Merge branch 'master' of github.com:quartiq/stabilizer

master
Ryan Summers 2021-03-01 14:22:23 +01:00
commit 46fb62e802
13 changed files with 136 additions and 112 deletions

View File

@ -5,7 +5,9 @@ on:
branches: [master, staging, trying]
pull_request:
branches: [master]
schedule:
# UTC
- cron: '48 4 * * *'
env:
CARGO_TERM_COLOR: always

1
Cargo.lock generated
View File

@ -724,6 +724,7 @@ dependencies = [
"dsp",
"embedded-hal",
"enum-iterator",
"generic-array 0.14.4",
"heapless",
"log",
"mcp23017",

View File

@ -46,6 +46,7 @@ dsp = { path = "dsp" }
ad9959 = { path = "ad9959" }
smoltcp-nal = "0.1.0"
miniconf = "0.1"
generic-array = "0.14"
[patch.crates-io.miniconf]
git = "https://github.com/quartiq/miniconf.git"

View File

@ -1,8 +1,9 @@
use core::f32::consts::PI;
use dsp::{atan2, cossin};
use dsp::{iir, iir_int};
use dsp::{PLL, RPLL};
use easybench::bench_env;
use generic_array::typenum::U4;
use dsp::{atan2, cossin, iir, iir_int, Lowpass, PLL, RPLL};
fn atan2_bench() {
let xi = (10 << 16) as i32;
@ -70,6 +71,18 @@ fn iir_bench() {
);
}
fn lowpass_bench() {
let mut dut = Lowpass::<U4>::default();
println!(
"Lowpass::<U4>::update(x, k): {}",
bench_env((0x32421, 14), |(x, k)| dut.update(*x, *k))
);
println!(
"Lowpass::<U4>::update(x, 14): {}",
bench_env(0x32421, |x| dut.update(*x, 14))
);
}
fn main() {
atan2_bench();
cossin_bench();
@ -77,4 +90,5 @@ fn main() {
pll_bench();
iir_int_bench();
iir_bench();
lowpass_bench();
}

View File

@ -15,7 +15,7 @@ fn write_cossin_table() {
.unwrap();
write!(
file,
"pub(crate) const COSSIN: [(u16, u16); 1 << COSSIN_DEPTH] = ["
"pub(crate) const COSSIN: [u32; 1 << COSSIN_DEPTH] = ["
)
.unwrap();
@ -26,15 +26,17 @@ fn write_cossin_table() {
const AMPLITUDE: f64 = u16::MAX as f64;
for i in 0..(1 << DEPTH) {
// use midpoint samples to save one entry in the LUT
let phase = (PI / 4. / (1 << DEPTH) as f64) * (i as f64 + 0.5);
// add one bit accuracy to cos due to 0.5 < cos(z) <= 1 for |z| < pi/4
let cos = ((phase.cos() - 0.5) * 2. * AMPLITUDE).round() as u16;
let sin = (phase.sin() * AMPLITUDE).round() as u16;
if i % 4 == 0 {
write!(file, "\n ").unwrap();
}
write!(file, " ({}, {}),", cos, sin).unwrap();
// Use midpoint samples to save one entry in the LUT
let (sin, cos) =
(PI / 4. * ((i as f64 + 0.5) / (1 << DEPTH) as f64)).sin_cos();
// Add one bit accuracy to cos due to 0.5 < cos(z) <= 1 for |z| < pi/4
// The -1 LSB is cancelled when unscaling with the biased half amplitude
let cos = ((cos * 2. - 1.) * AMPLITUDE - 1.).round() as u32;
let sin = (sin * AMPLITUDE).round() as u32;
write!(file, " {},", cos + (sin << 16)).unwrap();
}
writeln!(file, "\n];").unwrap();

View File

@ -8,6 +8,8 @@ pub trait ComplexExt<T, U> {
fn abs_sqr(&self) -> U;
fn log2(&self) -> T;
fn arg(&self) -> T;
fn saturating_add(&self, other: Self) -> Self;
fn saturating_sub(&self, other: Self) -> Self;
}
impl ComplexExt<i32, u32> for Complex<i32> {
@ -23,7 +25,7 @@ impl ComplexExt<i32, u32> for Complex<i32> {
/// ```
fn from_angle(angle: i32) -> Self {
let (c, s) = cossin(angle);
Self { re: c, im: s }
Self::new(c, s)
}
/// Return the absolute square (the squared magnitude).
@ -85,6 +87,20 @@ impl ComplexExt<i32, u32> for Complex<i32> {
fn arg(&self) -> i32 {
atan2(self.im, self.re)
}
fn saturating_add(&self, other: Self) -> Self {
Self::new(
self.re.saturating_add(other.re),
self.im.saturating_add(other.im),
)
}
fn saturating_sub(&self, other: Self) -> Self {
Self::new(
self.re.saturating_sub(other.re),
self.im.saturating_sub(other.im),
)
}
}
/// Full scale fixed point multiplication.

View File

@ -11,10 +11,10 @@ include!(concat!(env!("OUT_DIR"), "/cossin_table.rs"));
///
/// # Returns
/// The cos and sin values of the provided phase as a `(i32, i32)`
/// tuple. With a 7-bit deep LUT there is 1e-5 max and 6e-8 RMS error
/// tuple. With a 7-bit deep LUT there is 9e-6 max and 4e-6 RMS error
/// in each quadrature over 20 bit phase.
pub fn cossin(phase: i32) -> (i32, i32) {
// Phase bits excluding the three highes MSB
// Phase bits excluding the three highest MSB
const OCTANT_BITS: usize = 32 - 3;
// This is a slightly more compact way to compute the four flags for
@ -22,50 +22,55 @@ pub fn cossin(phase: i32) -> (i32, i32) {
let mut octant = (phase as u32) >> OCTANT_BITS;
octant ^= octant << 1;
// Mask off octant bits. This leaves the angle in the range [0, pi/4).
let mut phase = phase & ((1 << OCTANT_BITS) - 1);
let mut phase = phase;
if octant & 1 != 0 {
// phase = pi/4 - phase
phase = (1 << OCTANT_BITS) - 1 - phase;
phase = !phase;
}
let lookup = COSSIN[(phase >> (OCTANT_BITS - COSSIN_DEPTH)) as usize];
// 1/2 < cos(0 <= x <= pi/4) <= 1: Shift the cos
// values and scale the sine values as encoded in the LUT.
let mut cos = lookup.0 as i32 + u16::MAX as i32;
let mut sin = (lookup.1 as i32) << 1;
// Mask off octant bits. This leaves the angle in the range [0, pi/4).
phase &= (1 << OCTANT_BITS) - 1;
// 16 + 1 bits for cos/sin and 15 for dphi to saturate the i32 range.
const ALIGN_MSB: usize = 32 - 16 - 1;
phase >>= OCTANT_BITS - COSSIN_DEPTH - ALIGN_MSB;
let lookup = COSSIN[(phase >> ALIGN_MSB) as usize];
phase &= (1 << ALIGN_MSB) - 1;
// The phase values used for the LUT are at midpoint for the truncated phase.
// Interpolate relative to the LUT entry midpoint.
phase -= (1 << (ALIGN_MSB - 1)) - (octant & 1) as i32;
// Fixed point pi/4.
const PI4: i32 = (PI / 4. * (1 << (32 - ALIGN_MSB)) as f64) as i32;
// No rounding bias necessary here since we keep enough low bits.
let dphi = (phase * PI4) >> (32 - ALIGN_MSB);
phase -= 1 << (ALIGN_MSB - 1);
// Make room for the sign bit.
let dcos = (sin * dphi) >> (COSSIN_DEPTH + 1);
// Cancel the -1 bias that was conditionally introduced above.
// This lowers the DC spur from 2e-8 to 2e-10 magnitude.
// phase += (octant & 1) as i32;
// Fixed point pi/4.
const PI4: i32 = (PI / 4. * (1 << 16) as f64) as _;
// No rounding bias necessary here since we keep enough low bits.
let dphi = (phase * PI4) >> 16;
// 1/2 < cos(0 <= x <= pi/4) <= 1: Shift the cos
// values and scale the sine values as encoded in the LUT.
let mut cos = (lookup & 0xffff) as i32 + (1 << 16);
let mut sin = (lookup >> 16) as i32;
let dcos = (sin * dphi) >> COSSIN_DEPTH;
let dsin = (cos * dphi) >> (COSSIN_DEPTH + 1);
cos = (cos << (ALIGN_MSB - 1)) - dcos;
sin = (sin << (ALIGN_MSB - 1)) + dsin;
sin = (sin << ALIGN_MSB) + dsin;
// Unmap using octant bits.
if octant & 2 != 0 {
core::mem::swap(&mut sin, &mut cos);
}
if octant & 4 != 0 {
cos *= -1;
cos = -cos;
}
if octant & 8 != 0 {
sin *= -1;
sin = -sin;
}
(cos, sin)
@ -79,7 +84,7 @@ mod tests {
#[test]
fn cossin_error_max_rms_all_phase() {
// Constant amplitude error due to LUT data range.
const AMPLITUDE: f64 = ((1i64 << 31) - (1i64 << 15)) as _;
const AMPLITUDE: f64 = (1i64 << 31) as f64 - 0.85 * (1i64 << 15) as f64;
const MAX_PHASE: f64 = (1i64 << 32) as _;
let mut rms_err = (0f64, 0f64);
let mut sum_err = (0f64, 0f64);
@ -121,8 +126,8 @@ mod tests {
max_err.0 = max_err.0.max(err.0.abs());
max_err.1 = max_err.1.max(err.1.abs());
}
rms_err.0 /= MAX_PHASE;
rms_err.1 /= MAX_PHASE;
rms_err.0 /= (1 << PHASE_DEPTH) as f64;
rms_err.1 /= (1 << PHASE_DEPTH) as f64;
println!("sum: {:.2e} {:.2e}", sum.0, sum.1);
println!("demod: {:.2e} {:.2e}", demod.0, demod.1);
@ -131,18 +136,18 @@ mod tests {
println!("max: {:.2e} {:.2e}", max_err.0, max_err.1);
assert!(sum.0.abs() < 4e-10);
assert!(sum.1.abs() < 4e-10);
assert!(sum.1.abs() < 3e-8);
assert!(demod.0.abs() < 4e-10);
assert!(demod.1.abs() < 4e-10);
assert!(demod.1.abs() < 1e-8);
assert!(sum_err.0.abs() < 4e-10);
assert!(sum_err.1.abs() < 4e-10);
assert!(rms_err.0.sqrt() < 6e-8);
assert!(rms_err.1.sqrt() < 6e-8);
assert!(rms_err.0.sqrt() < 4e-6);
assert!(rms_err.1.sqrt() < 4e-6);
assert!(max_err.0 < 1.1e-5);
assert!(max_err.1 < 1.1e-5);
assert!(max_err.0 < 1e-5);
assert!(max_err.1 < 1e-5);
}
}

View File

@ -1,5 +1,7 @@
use super::tools::macc_i32;
use core::f64::consts::PI;
use serde::{Deserialize, Serialize};
use miniconf::StringSet;
use serde::Deserialize;
/// Generic vector for integer IIR filter.
/// This struct is used to hold the x/y input/output data vector or the b/a coefficient
@ -39,23 +41,12 @@ impl Coeff for Vec5 {
}
}
fn macc(y0: i32, x: &[i32], a: &[i32], shift: u32) -> i32 {
// Rounding bias, half up
let y0 = ((y0 as i64) << shift) + (1 << (shift - 1));
let y = x
.iter()
.zip(a)
.map(|(x, a)| *x as i64 * *a as i64)
.fold(y0, |y, xa| y + xa);
(y >> shift) as i32
}
/// Integer biquad IIR
///
/// See `dsp::iir::IIR` for general implementation details.
/// Offset and limiting disabled to suit lowpass applications.
/// Coefficient scaling fixed and optimized.
#[derive(Copy, Clone, Default, Deserialize, Serialize)]
#[derive(Copy, Clone, Default, Debug, StringSet, Deserialize)]
pub struct IIR {
pub ba: Vec5,
// pub y_offset: i32,
@ -85,7 +76,7 @@ impl IIR {
// Store x0 x0 x1 x2 y1 y2
xy[0] = x0;
// Compute y0 by multiply-accumulate
let y0 = macc(0, xy, &self.ba, IIR::SHIFT);
let y0 = macc_i32(0, xy, &self.ba, IIR::SHIFT);
// Limit y0
// let y0 = y0.max(self.y_min).min(self.y_max);
// Store y0 x0 x1 y0 y1 y2

View File

@ -1,12 +1,12 @@
use super::{Complex, ComplexExt, Lowpass, MulScaled};
use generic_array::typenum::U2;
use generic_array::ArrayLength;
#[derive(Clone, Default)]
pub struct Lockin {
state: [Lowpass<U2>; 2],
pub struct Lockin<N: ArrayLength<i32>> {
state: [Lowpass<N>; 2],
}
impl Lockin {
impl<N: ArrayLength<i32>> Lockin<N> {
/// Update the lockin with a sample taken at a given phase.
pub fn update(&mut self, sample: i32, phase: i32, k: u8) -> Complex<i32> {
// Get the LO signal for demodulation and mix the sample;

View File

@ -27,10 +27,10 @@ impl<N: ArrayLength<i32>> Lowpass<N> {
// Note T-DF-I and the zeros at Nyquist.
let mut x = x;
for y in self.y.iter_mut() {
let dy = x.saturating_sub(*y).saturating_add(1 << (k - 1)) >> k;
let dy = x.saturating_sub(*y) >> k;
*y += dy;
x = *y - (dy >> 1);
}
x
x.saturating_add((self.y.len() as i32) << (k - 1))
}
}

View File

@ -77,6 +77,17 @@ where
.fold(y0, |y, xa| y + xa)
}
pub fn macc_i32(y0: i32, x: &[i32], a: &[i32], shift: u32) -> i32 {
// Rounding bias, half up
let y0 = ((y0 as i64) << shift) + (1 << (shift - 1));
let y = x
.iter()
.zip(a)
.map(|(x, a)| *x as i64 * *a as i64)
.fold(y0, |y, xa| y + xa);
(y >> shift) as i32
}
/// Combine high and low i32 into a single downscaled i32, saturating the type.
pub fn saturating_scale(lo: i32, hi: i32, shift: u32) -> i32 {
debug_assert!(shift & 31 == shift);

View File

@ -2,14 +2,12 @@
#![no_std]
#![no_main]
use stm32h7xx_hal as hal;
use stabilizer::{hardware, hardware::design_parameters};
use dsp::{Accu, Complex, ComplexExt, Lockin, RPLL};
use generic_array::typenum::U4;
use hardware::{
Adc0Input, Adc1Input, Dac0Output, Dac1Output, InputStamper, AFE0, AFE1,
};
use stabilizer::{hardware, hardware::design_parameters};
#[rtic::app(device = stm32h7xx_hal::stm32, peripherals = true, monotonic = rtic::cyccnt::CYCCNT)]
const APP: () = {
@ -20,7 +18,7 @@ const APP: () = {
timestamper: InputStamper,
pll: RPLL,
lockin: Lockin,
lockin: Lockin<U4>,
}
#[init]
@ -88,22 +86,20 @@ const APP: () = {
.map(|t| t as i32);
let (pll_phase, pll_frequency) = c.resources.pll.update(
timestamp,
21, // frequency settling time (log2 counter cycles), TODO: expose
21, // phase settling time, TODO: expose
21, // frequency settling time (log2 counter cycles),
21, // phase settling time
);
// Harmonic index of the LO: -1 to _de_modulate the fundamental (complex conjugate)
let harmonic: i32 = -1; // TODO: expose
let harmonic: i32 = -1;
// Demodulation LO phase offset
let phase_offset: i32 = 0; // TODO: expose
let phase_offset: i32 = 0;
// Log2 lowpass time constant
let time_constant: u8 = 6; // TODO: expose
let time_constant: u8 = 6;
let sample_frequency = ((pll_frequency
// half-up rounding bias
// .wrapping_add(1 << design_parameters::SAMPLE_BUFFER_SIZE_LOG2 - 1)
>> design_parameters::SAMPLE_BUFFER_SIZE_LOG2)
as i32)
.wrapping_mul(harmonic);
@ -131,7 +127,7 @@ const APP: () = {
Quadrature,
}
let conf = Conf::FrequencyDiscriminator; // TODO: expose
let conf = Conf::FrequencyDiscriminator;
let output = match conf {
// Convert from IQ to power and phase.
Conf::PowerPhase => [(output.log2() << 24) as _, output.arg()],
@ -149,14 +145,13 @@ const APP: () = {
#[idle(resources=[afes])]
fn idle(_: idle::Context) -> ! {
loop {
// TODO: Implement network interface.
cortex_m::asm::wfi();
}
}
#[task(binds = ETH, priority = 1)]
fn eth(_: eth::Context) {
unsafe { hal::ethernet::interrupt_handler() }
unsafe { stm32h7xx_hal::ethernet::interrupt_handler() }
}
#[task(binds = SPI2, priority = 3)]
@ -166,7 +161,7 @@ const APP: () = {
#[task(binds = SPI3, priority = 3)]
fn spi3(_: spi3::Context) {
panic!("ADC0 input overrun");
panic!("ADC1 input overrun");
}
#[task(binds = SPI4, priority = 3)]

View File

@ -3,6 +3,7 @@
#![no_main]
use dsp::{Accu, Complex, ComplexExt, Lockin};
use generic_array::typenum::U2;
use hardware::{Adc1Input, Dac0Output, Dac1Output, AFE0, AFE1};
use stabilizer::{hardware, hardware::design_parameters};
@ -20,7 +21,7 @@ const APP: () = {
adc: Adc1Input,
dacs: (Dac0Output, Dac1Output),
lockin: Lockin,
lockin: Lockin<U2>,
}
#[init]
@ -44,24 +45,13 @@ const APP: () = {
}
}
/// Main DSP processing routine for Stabilizer.
/// Main DSP processing routine.
///
/// # Note
/// Processing time for the DSP application code is bounded by the following constraints:
/// See `dual-iir` for general notes on processing time and timing.
///
/// DSP application code starts after the ADC has generated a batch of samples and must be
/// completed by the time the next batch of ADC samples has been acquired (plus the FIFO buffer
/// time). If this constraint is not met, firmware will panic due to an ADC input overrun.
///
/// The DSP application code must also fill out the next DAC output buffer in time such that the
/// DAC can switch to it when it has completed the current buffer. If this constraint is not met
/// it's possible that old DAC codes will be generated on the output and the output samples will
/// be delayed by 1 batch.
///
/// Because the ADC and DAC operate at the same rate, these two constraints actually implement
/// the same time bounds, meeting one also means the other is also met.
///
/// TODO: Document
/// This is an implementation of an internal-reference lockin on the ADC1 signal.
/// The reference at f_sample/8 is output on DAC0 and the phase of the demodulated
/// signal on DAC1.
#[task(binds=DMA1_STR4, resources=[adc, dacs, lockin], priority=2)]
fn process(c: process::Context) {
let lockin = c.resources.lockin;
@ -71,29 +61,23 @@ const APP: () = {
c.resources.dacs.1.acquire_buffer(),
];
// DAC0 always generates a fixed sinusoidal output.
dac_samples[0]
.iter_mut()
.zip(DAC_SEQUENCE.iter())
.for_each(|(d, s)| *d = *s as u16 ^ 0x8000);
// Reference phase and frequency are known.
let pll_phase = 0;
let pll_phase = 0i32;
let pll_frequency =
1i32 << (32 - design_parameters::SAMPLE_BUFFER_SIZE_LOG2);
// Harmonic index of the LO: -1 to _de_modulate the fundamental
let harmonic: i32 = -1; // TODO: expose
// Harmonic index of the LO: -1 to _de_modulate the fundamental (complex conjugate)
let harmonic: i32 = -1;
// Demodulation LO phase offset
let phase_offset: i32 = (0.25 * i32::MAX as f32) as i32; // TODO: expose
let phase_offset: i32 = 1 << 30;
// Log2 lowpass time constant.
let time_constant: u8 = 8;
let sample_frequency = (pll_frequency as i32).wrapping_mul(harmonic);
let sample_phase = phase_offset
.wrapping_add((pll_phase as i32).wrapping_mul(harmonic));
let sample_phase =
phase_offset.wrapping_add(pll_phase.wrapping_mul(harmonic));
let output: Complex<i32> = adc_samples
.iter()
@ -109,15 +93,17 @@ const APP: () = {
.unwrap()
* 2; // Full scale assuming the 2f component is gone.
for value in dac_samples[1].iter_mut() {
*value = (output.arg() >> 16) as u16 ^ 0x8000;
// Convert to DAC data.
for (i, data) in DAC_SEQUENCE.iter().enumerate() {
// DAC0 always generates a fixed sinusoidal output.
dac_samples[0][i] = *data as u16 ^ 0x8000;
dac_samples[1][i] = (output.arg() >> 16) as u16 ^ 0x8000;
}
}
#[idle(resources=[afes])]
fn idle(_: idle::Context) -> ! {
loop {
// TODO: Implement network interface.
cortex_m::asm::wfi();
}
}