Dyad

Appearance

LIBRARY
Analog.Examples.ThermalRunaway.md

Analog.Examples.ThermalRunaway ​

Thermal runaway in a fixed-V_BE bipolar stage.

A demonstration of the electro-thermal feedback loop that the heat port and the temperature model make possible. The loop is:

  1. the saturation current climbs steeply with temperature, I_s(T) = I_s (T/T_{NOM})^{XTI} e^{(E_G/V_T)(T/T_{NOM}-1)}

  2. at fixed base-emitter voltage, collector current follows it directly

  3. more current means more dissipation, reported as LossPower

  4. the heat port carries that into the thermal network, raising T

  5. back to 1

Whether this settles or diverges is decided by the loop gain, and the term the user controls is the thermal resistance R_th. Small R_th — a good heatsink — and the junction barely warms. Large R_th and the stage walks itself up until the collector load finally limits the current.

Why fixed V_BE is the interesting bias ​

Nothing here pushes back. Drive the base from a stiff current source instead and I_c = Bf I_b pins the current regardless of temperature, so there is no runaway — with XTB = 0 the betas do not even drift. Add an emitter resistor and rising current lifts the emitter, cutting V_BE and closing the loop down. That degeneration resistor is why it appears in every real bias network, and this example is what it is defending against.

The circuit ​

   V_CC ──[R_C]──┬──── c              port ──── HeatCapacitor(C_th)
                 │     │                              │
   V_BE ──[R_B]──┼──── b                       ThermalConductor(G = 1/R_th)
                 │     │                              │
                gnd   gnd                      FixedTemperature(T_ambient)

R_B is deliberately small — ten ohms at tens of microamps is a fraction of a millivolt, so V_BE is still effectively pinned. It is present because an ideal source placed directly across the base-emitter junction also sits directly across C_jE, which is a structurally singular configuration.

Measured behavior ​

Sweeping R_th at the default bias, to 60 s:

R_th [K/W]30060090011002000
ΔT [K]5.78113.5120.0122.1126.1
I_c [mA]1.9674.984.387.593.4

So the boundary sits between 300 and 600 K/W. A linear criterion evaluated at ambient — R_th dP/dT > 1, with dP/dT ≈ 0.94 mW/K here — would predict about 1060 K/W and be roughly a factor of two too high. That is because dP/dT itself climbs steeply with temperature: by the time ΔT is a few tens of kelvin the gain is several times larger. Onset is really a saddle-node bifurcation, the point at which the cool solution stops existing, not the point where the gain at ambient reaches one.

Why the runaway stops ​

There is no destruction mechanism, and the load line ends it before anything else would. Dissipation is

a downward parabola in I_c peaking at I_c = V_{CC}/2R_C — 50 mA here, where V_CE = 5 V and P = 250 mW. Past that peak more current means less dissipation, so dP/dT turns negative and the loop becomes negative feedback. That is what arrests the runaway: the device crossing the power maximum, not V_CE reaching zero. At R_th = 2000 it ends at I_c = 93.4 mA, V_CE = 0.66 V, P = 63 mW, well down the falling side.

The same parabola makes the worst case interior in the drive as well:

V_BE0.780.820.850.88
ΔT [K]5.7869.250.628.3
P [mW]19.3230.6168.894.2

A real device would have crossed its junction temperature rating around V_BE = 0.82 and failed. This one settles at a new, much hotter operating point instead. The runaway is the transition, not an asymptote.

Usage ​

ElectricalComponents.Analog.Examples.ThermalRunaway(R_th=600.0, V_BE_drive=0.78, V_CC=10.0, R_C=100.0, C_th=2.0e-3, T_ambient=300.15)

Parameters: ​

NameDescriptionUnitsDefault value
R_thJunction-to-ambient thermal resistance. The knob that decides the outcome. The default runs away. Set R_th = 300 for the stable comparison.K/W600.0
V_BE_driveBase-emitter drive, held fixed. Dimensionless: it feeds a signal-level Constant.–0.78
V_CCCollector supply–10.0
R_CCollector loadΩ100.0
C_thThermal mass of the junction and its packageJ/K2.0e-3
T_ambientAmbient temperatureK300.15

Behavior ​

Source ​

dyad
"""
Thermal runaway in a fixed-`V_BE` bipolar stage.

A demonstration of the electro-thermal feedback loop that the heat port and the
temperature model make possible.  The loop is:

  1. the saturation current climbs steeply with temperature,
     `I_s(T) = I_s (T/T_{NOM})^{XTI} e^{(E_G/V_T)(T/T_{NOM}-1)}`
  2. at *fixed* base-emitter voltage, collector current follows it directly
  3. more current means more dissipation, reported as `LossPower`
  4. the heat port carries that into the thermal network, raising `T`
  5. back to 1

Whether this settles or diverges is decided by the loop gain, and the term the
user controls is the thermal resistance `R_th`.  Small `R_th` — a good heatsink —
and the junction barely warms.  Large `R_th` and the stage walks itself up until
the collector load finally limits the current.

## Why fixed `V_BE` is the interesting bias

Nothing here pushes back.  Drive the base from a stiff current source instead and
`I_c = Bf I_b` pins the current regardless of temperature, so there is no
runaway — with `XTB = 0` the betas do not even drift.  Add an emitter resistor
and rising current lifts the emitter, cutting `V_BE` and closing the loop down.
That degeneration resistor is *why* it appears in every real bias network, and
this example is what it is defending against.

## The circuit

```
   V_CC ──[R_C]──┬──── c              port ──── HeatCapacitor(C_th)
                 │     │                              │
   V_BE ──[R_B]──┼──── b                       ThermalConductor(G = 1/R_th)
                 │     │                              │
                gnd   gnd                      FixedTemperature(T_ambient)
```

`R_B` is deliberately small — ten ohms at tens of microamps is a fraction of a
millivolt, so `V_BE` is still effectively pinned.  It is present because an ideal
source placed directly across the base-emitter junction also sits directly across
`C_jE`, which is a structurally singular configuration.

## Measured behavior

Sweeping `R_th` at the default bias, to 60 s:

| `R_th` [K/W] | 300 | 600 | 900 | 1100 | 2000 |
| --- | --- | --- | --- | --- | --- |
| ΔT [K] | 5.78 | 113.5 | 120.0 | 122.1 | 126.1 |
| `I_c` [mA] | 1.96 | 74.9 | 84.3 | 87.5 | 93.4 |

So the boundary sits between 300 and 600 K/W.  A linear criterion evaluated at
ambient — `R_th dP/dT > 1`, with `dP/dT ≈ 0.94 mW/K` here — would predict about
1060 K/W and be roughly a factor of two too high.  That is because `dP/dT`
itself climbs steeply with temperature: by the time ΔT is a few tens of kelvin
the gain is several times larger.  Onset is really a saddle-node bifurcation,
the point at which the cool solution stops existing, not the point where the
gain at ambient reaches one.

## Why the runaway stops

There is no destruction mechanism, and the *load line* ends it before anything
else would.  Dissipation is
```math
P = I_c V_{CE} = I_c \\left( V_{CC} - I_c R_C \\right)
```
a downward parabola in `I_c` peaking at `I_c = V_{CC}/2R_C` — 50 mA here, where
`V_CE = 5 V` and `P = 250 mW`.  Past that peak **more current means less
dissipation**, so `dP/dT` turns negative and the loop becomes negative feedback.
That is what arrests the runaway: the device crossing the power maximum, not
`V_CE` reaching zero.  At `R_th = 2000` it ends at `I_c = 93.4 mA`,
`V_CE = 0.66 V`, `P = 63 mW`, well down the falling side.

The same parabola makes the worst case interior in the drive as well:

| `V_BE` | 0.78 | 0.82 | 0.85 | 0.88 |
| --- | --- | --- | --- | --- |
| ΔT [K] | 5.78 | 69.2 | 50.6 | 28.3 |
| P [mW] | 19.3 | 230.6 | 168.8 | 94.2 |

A real device would have crossed its junction temperature rating around
`V_BE = 0.82` and failed.  This one settles at a new, much hotter operating
point instead.  The runaway is the *transition*, not an asymptote.
"""
example component ThermalRunaway
  "The device under test, thermally coupled"
  transistor = ElectricalComponents.Analog.Semiconductors.NPN(use_heat_port = true, use_temperature_dependency = true) {
    "Dyad": {"placement": {"diagram": {"x1": 415, "x2": 515, "y1": 441, "y2": 541}}}
  }
  "Collector load resistor"
  load = ElectricalComponents.Analog.Basic.Resistor(R = R_C) {
    "Dyad": {"placement": {"diagram": {"x1": 285, "x2": 385, "y1": 226, "y2": 326}}}
  }
  "Small base series resistance; keeps the base off an ideal source"
  base_r = ElectricalComponents.Analog.Basic.Resistor(R = 10.0) {
    "Dyad": {"placement": {"diagram": {"x1": 285, "x2": 385, "y1": 441, "y2": 541}}}
  }
  "Collector supply"
  supply = ElectricalComponents.Analog.Sources.VoltageSource() {
    "Dyad": {"placement": {"diagram": {"x1": 155, "x2": 255, "y1": 125, "y2": 225}}}
  }
  "Supply level"
  supply_level = BlockComponents.Sources.Constant(k = V_CC) {"Dyad": {"placement": {"diagram": {"x1": 20, "x2": 120, "y1": 20, "y2": 120}}}}
  "Fixed base-emitter drive"
  bias = ElectricalComponents.Analog.Sources.VoltageSource() {
    "Dyad": {"placement": {"diagram": {"x1": 155, "x2": 255, "y1": 341, "y2": 441}}}
  }
  "Drive level"
  bias_level = BlockComponents.Sources.Constant(k = V_BE_drive) {
    "Dyad": {"placement": {"diagram": {"x1": 20, "x2": 120, "y1": 236, "y2": 336}}}
  }
  "Electrical reference"
  gnd = ElectricalComponents.Analog.Basic.Ground() {
    "Dyad": {"placement": {"diagram": {"x1": 542.5, "x2": 642.5, "y1": 230, "y2": 330}}}
  }
  "Junction and package thermal mass"
  mass = ThermalComponents.Components.HeatCapacitor(C = C_th, T0 = T_ambient) {
    "Dyad": {"placement": {"diagram": {"x1": 542.5, "x2": 642.5, "y1": 470, "y2": 570}}}
  }
  "Heatsink path to ambient; G is the reciprocal of the datasheet R_th"
  sink = ThermalComponents.Components.ThermalConductor(G = 1.0 / R_th) {
    "Dyad": {"placement": {"diagram": {"x1": 670, "x2": 770, "y1": 575, "y2": 675}}}
  }
  "Ambient reservoir"
  ambient = ThermalComponents.Sources.FixedTemperature(T = T_ambient) {
    "Dyad": {"placement": {"diagram": {"x1": 540, "x2": 640, "y1": 350, "y2": 450}}}
  }
  """
  Junction-to-ambient thermal resistance. The knob that decides the outcome.
  The default runs away. Set `R_th = 300` for the stable comparison.
  """
  parameter R_th::Real(units = "K/W") = 600.0
  "Base-emitter drive, held fixed. Dimensionless: it feeds a signal-level Constant."
  parameter V_BE_drive::Real = 0.78
  "Collector supply"
  parameter V_CC::Real = 10.0
  "Collector load"
  parameter R_C::Resistance = 100.0
  "Thermal mass of the junction and its package"
  parameter C_th::HeatCapacity = 2.0e-3
  "Ambient temperature"
  parameter T_ambient::Temperature = 300.15
relations
  initial transistor.V_BE = V_BE_drive
  initial transistor.V_BC = V_BE_drive - V_CC
  "Signal levels into the two sources"
  connect(supply_level.y, supply.V) {"Dyad": {"edges": [{"S": 1, "M": [{"x": 205, "y": 70}], "E": 2}], "path": {}}}
  connect(bias_level.y, bias.V) {"Dyad": {"edges": [{"S": 1, "M": [{"x": 205, "y": 286}], "E": 2}], "path": {}}}
  "Collector load"
  connect(supply.p, load.p) {
    "Dyad": {
      "edges": [{"S": 1, "M": [{"x": 140, "y": 175}, {"x": 140, "y": 276}], "E": 2}],
      "path": {}
    }
  }
  connect(load.n, transistor.c) {"Dyad": {"edges": [{"S": 1, "M": [{"x": 465, "y": 276}], "E": 2}], "path": {}}}
  "Fixed base drive through a small series resistance"
  connect(bias.p, base_r.p) {
    "Dyad": {
      "edges": [{"S": 1, "M": [{"x": 140, "y": 391}, {"x": 140, "y": 491}], "E": 2}],
      "path": {}
    }
  }
  connect(base_r.n, transistor.b) {"Dyad": {"edges": [{"S": 1, "M": [], "E": 2}], "path": {}}}
  "Emitter and substrate to ground"
  connect(transistor.e, gnd.g) {
    "Dyad": {
      "edges": [
        {"S": 1, "M": [{"x": 465, "y": 596}, {"x": 530, "y": 596}], "E": -1},
        {"S": -1, "M": [{"x": 592.5, "y": 175}], "E": 2}
      ],
      "junctions": [{"x": 530, "y": 175}],
      "path": {}
    }
  }
  connect(transistor.s, gnd.g) {
    "Dyad": {
      "edges": [
        {"S": 1, "M": [], "E": -1},
        {"S": -1, "M": [{"x": 530, "y": 175}, {"x": 592.5, "y": 175}], "E": 2}
      ],
      "junctions": [{"x": 530, "y": 491}],
      "path": {}
    }
  }
  connect(supply.n, gnd.g) {
    "Dyad": {"edges": [{"S": 1, "M": [{"x": 592.5, "y": 175}], "E": 2}], "path": {}}
  }
  connect(bias.n, gnd.g) {
    "Dyad": {
      "edges": [
        {"S": 1, "M": [{"x": 270, "y": 391}], "E": -1},
        {"S": -1, "M": [{"x": 592.5, "y": 175}], "E": 2}
      ],
      "junctions": [{"x": 270, "y": 175}],
      "path": {}
    }
  }
  "Thermal path: junction into the mass, mass to ambient through the heatsink"
  connect(transistor.port, mass.port) {
    "Dyad": {
      "edges": [
        {"S": 1, "M": [{"x": 400, "y": 521}, {"x": 400, "y": 625}], "E": -1},
        {"S": -1, "M": [], "E": 2}
      ],
      "junctions": [{"x": 592.5, "y": 625}],
      "path": {}
    }
  }
  connect(mass.port, sink.port_a) {
    "Dyad": {
      "edges": [{"S": 1, "M": [], "E": -1}, {"S": -1, "M": [], "E": 2}],
      "junctions": [{"x": 592.5, "y": 625}],
      "path": {}
    }
  }
  connect(sink.port_b, ambient.port) {
    "Dyad": {
      "edges": [{"S": 1, "M": [{"x": 785, "y": 625}, {"x": 785, "y": 400}], "E": 2}],
      "path": {}
    }
  }
metadata {
  "Dyad": {
    "icons": {"default": "dyad://ElectricalComponents/Example.svg"},
    "tests": {
      "case1": {
        "stop": 60,
        "abstol": 1e-8,
        "reltol": 0.000001,
        "expect": {"signals": ["transistor.T_dev", "transistor.LossPower"]}
      }
    }
  }
}
end
Flattened Source
dyad
"""
Thermal runaway in a fixed-`V_BE` bipolar stage.

A demonstration of the electro-thermal feedback loop that the heat port and the
temperature model make possible.  The loop is:

  1. the saturation current climbs steeply with temperature,
     `I_s(T) = I_s (T/T_{NOM})^{XTI} e^{(E_G/V_T)(T/T_{NOM}-1)}`
  2. at *fixed* base-emitter voltage, collector current follows it directly
  3. more current means more dissipation, reported as `LossPower`
  4. the heat port carries that into the thermal network, raising `T`
  5. back to 1

Whether this settles or diverges is decided by the loop gain, and the term the
user controls is the thermal resistance `R_th`.  Small `R_th` — a good heatsink —
and the junction barely warms.  Large `R_th` and the stage walks itself up until
the collector load finally limits the current.

## Why fixed `V_BE` is the interesting bias

Nothing here pushes back.  Drive the base from a stiff current source instead and
`I_c = Bf I_b` pins the current regardless of temperature, so there is no
runaway — with `XTB = 0` the betas do not even drift.  Add an emitter resistor
and rising current lifts the emitter, cutting `V_BE` and closing the loop down.
That degeneration resistor is *why* it appears in every real bias network, and
this example is what it is defending against.

## The circuit

```
   V_CC ──[R_C]──┬──── c              port ──── HeatCapacitor(C_th)
                 │     │                              │
   V_BE ──[R_B]──┼──── b                       ThermalConductor(G = 1/R_th)
                 │     │                              │
                gnd   gnd                      FixedTemperature(T_ambient)
```

`R_B` is deliberately small — ten ohms at tens of microamps is a fraction of a
millivolt, so `V_BE` is still effectively pinned.  It is present because an ideal
source placed directly across the base-emitter junction also sits directly across
`C_jE`, which is a structurally singular configuration.

## Measured behavior

Sweeping `R_th` at the default bias, to 60 s:

| `R_th` [K/W] | 300 | 600 | 900 | 1100 | 2000 |
| --- | --- | --- | --- | --- | --- |
| ΔT [K] | 5.78 | 113.5 | 120.0 | 122.1 | 126.1 |
| `I_c` [mA] | 1.96 | 74.9 | 84.3 | 87.5 | 93.4 |

So the boundary sits between 300 and 600 K/W.  A linear criterion evaluated at
ambient — `R_th dP/dT > 1`, with `dP/dT ≈ 0.94 mW/K` here — would predict about
1060 K/W and be roughly a factor of two too high.  That is because `dP/dT`
itself climbs steeply with temperature: by the time ΔT is a few tens of kelvin
the gain is several times larger.  Onset is really a saddle-node bifurcation,
the point at which the cool solution stops existing, not the point where the
gain at ambient reaches one.

## Why the runaway stops

There is no destruction mechanism, and the *load line* ends it before anything
else would.  Dissipation is
```math
P = I_c V_{CE} = I_c \\left( V_{CC} - I_c R_C \\right)
```
a downward parabola in `I_c` peaking at `I_c = V_{CC}/2R_C` — 50 mA here, where
`V_CE = 5 V` and `P = 250 mW`.  Past that peak **more current means less
dissipation**, so `dP/dT` turns negative and the loop becomes negative feedback.
That is what arrests the runaway: the device crossing the power maximum, not
`V_CE` reaching zero.  At `R_th = 2000` it ends at `I_c = 93.4 mA`,
`V_CE = 0.66 V`, `P = 63 mW`, well down the falling side.

The same parabola makes the worst case interior in the drive as well:

| `V_BE` | 0.78 | 0.82 | 0.85 | 0.88 |
| --- | --- | --- | --- | --- |
| ΔT [K] | 5.78 | 69.2 | 50.6 | 28.3 |
| P [mW] | 19.3 | 230.6 | 168.8 | 94.2 |

A real device would have crossed its junction temperature rating around
`V_BE = 0.82` and failed.  This one settles at a new, much hotter operating
point instead.  The runaway is the *transition*, not an asymptote.
"""
example component ThermalRunaway
  "The device under test, thermally coupled"
  transistor = ElectricalComponents.Analog.Semiconductors.NPN(use_heat_port = true, use_temperature_dependency = true) {
    "Dyad": {"placement": {"diagram": {"x1": 415, "x2": 515, "y1": 441, "y2": 541}}}
  }
  "Collector load resistor"
  load = ElectricalComponents.Analog.Basic.Resistor(R = R_C) {
    "Dyad": {"placement": {"diagram": {"x1": 285, "x2": 385, "y1": 226, "y2": 326}}}
  }
  "Small base series resistance; keeps the base off an ideal source"
  base_r = ElectricalComponents.Analog.Basic.Resistor(R = 10.0) {
    "Dyad": {"placement": {"diagram": {"x1": 285, "x2": 385, "y1": 441, "y2": 541}}}
  }
  "Collector supply"
  supply = ElectricalComponents.Analog.Sources.VoltageSource() {
    "Dyad": {"placement": {"diagram": {"x1": 155, "x2": 255, "y1": 125, "y2": 225}}}
  }
  "Supply level"
  supply_level = BlockComponents.Sources.Constant(k = V_CC) {"Dyad": {"placement": {"diagram": {"x1": 20, "x2": 120, "y1": 20, "y2": 120}}}}
  "Fixed base-emitter drive"
  bias = ElectricalComponents.Analog.Sources.VoltageSource() {
    "Dyad": {"placement": {"diagram": {"x1": 155, "x2": 255, "y1": 341, "y2": 441}}}
  }
  "Drive level"
  bias_level = BlockComponents.Sources.Constant(k = V_BE_drive) {
    "Dyad": {"placement": {"diagram": {"x1": 20, "x2": 120, "y1": 236, "y2": 336}}}
  }
  "Electrical reference"
  gnd = ElectricalComponents.Analog.Basic.Ground() {
    "Dyad": {"placement": {"diagram": {"x1": 542.5, "x2": 642.5, "y1": 230, "y2": 330}}}
  }
  "Junction and package thermal mass"
  mass = ThermalComponents.Components.HeatCapacitor(C = C_th, T0 = T_ambient) {
    "Dyad": {"placement": {"diagram": {"x1": 542.5, "x2": 642.5, "y1": 470, "y2": 570}}}
  }
  "Heatsink path to ambient; G is the reciprocal of the datasheet R_th"
  sink = ThermalComponents.Components.ThermalConductor(G = 1.0 / R_th) {
    "Dyad": {"placement": {"diagram": {"x1": 670, "x2": 770, "y1": 575, "y2": 675}}}
  }
  "Ambient reservoir"
  ambient = ThermalComponents.Sources.FixedTemperature(T = T_ambient) {
    "Dyad": {"placement": {"diagram": {"x1": 540, "x2": 640, "y1": 350, "y2": 450}}}
  }
  """
  Junction-to-ambient thermal resistance. The knob that decides the outcome.
  The default runs away. Set `R_th = 300` for the stable comparison.
  """
  parameter R_th::Real(units = "K/W") = 600.0
  "Base-emitter drive, held fixed. Dimensionless: it feeds a signal-level Constant."
  parameter V_BE_drive::Real = 0.78
  "Collector supply"
  parameter V_CC::Real = 10.0
  "Collector load"
  parameter R_C::Resistance = 100.0
  "Thermal mass of the junction and its package"
  parameter C_th::HeatCapacity = 2.0e-3
  "Ambient temperature"
  parameter T_ambient::Temperature = 300.15
relations
  initial transistor.V_BE = V_BE_drive
  initial transistor.V_BC = V_BE_drive - V_CC
  "Signal levels into the two sources"
  connect(supply_level.y, supply.V) {"Dyad": {"edges": [{"S": 1, "M": [{"x": 205, "y": 70}], "E": 2}], "path": {}}}
  connect(bias_level.y, bias.V) {"Dyad": {"edges": [{"S": 1, "M": [{"x": 205, "y": 286}], "E": 2}], "path": {}}}
  "Collector load"
  connect(supply.p, load.p) {
    "Dyad": {
      "edges": [{"S": 1, "M": [{"x": 140, "y": 175}, {"x": 140, "y": 276}], "E": 2}],
      "path": {}
    }
  }
  connect(load.n, transistor.c) {"Dyad": {"edges": [{"S": 1, "M": [{"x": 465, "y": 276}], "E": 2}], "path": {}}}
  "Fixed base drive through a small series resistance"
  connect(bias.p, base_r.p) {
    "Dyad": {
      "edges": [{"S": 1, "M": [{"x": 140, "y": 391}, {"x": 140, "y": 491}], "E": 2}],
      "path": {}
    }
  }
  connect(base_r.n, transistor.b) {"Dyad": {"edges": [{"S": 1, "M": [], "E": 2}], "path": {}}}
  "Emitter and substrate to ground"
  connect(transistor.e, gnd.g) {
    "Dyad": {
      "edges": [
        {"S": 1, "M": [{"x": 465, "y": 596}, {"x": 530, "y": 596}], "E": -1},
        {"S": -1, "M": [{"x": 592.5, "y": 175}], "E": 2}
      ],
      "junctions": [{"x": 530, "y": 175}],
      "path": {}
    }
  }
  connect(transistor.s, gnd.g) {
    "Dyad": {
      "edges": [
        {"S": 1, "M": [], "E": -1},
        {"S": -1, "M": [{"x": 530, "y": 175}, {"x": 592.5, "y": 175}], "E": 2}
      ],
      "junctions": [{"x": 530, "y": 491}],
      "path": {}
    }
  }
  connect(supply.n, gnd.g) {
    "Dyad": {"edges": [{"S": 1, "M": [{"x": 592.5, "y": 175}], "E": 2}], "path": {}}
  }
  connect(bias.n, gnd.g) {
    "Dyad": {
      "edges": [
        {"S": 1, "M": [{"x": 270, "y": 391}], "E": -1},
        {"S": -1, "M": [{"x": 592.5, "y": 175}], "E": 2}
      ],
      "junctions": [{"x": 270, "y": 175}],
      "path": {}
    }
  }
  "Thermal path: junction into the mass, mass to ambient through the heatsink"
  connect(transistor.port, mass.port) {
    "Dyad": {
      "edges": [
        {"S": 1, "M": [{"x": 400, "y": 521}, {"x": 400, "y": 625}], "E": -1},
        {"S": -1, "M": [], "E": 2}
      ],
      "junctions": [{"x": 592.5, "y": 625}],
      "path": {}
    }
  }
  connect(mass.port, sink.port_a) {
    "Dyad": {
      "edges": [{"S": 1, "M": [], "E": -1}, {"S": -1, "M": [], "E": 2}],
      "junctions": [{"x": 592.5, "y": 625}],
      "path": {}
    }
  }
  connect(sink.port_b, ambient.port) {
    "Dyad": {
      "edges": [{"S": 1, "M": [{"x": 785, "y": 625}, {"x": 785, "y": 400}], "E": 2}],
      "path": {}
    }
  }
metadata {
  "Dyad": {
    "icons": {"default": "dyad://ElectricalComponents/Example.svg"},
    "tests": {
      "case1": {
        "stop": 60,
        "abstol": 1e-8,
        "reltol": 0.000001,
        "expect": {"signals": ["transistor.T_dev", "transistor.LossPower"]}
      }
    }
  }
}
end


Test Cases ​

Test Case case1 ​

julia
plt

julia
plt

  • Examples

  • Experiments

  • Analyses