Truepath CFD

Thermal & CFD consulting · Houston, TX

Thermal models you can trust.

Joule-heating and conjugate heat transfer simulation for multi-zone semiconductor heaters, electronics cooling and liquid cold plates. When your model and your thermocouples disagree, we find out why.

Book a 20-minute scope call See the work

Designed and built high-power ceramic heaters for semiconductor equipment · 9 patents and applications

  • 0.045 °C wafer uniformity (simulated)
  • Cold plate: within 2.5% of published data

Selected work

The public benchmark uses published data only, and the inventor design study is published with the owner’s permission. Sample cases are simulated design studies; full details are available on request.

Methodology calibrated to R² > 0.9 on more than 30 thermocouples on previous builds.

Sample Case · Diagnostic Sprint

0.045 °C

Design Competition: Truepath Expert vs AI Agents

Competition
A team of AI design and manufacturing engineers competed against a similar team led by a Truepath expert, designing the same 300 mm, 8-zone alumina heater.
Finding
Expert direction held the wafer to 0.045 °C hottest-to-coldest vs 0.152 °C: 3.4× more uniform (simulated, 1 and 10 torr).
Outcome
It also exposed the next bottleneck, print-thickness variation within a zone, which expert direction brought into view.

Download the case tile (PDF) · Book a call

Sample Case · Model Audit

Cold plate: datasheet matched within 2.5%, and a 45% hot spot it doesn’t show

Challenge
Build the model from only a public drawing and datasheet, without the inputs that decide accuracy: contact resistance, tube wall, and where the test sensor sat.
Finding
Once those unknowns were pinned down, the model matched the published curve within 2.5% at all four flow rates and showed the hot spot running up to 45% above the datasheet value (simulated).
Outcome
A datasheet match can still miss your hot spot. A Model Audit recalibrates your model against your test data before you size the loop.

Download the case tile (PDF) · Book a call

Sample Case · Diagnostic Sprint

300 mm wafer heater: about 20× more uniform

Challenge
Thin-film deposition at 150 °C needs the 300 mm wafer held to 0.04 °C. Before design optimization, this generic 8-zone alumina heater reached 0.86 °C hottest-to-coldest.
Finding
A flat heater face is not a flat wafer. Uniformity has to be proven at the wafer, across the gas gap.
Outcome
Wafer held to 0.04 °C hottest-to-coldest across 300 mm at 150 °C, about 20× more uniform than before design optimization (simulated).

Download the case tile (PDF) · Book a call

Sample Case · Calibrated Design Loop

Convection heater: 79% lighter, 345 °C air in about 1 s

Inventor-owned design (David Y. Guo, patent pending). Study by Truepath CFD™, published with the owner’s permission.

Challenge
Hot air in about a second from a small 7.4 V battery pack, puff after puff.
Finding
Simulated transient Joule heating with airflow, over 3 design rounds, cut heater mass from 18.74 g to 4.00 g (79% less).
Outcome
345 °C peak air in 1.1 s on the first puff (simulated). A Calibrated Design Loop matches the model to your test data before you build.

Download the case tile (PDF) · Book a call

0.045 °C

Wafer hottest-to-coldest in a simulated 300 mm design study, 3.4× more uniform than the AI-alone design.

2.5%

Cold-plate model within 2.5% of published data at four flow rates.

9

Patents and applications in heaters, airflow and heater control. See the list.

Send your current uniformity map, and we'll tell you what's limiting it.

Or send a model and test data that disagree.

Book a 20-minute scope call Email the map

Most thermal models are easy to build and hard to trust. Once the simulation and the thermocouple map disagree, the model is no use for the decision that matters: which trace, zone, gap or flow change to make before you tool. That's the problem we work on.

Services

For thermal, R&D, process and mechanical leads on heaters, electrostatic chucks, pedestals, high-power electronics and liquid cold plates. Best fit when the next decision is a trace, a gap, a cold plate or a measurement, and a wrong one costs a spin.

Start here

Thermal Model Check

An independent expert review of one existing thermal model, or one test-vs-simulation mismatch. You get a 2-page memo in 5 business days: likely root causes ranked by impact, assumptions to check and a recommended next step. New simulations are a separate project.

$2,500 fixed

Net 15 against your PO. The full fee is credited toward a Model Audit booked within 30 days of memo delivery.

01

Model Audit

The fix-it step for a model that doesn't match your tests. One model recalibrated against one test dataset. You get a corrected, calibrated model and a short report on what changed and why.

From $5,000 fixed. Larger scopes quoted in the SOW.

02

Diagnostic Sprint

One geometry, the real boundary conditions and a written verdict: will this layout hit its temperature gate, and the most effective fix if it won't. For heaters, that includes wafer uniformity.

From $6,500 fixed. 1–2 weeks. The usual first full study.

03

Calibrated Design Loop

Match the model to your thermal map or thermocouples, then change traces, zones or flow hardware against that model, including cold-plate design and optimization. It stops when the predicted window is narrow enough to tool.

From $12,000. Run by milestone.

04

Production Support

Drawings-ready recommendations, rating notes and a measurement list your lab can run on equipment it already has. For liquid loops: flow and pump-head sizing from the plate's flow curves. On-site setup is a separate trip, by agreement.

From $18,000. Run by milestone.

01 · Model Audit

from $5,000

How a first job starts

  1. 20-minute scope call: temperature gate, geometry class, what you already measured.
  2. One-page statement of work: fee, dates, deliverables, out of scope.
  3. Kickoff, then CAD, materials, duty cycle and any test data. Mutual NDA before proprietary geometry is exchanged.

02 · Diagnostic Sprint

from $6,500

What you receive

  • Annotated temperature and power plots a non-specialist can brief upward.
  • A short decision memo: keep, change allocation, change geometry, or stop and measure.
  • Native plots plus CAD notes so your team keeps the geometry.
  • A next-step list with effort, and a measurement list for equipment you already have.

Mid-sprint working session: we review results together halfway, so the last day is a decision, not a reveal.

Fixed project fees, scoped on a short call. Fees move only when scope moves.

Engineering judgment, with AI where it helps

AI-assisted work is available at the client's request. The judgment stays with a Truepath expert: which unknowns matter, whether the model matches the data, and the recommendation we stand behind.

About

Truepath CFD™ is an independent thermal practice in Houston, TX, led by David Guo. We build and audit transient Joule-heating and conjugate heat transfer models of multi-zone ceramic heaters, wafer heaters, pedestals, high-power electronics and liquid cold plates. We check them against your test data, and when they don't match, we find out why.

David's models track how heater resistance changes with temperature, so predictions hold at operating conditions.

David Guo, Principal, Truepath CFD™

David Guo, Principal

  • Designed and built high-power ceramic heaters for semiconductor equipment.
  • Chief engineer on consumer rapid-heating products.
  • Liquid cold plates: model matched published data within 2.5%.
  • Simulated flow, pressure-drop and coolant studies.
  • M.S. Physics, Cal State LA. B.A. Economics, UC Berkeley.
  • 9 patents and applications.
  • English and Mandarin.

Where to start

If you have a thermal model and test data that disagree, start with a Thermal Model Check. If you have a layout and a temperature gate, start with a Diagnostic Sprint. Send the geometry class, the temperature gate and what you've already measured.

Book a 20-minute scope call