Moats Engineering ThermalArchitect
Launch Thermal Architect

Thermal analysis and design software · Moats Engineering

Thermal analysis without the usual friction.

Thermal Architect is engineering software for building, solving, and understanding thermal systems, from simple resistance networks to complex electronics and fluid thermal models.

Launch Thermal Architect Explore the capabilities

Thermal Architect is in a controlled alpha and accounts are by invitation. If you don't have one yet, ask Moats Engineering for access.

Models
Heat paths, circuit boards and cooling loops, in one model
Solves
Steady state and transient
Studies
Sensitivity, parameter sweeps, uncertainty, optimization
Output
Engineering reports, ready to share
Thermal Architect showing a solved laptop cooling model: processors, a vapor chamber, heat pipes and fin stacks on the canvas, each labeled with its temperature, and the selected heat pipe's properties on the right.
Fig. 1A laptop cooling model, solved. Each part shows its temperature and each path the heat it carries. Every screenshot on this page is the application itself.

Where it fits

Thermal analysis shouldn't require fighting the tool.

A thermal question rarely stays in one place. A junction temperature starts as a hand calculation, moves into a spreadsheet when the stack-up grows, picks up a script for the tolerance study, and finishes in a document someone retypes the numbers into.

Detailed CFD and finite-element analysis are the right tools when the geometry or the flow field is the question. A lot of design questions come earlier than that. They need a model that can be built in an afternoon and still hold up in a design review.

Thermal Architect is built for that middle ground: one model that carries the analysis, the studies and the report.

Hand calculations and spreadsheets

Fast for one path and one operating point. Harder to extend, check and hand over.

Thermal Architect

Heat paths, circuit boards and cooling loops, with established heat transfer methods built in. Studies and reports run on the same model.

3-D CFD and FEA

Resolved geometry and flow fields, for when that level of detail is what the decision depends on.

It isn't a CFD or finite-element package, and doesn't try to be. When a decision depends on detailed geometry or airflow, those tools are the right next step.

Workflow

One model, carried through the whole question.

Build it once, then keep asking it things.

  1. 01

    Build

    Place components on a canvas and connect them with thermal paths. Each path is described the way you would naturally specify it: a material and thickness, an interface material, a heat sink, a heat pipe.

    Canvas with linked thermal nodes; link widths show heat flow.
  2. 02

    Parameterize

    Give the important dimensions, powers and operating conditions names, then define the rest of the model from them. Change one value and everything that depends on it follows.

    A table of named design parameters, some defined from others.
  3. 03

    Solve

    Run steady state or transient. Set requirements, such as a maximum temperature with margin, and every result tells you whether the design meets them.

    Results panel with a requirement check and a table of temperatures.
  4. 04

    Understand

    See temperatures and heat flow on the model itself, and a breakdown of which parts of the heat path account for the temperature rise.

    Bar chart breaking the temperature rise down by element of the heat path.
  5. 05

    Iterate

    See which change would help most, try it, and compare. No rebuilding the model for each variation.

    Recommendation cards showing what each possible improvement would buy.

Capabilities

Built for real thermal engineering.

What the current release does, grouped by the job it does.

Heat transfer

  • Conduction
  • Natural and forced convection
  • Radiation
  • Contact and interface resistance
  • Thermal interface materials
  • Heat pipes and vapor chambers
  • Heat sinks and fins
  • Spreading resistance
  • Boiling and condensation

Analysis

  • Steady state and transient
  • Temperature breakdown and recommendations
  • Sensitivity ranking
  • Parameter sweeps
  • Uncertainty (Monte Carlo)
  • What-if comparisons
  • Trade-offs and optimization
  • Worst-case operating conditions
  • Correlation to test data

Electronics and hardware

  • Circuit board thermal models
  • Layout import from KiCad, Altium and IPC-2581
  • Copper, vias and board construction
  • Component power and junction temperatures
  • Library of common component packages
  • Heat sinks, interfaces and enclosures in the same model

Fluid systems

  • Liquid and air cooling loops
  • Pumps, fans, pipes and valves
  • Cold plates and heat exchangers
  • Solved together with the thermal model
  • Built-in fluid properties

Engineering workflow

  • Visual model building
  • Named design parameters
  • Requirements checked on every solve
  • Material library, plus your own materials
  • Reusable subsystem templates
  • Saved projects
  • Engineering reports

Everything listed is available in the current alpha.

In the application

From first sketch to report.

A typical electronics cooling problem, start to finish: a processor on a forced-air heat sink, with a temperature requirement to meet.

A thermal model of a processor on a heat sink, with each part labeled with its temperature and each path with its heat flow.
Fig. 2Build. The parts and the paths between them, with temperatures and heat flow after a solve.
Results panel with a requirement check above a table of temperatures.
Fig. 3Solve. Requirements are checked automatically, so the result says whether the design meets them before anyone reads the table.
Heat sink properties: material, airflow and geometry, with several fields defined from named parameters.
Fig. 4Define the physics. The heat sink is described by its geometry, material and airflow. The highlighted fields come from named parameters.
A report page with the contents, a results summary and a requirement compliance table.
A report page with heat flow by path and a chart of the temperature rise from ambient air to the processor.
Fig. 5Report. Generated from the same model: requirement compliance, results, and the temperature rise from ambient to the hottest part.

Design Parameters

Turn your model into a design space.

Dimensions, powers, material properties and operating conditions can be named once and used anywhere in the model. A heat sink's width, height and fin count become parameters, and so can the size of the chip they depend on.

Change one value and everything defined from it follows. Units are checked as you type, so mismatched quantities are caught before they reach a result.

That turns “what happens if I change this?” into a quick study instead of a rebuild. Sweep one parameter, or two against each other, and every point is a full solve.

Example · parameters defined from parameters

CHIP_WIDTH40 mm
CHIP_DEPTH40 mm
CHIP_AREA= CHIP_WIDTH * CHIP_DEPTH1600 mm²
HS_OVERHANG15 mm
HS_WIDTH= CHIP_WIDTH + 2 * HS_OVERHANG70 mm
A table of processor temperature for every combination of chip size and fan speed, colored from cool to hot, with cells over the limit outlined.
Fig. 6Two parameters swept against each other, chip size and fan speed, with every cell a full solve. Outlined cells miss the design limit. Here, fan speed matters far more than chip size.

Analysis

From a single answer to understanding the design space.

Studies run on the model you've already built, organized by the question you're asking: what drives the result, what should change, and will it hold up.

Steady state and transient
Operating temperatures, or how they change over time as loads cycle on and off.
Temperature breakdown
Which parts of the heat path account for the temperature rise, and what each possible fix would buy.
Sensitivity and sweeps
Which inputs matter most, and how the result responds as they change.
Uncertainty
How tolerances and variation spread the result, and how likely a limit is to be exceeded.
Trade-offs and optimization
The best balance between temperature and cost, mass or another objective.
Worst case and test correlation
The worst operating condition and the margin left, and tuning the model to match measured data.
The Analysis workspace, titled What do you want to know?, listing studies under Explain, Improve, Verify and Document, with requirements, inputs and baseline results summarized above.
Fig. 7The Analysis workspace. Requirements and inputs are set up once and shared by every study.
Temperature breakdown: one bar per element of the heat path, with recommendation cards below.
Fig. 8Temperature breakdown. One element accounts for most of the rise, and each recommended fix is checked with a full re-solve.
Sensitivity chart ranking every input by its effect on the hottest part.
Fig. 9Sensitivity. Every input varied in turn and ranked by its effect on the hottest part.

Electronics

Model the electronics that actually generate the heat.

Heat spreads across a circuit board very differently from the way it moves through it, so where a part sits on the board matters. Thermal Architect models each board in detail and connects it to the heat sinks, cold plates and enclosure around it, so the board and the system are solved together.

Layouts can be imported from KiCad, Altium or IPC-2581. Anything the file doesn't contain, such as a part's power, is shown as an assumption you can review and change.

  • Component placement on either side of the board
  • Component power, entered directly or imported
  • Board shape, layers, copper and thermal vias
  • Junction temperature for every part, against its limit
  • How much heat leaves through the board and how much through the package
The board designer: board construction settings on the left, the component layout with its solved temperature field in the middle, and the selected component's power and thermal properties on the right.
Fig. 10A power electronics board in the board designer: component layout, board construction and the solved temperature field.

Engineering basis

Built around engineering fundamentals.

Underneath the canvas is what an engineer would sketch on paper: a network of thermal resistances and capacitances, with energy balanced at every node. Convection, contact and two-phase heat transfer use established methods from the heat transfer literature.

The assumptions are written down. The theory manual explains how each part of a model is calculated and where the methods apply, and a separate page lists what the software does not model.

Heat transfer
Established, published methods, with sources cited
Properties
Material and fluid libraries, plus your own materials
Time
Steady-state and transient solutions
Traceability
Every report records the model and settings it came from

Theory manual Known limitations

Who it's for

For engineers who need answers, not another workflow to manage.

  • Mechanical engineers

    Evaluate thermal behavior during design iterations.

  • Thermal engineers

    Build detailed thermal networks and investigate system behavior.

  • Electronics engineers

    Understand component, PCB and enclosure thermal performance.

  • Hardware teams

    Explore thermal design decisions before committing to detailed simulation.

  • Consultants

    Build reusable models and communicate results clearly.

Start building your thermal model.

Explore Thermal Architect and see how quickly you can move from an engineering question to a useful answer. Ready-made examples are included to start from.

Launch Thermal Architect Contact Moats Engineering

Controlled alpha · accounts by invitation · runs in the browser at app.moatsengineering.com

Questions it's built to answer

  • How hot does this part run?
  • Which part of the heat path is the bottleneck?
  • What happens if the ambient rises or the fan slows?
  • How much margin is left once tolerances are included?
  • What goes into the design review?