Every instrument in the panel lies to you in some specific, predictable way โ and instrument flying is largely the art of knowing which one is lying right now. This page lets you break the pitot-static system and watch the gauges respond, fly the altimeter into pressure and temperature traps, drift the gyros, and take the magnetic compass through the turning and acceleration errors that make it so hard to trust.
The airspeed indicator, altimeter and VSI are all just pressure gauges plumbed to two openings: the pitot tube (ram + static pressure, from flying into the air) and the static port (ambient pressure alone). The ASI reads the difference between them; the altimeter reads static pressure against a set datum; the VSI reads how fast static pressure is changing. Ice or a mud dauber in either hole, and specific instruments fail in specific ways. Pick a blockage below, then climb or descend and watch which needles tell the truth. The middle row is the X-ray view: every needle is just a capsule flexing โ the ASI's diaphragm with pitot air inside and static around it, the altimeter's sealed vacuum wafers, and the VSI's diaphragm racing its own leaky case. The drifting dots are the air molecules in each case (density exaggerated so you can watch it): climb and the air visibly thins; the wafers have none at all โ that's the vacuum the whole altimeter is built around.
The killer case is pitot + drain blocked โ trapped ram pressure turns the ASI into a crude altimeter. Climb and it reads faster; the instinctive pitch-up to slow down is exactly wrong. With a blocked static port, the altimeter freezes, the VSI insists you're level, and the ASI reads slow above the blockage altitude and fast below it. The escape is the alternate static source (or, in a pinch, breaking the VSI glass โ it's the least useful gauge on the line). Cabin static reads slightly low, so expect the altimeter and ASI to read slightly high on alternate air.
An altimeter never measures height โ it measures pressure and reports the altitude that pressure would mean in a standard atmosphere with the setting you dialed in. Fly toward lower pressure without resetting, or fly in air colder than standard, and every pressure level sits lower than the instrument assumes: the altimeter over-reads while you sink toward the rocks. "From high to low โ pressure or temperature โ look out below."
Rules of thumb: 1 in Hg โ 1,000 ft of altimeter error, and roughly 4 ft per 1,000 ft per ยฐC of ISA deviation (4% per 10 ยฐC). Cold is the sneaky one โ the altimeter setting fixes the error at the surface but says nothing about the shrunken column above it, which is why cold-temperature-restricted airports require you to add a correction to approach altitudes when it's cold enough. 91.121 requires a current altimeter setting within 100 NM for a reason.
Gyro instruments trade the compass's twitchiness for a different flaw: a gyro holds its direction in inertial space, and the Earth doesn't sit still underneath it. Add bearing friction, and every spinning instrument slowly wanders. The attitude indicator has its own trap โ its self-erecting mechanism aligns to apparent gravity, which tilts backward every time you accelerate.
This is just the Earth rotating under a rigid gyro โ 15ยฐ/hr ร sin(latitude), the full 15ยฐ/hr at the pole and zero at the equator. A latitude nut inside the instrument pre-torques away the error for one latitude, and random friction precession piles on top in either direction โ which is why the habit is what matters: reset the heading indicator to the magnetic compass every 10โ15 minutes, in straight, level, unaccelerated flight (section 4 explains why only then).
Gravity plus inertia makes "down" tilt aft during acceleration. The AI's erection system slowly chases that false vertical, so a vacuum AI can creep a couple degrees nose-up on a long takeoff roll or go-around โ but your vestibular system feels the whole tilt, instantly. That's the somatogravic illusion: accelerating into a black night, you feel a steep pitch-up, push forward, and fly into the ground. The defense is believing the instruments โ which is also why it's worth knowing the AI leans the same direction, just far less. Old units also show small errors after steep turns (a slight climb indication on rollout) and precess after aerobatic attitudes.
The Earth's field doesn't run parallel to the ground โ it dives into it, steeply at high latitudes (about 60โ70ยฐ down across the U.S.). The compass card is weighted to stay level and ignore that vertical pull, but the moment you bank or accelerate, the card tilts, the vertical component of the field grabs it, and the reading swings โ with nothing actually turning. Turning errors: UNOS โ undershoot rolling out on north, overshoot on south. Acceleration errors: ANDS โ accelerate and it shows north, decelerate and it shows south (east/west headings only). Fly both below.
Practical rollout numbers (northern hemisphere): rolling out on north, start the rollout when the compass still has about latitude + ยฝ bank degrees to go; on south, let it swing past by about latitude โ ยฝ bank. Near east and west the turning error passes through zero โ and that's where ANDS takes over instead. The card is only trustworthy in straight, level, unaccelerated flight, which is exactly the condition for setting the heading indicator from it. One more trap: the wet compass card reads backwards compared to your DG โ 33 sits to the right of N in the window โ so turn toward the small numbers feels wrong the first time.