Thomas McKee Design Engineering Book a call →
Case study · Ooni Volt 2

Designing how an electric oven performs, from the heating elements to the firmware.

The Volt 2 is Ooni's compact electric indoor oven, reaching 450 °C from a domestic plug. I was lead engineer for its function and performance across both hardware and firmware: the inner cavity, heating elements, shells, insulation and thermocouples, and the application firmware and UI that control them. It won a Red Dot Design Award in 2026.

Start your project → See the process
+41%
usable temperature range vs the previous generation
70%
better stone temperature distribution
23%
faster heat-up, with a 64% larger window
>95%
yield on the deep-drawn cavity parts
The design process

How the project ran, end to end

The Volt 2 was a performance project first. It had to heat up faster and cook better than the oven before it, in a smaller box with far more glass, and each of those targets worked against the others. Hardware and firmware were developed together in one loop, so no single part had to solve the problem on its own. This is how I worked through it.

  1. Define the performance targets

    Before any geometry or code, I wrote the product and firmware performance requirements: heat-up time, usable temperature range, thermocouple accuracy, holding 20 °C for proofing, and the cooking styles it had to nail.

    Requirements · technical documentation · test planning

  2. Understand the physics first

    Hand calculations on element surface loading, radiant view factors and conduction into the stone told me what was physically possible under a 110 V supply before I committed to a design.

    Hand calculation · heat transfer · thermal engineering

  3. Build rigs and measure

    Instrumented rigs with thermocouples across the stone, logged on multiple channels, gave me real data on the old elements and every new idea.

    Test rig design · instrumentation · multi-channel data logging

  4. Validate a model, then let it do the searching

    A thermal simulation matched to the rig data let me test element profiles in software, and a Python model of the whole oven let me develop control logic without waiting on physical heat-ups.

    SolidWorks Flow Simulation · Python · mathematical modelling

  5. Design the hardware and firmware together

    Elements, cavity, control logic and UI were iterated as one system, with firmware revisions carrying both the logic flow and a parameter table.

    Application firmware · state machines · UI and UX definition

  6. Design for manufacture and work the tooling

    The deep-drawn stainless cavity was designed around its process from the first sketch, then taken through tooling trials with the supplier, much of it on the factory floor.

    DFMA · sheet metal and deep drawing · supplier liaison

  7. Verify, then support production

    I led the test teams in design verification, covering error states, misuse, touch temperature, water ingress and longevity, and supported the parts into mass production.

    Design verification · compliance · production support
1.1 · Heating elements 1.2 · Thermal model & firmware 1.3 · Cavity DFMA
Skills used

What this project called on

Engineering
Thermal engineeringHeat transferHand calculationMathematical modellingControl systemsApplication firmwareUI & UX definitionRequirements & verificationDFMATolerancingMaterials selectionCompliance
Tools
SolidWorksSolidWorks Flow SimulationPythonMulti-channel data loggingThermal imagingTest rig instrumentation
Manufacturing
Deep drawingSheet metal pressingTubular heating elementsTooling trialsSpringback compensationOverseas mass production
Case study 1.1 · Heating elements

Designing the Volt 2 heating elements for improved performance under high constraints

Problem

The original Volt used off-the-shelf circular elements, and they caused a set of coupled problems. Poor temperature distribution across the stone scorched bases in the shape of the element beneath them, and the radiation only reached the top of the pizza and not the side, leaving a pale, undercooked crust.

The constraints were tight on every side. Element power was capped by a 110 V (US) domestic supply. Cavity height was fixed by other functional requirements. Each element had to be manufacturable as a single, continuous bent tube, with a large enough circumference for the radiation to reach the side of the pizza, yet short enough to keep a low energy density and a rapid response.

Defined failure modes: uneven stone temperature · pale crust edge · cold spot at the terminations · stone thermal shock.
Fig. 2, Off-the-shelf circular element, simulated stone temperature
Fig. 3, Scorched base
Fig. 4, Pale crust edge
Process

From hand calculations to a validated simulation

  1. 01Turn the complaints into failure modes. Writing the problems down as four defined failure modes meant every design could be judged against the same criteria.
  2. 02Size it from first principles. With hand calculations I worked out the element surface loading, the radiant view factor to the stone and the pizza edge, and the conduction path into the stone. Surface loading q″ = P / πdL sets the sheath temperature, so a longer element at the same power spreads the heat and lowers the scorch risk. From there I drafted the first element designs.
  3. 03Build a rig and measure. I bonded thermocouples across the stone on a grid and logged them on multiple channels at once, starting with the off-the-shelf element as a baseline.
  4. 04Validate the simulation against the rig. I modelled the cavity in SolidWorks Flow Simulation with conduction, convection and radiation active, the element as a heat-generating solid and the stone as a conducting solid. Once the simulated stone field matched the measured one in both shape and magnitude, I could trust it.
  5. 05Optimise in software. With a validated model I tested profile after profile across serpentine pitch, perimeter offset, corner radius and the split of duty between top and bottom, checking heat-up time and element reactivity each time. The top element was optimised separately for cooking. Using two elements for their different strengths was the change that unlocked it, where the previous generation had tried to heat the stone and cook with one shape.
  6. 06Prove it on the plate. The final elements went into prototype ovens for cook testing and user testing across every style.
Fig. 6, Notebook hand calculations: element length against temperature rise
Fig. 7, Testing the off-the-shelf element on the instrumented rig
Fig. 9, Simulation results for the off-the-shelf element against the new profiles
Fig. 10, Transient thermal simulation
Results

70% better stone temperature distribution and markedly improved cooking performance

70%better stone distribution
0expensive prototypes before the final design
100%of user testers saw better cooking

The results were positive both objectively and subjectively: a far more even stone temperature across the cooking plane, and user testing that returned 100% agreement on improved cooking performance, all while meeting the heat-up and reaction targets. Bases cook evenly, the pale ring around the edge has gone, and the radiant intensity at the top is high enough for leopard spotting on a Neapolitan, while the same oven can still bake an even Detroit, or a Victoria sponge.

Fig. 13, Final simulation of the optimised elements
Fig. 14, Base, evenly cooked
Fig. 15, Leopard-spotted Neapolitan crust
Case study 1.2 · Thermal model & control firmware

Optimising heat-up and cooking performance despite a bigger window and a bigger cavity

Problem

The Volt 2 has a 64% larger viewing window in a 30% smaller footprint. Glass is a poor insulator and lets radiation straight through, and the smaller envelope left less room for insulation, so the oven started in a worse thermal position than the previous generation. The brief still asked for faster heat-up.

The bottom element heats the stone efficiently because its radiation is trapped, but on its own it scorches the base. The top element is less efficient but crucial for a good cook. Getting the most out of both meant optimising how the firmware shared the power between them, and that work later became the basis for Pizza Intelligence™.

Fig. 17, Volt 2 against Volt 1: larger window, smaller footprint
Fig. 18, Element scheduling trials for optimum heat-up
Process

Writing my own simulation, then developing the firmware against it

  1. 01Write the spec before the code. I wrote the product and firmware performance document covering heat-up time, thermocouple accuracy tolerance, holding 20 °C for proofing, and subjective targets such as New York style cooking.
  2. 02Model the oven from first principles. No off-the-shelf software would let me test control logic, so I wrote my own. Through mathematical modelling I treated the oven as two lumped thermal masses, the stone as one slow node and the element, shell and cavity air as a faster one, with radiation linearised about the operating point. The energy balance reduces to a single second-order equation in stone temperature, d²T/dt² + a·dT/dt + b·T = c.
  3. 03Fit it to real data. I fitted the coefficients for three states, top element on, bottom element on and idle, using least-squares fitting against logged thermocouple data, until the model reproduced the real oven closely enough to develop and debug logic on.
  4. 04Run the firmware against the model. I coded the model in Python and ran it as a loop against the control firmware. Each firmware revision carried a logic flow and a parameter table, revision-controlled together. I debugged the logic with standard parameters first, then tuned performance, and found edge cases and first-pass parameters before any lengthy heat-up tests.
  5. 05Fix a late hardware problem in firmware. Late in development the stone thermistor failed food safety, and its replacement read so poorly that one reading matched several different stone temperatures, so no polynomial could map it. Changing tooling would have cost time and money, so I solved it in firmware with a state-aware transform, tested rapidly on the model. It became a building block of Pizza Intelligence™.
  6. 06Verify safety and the edge cases. I defined error codes, misuse behaviour, a hot-surface warning, a failsafe on sensor loss and verified control limits, and led the test teams through design verification.
Fig. 19, The model coded up, run as a loop against the firmware
Fig. 20, Model against empirical data
Fig. 23, Thermistor reading against true stone temperature
Fig. 24, Hand-built prototype used for firmware development
Results

A wider usable range and a faster heat-up, in a smaller box with more glass

+41%usable range, 20 to 450 °C
23%faster heat-up
50%faster than competitors at the same max
−30%footprint

The oven holds a steady 20 °C for proofing dough at one end of its range and reaches 450 °C for Neapolitan at the other, with modes for New York, Detroit, grill and a domestic oven, plus smart presets for beginners and full control for experts in one UI. Pizza Intelligence™ (patent pending) works out when a pizza has gone in from the stone temperature signature alone, then adapts in real time, and launch detection is reliable across every preset. It also finds its temperature again after a power cycle or a mid-cook mode change.

All of it shipped on time and on budget with no tooling change. The Volt 2 sits at 4.5 stars across 374 reviews, and every one of the 21% of reviews that mention cooking results or performance is five stars.

Fig. 26, Usable temperature range, 20 to 450 °C
Fig. 29, Stone temperature signature, stage 3 is the launch event
Case study 1.3 · Cavity DFMA & environment

Inner cavity DFMA, and designing for mixing steam and electronics at 450 °C

Problem

The inner cavity is a deep-drawn and pressed stainless steel assembly. It has to locate the elements and the thermocouple accurately, seal against steam, survive thermal cycling to 450 °C, and be assembled quickly at volume.

It also sits in a demanding environment. Cooking produces steam, and the electronics sit right next to the cavity, so the design had to keep water away from the electronics, keep them cool, and keep every surface a user can touch within the touch-temperature limits.

Process

Designing around the process, then solving it on the factory floor

  1. 01Sketch the architecture around the process. Early sketches worked out the pressings, the fixings and how the elements would be held, settling on two pressings and a wrap.
  2. 02Apply DFM from the first concept. One line of draw where possible, generous flow radii, features pressed rather than welded, and datums chosen so tolerance builds up away from the element mounts.
  3. 03Design for the environment. Material selection for high heat and steam, and a steam chimney to route moisture away from the control cavity.
  4. 04Verify touch temperatures. I used the thermal model and thermal imaging to check every touchable surface against the standard.
  5. 05Work the tooling with the supplier. Springback was one of the hardest problems on the project, and we compensated for it in the tool together. I spent time on the factory floor through slow-build tooling trials and inspection, diagnosing and fixing tooling issues, then took the parts into mass production.
Fig. 32, Early design sketch: pressings, fixings, element accommodation
Fig. 33, Assembly concept: two pressings and the wrap
Fig. 31, Section view through the cavity assembly
Results

A high-yield cavity that locates everything without adjustment

>95%yield on parts
0water ingress
5 yrguarantee

Over 95% yield on the parts, cheaper to make and quicker to assemble, with the elements and thermocouple locating without any adjustment. The oven is compliant on touch temperature, has no water ingress, passed longevity testing and carries a five-year guarantee.

Fig. 35, Production tooling for my parts
Fig. 36, Production line
Fig. 37, Thermal imaging for touch-temperature compliance
How I could help you

The same process, on your product

Measure first, validate a model against the real thing, then let the model do the searching. It's how I approach Design Verification, and it runs through every phase of Product Development. If you have a coupled thermal, control or firmware and mechanical problem, it's home territory for me.

Let's talk

Got a hard performance problem?

Book a free intro call and tell me about it. If it's the kind of thermal, control or crossover challenge the Volt 2 was, it's exactly what I love to work on.

Book a free intro call → See more work