A line follower robot doesn't need to fight physics. It needs to cheat it, legally.

That's what a suction system does. Bolt a fan under your chassis, seal the edges, and you create a zone of low pressure that presses your robot into the track. No extra weight. No extra motor torque wasted on acceleration. Just more grip, exactly when you need it most: mid-corner, at full speed, when a normal line follower would already be sliding off the track.

This is the same trick Formula 1 used in 1978, and then banned within a season. Some robotics competitions have quietly done the same thing.

If you're building a suction LFR from scratch instead of buying a finished bot, this is the complete list: chassis, impeller, speed controller, sensors, the actual science behind why it works, and where these robots go to race. Grab a coffee. This one's long enough to matter.

Why Suction Changes the Whole Equation

Every line follower lives or dies by one equation: friction = coefficient of friction × normal force (F = μN). Normal force is basically how hard your wheels push down on the track. More normal force, more grip, higher cornering speed before you slide off the line.

The obvious way to increase normal force is to add weight. The problem: weight also increases the mass you need to accelerate and decelerate. You gain grip and lose responsiveness in the same move. It's a trade-off every LFR builder eventually hits a wall with.

Suction breaks that trade-off. An impeller (a small centrifugal fan) spins under a sealed chassis and pulls air out from beneath it. That drop in pressure creates a downward force. It doesn't come from added mass. It comes from the pressure difference between the air pressing down from above and the near-vacuum underneath. The result: more normal force, more grip, faster cornering, and the robot never gets any heavier.

A well-built suction chassis can generate downforce several times the robot's own weight. Techgeeks' own suction chassis and impeller modules are rated for 600g+ of downforce on a robot that weighs around 150g: call it 4x its own bodyweight, pressing straight down, on demand.

The Same Trick Formula 1 Used, Then Banned

If this sounds familiar, it should. Racing has been here before.

In 1970, the Chaparral 2J (nicknamed the "sucker car") mounted two fans at the rear, driven by a separate engine, to suck air from underneath the chassis and generate downforce that didn't depend on the car's speed. It was so effective that rival teams lobbied to have it banned, and it was gone from competition within a year.

Brabham's Gordon Murray revisited the idea in 1978 with the BT46B, built to counter the ground-effect Lotus 79 that was dominating the season. Instead of shaping the underbody like an inverted wing (the approach Lotus was using), Murray put a large fan at the rear of the car, officially there to cool the engine. Unofficially, it pulled air from underneath the car and created a low-pressure zone that glued the BT46B to the track. Because the fan's official job was cooling, and cooling accounted for more than half its airflow, the car stayed within the letter of the rules.

The result was downforce that didn't need speed to build up. Grip was simply there, all the time, even under braking and mid-corner, unlike ground-effect cars that needed airflow to work. It won its only race, at the 1978 Swedish Grand Prix. It never raced again. Rival teams pushed back hard enough that it got pulled from competition before regulators formally closed the loophole.

The hardest engineering problem wasn't the fan. It was the seal. Murray later described skirts that had to be tuned precisely: too tight and the wearing strip would fail before the race ended; too loose and suction would lift the skirt and the seal would fail entirely, dumping all the downforce at once.

That's the exact same problem you'll fight on a desktop-sized suction LFR. A skirt that doesn't seal properly leaks air in from the sides, your vacuum collapses, and you lose grip precisely when you're carrying the most speed into a corner. Scale changes. Physics doesn't.

The twist: F1 banned fan cars outright. Some line follower competitions have started doing the same thing. The rulebook for the International Greenwich Olympiad's line follower event explicitly prohibits turbines. It's a small sign of where this is heading. Suction is currently a legitimate edge in most competitions, but check your specific rulebook before you build. It's not universal anymore.

Where These Robots Actually Race

Line following is one of the most widely run robotics events in the world, and suction builds show up at the top end of nearly every serious one:

• DD Robocon / ABU Robocon: India's national qualifier into the Asia-Pacific Robot Contest, and the event most Indian college teams (IIT Bombay Techfest, VIT Gravitas, NIT Trichy Pragyan) are building toward.

• Technoxian World Robotics Championship: runs a dedicated Fastest Line Follower category and draws international teams competing head-to-head on the same track.

• Robotex International (Estonia): one of the largest line-following events globally, with mirrored parallel tracks and strict weight and size caps that make every gram of "free" grip worth fighting for.

• Eurobot: Europe's long-running student robotics contest, open to international teams beyond the continent.

What you'll notice watching finals footage from any of these: the top times aren't separated by better code. Most competitive teams are running some form of PID control by now. They're separated by who can carry more speed through a corner without breaking traction, and that's a mechanical problem before it's a software one. That's the entire reason suction exists as a category.

What You'll Actually Need: The Shopping List

Here's the real component breakdown, in the order you should actually buy things. Prices below are current Techgeeks listings so you can build an accurate budget before you start.

1. Chassis

Don't build a suction system on a chassis that wasn't designed for it. Sealing an afterthought fan onto a flat acrylic plate almost never holds vacuum. You need mounts and geometry built around the impeller from the start.

• Techgeeks Advanced Suction Chassis: ₹1,499. 120mm × 180mm × 40mm, ~45g, PLA+/PETG with sealed skirt geometry built in and a universal sensor and PCB tray.

• If you're just starting and want to test PID and sensor tuning before adding suction, the Basic LFR Chassis (₹499) is the cheaper way to learn the fundamentals first.

2. Impeller + Speed Controller (Your ESC)

This is the part that actually does the sucking, and it's the single biggest upgrade over a standard LFR.

• Suction Impeller Setup: ₹999. Grade-A 8520 coreless motor plus centrifugal impeller and housing, rated for 600g+ downforce, and it already includes a USC (Unidirectional Speed Controller) for PWM regulation.

• The USC matters more than it sounds like it should. A standard H-bridge driver like a TB6612FNG or L298N is built for bidirectional motors and runs oversized, hotter, and noisier at high switching frequencies than a purpose-built unidirectional controller. An impeller only ever spins one direction. A standalone USC (₹449) is what you'd add if you're wiring a second impeller or replacing one, since the impeller kit already ships with its own.

3. Sensor Array

• ARC8, 8-Channel IR Array: ₹999. 3g, 10 kHz sampling, this is the default choice for a straight speed-focused suction build using weighted-average PID.

• ARC16, 16-Channel 2D IR Array: ₹1,499. Only worth the upgrade if your track has junctions, loops, or maze elements the robot needs to recognize. Its MUX architecture reads T-junctions and intersections that an 8-channel array can't tell apart from a straight line.

4. Controller Board

• Blueprint 01 Robot Controller PCB: ₹799. Arduino Nano mount, integrated TB6612FNG driver for your drive motors, on-board 5V regulation, and an 8-channel ADC rail. This replaces a breadboard, which is worth doing on its own: unstable power rails and signal noise from loose breadboard wiring are the single most common cause of "my PID works on the bench but not on the track."

5. Motors & Wheels

• N20 High-Traction Wheels: ₹349 for a pair. Comes in three diameters, from 20×20mm for high-RPM, low-resistance builds, up to 30×30mm for torque-heavy setups.

• Drive motors aren't sold standalone in the current Techgeeks catalog. They come bundled in the Mark 2 kit (12V, 2000 RPM N20 gear motors). For a scratch build, source 12V N20 gear motors rated 1000RPM+ from a general electronics supplier; higher RPM suits the extra grip a suction chassis gives you.

6. Battery

A suction build needs more electrical headroom than a standard LFR, because the impeller draws current independently of your drive motors.

• 3S 12.6V LiPo, ideally with an isolated power rail for the impeller separate from your drive motors and logic board. This isn't sold as a standalone Techgeeks item currently. Source a 3S LiPo rated for at least 2A continuous draw (the 8520 coreless motor alone pulls around 2A under load) from a hobby RC supplier.

Quick reference table

*Motors and battery aren't in Techgeeks' current catalog as standalone items. Prices are approximate and vary by vendor.

Controlling the Impeller Without Frying It

The impeller motor doesn't want to go from 0 to full duty cycle instantly. That current spike is a common way to cook a coreless motor or arc a connector on your first test run. Ramp it.

Recommended duty-cycle ceilings, based on Techgeeks' own thermal testing: 0–35 out of 255 on a 2S LiPo, 0–25 out of 255 on a 3S LiPo. Going past that window on a small coreless motor is where things start overheating.

// Simplified impeller ramp-up sequence (Arduino core)
// Full register-level Timer2 PWM driver ships with the impeller kit's GitHub repo

const int IMPELLER_PIN = 11;   // D11 / OC2A
const int MAX_DUTY_2S  = 35;   // thermal-safe ceiling on 2S, out of 255
const int MAX_DUTY_3S  = 25;   // thermal-safe ceiling on 3S, out of 255

void setup() {
  pinMode(IMPELLER_PIN, OUTPUT);
}

void rampImpeller(int targetDuty) {
  for (int duty = 0; duty <= targetDuty; duty++) {
    analogWrite(IMPELLER_PIN, duty);
    delay(15);   // soft-start: spares the motor and connectors
  }
}

void loop() {
  rampImpeller(MAX_DUTY_3S);  // switch to MAX_DUTY_2S if you're running 2S
}

// Simplified impeller ramp-up sequence (Arduino core)
// Full register-level Timer2 PWM driver ships with the impeller kit's GitHub repo

const int IMPELLER_PIN = 11;   // D11 / OC2A
const int MAX_DUTY_2S  = 35;   // thermal-safe ceiling on 2S, out of 255
const int MAX_DUTY_3S  = 25;   // thermal-safe ceiling on 3S, out of 255

void setup() {
  pinMode(IMPELLER_PIN, OUTPUT);
}

void rampImpeller(int targetDuty) {
  for (int duty = 0; duty <= targetDuty; duty++) {
    analogWrite(IMPELLER_PIN, duty);
    delay(15);   // soft-start: spares the motor and connectors
  }
}

void loop() {
  rampImpeller(MAX_DUTY_3S);  // switch to MAX_DUTY_2S if you're running 2S
}

Advanced builders skip analogWrite() entirely and drive Timer2 directly at register level for tighter, jitter-free control with none of the Arduino HAL's latency. That's the setup documented in the impeller kit's GitHub repo if you want to go further than this simplified version.

Where Most First Suction Builds Go Wrong

• Treating the skirt as an afterthought. The seal is the whole system. A gap you can barely see with your eyes is enough to collapse the vacuum. This is literally the same failure mode that made Gordon Murray tune BT46B skirts by hand at Brands Hatch.

• Wiring the impeller off the same rail as your logic board. Current draw spikes from the impeller will brown out your microcontroller mid-run. Isolate the rail.

• Skipping the ramp-up. Slamming straight to max duty cycle at power-on is the fastest way to damage a coreless motor.

• Adding downforce without upgrading grip. More normal force only helps if your wheels can actually use it. Cheap, hard rubber wheels will still slip regardless. This is why traction-focused wheel compounds matter more on a suction build than on a standard one.

• Ignoring your weight budget. Every gram you add fights the downforce-to-weight ratio that makes suction worth doing in the first place. A heavier suction bot is often slower than a lighter one with less downforce.

DIY vs. Ready-Built: What You're Actually Paying For

Piecing this together yourself, using the components above, lands somewhere around ₹5,800–₹6,700 depending on where you source your motors and battery.

Compare that to the Mark 2 Suction Kit at ₹6,999, or the pre-assembled and pre-tuned Mark 2 (tuned) at ₹7,499.

That gap has narrowed to almost nothing: as little as a few hundred rupees, and in some sourcing scenarios the kit is basically the same price as buying the parts separately. What used to be a real discount for going DIY is now mostly just the experience of doing it yourself. What the kit price still buys you: components already matched to each other (impeller sized correctly for the chassis' skirt volume, motors geared for the wheel size), a firmware baseline that's already tuned instead of a blank sketch, and a build guide that skips the trial-and-error most first-time suction builders go through on skirt sealing alone.

If you want the learning experience and don't mind debugging your own seal, build from scratch. You'll still come out slightly ahead on price and a lot ahead on understanding. But if the DIY discount was the main reason you were considering it, that math doesn't really hold anymore. At these prices, the kit is the more rational default unless the build itself is the point.

Conclusion

A suction line follower isn't a bigger motor or a smarter algorithm bolted onto the same old chassis. It's a different way of generating grip entirely, borrowed from a piece of F1 history that lasted exactly one race. Get the skirt sealing right, keep the impeller on its own power rail, and the downforce takes care of the rest.

So, are you building this from scratch for the learning curve, or going with a pre-tuned kit to get track time sooner?

Related reading: Chaparral 2J and the origins of fan-car downforce · Brabham BT46B on Wikipedia