Zwift Setup & Equipment Guides
Cooling, power measurement, trainer modes and the connection stack, in the order they actually affect your riding. Most of what gets sold as an indoor upgrade changes nothing. A few cheap things change a lot.
What actually matters, in order
The bike in your room does almost nothing for your Zwift speed. Zwift models the frame and wheels you picked in the in-game garage, not the ones bolted to your trainer, so a lighter wheelset on the indoor bike buys nothing. The ranking is short:
- Airflow. A fan you would describe as too big is the correct size.
- A power number that is consistent. Consistent beats accurate.
- A connection that does not drop.
- Sweat management, which costs two minutes a session.
- A screen and sound you can actually use.
- A position that matches your outdoor fit.
Everything below that line, the aero cockpit and the ceramic bearings and the expensive tape, is comfort rather than speed. The shopping list version is in the Zwift setup guide.
Cooling is a power item, not a comfort item
Cycling is an inefficient engine. Roughly a fifth to a quarter of the energy you burn comes out of the pedals as power; the rest leaves as heat. A rider holding 250 W sheds close to a kilowatt of it for as long as the effort lasts.
Outdoors that is solved for you: at 30 km/h you sit in a 30 km/h wind that strips sweat off your skin as fast as you make it. Indoors your road speed is zero, the air does not move, it warms, and it fills with the moisture you just produced. Sweat that runs off you instead of evaporating has cooled you almost not at all, because the phase change removes the heat, not the wetness.
So core temperature climbs. Your body sends blood to the skin to dump it, leaving less volume for the working muscles. Heart rate drifts up at the same power and the power you can hold falls. That is why a workout that felt fine at 18°C (64°F) turns into a suffer-fest at 24°C (75°F). A big fan is the only upgrade that reliably moves your sustainable power. Buy the fan first, then buy the second fan.
Sweat is corrosive and it lands in the worst place
Sweat is salt water with a little acid in it. On the road it blows backwards off you. On a trainer you are stationary and bent forward, so it lands on the stem, the top cap, the headset bearings underneath, the bar tape and the top tube.
Those spots have everything corrosion needs: steel bolts threaded into aluminium, a bearing race in a dark sealed space that never dries out, and salt, which keeps pulling moisture from the room long after you stop. Left alone over a winter of four sessions a week, that means crusty bar tape, seized stem bolts, a notchy headset and bubbling paint.
The two-minute routine
- Before the ride, drape a thin towel over the bars, stem and top tube, and keep a second one for your face.
- After it, wipe those parts with a damp cloth and dry them. Salt dissolves in plain water, and a solvent only strips the finish.
- Once a season, undo the top cap and stem bolts, grease the threads and reinstall to torque. Wipe the trainer and the mat too.
The long version is in towels during Zwifting.
Where the power number comes from
Zwift does not know what you are doing. It knows what one device told it.
A wheel-on trainer usually has no strain gauge. It reads power off a modelled resistance curve driven by roller speed, so everything in the tyre-to-roller contact becomes a variable in your power: pressure, compound, clamp tension, rubber temperature. Skip the spindown calibration and the same workout reads differently on Tuesday and Thursday.
A direct-drive trainer removes the tyre: your chain drives its cassette straight onto its own measurement. That sits downstream of your drivetrain, so chain friction is already subtracted and a crank meter on the same bike reads a little higher. Both can be right.
A power meter in the pedals, crank or spider measures closest to where your legs make force. Left-only units double one leg, so an imbalance becomes a permanent offset. Different measuring points and different smoothing are why two devices can both sit inside their quoted accuracy and still be fifteen watts apart at threshold.
That gap matters the moment you compare yourself to anyone else: a race result, a category boundary, a friend's climb time. A source reading three percent high is free watts you never made. Feed it into the Alpe du Zwift calculator and you get a time you cannot ride, or into the racing category calculator and you land in the wrong bracket.
It matters much less if you only compare yourself to yourself. A consistently wrong number is still useful: intervals, an FTP test and week-to-week progress all work as long as the bias holds still. Pick one source, set it in Zwift, and stop switching mid-block.
ERG mode or simulation mode
In ERG mode you name a power and the trainer holds it, adjusting resistance so gear and cadence barely matter. For structured intervals that is excellent: no pacing decisions to get wrong, no drift below target.
On a climb it is the wrong tool. A climb is a gradient problem: speed comes from power against your weight on a slope, and the skill is picking a gear and cadence you can hold for fifty minutes. ERG deletes that. Hairpins feel like straights, and if your cadence sags the trainer adds resistance to defend the target, which sags it further. That is the spiral of death.
In simulation mode the game sends the gradient, the trainer turns it into resistance, and you shift like a cyclist. Alpe du Zwift, 12.2 km and 1,036 m of gain at roughly 8.5% across 21 hairpins, should be ridden that way, and so should Ven-Top at about 20.9 km and 1,534 m. There is a pacing walkthrough in mastering Ven-Top.
One setting to know: Trainer Difficulty scales how much of the modelled gradient the trainer applies, and it defaults to 50%. It changes how the climb feels, not your speed in the game.
The connection stack and how it fails
Three ways your trainer talks to Zwift. ANT+ broadcasts through a USB dongle and any number of receivers can listen at once. Bluetooth connects directly to the device running Zwift, and plenty of hardware holds only one link at a time. Zwift Companion bridges Bluetooth sensors from your phone to the machine running the game.
The failure modes are boring, repeatable, and they cost people races:
- Something else grabbed the Bluetooth channel first: the manufacturer's app in the background, a bike computer set to auto-connect, a second phone on the shelf.
- The Companion bridge dies mid-effort. The screen locks, battery saver kills the background link, or phone and computer sit on different Wi-Fi bands. Your power reads zero and you drift backwards.
- The ANT+ dongle sits in a USB 3.0 port behind a metal tower. USB 3.0 radiates noise in the 2.4 GHz band ANT+ uses, so use USB 2.0 and an extension cable.
- A firmware update prompt three minutes before the pen closes.
So pair straight to the device running Zwift and keep Companion for the controls, not the data. Close the manufacturer app, turn off auto-connect on your head unit, and pair well before the start, not in the pen. Latency is a separate problem, covered in Zwift network performance.
Screen, sound, and where the fan actually points
Put the screen at eye level or slightly below, close enough to read your power without leaning forward; if you have to squint at the number you will ride the scenery instead of the target. Sound has to beat the trainer and the fan, and a cheap speaker usually wins, because headphones and a face full of sweat are a short-lived combination.
The fan needs to hit your face and chest, not your shins. Blood runs close to the surface at the face, neck and chest, and that is where the sweat film is thickest, so that is where moving air removes the most heat. Mount it at bar height or above and angle it down. A second fan low and behind, aimed at your back, is the best cheap follow-up there is. Open a door or a window too, or the fan just recirculates warm, saturated air.
The articles below go deeper. The FAQ answers common trainer and calculator questions, and the rest sits on the blog.