Direct Drive vs Belt vs Gear: What Actually Changes

All three drive types produce force at the rim. What separates them is what sits between the motor and your hands. A gear wheel routes the motor through a helical gear train, a belt wheel routes it through a toothed belt and pulleys, and a direct drive base bolts the rim straight onto the motor shaft. Every difference people argue about, notchiness, noise, detail, maintenance and price, follows from that one structural fact.

What is physically different between the three drive types?

Force feedback drive type describes the mechanical path from the motor to the steering rim. There are only three arrangements in consumer sim racing, and each one is a different answer to the same engineering problem: small electric motors spin fast and produce little torque, while a steering wheel turns slowly and needs a lot of it.

A gear driven wheel solves that with reduction gearing. The Logitech G920 at $199.99 and its PlayStation twin the Logitech G29 both use two small motors driving a helical gear set. Gearing multiplies torque and divides speed, so a cheap high-speed motor becomes usable. The cost is that gear teeth have to mesh, and meshing teeth have clearance between them.

A belt driven wheel replaces the gear teeth with a toothed belt running over pulleys. The Thrustmaster T300RS GT at $399.99 uses a brushless motor and a belt reduction, and the belt is the key part: rubber and fibre have compliance, so the abrupt tooth-to-tooth handoff of a gear train is smoothed out. Hybrid wheels such as the Thrustmaster T248 use a belt for part of the reduction and gears for the rest, which lands between the two in both feel and price.

A direct drive base removes the reduction stage entirely. A large, slow, high-pole-count motor produces the torque directly, and the wheel rim bolts to its shaft. There is no belt to stretch, no gear teeth to mesh and no backlash to take up. The MOZA R3 at 3.9 Nm and the MOZA R9 V3 at 9 Nm are the same architecture at different motor sizes.

Why is detail resolution the real difference, not peak strength?

The number every product page leads with is peak torque in newton metres, and it is the least useful figure for choosing between drive types. Peak torque is the maximum rotational force the motor can apply at the shaft. It tells you how hard the wheel can pull, not how clearly it can tell you two different things at once.

What you actually feel while driving is a stack of overlapping signals: tyre load building through a corner, the front axle starting to slide, kerb serration, a bump in the surface, engine vibration, and the self-aligning torque that tries to straighten the wheel. Those signals differ in magnitude by more than an order of magnitude. A kerb strike might be ten times the force of the front tyres beginning to let go, and the tyre signal is the one that changes your lap time.

A gear train degrades that stack in two ways. Backlash swallows small reversals below the size of the tooth clearance, so the quietest signals never make it to your hands at all. And the inertia of the geartrain, multiplied by the square of the reduction ratio, acts like a low pass filter, rounding off fast transients. That is why a gear wheel can be strong and still feel vague: it is strong at the large forces and blind to the small ones.

Direct drive inverts that. Because there is no reduction stage, the motor's own resolution reaches the rim unmodified. This is why a MOZA R5 at 5.5 Nm at $379.99 can read finer than a wheel producing similar total force through gears, and why experienced drivers habitually run high torque bases at 30 to 50 percent gain. They are not buying strength. They are buying the ability to set the ceiling high enough that the small forces never get compressed against it.

The one sentence version. Peak torque sets the ceiling. Drive type sets how many distinct things you can feel underneath that ceiling. Spend on the second one.

How do the three types compare on paper?

The table below uses published manufacturer figures where a manufacturer publishes them. Where a maker does not publish a torque rating, the cell says so rather than repeating a community estimate as a specification.

Why can nobody quote a newton metre figure for a Logitech gear wheel?

Because Logitech does not publish one. Not for the G29, not for the G920, and not for the G923 at $299.99. The specification sheets list 900 degrees of rotation, a dual motor helical gear drive, and on the G923 the TRUEFORCE audio-driven feedback layer. There is no torque rating anywhere in the official documentation.

Numbers do circulate. People clamp a wheel down, hang a known mass off a lever arm and calculate a figure, and those measurements land in the low single digits of newton metres. That is useful context and it is also not a manufacturer specification. Measurements vary with firmware, with the game, with temperature after twenty minutes of running, and with where on the rim the load is applied. Anyone quoting a confident decimal figure for a G29 is repeating a number nobody at Logitech ever signed off.

The practical consequence for a buyer is that you cannot cross-shop a gear wheel against a direct drive base on the torque column. Compare them on drive architecture, on rotation range, on the platforms they authenticate against, and on what the pedals in the box actually measure. The wheelbase torque chart lists every model with a published figure and leaves the rest blank on purpose.

What does the centre dead zone actually feel like?

Backlash is easiest to notice in a slow, straight-line moment rather than a corner. Sit stationary on the grid, turn the wheel a few degrees left, then a few degrees right, and pay attention to the transition through centre. On a gear wheel there is a small window where the rim moves and nothing pushes back, followed by a distinct step as the teeth re-engage.

On track it shows up as a hesitation in corrections. Catching a rear slide means a rapid reversal of steering direction, exactly the input that has to cross centre, and the wheel is momentarily uninformative during that crossing. It also shows up on long straights, where holding a car steady in a crosswind involves constant tiny reversals and the wheel keeps stepping between them.

A belt drive reduces this because the belt takes up load progressively rather than tooth-to-tooth. A direct drive base has no crossing to make. That is the single change most people notice first when moving from a G920 to something like a Thrustmaster T598 at $429.99, and it is more immediately obvious than any increase in strength.

What breaks, and what maintenance does each type need?

Gear wheels wear. Plastic or composite gear teeth accumulate wear over hundreds of hours, and the symptom is a growing notch or a grinding sound near centre. It is repairable with a gear and grease kit if you are willing to open the housing, and the parts are cheap because so many of these wheels exist. Expect this as a maintenance item on a wheel used daily for a few years, not as a defect.

Belt wheels have the belt itself, which stretches slowly and can slip if a pulley collects debris. Belts are replaceable and the job is more involved than a gear regrease. Belt wheels also usually run a cooling fan, which is a second wear part and the main noise source once the belt is quiet.

Direct drive bases have essentially no wear parts in the drivetrain, because there is no drivetrain. What they do have is heat. A base run hard for a two hour stint will get warm, and firmware will reduce output to protect the motor if it exceeds its thermal limit. On paper this makes direct drive the lowest maintenance option in the category, which is a genuine argument for it beyond feel.

The failure mode people underestimate is not the base at all, it is the mount. A frame that flexes converts feedback into frame movement, and the harder the base the more of the signal gets lost that way. If you are moving up a torque band, budget for the cockpit in the same purchase. A frame like the Playseat Trophy at $599.00 exists specifically because a 9 Nm base on a folding frame is a base you cannot use at full output.

Safety, and this is not boilerplate. A direct drive base applies its full rated torque with no gearing to absorb it, and it can sprain a wrist or a thumb. Set the torque ceiling in the driver software before the first power-up, never hook your thumbs through the rim spokes, stand clear during the calibration sweep when the wheel turns to full lock on its own, and keep the emergency stop reachable on bases that have one. Frames matter too: a mid direct drive base and a steel cockpit together weigh enough that moving the assembly is a two person job, so unbolt the base before you move the rig.

Which drive type should you actually buy?

If your budget genuinely stops around $200 and you are on a console, a gear wheel is still the correct answer, because it comes with pedals and it authenticates on every platform. Buy it knowing what you are getting: a working force feedback wheel with a coarse signal and an audible gear train.

If you have $250 to $450 and you are on PC, skip belt and buy entry direct drive. The MOZA R5 bundle includes base, rim and pedals, and it clamps or bolts to a stand without drama. On console, the Thrustmaster T598 and the Fanatec CSL DD family are the two routes that authenticate, and the Fanatec route means buying largely direct rather than from a marketplace.

If you have a bolted cockpit already, the 9 to 12 Nm bracket is where most people stop for good. The MOZA R12 V2 at $399.99 and the Thrustmaster T818 both sit there, and either one will outlast several sets of pedals. Above that, the MOZA R21 and the Simucube 2 family are for people whose rig is aluminium profile and whose limiting factor is genuinely themselves.

Work out what your current frame can hold before you commit, using the wheelbase torque calculator, and if you are choosing between two specific bases the head to head in direct drive vs belt drive covers the money question in more detail. The full ranked list lives in the best direct drive wheelbases, and the model-by-model figures for the base you are considering are on its individual spec page.