A single planetary set can give you a few ratios. Two sets, wired together sharing a sun shaft, give you a usable spread that covers launch to cruise. This is the Simpson arrangement — the architecture inside most transmissions you've opened.
A single planetary has sun, ring, and carrier. Lock one, drive another, the third is forced to be output.
Every automatic transmission obeys the Willis equation. One planetary set has exactly three connection points to the outside world. Lock one of them to the case (ground it), connect an input shaft to another, and the third is mathematically forced to be your output. You get three useful ratios from one physical set:
| Grounded Member | Result | Speed Ratio |
|---|---|---|
| Carrier fixed | Sun and ring counter-rotate. Reverse direction. | Reverses output |
| Ring fixed | Sun drives carrier at deep reduction. Best launch torque. | 3:1 reduction (typical) |
| Sun fixed | Ring drives carrier at mild reduction. | 1.5:1 reduction (typical) |
The problem: One set alone cannot cover the full range from launch (needing 3:1 or deeper) to cruise (needing 0.7:1 overdrive). You would need a third ratio between them, and one set cannot provide it cleanly without adding messy mechanical complexity.
A single planetary gives you reduction in two modes and reversal in one. You have no overdrive. You have no intermediate stepping. One set is incomplete for a real vehicle.
Bolt the front set's sun to the rear set's sun. Connect the front carrier to the rear ring. Now you can select six useful states from two pieces of hardware.
The genius of the Simpson arrangement, used in the Ford C4, GM Turbo Hydramatic, and Chrysler Torqueflite, is simple: take two planetary sets and connect them with a shared sun shaft. The front set's carrier then feeds into the rear set's ring.
Why this works: By choosing which member to ground on which set, you create cascading reduction. The front set might give you 1200 rpm output from 3600 rpm input, then that becomes input to the rear set, which can further reduce (or overdrive) it.
Example walk-through at 3600 rpm engine input:
| Gear | Front Set State | Rear Set State | Carrier B Output | Ratio |
|---|---|---|---|---|
| 1st | Sun driven, Ring grounded | Ring (floating), Sun driven | 1067 rpm | 3.37 : 1 |
| 2nd | Sun driven, Ring grounded | Ring (floating), Sun driven | 1200 rpm | 3.00 : 1 |
| 3rd | Sun locked to ring (1:1) | Whole assembly locked (1:1) | 3600 rpm | 1.00 : 1 |
| Overdrive | Carrier grounded, Ring floating | Ring grounded, Sun output | 4500 rpm | 0.80 : 1 |
| Reverse | Sun driven, Carrier grounded | Ring grounded | −1800 rpm | Reverses |
Notice: the step from 1st to 2nd is only 3.37 ÷ 3.00 = 1.12 — a smooth, tight progression. The step from 3rd to 4th is huge, but that's intentional: at cruise speed the engine is already wound down, so the jump to overdrive feels like relief, not a shock.
Part count. Efficiency. Packaging. Smoothness. Pick any metric—Simpson delivers.
Two full planetary sets, shared sun shaft, common carrier, shared ring. That's roughly the same parts as three totally separate sets, but with half the mounting complexity.
Because you can fine-tune the tooth counts on each set independently, you control gap size between gears. Too large a step = harshness. Too small = more gears needed.
The Turbo Hydramatic stayed in production for decades running hundreds of thousands of vehicles. Simpson architecture is not theoretical—it's proven metal.
If one shift fails, you know exactly which member should be grounded and which should be driven. Pressure test that path first. Systematic, not guesswork.
Simpson's elegance comes with a trade-off: you are limited to about 4 forward gears before the hydraulic control becomes hairy. Modern 6, 8, and 10-speed boxes use Ravigneaux sets (two suns, one ring, one carrier) or chain-linked multiple sets. More complexity, but same basic Willis equation underneath.
Your transmission is not shifting between different gearboxes. It is re-wiring the same hardware into different states.
Most people imagine a transmission like a stack of coins: you pull out one and put in another, each with fixed tooth counts. That is wrong and will mislead every diagnosis you try.
Instead, imagine a transmission like a patch panel: you have three ports (sun, ring, carrier on each set), and you have five clutches/brakes. The transmission control unit is rapidly engaging and disengaging clutches to change which port is grounded, which is driven, and which is read as output. Every shift is a state change, not a different machine.
When diagnostics fail on one shift, ask: which member should be grounded? Is that clutch applying? Is there pressure reaching it? Is the friction material worn? That path-based thinking will save you hours of guessing.
Animation shows you motion. Pressure data shows you control. Understanding both makes you unstoppable at diagnosis.
When you watch an animated breakdown of a Simpson arrangement (like Sabin Civil Engineering's Allison 6-speed demonstration), you're seeing something critical: the planets orbit and spin simultaneously. That's the key insight most mechanics miss.
In a single planetary set, the planet gears always orbit around the sun. They're not stationary. When the ring is held fixed, the sun spins, and the planet carrier (your output) spins too—all three are in motion. The math (Willis equation) tells you the ratio of their motions, but animation shows you the direction and speed.
In a Simpson arrangement:
Let's walk through what's actually moving when you're in 1st:
When you apply 1st gear clutches:
Now here's where pressure data tells you what animation cannot: which physical clutches are actually engaged. A transmission has (typically) three to five clutches. Each one grounds a specific member. When you shift, you're not selecting "1st gear"—you're selecting "apply clutch A and B, release C." The hydraulic system does this by sending pressure to specific apply piston ports.
In a Ford C4 (Simpson-based 3-speed):
| Gear | Clutch A (front ring) | Clutch B (rear carrier) | Direct Clutch |
|---|---|---|---|
| 1st | APPLY | APPLY | OFF |
| 2nd | APPLY | OFF | OFF |
| 3rd | OFF | OFF | APPLY |
| Reverse | APPLY | OFF | OFF |
When you connect a scanner and watch live pressure during a shift from 1st to 2nd:
This is why pressure testing beats guessing:
Watch the animation first. You see the front ring held, the front carrier orbiting, feeding into the rear set. You pause at 1st gear and trace it: "Okay, at 3000 rpm engine, front reduces me to 890, then rear reduces that to 264. Two cascades. That's 3.37:1." You see why 1st exists: you need that deep ratio to launch the car from stop.
Now connect your scanner. You watch the C-line (front ring) pressure and the B-line (rear carrier) pressure as you drive. You feel the shift timing. You note whether the engine flares (slips) or does clean steps.
Finally, you know what to fix. If B-line pressure is low, you know Clutch B is the problem. You're not "replacing the transmission." You're replacing the rear carrier pack (or the seals, or the frictions). You know exactly where to go.
Most transmission shops see a "4-speed slipping." You see: "Front set ring clutch is worn, based on C-line pressure not holding." You already know the fix before you drop the pan. That's not magic—that's Simpson architecture knowledge plus pressure reading. Animation teaches you the first part. Your pressure gauge teaches you the reality.
The Simpson arrangement is not magic. It is elegant engineering that lets two sets of hardware do the work of three, with predictable, controllable states.
When you understand that every shift is a state change in a four-clutch (or five-clutch) control pattern, you stop guessing. You test the grounding path. You verify pressure. You find the failure fast.
Your discovery—that a transmission is a torque translator, not a mystery box—becomes powerful when you combine it with Simpson's architecture insight. The machine is systematic. The failures are systematic. The fixes follow logic, not luck.
Take a real transmission with a known shift problem. Map it to the Simpson diagram. Identify which member should be grounded and which clutch does the grounding. Test that circuit first. Document your finding. Add it to the field guide so future mechanics see your thinking.