piotr/edgetx-gx12
1
0
Fork 0
Code Issues 2 Pull Requests Actions Packages Projects Releases Wiki Activity

Avoids splash + checks at startup on wdt reboot

Browse Source
This commit is contained in:
bsongis
2011-09-28 21:32:10 +00:00
parent 3fc4203261
commit 5897553265
1 changed files with 33 additions and 22 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/gruvin9x.cpp
+33 -22
View File
@@ -523,7 +523,7 @@ void doSplash()
}
}
void checkMem()
void checkLowEEPROM()
{
if(g_eeGeneral.disableMemoryWarning) return;
if(EeFsGetFree() < 200)
@@ -845,25 +845,33 @@ void getADC_bandgap()
{
#if defined(PCBSTD)
ADMUX=0x1E|ADC_VREF_TYPE; // Switch MUX to internal 1.22V reference
_delay_us(5); // short delay to stablise reference voltage
_delay_us(10); // short delay to stablise reference voltage
ADCSRA|=0x40; while ((ADCSRA & 0x10)==0); ADCSRA|=0x10; // grab a sample
ADCSRA|=0x40; while ((ADCSRA & 0x10)==0); ADCSRA|=0x10; // grab a sample
ADCSRA|=0x40; while ((ADCSRA & 0x10)==0); ADCSRA|=0x10; // again becasue first one is usually inaccurate
BandGap=ADCW;
#else
//BandGap=225; // gruvin: 1.1V internal Vref doesn't seem to work on the ATmega2561. :/ Weird.
// See http://www.avrfreaks.net/index.php?name=PNphpBB2&file=viewtopic&p=847208#847208
// In the end, simply using a longer delay (presumably to account for the higher
// impedance Vbg internal source) solved the problem. NOTE: Does NOT adversely affect PPM_out latency.
#if defined (PCBV4)
#elif defined (PCBV4)
// For PCB V4, use our own 1.2V, external reference (connected to ADC3)
ADCSRB &= ~(1<<MUX5);
#endif
ADMUX=0x1E|ADC_VREF_TYPE; // Switch MUX to internal 1.1V reference
_delay_us(300); // this somewhat costly delay is the only remedy for stable results on the Atmega2560/1 chips
ADCSRA|=0x40; while ((ADCSRA & 0x10)==0); ADCSRA|=0x10; // again becasue first one is usually inaccurate
ADMUX=0x03|ADC_VREF_TYPE; // Switch MUX to internal 1.1V reference
_delay_us(10); // tiny bit of stablisation time needed to allow capture-hold capacitor to charge
// For times over-sample with no divide, x2 to end at a half averaged, x8. DON'T ASK mmmkay? :P This is how I want it.
ADCSRA|=0x40; while ((ADCSRA & 0x10)==0); ADCSRA|=0x10;
BandGap=ADCW;
#if defined (PCBV4)
ADCSRA|=0x40; while ((ADCSRA & 0x10)==0); ADCSRA|=0x10;
BandGap+=ADCW;
ADCSRA|=0x40; while ((ADCSRA & 0x10)==0); ADCSRA|=0x10;
BandGap+=ADCW;
ADCSRA|=0x40; while ((ADCSRA & 0x10)==0); ADCSRA|=0x10;
BandGap+=ADCW;
BandGap *= 2;
ADCSRB |= (1<<MUX5);
#endif
#else
ADMUX=0x1E|ADC_VREF_TYPE; // Switch MUX to internal 1.1V reference
_delay_us(400); // this somewhat costly delay is the only remedy for stable results on the Atmega2560/1 chips
ADCSRA|=0x40; while ((ADCSRA & 0x10)==0); ADCSRA|=0x10; // take sample
BandGap=ADCW;
#endif
}
@@ -2218,16 +2226,19 @@ int main(void)
eeReadAll();
uint8_t cModel = g_eeGeneral.currModel;
checkQuickSelect();
doSplash();
checkMem();
if (~MCUCSR & (1 << WDRF)) {
checkQuickSelect();
getADC_single();
checkTHR();
doSplash();
checkLowEEPROM();
checkSwitches();
checkAlarm();
getADC_single();
checkTHR();
checkSwitches();
checkAlarm();
}
clearKeyEvents(); //make sure no keys are down before proceeding