mirror of
https://github.com/TomHarte/CLK.git
synced 2024-11-22 12:33:29 +00:00
Adds an extended rationale for current implementation.
Also strips some cruft of prior guesses.
This commit is contained in:
parent
1ae3799aee
commit
9bbccd89d3
@ -63,6 +63,7 @@ uint8_t IWM::read(int address) {
|
||||
|
||||
if(data_register_ & 0x80) {
|
||||
// printf("\n\nIWM:%02x\n\n", data_register_);
|
||||
// printf(".");
|
||||
data_register_ = 0;
|
||||
}
|
||||
// LOG("Reading data register: " << PADHEX(2) << int(result));
|
||||
|
@ -23,26 +23,37 @@ void DriveSpeedAccumulator::post_sample(uint8_t sample) {
|
||||
if(sample_pointer_ == samples_.size()) {
|
||||
sample_pointer_ = 0;
|
||||
|
||||
// Treat 33 as a zero point and count zero crossings; then approximate
|
||||
// the RPM from the frequency of those.
|
||||
int samples_over = 0;
|
||||
int sum = 0;
|
||||
const uint8_t centre = 33;
|
||||
for(size_t c = 0; c < 512; ++c) {
|
||||
if(samples_[c] > centre) ++ samples_over;
|
||||
sum += samples_[c];
|
||||
}
|
||||
// The below fits for a function like `a + bc`; it encapsultes the following
|
||||
// beliefs:
|
||||
//
|
||||
// (i) motor speed is proportional to voltage supplied;
|
||||
// (ii) with pulse-width modulation it's therefore proportional to the duty cycle;
|
||||
// (iii) the Mac pulse-width modulates whatever it reads from the disk speed buffer;
|
||||
// (iv) ... subject to software pulse-width modulation of that pulse-width modulation.
|
||||
//
|
||||
// So, I believe current motor speed is proportional to a low-pass filtering of
|
||||
// the speed buffer. Which I've implemented very coarsely via 'large' bucketed-averages,
|
||||
// noting also that exact disk motor speed is always a little approximate.
|
||||
|
||||
// Sum all samples.
|
||||
// TODO: if the above is the correct test, do it on sample receipt rather than
|
||||
// bothering with an intermediate buffer.
|
||||
int sum = 0;
|
||||
for(auto s: samples_) {
|
||||
sum += s;
|
||||
}
|
||||
|
||||
// The below fits for a function like `a + bc`.
|
||||
const float rotation_speed = (float(sum) * 0.052896440564137f) - 259.0f;
|
||||
// The formula below was derived from observing values the Mac wrote into its
|
||||
// disk-speed buffer. Given that it runs a calibration loop before doing so,
|
||||
// I cannot guarantee the accuracy of these numbers beyond being within the
|
||||
// range that the computer would accept.
|
||||
const float normalised_sum = float(sum) / float(samples_.size());
|
||||
const float rotation_speed = (normalised_sum * 27.08f) - 259.0f;
|
||||
|
||||
for(int c = 0; c < number_of_drives_; ++c) {
|
||||
drives_[c]->set_rotation_speed(rotation_speed);
|
||||
}
|
||||
// printf("RPM: %0.2f (%d over; %d sum)\n", rotation_speed, samples_over, sum);
|
||||
// printf("RPM: %0.2f (%d sum)\n", rotation_speed, sum);
|
||||
}
|
||||
}
|
||||
|
||||
|
@ -33,7 +33,7 @@ class DriveSpeedAccumulator {
|
||||
void add_drive(Apple::Macintosh::DoubleDensityDrive *drive);
|
||||
|
||||
private:
|
||||
std::array<uint8_t, 512> samples_;
|
||||
std::array<uint8_t, 20> samples_;
|
||||
std::size_t sample_pointer_ = 0;
|
||||
Apple::Macintosh::DoubleDensityDrive *drives_[2] = {nullptr, nullptr};
|
||||
int number_of_drives_ = 0;
|
||||
|
Loading…
Reference in New Issue
Block a user