diff options
| -rw-r--r-- | Developer-documentation/Movement in RepRapFirmware.odt | bin | 0 -> 25345 bytes | |||
| -rw-r--r-- | src/Movement/InputShaper.cpp | 126 | ||||
| -rw-r--r-- | src/Movement/InputShaper.h | 16 |
3 files changed, 61 insertions, 81 deletions
diff --git a/Developer-documentation/Movement in RepRapFirmware.odt b/Developer-documentation/Movement in RepRapFirmware.odt Binary files differnew file mode 100644 index 00000000..bc55571c --- /dev/null +++ b/Developer-documentation/Movement in RepRapFirmware.odt diff --git a/src/Movement/InputShaper.cpp b/src/Movement/InputShaper.cpp index 1e8b3680..e3d07042 100644 --- a/src/Movement/InputShaper.cpp +++ b/src/Movement/InputShaper.cpp @@ -27,22 +27,18 @@ constexpr ObjectModelTableEntry InputShaper::objectModelTable[] = // 0. InputShaper members { "damping", OBJECT_MODEL_FUNC(self->zeta, 2), ObjectModelEntryFlags::none }, { "frequency", OBJECT_MODEL_FUNC(self->frequency, 2), ObjectModelEntryFlags::none }, -#if SUPPORT_DAA - { "minimumAcceleration", OBJECT_MODEL_FUNC(self->daaMinimumAcceleration, 1), ObjectModelEntryFlags::none }, -#endif + { "minimumAcceleration", OBJECT_MODEL_FUNC(self->minimumAcceleration, 1), ObjectModelEntryFlags::none }, { "type", OBJECT_MODEL_FUNC(self->type.ToString()), ObjectModelEntryFlags::none }, }; -constexpr uint8_t InputShaper::objectModelTableDescriptor[] = { 1, 3 + SUPPORT_DAA }; +constexpr uint8_t InputShaper::objectModelTableDescriptor[] = { 1, 4 }; DEFINE_GET_OBJECT_MODEL_TABLE(InputShaper) InputShaper::InputShaper() noexcept : frequency(DefaultFrequency), zeta(DefaultDamping), -#if SUPPORT_DAA - daaMinimumAcceleration(DefaultDAAMinimumAcceleration), -#endif + minimumAcceleration(DefaultMinimumAcceleration), type(InputShaperType::none), numImpulses(1) { @@ -59,13 +55,11 @@ GCodeResult InputShaper::Configure(GCodeBuffer& gb, const StringRef& reply) THRO seen = true; frequency = gb.GetLimitedFValue('F', MinimumInputShapingFrequency, MaximumInputShapingFrequency); } -#if SUPPORT_DAA if (gb.Seen('L')) { seen = true; - daaMinimumAcceleration = max<float>(gb.GetFValue(), 1.0); // very low accelerations cause problems with the maths + minimumAcceleration = max<float>(gb.GetFValue(), 1.0); // very low accelerations cause problems with the maths } -#endif if (gb.Seen('S')) { seen = true; @@ -96,15 +90,12 @@ GCodeResult InputShaper::Configure(GCodeBuffer& gb, const StringRef& reply) THRO { case InputShaperType::none: numImpulses = 1; - shapingTime = 0.0; break; #if SUPPORT_DAA case InputShaperType::DAA: + durations[0] = 1.0/dampedFrequency; numImpulses = 1; - times[1] = 1.0/dampedFrequency; - times[0] = 0.5 * times[1]; - shapingTime = 0.0; break; #endif @@ -114,10 +105,8 @@ GCodeResult InputShaper::Configure(GCodeBuffer& gb, const StringRef& reply) THRO coefficients[0] = 1.0/j; coefficients[1] = coefficients[0] + 2.0 * k/j; } - times[1] = 1.0/dampedFrequency; - times[0] = 0.5 * times[1]; + durations[0] = durations[1] = 0.5/dampedFrequency; numImpulses = 3; - shapingTime = times[1] * StepTimer::StepClockRate; break; case InputShaperType::ZVDD: // see https://www.researchgate.net/publication/316556412_INPUT_SHAPING_CONTROL_TO_REDUCE_RESIDUAL_VIBRATION_OF_A_FLEXIBLE_BEAM @@ -127,58 +116,57 @@ GCodeResult InputShaper::Configure(GCodeBuffer& gb, const StringRef& reply) THRO coefficients[1] = coefficients[0] + 3.0 * k/j; coefficients[2] = coefficients[1] + 3.0 * fsquare(k)/j; } - times[1] = 1.0/dampedFrequency; - times[0] = 0.5 *times[1]; - times[2] = times[0] + times[1]; + durations[0] = durations[1] = durations[2] = 0.5/dampedFrequency; numImpulses = 4; - shapingTime = times[2] * StepTimer::StepClockRate; break; case InputShaperType::EI2: // see http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.465.1337&rep=rep1&type=pdf. United States patent #4,916,635. { const float zetaSquared = fsquare(zeta); const float zetaCubed = zetaSquared * zeta; - coefficients[0] = 0.16054 + 0.76699 * zeta + 2.26560 * zetaSquared + -1.22750 * zetaCubed; - //TODO avoid the extra addition - coefficients[1] = coefficients[0] + 0.33911 + 0.45081 * zeta + -2.58080 * zetaSquared + 1.73650 * zetaCubed; - coefficients[2] = coefficients[1] + 0.34089 + -0.61533 * zeta + -0.68765 * zetaSquared + 0.42261 * zetaCubed; - times[0] = (0.49890 + 0.16270 * zeta + -0.54262 * zetaSquared + 6.16180 * zetaCubed)/dampedFrequency; - times[1] = (0.99748 + 0.18382 * zeta + -1.58270 * zetaSquared + 8.17120 * zetaCubed)/dampedFrequency; - times[2] = (1.49920 + -0.09297 * zeta + -0.28338 * zetaSquared + 1.85710 * zetaCubed)/dampedFrequency; + coefficients[0] = (0.16054) + (0.76699) * zeta + (2.26560) * zetaSquared + (-1.22750) * zetaCubed; + coefficients[1] = (0.16054 + 0.33911) + (0.76699 + 0.45081) * zeta + (2.26560 - 2.58080) * zetaSquared + (-1.22750 + 1.73650) * zetaCubed; + coefficients[2] = (0.16054 + 0.33911 + 0.34089) + (0.76699 + 0.45081 - 0.61533) * zeta + (2.26560 - 2.58080 - 0.68765) * zetaSquared + (-1.22750 + 1.73650 + 0.42261) * zetaCubed; + + // TODO is it really worth using precise durations instead of setting them all to 0.5/frequency? If we don't, then we can condense the shaping when the acceleration time is short. + durations[0] = ((0.49890) + ( 0.16270 ) * zeta + ( -0.54262) * zetaSquared + ( 6.16180) * zetaCubed)/dampedFrequency; + durations[1] = ((0.99748 - 0.49890) + ( 0.18382 - 0.16270) * zeta + (-1.58270 + 0.54262) * zetaSquared + (8.17120 - 6.16180) * zetaCubed)/dampedFrequency; + durations[2] = ((1.49920 - 0.99748) + (-0.09297 - 0.18382) * zeta + (-0.28338 + 1.58270) * zetaSquared + (1.85710 - 8.17120) * zetaCubed)/dampedFrequency; } numImpulses = 4; - shapingTime = times[2] * StepTimer::StepClockRate; break; case InputShaperType::EI3: // see http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.465.1337&rep=rep1&type=pdf. United States patent #4,916,635 { const float zetaSquared = fsquare(zeta); const float zetaCubed = zetaSquared * zeta; - coefficients[0] = 0.11275 + 0.76632 * zeta + 3.29160 * zetaSquared + -1.44380 * zetaCubed; - //TODO avoid the extra addition - coefficients[1] = coefficients[0] + 0.23698 + 0.61164 * zeta + -2.57850 * zetaSquared + 4.85220 * zetaCubed; - coefficients[2] = coefficients[1] + 0.30008 + -0.19062 * zeta + -2.14560 * zetaSquared + 0.13744 * zetaCubed; - coefficients[3] = coefficients[2] + 0.23775 + -0.73297 * zeta + 0.46885 * zetaSquared + -2.08650 * zetaCubed; - times[0] = (0.49974 + 0.23834 * zeta + 0.44559 * zetaSquared + 12.4720 * zetaCubed)/dampedFrequency; - times[1] = (0.99849 + 0.29808 * zeta + -2.36460 * zetaSquared + 23.3990 * zetaCubed)/dampedFrequency; - times[2] = (1.49870 + 0.10306 * zeta + -2.01390 * zetaSquared + 17.0320 * zetaCubed)/dampedFrequency; - times[3] = (1.99960 + -0.28231 * zeta + 0.61536 * zetaSquared + 5.40450 * zetaCubed)/dampedFrequency; + coefficients[0] = (0.11275) + 0.76632 * zeta + (3.29160) * zetaSquared + (-1.44380) * zetaCubed; + coefficients[1] = (0.11275 + 0.23698) + (0.76632 + 0.61164) * zeta + (3.29160 - 2.57850) * zetaSquared + (-1.44380 + 4.85220) * zetaCubed; + coefficients[2] = (0.11275 + 0.23698 + 0.30008) + (0.76632 + 0.61164 - 0.19062) * zeta + (3.29160 - 2.57850 - 2.14560) * zetaSquared + (-1.44380 + 4.85220 + 0.13744) * zetaCubed; + coefficients[3] = (0.11275 + 0.23698 + 0.30008 + 0.23775) + (0.76632 + 0.61164 - 0.19062 - 0.73297) * zeta + (3.29160 - 2.57850 - 2.14560 + 0.46885) * zetaSquared + (-1.44380 + 4.85220 + 0.13744 - 2.08650) * zetaCubed; + + // TODO is it really worth using precise durations instead of setting them all to 0.5/frequency? If we don't, then we can condense the shaping when the acceleration time is short. + durations[0] = ((0.49974) + (0.23834) * zeta + (0.44559) * zetaSquared + (12.4720) * zetaCubed)/dampedFrequency; + durations[1] = ((0.99849 - 0.49974) + (0.29808 - 0.23834) * zeta + (-2.36460 - 0.44559) * zetaSquared + (23.3990 - 12.4720) * zetaCubed)/dampedFrequency; + durations[2] = ((1.49870 - 0.99849) + (0.10306 - 0.29808) * zeta + (-2.01390 + 2.36460) * zetaSquared + (17.0320 - 23.3990) * zetaCubed)/dampedFrequency; + durations[3] = ((1.99960 - 1.49870) + (-0.28231 - 0.10306) * zeta + (0.61536 + 2.01390) * zetaSquared + (5.40450 - 17.0320) * zetaCubed)/dampedFrequency; } numImpulses = 5; - shapingTime = times[3] * StepTimer::StepClockRate; break; } + totalDuration = 0.0; float tLostAtStart = 0.0; float tLostAtEnd = 0.0; for (uint8_t i = 0; i < numImpulses - 1; ++i) { - const float segTime = (i == 0) ? times[0] : times[i] - times[i - 1]; - tLostAtStart += (1.0 - coefficients[i]) * segTime; - tLostAtEnd += coefficients[i] * segTime; + totalDuration += durations[i]; + tLostAtStart += (1.0 - coefficients[i]) * durations[i]; + tLostAtEnd += coefficients[i] * durations[i]; } clocksLostAtStart = tLostAtStart * StepTimer::StepClockRate; clocksLostAtEnd = tLostAtEnd * StepTimer::StepClockRate; + totalShapingClocks = totalDuration * StepTimer::StepClockRate; reprap.MoveUpdated(); } @@ -188,13 +176,7 @@ GCodeResult InputShaper::Configure(GCodeBuffer& gb, const StringRef& reply) THRO } else { - reply.printf("%s input shaping at %.1fHz damping factor %.2f", type.ToString(), (double)frequency, (double)zeta); -#if SUPPORT_DAA - if (type == InputShaperType::DAA) - { - reply.catf(", min. acceleration %.1f", (double)daaMinimumAcceleration); - } -#endif + reply.printf("%s input shaping at %.1fHz damping factor %.2f, min. acceleration %.1f", type.ToString(), (double)frequency, (double)zeta, (double)minimumAcceleration); } return GCodeResult::ok; } @@ -209,7 +191,7 @@ InputShaperPlan InputShaper::PlanShaping(DDA& dda, BasicPrepParams& params, bool case InputShaperType::DAA: { // Try to reduce the acceleration/deceleration of the move to cancel ringing - const float idealPeriod = times[1]; // this the full period + const float idealPeriod = durations[0]; // for DAA this the full period float proposedAcceleration = dda.acceleration, proposedAccelDistance = dda.beforePrepare.accelDistance; bool adjustAcceleration = false; @@ -257,7 +239,7 @@ InputShaperPlan InputShaper::PlanShaping(DDA& dda, BasicPrepParams& params, bool { if (proposedAccelDistance + proposedDecelDistance <= dda.totalDistance) { - if (proposedAcceleration < daaMinimumAcceleration || proposedDeceleration < daaMinimumAcceleration) + if (proposedAcceleration < minimumAcceleration || proposedDeceleration < minimumAcceleration) { break; } @@ -276,7 +258,7 @@ InputShaperPlan InputShaper::PlanShaping(DDA& dda, BasicPrepParams& params, bool { proposedAcceleration = (twiceTotalDistance - ((3 * dda.startSpeed + dda.endSpeed) * idealPeriod))/(2 * fsquare(idealPeriod)); proposedDeceleration = (twiceTotalDistance - ((dda.startSpeed + 3 * dda.endSpeed) * idealPeriod))/(2 * fsquare(idealPeriod)); - if ( proposedAcceleration < daaMinimumAcceleration || proposedDeceleration < daaMinimumAcceleration + if ( proposedAcceleration < minimumAcceleration || proposedDeceleration < minimumAcceleration || proposedAcceleration > dda.acceleration || proposedDeceleration > dda.deceleration ) { @@ -292,7 +274,7 @@ InputShaperPlan InputShaper::PlanShaping(DDA& dda, BasicPrepParams& params, bool { // Change it into an accelerate-only move, accelerating as slowly as we can proposedAcceleration = (fsquare(dda.endSpeed) - fsquare(dda.startSpeed))/twiceTotalDistance; - if (proposedAcceleration < daaMinimumAcceleration) + if (proposedAcceleration < minimumAcceleration) { break; // avoid very small accelerations because they can be problematic } @@ -305,7 +287,7 @@ InputShaperPlan InputShaper::PlanShaping(DDA& dda, BasicPrepParams& params, bool { // Change it into a decelerate-only move, decelerating as slowly as we can proposedDeceleration = (fsquare(dda.startSpeed) - fsquare(dda.endSpeed))/twiceTotalDistance; - if (proposedDeceleration < daaMinimumAcceleration) + if (proposedDeceleration < minimumAcceleration) { break; // avoid very small accelerations because they can be problematic } @@ -343,13 +325,13 @@ InputShaperPlan InputShaper::PlanShaping(DDA& dda, BasicPrepParams& params, bool params.SetFromDDA(dda); // Set the plan to what we would like to do, if possible - plan.shapeAccelStart = params.accelClocks + clocksLostAtStart >= shapingTime + plan.shapeAccelStart = params.accelClocks + clocksLostAtStart >= totalShapingClocks && ((dda.GetPrevious()->state != DDA::DDAState::frozen && dda.GetPrevious()->state != DDA::DDAState::executing) || !dda.GetPrevious()->flags.wasAccelOnlyMove); - plan.shapeAccelEnd = params.accelClocks + clocksLostAtEnd >= shapingTime + plan.shapeAccelEnd = params.accelClocks + clocksLostAtEnd >= totalShapingClocks && params.decelStartDistance > params.accelDistance; - plan.shapeDecelStart = params.decelClocks + clocksLostAtStart >= shapingTime + plan.shapeDecelStart = params.decelClocks + clocksLostAtStart >= totalShapingClocks && params.decelStartDistance > params.accelDistance; - plan.shapeDecelEnd = params.decelClocks + clocksLostAtEnd >= shapingTime + plan.shapeDecelEnd = params.decelClocks + clocksLostAtEnd >= totalShapingClocks && (dda.GetNext()->GetState() != DDA::DDAState::provisional || !dda.GetNext()->IsDecelerationMove()); // debugPrintf("Original plan %03x ", (unsigned int)plan.all); @@ -357,7 +339,7 @@ InputShaperPlan InputShaper::PlanShaping(DDA& dda, BasicPrepParams& params, bool // See if we can shape the acceleration if (plan.shapeAccelStart || plan.shapeAccelEnd) { - if (plan.shapeAccelStart && plan.shapeAccelEnd && params.accelClocks < 2 * shapingTime) + if (plan.shapeAccelStart && plan.shapeAccelEnd && params.accelClocks < 2 * totalShapingClocks) { // Acceleration segment is too short to shape both the start and the end plan.shapeAccelStart = plan.shapeAccelEnd = false; @@ -398,7 +380,7 @@ InputShaperPlan InputShaper::PlanShaping(DDA& dda, BasicPrepParams& params, bool // See if we can shape the deceleration if (plan.shapeDecelStart || plan.shapeDecelEnd) { - if (plan.shapeDecelStart && plan.shapeDecelEnd && params.decelClocks < 2 * shapingTime) + if (plan.shapeDecelStart && plan.shapeDecelEnd && params.decelClocks < 2 * totalShapingClocks) { // Deceleration segment is too short to shape both the start and the end plan.shapeDecelStart = plan.shapeDecelEnd = false; @@ -465,14 +447,14 @@ MoveSegment *InputShaper::GetAccelerationSegments(DDA& dda, BasicPrepParams& par ++numAccelSegs; endAccelSegs = MoveSegment::Allocate(endAccelSegs); const float acceleration = dda.acceleration * (1.0 - coefficients[i]); - const float segTime = ((i == 0) ? times[0] : times[i] - times[i - 1]); + const float segTime = durations[i]; segStartSpeed -= acceleration * segTime; const float uDivA = (segStartSpeed * StepTimer::StepClockRate)/acceleration; const float twoDistDivA = (2 * StepTimer::StepClockRateSquared * dda.totalDistance)/acceleration; endAccelSegs->SetNonLinear(endDistance/dda.totalDistance, segTime * StepTimer::StepClockRate, uDivA, twoDistDivA, acceleration/StepTimer::StepClockRateSquared); endDistance -= (segStartSpeed + (0.5 * acceleration * segTime)) * segTime; } - accumulatedSegTime += times[numImpulses - 2]; + accumulatedSegTime += totalDuration; } float startDistance = 0.0; @@ -486,7 +468,7 @@ MoveSegment *InputShaper::GetAccelerationSegments(DDA& dda, BasicPrepParams& par ++numAccelSegs; MoveSegment *seg = MoveSegment::Allocate(nullptr); const float acceleration = dda.acceleration * coefficients[i]; - const float segTime = ((i == 0) ? times[0] : times[i] - times[i - 1]); + const float segTime = durations[i]; const float uDivA = (startSpeed * StepTimer::StepClockRate)/acceleration; const float twoDistDivA = (2 * StepTimer::StepClockRateSquared * dda.totalDistance)/acceleration; startDistance += (startSpeed + (0.5 * acceleration * segTime)) * segTime; @@ -501,7 +483,7 @@ MoveSegment *InputShaper::GetAccelerationSegments(DDA& dda, BasicPrepParams& par } startSpeed += acceleration * segTime; } - accumulatedSegTime += times[numImpulses - 2]; + accumulatedSegTime += totalDuration; } // Do the constant acceleration part @@ -549,14 +531,14 @@ MoveSegment *InputShaper::GetDecelerationSegments(DDA& dda, BasicPrepParams& par ++numDecelSegs; endDecelSegs = MoveSegment::Allocate(endDecelSegs); const float acceleration = -dda.deceleration * (1.0 - coefficients[i]); - const float segTime = ((i == 0) ? times[0] : times[i] - times[i - 1]); + const float segTime = durations[i]; segStartSpeed -= acceleration * segTime; const float uDivA = (segStartSpeed * StepTimer::StepClockRate)/acceleration; const float twoDistDivA = (2 * StepTimer::StepClockRateSquared * dda.totalDistance)/acceleration; endDecelSegs->SetNonLinear(endDistance/dda.totalDistance, segTime * StepTimer::StepClockRate, uDivA, twoDistDivA, acceleration/StepTimer::StepClockRateSquared); endDistance -= (segStartSpeed + (0.5 * acceleration * segTime)) * segTime; } - accumulatedSegTime += times[numImpulses - 2]; + accumulatedSegTime += totalDuration; } float startDistance = params.decelStartDistance; @@ -570,7 +552,7 @@ MoveSegment *InputShaper::GetDecelerationSegments(DDA& dda, BasicPrepParams& par ++numDecelSegs; MoveSegment *seg = MoveSegment::Allocate(nullptr); const float acceleration = -dda.deceleration * coefficients[i]; - const float segTime = ((i == 0) ? times[0] : times[i] - times[i - 1]); + const float segTime = durations[i]; const float uDivA = (startSpeed * StepTimer::StepClockRate)/acceleration; const float twoDistDivA = (2 * StepTimer::StepClockRateSquared * dda.totalDistance)/acceleration; startDistance += (startSpeed + (0.5 * acceleration * segTime)) * segTime; @@ -585,7 +567,7 @@ MoveSegment *InputShaper::GetDecelerationSegments(DDA& dda, BasicPrepParams& par } startSpeed += acceleration * segTime; } - accumulatedSegTime += times[numImpulses - 2]; + accumulatedSegTime += totalDuration; } // Do the constant deceleration part @@ -646,7 +628,7 @@ float InputShaper::GetExtraAccelStartDistance(const DDA& dda) const noexcept float u = dda.startSpeed; for (int seg = 0; seg < numImpulses - 1; ++seg) { - const float segTime = (seg == 0) ? times[0] : times[seg] - times[seg - 1]; + const float segTime = durations[seg]; const float speedChange = coefficients[seg] * dda.acceleration * segTime; extraDistance += (1.0 - coefficients[seg]) * (u + 0.5 * speedChange) * segTime; u += speedChange; @@ -661,7 +643,7 @@ float InputShaper::GetExtraAccelEndDistance(const DDA& dda) const noexcept float v = dda.topSpeed; for (int seg = numImpulses - 2; seg >= 0; --seg) { - const float segTime = (seg == 0) ? times[0] : times[seg] - times[seg - 1]; + const float segTime = durations[seg]; const float speedChange = (1.0 - coefficients[seg]) * dda.acceleration * segTime; extraDistance += coefficients[seg] * (v - 0.5 * speedChange) * segTime; v -= speedChange; @@ -676,7 +658,7 @@ float InputShaper::GetExtraDecelStartDistance(const DDA& dda) const noexcept float u = dda.topSpeed; for (int seg = 0; seg < numImpulses - 1; ++seg) { - const float segTime = (seg == 0) ? times[0] : times[seg] - times[seg - 1]; + const float segTime = durations[seg]; const float speedChange = coefficients[seg] * dda.deceleration * segTime; extraDistance += (1.0 - coefficients[seg]) * (u - 0.5 * speedChange) * segTime; u -= speedChange; @@ -691,7 +673,7 @@ float InputShaper::GetExtraDecelEndDistance(const DDA& dda) const noexcept float v = dda.endSpeed; for (int seg = numImpulses - 2; seg >= 0; --seg) { - const float segTime = (seg == 0) ? times[0] : times[seg] - times[seg - 1]; + const float segTime = durations[seg]; const float speedChange = (1.0 - coefficients[seg]) * dda.deceleration * segTime; extraDistance += coefficients[seg] * (v + 0.5 * speedChange) * segTime; v += speedChange; diff --git a/src/Movement/InputShaper.h b/src/Movement/InputShaper.h index 02dd5867..fd145216 100644 --- a/src/Movement/InputShaper.h +++ b/src/Movement/InputShaper.h @@ -73,20 +73,18 @@ private: float GetExtraDecelStartDistance(const DDA& dda) const noexcept; float GetExtraDecelEndDistance(const DDA& dda) const noexcept; + static constexpr unsigned int MaxImpulses = 5; static constexpr float DefaultFrequency = 40.0; static constexpr float DefaultDamping = 0.1; -#if SUPPORT_DAA - static constexpr float DefaultDAAMinimumAcceleration = 10.0; -#endif + static constexpr float DefaultMinimumAcceleration = 10.0; float frequency; // the undamped frequency float zeta; // the damping ratio, see https://en.wikipedia.org/wiki/Damping. 0 = undamped, 1 = critically damped. -#if SUPPORT_DAA - float daaMinimumAcceleration; // the minimum value that we reduce acceleration to (DAA only) -#endif - float coefficients[4]; // the coefficients of all the impulses, except the last which is 1 - float times[4]; // the time in seconds for the second and subsequent impulses - float shapingTime; // the time needed to send all the impulses, in step clocks + float minimumAcceleration; // the minimum value that we reduce acceleration to (DAA only) + float coefficients[MaxImpulses - 1]; // the coefficients of all the impulses, except the last which is 1 + float durations[MaxImpulses - 1]; // the duration in seconds of each impulse except the last + float totalDuration; // the total input shaping time in seconds + float totalShapingClocks; // the total input shaping time in step clocks float clocksLostAtStart, clocksLostAtEnd; // the acceleration time lost due to input shaping. Multiply by 2 if shaping is used at both the start and end of acceleration. InputShaperType type; uint8_t numImpulses; |