/* * TipThermoModel.cpp * * Created on: 7 Oct 2019 * Author: ralim */ #include "TipThermoModel.h" #include "Settings.h" #include "BSP.h" #include "power.hpp" #include "../../configuration.h" #include "main.hpp" /* * The hardware is laid out as a non-inverting op-amp * There is a pullup of 39k(TS100) from the +ve input to 3.9V (1M pulup on TS100) * * The simplest case to model this, is to ignore the pullup resistors influence, and assume that its influence is mostly constant * -> Tip resistance *does* change with temp, but this should be much less than the rest of the system. * * When a thermocouple is equal temperature at both sides (hot and cold junction), then the output should be 0uV * Therefore, by measuring the uV when both are equal, the measured reading is the offset value. * This is a mix of the pull-up resistor, combined with tip manufacturing differences. * * All of the thermocouple readings are based on this expired patent * - > https://patents.google.com/patent/US6087631A/en * * This was bought to my attention by */ uint32_t TipThermoModel::convertTipRawADCTouV(uint16_t rawADC) { // This takes the raw ADC samples, converts these to uV // Then divides this down by the gain to convert to the uV on the input to the op-amp (A+B terminals) // Then remove the calibration value that is stored as a tip offset uint32_t vddRailmVX10 = 33000; //The vreg is +-2%, but we have no higher accuracy available // 4096 * 8 readings for full scale // Convert the input ADC reading back into mV times 10 format. uint32_t rawInputmVX10 = (rawADC * vddRailmVX10) / (4096 * 8); uint32_t valueuV = rawInputmVX10 * 100; // shift into uV //Now to divide this down by the gain valueuV /= OP_AMP_GAIN_STAGE; if (systemSettings.CalibrationOffset) { //Remove uV tipOffset if (valueuV >= systemSettings.CalibrationOffset) valueuV -= systemSettings.CalibrationOffset; else valueuV = 0; } // Bias removal (Compensating for a temperature related non-linearity // This uses the target temperature for the tip to calculate a compensation value for temperature related bias // This is not entirely ideal as this means we will be wrong on heat up, but will settle to the correct value // This will cause us to underread on the heatup until we reach the target temp // Compensation (uV)== ((((80+150*(target_temp_c_x10-1000)/3000)*vddRailmVX10)/4096)*100)/GAIN // Reordered with Wolframalpha if (currentTempTargetDegC > 0) { uint32_t compensation = (20625 * ((currentTempTargetDegC * 10) + 600)) / 512; compensation /= OP_AMP_GAIN_STAGE; if (valueuV > compensation) { valueuV -= compensation; } } return valueuV; } uint32_t TipThermoModel::convertTipRawADCToDegC(uint16_t rawADC) { return convertuVToDegC(convertTipRawADCTouV(rawADC)); } #ifdef ENABLED_FAHRENHEIT_SUPPORT uint32_t TipThermoModel::convertTipRawADCToDegF(uint16_t rawADC) { return convertuVToDegF(convertTipRawADCTouV(rawADC)); } #endif //Table that is designed to be walked to find the best sample for the lookup //Extrapolate between two points // [x1, y1] = point 1 // [x2, y2] = point 2 // x = input value // output is x's extrapolated y value int32_t LinearInterpolate(int32_t x1, int32_t y1, int32_t x2, int32_t y2, int32_t x) { return y1 + (((((x - x1) * 1000) / (x2 - x1)) * (y2 - y1))) / 1000; } const uint16_t uVtoDegC[] = { // // 0, 0, // 175, 10, // 381, 20, // 587, 30, // 804, 40, // 1005, 50, // 1007, 60, // 1107, 70, // 1310, 80, // 1522, 90, // 1731, 100, // 1939, 110, // 2079, 120, // 2265, 130, // 2470, 140, // 2676, 150, // 2899, 160, // 3081, 170, // 3186, 180, // 3422, 190, // 3622, 200, // 3830, 210, // 4044, 220, // 4400, 230, // 4691, 240, // 4989, 250, // 5289, 260, // 5583, 270, // 5879, 280, // 6075, 290, // 6332, 300, // 6521, 310, // 6724, 320, // 6929, 330, // 7132, 340, // 7356, 350, // 7561, 360, // 7774, 370, // 7992, 380, // 8200, 390, // 8410, 400, // 8626, 410, // 8849, 420, // 9060, 430, // 9271, 440, // 9531, 450, // 9748, 460, // 10210, 470, // 10219, 480, // 10429, 490, // 10649, 500, // }; uint32_t TipThermoModel::convertuVToDegC(uint32_t tipuVDelta) { if (tipuVDelta) { int noItems = sizeof(uVtoDegC) / (2 * sizeof(uint16_t)); for (int i = 1; i < (noItems - 1); i++) { //If current tip temp is less than current lookup, then this current loopup is the higher point to interpolate if (tipuVDelta < uVtoDegC[i * 2]) { return LinearInterpolate(uVtoDegC[(i - 1) * 2], uVtoDegC[((i - 1) * 2) + 1], uVtoDegC[i * 2], uVtoDegC[(i * 2) + 1], tipuVDelta); } } return LinearInterpolate(uVtoDegC[(noItems - 2) * 2], uVtoDegC[((noItems - 2) * 2) + 1], uVtoDegC[(noItems - 1) * 2], uVtoDegC[((noItems - 1) * 2) + 1], tipuVDelta); } return 0; } #ifdef ENABLED_FAHRENHEIT_SUPPORT uint32_t TipThermoModel::convertuVToDegF(uint32_t tipuVDelta) { return convertCtoF(convertuVToDegC(tipuVDelta)); } uint32_t TipThermoModel::convertCtoF(uint32_t degC) { //(Y °C × 9/5) + 32 =Y°F return 32 + ((degC * 9) / 5); } uint32_t TipThermoModel::convertFtoC(uint32_t degF) { //(Y°F − 32) × 5/9 = Y°C if (degF < 32) return 0; return ((degF - 32) * 5) / 9; } #endif uint32_t TipThermoModel::getTipInC(bool sampleNow) { int32_t currentTipTempInC = TipThermoModel::convertTipRawADCToDegC(getTipRawTemp(sampleNow)); currentTipTempInC += getHandleTemperature() / 10; //Add handle offset // Power usage indicates that our tip temp is lower than our thermocouple temp. // I found a number that doesn't unbalance the existing PID, causing overshoot. // This could be tuned in concert with PID parameters... currentTipTempInC -= x10WattHistory.average() / 25; if (currentTipTempInC < 0) return 0; return currentTipTempInC; } #ifdef ENABLED_FAHRENHEIT_SUPPORT uint32_t TipThermoModel::getTipInF(bool sampleNow) { uint32_t currentTipTempInF = TipThermoModel::convertTipRawADCToDegF( getTipRawTemp(sampleNow)); currentTipTempInF += convertCtoF(getHandleTemperature() / 10); //Add handle offset currentTipTempInF += x10WattHistory.average() / 45; // 25 * 9 / 5, see getTipInC return currentTipTempInF; } #endif uint32_t TipThermoModel::getTipMaxInC() { uint32_t maximumTipTemp = TipThermoModel::convertTipRawADCToDegC(0x7FFF - (80 * 5)); //back off approx 5 deg c from ADC max maximumTipTemp += getHandleTemperature() / 10; //Add handle offset return maximumTipTemp - 1; }