped being true when battery-only sales cooled.
The technical teardown.
The wide-end answer, and the one the industry is actually converging on now, is close enough to my Fiero sketch that it's a little unsettling: keep the electric drive to the wheels, and reframe the energy source instead. An engine — gas, diesel, or otherwise — runs as a generator, feeding electric or hydraulic motors that do the actual work of turning the wheels, with no mechanical driveshaft connecting the engine to the road at all. Horse Powertrain's X-Range platform is built explicitly to let automakers convert a battery-EV platform into exactly this kind of extended-range architecture with minimal retooling. Toyota's hybrid lineup, already profitable and already immune to range anxiety, is the market proving the same point from the other direction, having never fully left the electric-drive-plus-engine-as-support architecture in the first place.
Here's the analytical move stated plainly: when a system fails, don't assume the whole system was wrong — separate the part that actually failed from the part that was just riding along with it, and only rebuild the part that failed. The EV architecture has two separable pieces: an energy-storage system (the battery) and a drive system (electric motors at the wheels). The battery struggled. The motors didn't. A narrow-end response throws out both and returns to what worked before either existed. A wide-end response keeps the motors — the part that was never the problem — and reopens the framing on the energy source alone.
Neither the extended-range platform nor the traditional hybrid is "going back" to the pre-EV architecture, and I want to be precise about that rather than overclaim it as a straight win for electric drive. There are real costs to this approach that a pure battery-EV, if the battery problem gets solved, doesn't carry: two propulsion systems instead of one, an engine's worth of maintenance that a pure EV skips entirely, and a fuel-burning component that a fully-charged pure EV doesn't need at all. The wide-end reframe isn't free. It's a deliberate trade — accepting the complexity and maintenance burden of keeping an engine on board, in exchange for solving the range and charging problem immediately rather than waiting on battery technology to catch up. That's a real trade-off, not a hidden win, and it's worth stating that plainly rather than pretending the reframe has no cost.
The same logic shows up in heavy equipment, where hydraulic-electric hybrid drivetrains let an engine run at a constant, efficient load while hydraulic or electric motors handle the variable, start-stop demands of the actual work — same reframe, different industry, and one more data point that the pattern isn't specific to passenger cars.
Forensic signature.
The tell that an industry is solving at the narrow end is a wave of automakers announcing they're "pausing EV plans" and quietly restarting pure gas-engine development from where it left off — treating the battery's problems as proof the whole electric-drive concept failed. The signature of the wide-end move is what's actually happening instead: billions in EV investment being redirected, not abandoned, toward architectures that keep the electric motors and change what feeds them.
Aftermath, and a forward prediction.
I didn't build that Fiero. It was a homework assignment, graded, filed, and mostly forgotten until an industry I'd been watching for other reasons started arriving at a very similar answer for very similar reasons — a fuel source that couldn't do the job on its own, forcing the question of whether the engine needed to be connected to the wheels at all. The lesson I'd draw from that isn't "I predicted the future." It's that a well-built constraint, even an artificial one handed to a student for a grade, can force the same wide-end reframe that a multi-billion-dollar market correction eventually forces on an entire industry. The constraint doesn't need to be real to do useful work. It just needs to make the narrow-end answer unavailable.
This closes out the four cases I set out to walk through — a legacy engine frame, a Cold War spaceplane, a falling rocket booster, and a compost-gas homework assignment that turned out to be closer to the future than I knew at the time. What ties them together isn't the industry or the decade. It's the same question, asked in four different rooms: when you hit a wall, is the wall the problem, or is it where an earlier decision, made by someone who's since moved on, finally became visible?
I'd love to hear where you've seen this pattern in your own work — the moment a system failed and the fix wasn't rebuilding what broke, but noticing that only part of it ever needed to change. That question drives everything I write here.
Herbert Roberts, P.E. is a licensed professional engineer with 30+ years in aviation research and development.
FEB (Formen Engpass Barriere)™ is a pending trademark of Inventor’s Mind Press, naming the practice of reframing a problem from the wide end of the funnel — reshaping the framing that produces a constraint, rather than working the constraint itself.