kopia lustrzana https://github.com/threeme3/usdx
Changed ATT2 range (down to 2) to prevent Hilbert transform from overload. Add new AGC algorithm. Add volume control becomes gain control when AGC is off. Change accuracy and easier to read smeter (TNX Alan, M0PUB).
guido
rodzic
23d7d4e522
commit
25c3b35356
141
QCX-SSB.ino
141
QCX-SSB.ino
|
|
@ -13,8 +13,8 @@
|
|||
#define LCD_D7 3 //PD3 (pin 5)
|
||||
#define LCD_EN 4 //PD4 (pin 6)
|
||||
#define FREQCNT 5 //PD5 (pin 11)
|
||||
#define ROT_A 6 //PD6 (pin 12)
|
||||
#define ROT_B 7 //PD7 (pin 13)
|
||||
#define ROT_A 7 //PD6 (pin 12)
|
||||
#define ROT_B 6 //PD7 (pin 13)
|
||||
#define RX 8 //PB0 (pin 14)
|
||||
#define SIDETONE 9 //PB1 (pin 15)
|
||||
#define KEY_OUT 10 //PB2 (pin 16)
|
||||
|
|
@ -893,8 +893,8 @@ public:
|
|||
|
||||
#define FAST __attribute__((optimize("Ofast")))
|
||||
|
||||
#define F_XTAL 27005000 // Crystal freq in Hz, nominal frequency 27004300
|
||||
//#define F_XTAL 25004000 // Alternate SI clock
|
||||
//#define F_XTAL 27005000 // Crystal freq in Hz, nominal frequency 27004300
|
||||
#define F_XTAL 25004000 // Alternate SI clock
|
||||
//#define F_XTAL 20004000 // A shared-single 20MHz processor/pll clock
|
||||
volatile uint32_t fxtal = F_XTAL;
|
||||
|
||||
|
|
@ -1568,22 +1568,73 @@ volatile bool cw_event = false;
|
|||
volatile bool agc = true;
|
||||
volatile uint8_t nr = 0;
|
||||
volatile uint8_t att = 0;
|
||||
volatile uint8_t att2 = 0;
|
||||
volatile uint8_t att2 = 2; // M0PUB: Minimum att2 increased from 0 to 2, to prevent numeric overflow on strong signals
|
||||
volatile uint8_t _init;
|
||||
|
||||
//static uint32_t gain = 1024;
|
||||
/* M0PUB: Old AGC algorithm which only increases gain, but does not decrease it for very strong signals.
|
||||
// Maximum possible gain is x32 (in practice, x31) so AGC range is x1 to x31 = 30dB approx.
|
||||
// Decay time is fine (about 1s) but attack time is much slower than I like.
|
||||
// For weak/medium signals it aims to keep the sample value between 1024 and 2048.
|
||||
static int16_t gain = 1024;
|
||||
inline int16_t process_agc(int16_t in)
|
||||
{
|
||||
//int16_t out = ((uint32_t)(gain) >> 20) * in;
|
||||
//gain = gain + (1024 - abs(out) + 512);
|
||||
int16_t out = (gain >= 1024) ? (gain >> 10) * in : in;
|
||||
//if(gain >= 1024) out = (gain >> 10) * in; // net gain >= 1
|
||||
//else if(gain >= 16) out = ((gain >> 4) * in) >> 6; // net gain < 1
|
||||
//else out = (gain * in) >> 10;
|
||||
int16_t accum = (1 - abs(out >> 10));
|
||||
if((INT16_MAX - gain) > accum) gain = gain + accum;
|
||||
if(gain < 1) gain = 1;
|
||||
return out;
|
||||
} */
|
||||
|
||||
// M0PUB: Experimental new AGC algorithm.
|
||||
// ASSUMES: Input sample values are constrained to a maximum of +/-4096 to avoid integer overflow in earlier
|
||||
// calculations.
|
||||
//
|
||||
// This algorithm aims to keep signals between a peak sample value of 1024 - 1536, with fast attack but slow
|
||||
// decay.
|
||||
//
|
||||
// The variable centiGain actually represents the applied gain x 128 - i.e. the numeric gain applied is centiGain/128
|
||||
//
|
||||
// Since the largest valid input sample has a value of +/- 4096, centiGain should never be less than 32 (i.e.
|
||||
// a 'gain' of 0.25). The maximum value for centiGain is 32767, and hence a gain of 255. So the AGC range
|
||||
// is 0.25:255, or approx. 60dB.
|
||||
//
|
||||
// Variable 'slowdown' allows the decay time to be slowed down so that it is not directly related to the value
|
||||
// of centiCount.
|
||||
|
||||
static int16_t centiGain = 128;
|
||||
#define DECAY_FACTOR 400 // AGC decay occurs <DECAY_FACTOR> slower than attack.
|
||||
static uint16_t decayCount = DECAY_FACTOR;
|
||||
#define HI(x) ((x) >> 8)
|
||||
#define LO(x) ((x) & 0xFF)
|
||||
|
||||
inline int16_t process_agc(int16_t in)
|
||||
{
|
||||
static bool small = true;
|
||||
int16_t out;
|
||||
|
||||
if (centiGain >= 128)
|
||||
out = (centiGain >> 5) * in; // net gain >= 1
|
||||
else
|
||||
out = (centiGain >> 2) * (in >> 3); // net gain < 1
|
||||
out >>= 2;
|
||||
|
||||
if (HI(abs(out)) > HI(1536)) {
|
||||
centiGain -= (centiGain >> 4); // Fast attack time when big signal encountered (relies on CentiGain >= 16)
|
||||
} else {
|
||||
if (HI(abs(out)) > HI(1024))
|
||||
small = false;
|
||||
if (--decayCount == 0) { // But slow ramp up of gain when signal disappears
|
||||
if (small) { // 400 samples below lower threshold - increase gain
|
||||
if (centiGain < (INT16_MAX-(INT16_MAX >> 4)))
|
||||
centiGain += (centiGain >> 4);
|
||||
else
|
||||
centiGain = INT16_MAX;
|
||||
}
|
||||
decayCount = DECAY_FACTOR;
|
||||
small = true;
|
||||
}
|
||||
}
|
||||
|
||||
return out;
|
||||
}
|
||||
|
||||
|
|
@ -1730,7 +1781,10 @@ volatile int16_t i, q;
|
|||
inline int16_t slow_dsp(int16_t ac)
|
||||
{
|
||||
static uint8_t absavg256cnt;
|
||||
if(!(absavg256cnt--)){ _absavg256 = absavg256; absavg256 = 0;
|
||||
if(!(absavg256cnt--)){
|
||||
_absavg256 = absavg256;
|
||||
absavg256 = 0;
|
||||
|
||||
//#define AUTO_ADC_BIAS 1
|
||||
#ifdef AUTO_ADC_BIAS
|
||||
if(param_b < 0){
|
||||
|
|
@ -1767,8 +1821,16 @@ inline int16_t slow_dsp(int16_t ac)
|
|||
//ac = ac * (F_SAMP_RX/R) / _UA; // =ac*3.5 -> skip
|
||||
} // needs: p.12 https://www.veron.nl/wp-content/uploads/2014/01/FmDemodulator.pdf
|
||||
else { ; } // USB, LSB, CW
|
||||
if(agc) ac = process_agc(ac);
|
||||
ac = ac >> (16-volume);
|
||||
|
||||
if(agc) {
|
||||
ac = process_agc(ac);
|
||||
ac = ac >> (16-volume);
|
||||
} else {
|
||||
if (volume <= 9) // M0PUB: if no AGC allow volume control to boost weak signals
|
||||
ac = ac >> (9-volume);
|
||||
else
|
||||
ac = ac << (volume-9);
|
||||
}
|
||||
if(nr) ac = process_nr(ac);
|
||||
|
||||
if(filt) ac = filt_var(ac) << 2;
|
||||
|
|
@ -1909,7 +1971,9 @@ void sdr_rx_q()
|
|||
// Process Q (down-sampled) samples
|
||||
static int16_t v[14];
|
||||
q = v[7];
|
||||
qh = ((v[0] - ac2) * 2 + (v[2] - v[12]) * 8 + (v[4] - v[10]) * 21 + (v[6] - v[8]) * 15) / 128 + (v[6] - v[8]) / 2; // Hilbert transform, 40dB side-band rejection in 400..1900Hz (@4kSPS) when used in image-rejection scenario; (Hilbert transform require 5 additional bits)
|
||||
// qh = ((v[0] - ac2) * 2 + (v[2] - v[12]) * 8 + (v[4] - v[10]) * 21 + (v[6] - v[8]) * 15) / 128 + (v[6] - v[8]) / 2; // Old Hilbert transform, 40dB side-band rejection in 400..1900Hz (@8kSPS) when used in image-rejection scenario; (Hilbert transform require 5 additional bits)
|
||||
qh = ((v[0] - ac2) + (v[2] - v[12]) * 4) / 64 + ((v[4] - v[10]) + (v[6] - v[8])) / 8 + ((v[4] - v[10]) * 5 - (v[6] - v[8])) / 128 + (v[6] - v[8]) / 2; // New Hilbert transform, 40dB side-band rejection in 400..1900Hz (@8kSPS).
|
||||
// Same coefficients as before, but refactored to reduce numeric overflow.
|
||||
for(uint8_t j = 0; j != 13; j++) v[j] = v[j + 1]; v[13] = ac2;
|
||||
}
|
||||
rx_state = 0; return;
|
||||
|
|
@ -2232,38 +2296,48 @@ uint16_t analogSampleMic()
|
|||
volatile bool change = true;
|
||||
volatile int32_t freq = 7074000;
|
||||
|
||||
int8_t smode = 1;
|
||||
// M0PUB: We measure the average amplitude of the signal (see slow_dsp()) but the S-meter should be based on RMS value.
|
||||
// So we multiply by 0.707/0.639 in an attempt to roughly compensate, although that only really works if the input
|
||||
// is a sine wave
|
||||
|
||||
float dbm_max;
|
||||
float smeter(float ref = 5) //= 10*log(8000/2400)=5 ref to 2.4kHz BW. plus some other calibration factor
|
||||
int8_t smode = 1;
|
||||
float dbm_max = -140.0;
|
||||
|
||||
float smeter(float ref = 0) // M0PUB: ref was 5 (= 10*log(8000/2400)) but I don't think that is correct?
|
||||
{
|
||||
if(smode == 0){ // none, no s-meter
|
||||
return 0;
|
||||
}
|
||||
float rms = _absavg256 / 256.0; //sqrt(256.0);
|
||||
|
||||
//if(dsp_cap == SDR) rms = (float)rms * 1.1 * (float)(1 << att2) / (1024.0 * (float)R * 4.0 * 100.0 * 40.0); // 2 rx gain stages: rmsV = ADC value * AREF / [ADC DR * processing gain * receiver gain * audio gain]
|
||||
if(dsp_cap == SDR) rms = (float)rms * 1.1 * (float)(1 << att2) / (1024.0 * (float)R * 4.0 * 820.0 * 3.0/*??*/); // 1 rx gain stage: rmsV = ADC value * AREF / [ADC DR * processing gain * receiver gain * audio gain]
|
||||
if(dsp_cap == SDR) rms = (float)rms * 1.1 * (float)(1 << att2) / (1024.0 * (float)R * 8.0 * 500.0 * 0.639 / 0.707); // M0PUB updated version: 1 rx gain stage: rmsV = ADC value * AREF / [ADC DR * processing gain * receiver gain * "RMS compensation"]
|
||||
else rms = (float)rms * 5.0 * (float)(1 << att2) / (1024.0 * (float)R * 2.0 * 100.0 * 120.0 / 1.750);
|
||||
float dbm = (10.0 * log10((rms * rms) / 50.0) + 30.0) - ref; //from rmsV to dBm at 50R
|
||||
|
||||
dbm_max = max(dbm_max, dbm);
|
||||
static uint8_t cnt;
|
||||
cnt++;
|
||||
if((cnt % 8) == 0){
|
||||
if((cnt % 32) == 0){ // M0PUB: slowed down display slightly
|
||||
Serial.print((int)dbm_max);
|
||||
Serial.print(" ");
|
||||
|
||||
if(smode == 1){ // dBm meter
|
||||
lcd.noCursor(); lcd.setCursor(9, 0); lcd.print((int16_t)dbm_max); lcd.print(F("dBm "));
|
||||
}
|
||||
if(smode == 2){ // S-meter
|
||||
uint8_t s = (dbm_max < -63) ? ((dbm_max - -127) / 6) : (uint8_t)(dbm_max - -63 + 10) % 10; // dBm to S
|
||||
lcd.noCursor(); lcd.setCursor(14, 0); if(s < 10){ lcd.print('S'); } lcd.print(s);
|
||||
uint8_t s = (dbm_max < -63) ? ((dbm_max - -127) / 6) : (((uint8_t)(dbm_max - -73)) / 10) * 10; // dBm to S (M0PUB: modified to work correctly above S9)
|
||||
lcd.noCursor(); lcd.setCursor(14, 0); if(s < 10){ lcd.print('S'); } lcd.print(s);
|
||||
Serial.println(s);
|
||||
}
|
||||
dbm_max = -174.0 + 34.0;
|
||||
}
|
||||
if(smode == 3){ // S-bar
|
||||
int8_t s = (dbm < -63) ? ((dbm - -127) / 6) : (uint8_t)(dbm - -63 + 10) % 10; // dBm to S
|
||||
lcd.noCursor(); lcd.setCursor(12, 0);
|
||||
char tmp[5];
|
||||
for(uint8_t i = 0; i != 4; i++){ tmp[i] = max(2, min(5, s + 1)); s = s - 3; } tmp[4] = 0;
|
||||
lcd.print(tmp);
|
||||
if(smode == 3){ // S-bar. M0PUB: converted to use dbm_max as well - previously just used dbm
|
||||
int8_t s = (dbm_max < -63) ? ((dbm_max - -127) / 6) : (((int8_t)(dbm_max - -73)) / 10) * 10; // dBm to S (M0PUB: modified to work correctly above S9)
|
||||
lcd.noCursor(); lcd.setCursor(12, 0);
|
||||
char tmp[5];
|
||||
for(uint8_t i = 0; i != 4; i++){ tmp[i] = max(2, min(5, s + 1)); s = s - 3; } tmp[4] = 0;
|
||||
lcd.print(tmp);
|
||||
}
|
||||
dbm_max = dbm_max - 2.0; // Implement peak hold/decay for all meter types
|
||||
}
|
||||
return dbm;
|
||||
}
|
||||
|
|
@ -2559,7 +2633,7 @@ void paramAction(uint8_t action, uint8_t id = ALL) // list of parameters
|
|||
case AGC: paramAction(action, agc, F("1.6"), F("AGC"), offon_label, 0, 1, false); break;
|
||||
case NR: paramAction(action, nr, F("1.7"), F("NR"), NULL, 0, 8, false); break;
|
||||
case ATT: paramAction(action, att, F("1.8"), F("ATT"), att_label, 0, 7, false); break;
|
||||
case ATT2: paramAction(action, att2, F("1.9"), F("ATT2"), NULL, 0, 16, false); break;
|
||||
case ATT2: paramAction(action, att2, F("1.9"), F("ATT2"), NULL, 2, 16, false); break; // M0PUB: Minimum att2 increased from 0 to 2, to prevent numeric overflow on strong signals
|
||||
case SMETER: paramAction(action, smode, F("1.10"), F("S-meter"), smode_label, 0, _N(smode_label) - 1, false); break;
|
||||
case CWDEC: paramAction(action, cwdec, F("2.1"), F("CW Decoder"), offon_label, 0, 1, false); break;
|
||||
case CWTONE: paramAction(action, cw_tone, F("2.2"), F("CW Tone"), cw_tone_label, 0, 1, false); break;
|
||||
|
|
@ -2715,7 +2789,7 @@ void setup()
|
|||
//ssb_cap = 0; dsp_cap = 0; // force standard QCX capability
|
||||
//ssb_cap = 1; dsp_cap = 0; // force SSB and standard QCX-RX capability
|
||||
//ssb_cap = 1; dsp_cap = 1; // force SSB and DSP capability
|
||||
//ssb_cap = 1; dsp_cap = 2; // force SSB and SDR capability
|
||||
ssb_cap = 1; dsp_cap = 2; // !!! TEMP !!! force SSB and SDR capability
|
||||
|
||||
show_banner();
|
||||
lcd.setCursor(7, 0); lcd.print(F(" R")); lcd.print(F(VERSION)); lcd_blanks();
|
||||
|
|
@ -2921,7 +2995,7 @@ void loop()
|
|||
delay(1);
|
||||
#endif
|
||||
|
||||
if(millis() > sec_event_time){
|
||||
if (millis() > sec_event_time) {
|
||||
sec_event_time = millis() + 1000; // schedule time next second
|
||||
|
||||
//#define LCD_REINIT
|
||||
|
|
@ -3312,3 +3386,4 @@ keyer dash-dot
|
|||
tiny-click removal, DC offset correction
|
||||
|
||||
*/
|
||||
|
||||
|
|
|
|||
Ładowanie…
Reference in New Issue