On this page
Closely spaced heated regions can produce a smoother temperature field away from their source plane than widely spaced regions. The relevant distance is not only the substrate thickness: it is the distance relative to the lateral pattern scale. A simple spatial-mode calculation helps explain that relationship without claiming that heater pitch alone guarantees a uniform load temperature.
Key design decisions
- Define the spatial period of the measured temperature pattern, not only the artwork gap.
- Separate ripple amplitude from mean temperature and total heat flow.
- Use the actual opposite-face and edge boundaries when the semi-infinite approximation is inadequate.
1. Name the repeating temperature feature
For parallel heater runs, artwork pitch usually means a center-to-center repeat distance. The thermal pattern can have that period, but it can also contain broader variations from busbars, edge losses or unequal heating. Start with the temperature feature being evaluated and state its full spatial period along a defined line.
Distinguish amplitude from peak-to-peak variation. A sinusoidal ripple with amplitude 2 kelvins spans 4 kelvins from its minimum to maximum. A useful thermal requirement should identify the observation plane, region and metric. An average across several runs can conceal the very pattern-scale variation that the spacing review is intended to address.
2. Separate the average heat path from the spatial disturbance
Represent temperature as a mean component plus a zero-mean spatial variation. The mean component still needs an energy balance connecting heater power to the load and surroundings. Removing the ripple mathematically does not remove the power that must cross interfaces or leave the assembly.
The model below follows one ripple component in a homogeneous region without internal heat generation. It does not solve the mean surface temperature or the installed heater's efficiency. A highly uniform load surface can remain far above or below its intended mean temperature, and a low-ripple result says nothing by itself about excessive element temperature beneath an interface.
3. Use a steady spatial mode with an explicit boundary
Assume the source plane at depth zero has a specified sinusoidal temperature variation across an otherwise homogeneous, isotropic conducting region. Consider an interior region far from lateral edges, with sufficient depth that the disturbance can decay without encountering a nearby opposite boundary. At steady state and constant conductivity, the source-free temperature disturbance satisfies Laplace's equation.
A solution is a sinusoid along the source plane multiplied by an exponential decay into the material. The lateral and depth second derivatives cancel, which directly checks the governing equation. This is a steady spatial result; its exponential contains distance divided by wavelength, not elapsed time or thermal diffusivity.
θ(x,z) = A0 cos(qx) exp(−qz); q = 2π/λ; A(z)/A0 = exp(−2πz/λ)
- θ is temperature deviation from the separately determined mean field in K.
- A0 is source-plane ripple amplitude in K; A(z) is its amplitude at depth z.
- x, z and full spatial wavelength λ use the same length unit; q has inverse-length units.
Steady source-free constant-property isotropic conduction, sinusoidal prescribed temperature at the source plane and a semi-infinite or suitably remote-boundary region. No interface jump, hole, internal source or lateral edge effect is included.
4. Compare two periods at the same source amplitude
Assume a source-plane ripple amplitude of 20 kelvins and an observation depth of 0.25 millimetre. For a full period of 0.50 millimetre, the depth-to-period ratio is one half and the attenuation is exp of negative pi, approximately 0.04321. The predicted amplitude is about 0.864 kelvin, or 1.729 kelvins peak to peak.
For a period of 1.00 millimetre at the same depth and source amplitude, attenuation is approximately 0.20788. The predicted amplitude is about 4.158 kelvins, or 8.315 kelvins peak to peak. These are assumed mathematical boundary temperatures, not thermal results for a company heater. The comparison isolates spatial scale while holding the imposed amplitude equal.
Real artwork changes may also alter source amplitude because resistance, power per run and contact conditions change. If those quantities differ, the two computed attenuation ratios cannot by themselves rank the finished designs. Measure or model the changed source boundary as well.
5. Express a smoothing target as a depth-to-period ratio
If the ideal model is appropriate and the target amplitude fraction is f between zero and one, the required depth-to-period ratio is at least ln of one divided by f, divided by two pi. For a ten-percent fraction, that ratio is approximately 0.3665. A one-millimetre period therefore needs about 0.3665 millimetre of depth in this particular model to reach that fraction.
This is not a universal minimum ceramic thickness. A finite load face, a bonded metal layer or a convective boundary changes the solution. Added thickness can also increase the mean through-thickness temperature drop and participating thermal mass. Use the ratio as a screening comparison, then solve the actual stack before selecting thickness or claiming a uniformity limit.
z/λ ≥ ln(1/f)/(2π), for 0 < f < 1
- f is the allowable amplitude ratio A(z)/A0, not an absolute temperature tolerance.
- z/λ is dimensionless; for f = 0.10 its minimum is approximately 0.36647.
Same prescribed-temperature spatial-mode and remote-boundary assumptions as the preceding equation. This ratio does not address mean temperature or transient response.
6. Do not confuse prescribed temperature with prescribed power
Conductivity cancels from the normalized attenuation ratio for the stated temperature boundary. That does not mean conductivity is unimportant to heater design. If the boundary instead specifies a sinusoidal heat-flux component of amplitude Q1, Fourier's law gives source-temperature amplitude Q1 divided by conductivity times q in the same semi-infinite model.
Thus changing conductivity can change the starting temperature ripple for the same imposed flux. An electrical pattern generally generates power rather than enforcing a temperature waveform, so its thermal boundary must be derived from the coupled assembly. Keep mean flux and the zero-mean flux modulation separate; the latter may have signed deviations without requiring negative total heating anywhere.
7. Identify what interrupts the ideal smoothing path
Map the actual stack and outline before using an exponential ratio at a functional surface. The physical boundary that matters most may be an interface or nearby opening rather than the nominal ceramic grade.
| Actual feature | Why the simple solution changes | Required comparison |
|---|---|---|
| Nearby opposite load face | The depth boundary reflects the thermal field | Include its temperature, flux or heat-transfer condition |
| Bondline or incomplete contact | A spatially varying interface drop is added | Map contact coverage and interface properties |
| Slot or hole between heater runs | The lateral conduction domain is interrupted | Use the actual outline in the field calculation |
| Busbar or unheated edge region | Additional broad spatial components are present | Evaluate the full useful zone, not one central period |
| Temperature-dependent resistance or conductivity | Source and transport properties are coupled | Check the operating-point model and nonlinear sensitivity |
8. Verify the spatial scale at the functional plane
Use a measurement method whose spatial response can distinguish the expected pattern variation. If the measurement averages over a whole period, it can suppress the ripple independently of any physical heat spreading. Retain position registration between the heater artwork and temperature observations so maxima over traces and spaces can be compared.
Document the operating point, load contact and observation plane with the final result. If a model is used, preserve the source boundary and show that edges, interfaces and the opposite face were treated appropriately. The useful output is a pattern-and-stack-specific uniformity decision, with mean temperature and transient requirements checked separately.
Review heater pitch and load-plane uniformity
Provide both the generating pattern and the plane where temperature uniformity matters.
- Heater artwork with center-to-center pitch, trace widths, busbars, edges and openings.
- Complete stack, conductivity basis, interface conditions and depth to the functional observation plane.
- Mean power and temperature requirements, spatial ripple metric and any measured source-plane variation.
- Operating-point maps, spatial measurement response, load contact and warm-up requirements.
The drawing-upload form loads as you reach this section.