diff --git a/firmware/src/imu/drivers/bmi160/math.rs b/firmware/src/imu/drivers/bmi160/math.rs
index 3184fda8..2cc19d04 100644
--- a/firmware/src/imu/drivers/bmi160/math.rs
+++ b/firmware/src/imu/drivers/bmi160/math.rs
@@ -1,5 +1,7 @@
-const RAD_PER_DEG: f32 = 2. * core::f32::consts::PI / 360.;
+const RAD_PER_DEG: f32 = core::f32::consts::PI / 180.;
 const DEG_PER_RAD: f32 = 1. / RAD_PER_DEG;
+const ACCEL_PER_G: f32 = 9.81;
+const G_PER_ACCEL: f32 = 1. / ACCEL_PER_G;
 
 // TODO: This whole module needs unit tests
 
@@ -13,6 +15,7 @@ pub enum GyroFsr {
 	D250,
 	D125,
 }
+
 #[allow(dead_code)]
 impl GyroFsr {
 	/// The default FSR when the IMU is reset
@@ -85,6 +88,98 @@ impl GyroFsr {
 	pub const fn rad_per_lsb(self) -> f32 {
 		self.dps_per_lsb() * RAD_PER_DEG
 	}
+
+	/// The bmi160 returns the data from the gyro as an `i16`, we must use the Full
+	/// Scale Range to convert to a float rad/s
+	pub const fn discrete_to_velocity(self, discrete: i16) -> f32 {
+		discrete as f32 * self.rad_per_lsb()
+	}
+}
+
+/// The Full Scale Range of the accelerometer. For example, G2 means +/- 19.6 meters/sec^2
+#[derive(Eq, PartialEq, Copy, Clone, Debug)]
+#[allow(dead_code)]
+pub enum AccelFsr {
+	G2,
+	G4,
+	G8,
+	G16,
+}
+
+#[allow(dead_code)]
+impl AccelFsr {
+	/// The default FSR when the IMU is reset
+	pub const DEFAULT: Self = Self::G2;
+	pub const fn from_reg(v: u8) -> Result<Self, InvalidBitPattern> {
+		Ok(match v {
+			0b0011 => Self::G2,
+			0b0101 => Self::G4,
+			0b1000 => Self::G8,
+			0b1100 => Self::G16,
+			_ => return Err(InvalidBitPattern),
+		})
+	}
+
+	pub const fn to_reg(self) -> u8 {
+		match self {
+			Self::G2 => 0b0011,
+			Self::G4 => 0b0101,
+			Self::G8 => 0b1000,
+			Self::G16 => 0b1100,
+		}
+	}
+
+	pub const fn as_u16(self) -> u16 {
+		match self {
+			Self::G2 => 2,
+			Self::G4 => 4,
+			Self::G8 => 8,
+			Self::G16 => 16,
+		}
+	}
+
+	pub const fn from_u16(v: u16) -> Result<Self, InvalidNum> {
+		// TODO: I'm not confident this is performant
+		Ok(match v {
+			v if v == Self::G2.as_u16() => Self::G2,
+			v if v == Self::G4.as_u16() => Self::G4,
+			v if v == Self::G8.as_u16() => Self::G8,
+			v if v == Self::G16.as_u16() => Self::G16,
+			_ => return Err(InvalidNum),
+		})
+	}
+
+	/// least signficant bits per g
+	pub const fn lsb_per_g(self) -> f32 {
+		let range: f32 = self.as_u16() as _;
+		// Add 1 because there is MAX+1 numbers due to `0`
+		const TMP: f32 = i16::MAX as f32 + 1.;
+		TMP / range
+	}
+
+	/// g per least significant bit
+	pub const fn g_per_lsb(self) -> f32 {
+		let range: f32 = self.as_u16() as _;
+		// Add 1 because there is MAX+1 numbers due to `0`
+		const TMP: f32 = 1. / (i16::MAX as f32 + 1.);
+		range * TMP
+	}
+
+	/// least significant bits per accel
+	pub const fn lsb_per_accel(self) -> f32 {
+		self.lsb_per_g() * ACCEL_PER_G
+	}
+
+	/// g per least significant bit
+	pub const fn accel_per_lsb(self) -> f32 {
+		self.g_per_lsb() * G_PER_ACCEL
+	}
+
+	/// The bmi160 returns the data from the accel as an `i16`, we must use the Full
+	/// Scale Range to convert to a float m/s^2
+	pub const fn discrete_to_accel(self, discrete: i16) -> f32 {
+		discrete as f32 * self.accel_per_lsb()
+	}
 }
 
 #[derive(Debug)]
@@ -92,9 +187,3 @@ pub struct InvalidBitPattern;
 
 #[derive(Debug)]
 pub struct InvalidNum;
-
-/// The bmi160 returns the data from the gyro as an `i16`, we must use the Full Scale Range to
-/// convert to a float.
-pub const fn discrete_to_radians(fsr: GyroFsr, discrete: i16) -> f32 {
-	discrete as f32 * fsr.rad_per_lsb()
-}
diff --git a/firmware/src/imu/drivers/bmi160/mod.rs b/firmware/src/imu/drivers/bmi160/mod.rs
index ec8bffd9..d4efc037 100644
--- a/firmware/src/imu/drivers/bmi160/mod.rs
+++ b/firmware/src/imu/drivers/bmi160/mod.rs
@@ -1,6 +1,6 @@
 mod math;
 
-use self::math::{discrete_to_radians, GyroFsr};
+use self::math::GyroFsr;
 use crate::aliases::I2c;
 use crate::imu::{FusedData, Imu, Quat};
 use crate::utils::{self, nb2a};
@@ -86,9 +86,9 @@ impl<I: I2c> Bmi160<I> {
 		// TODO: Implement sensor fusion and temperature compensation
 		// TODO: This should be integrated to position, lol
 		Ok(nalgebra::UnitQuaternion::from_euler_angles(
-			discrete_to_radians(FSR, gyro_vel_euler.x),
-			discrete_to_radians(FSR, gyro_vel_euler.y),
-			discrete_to_radians(FSR, gyro_vel_euler.z),
+			FSR.discrete_to_velocity(gyro_vel_euler.x),
+			FSR.discrete_to_velocity(gyro_vel_euler.y),
+			FSR.discrete_to_velocity(gyro_vel_euler.z),
 		))
 	}
 }