Thermal Management
Not all heat is equal: why thermal architecture is getting smarter
By Brian Cumpston, senior director of application engineering, Ventiva T
he old model of pushing air from one end of a system to the other is struggling to keep pace with the demands of AI-era electronics. A more purposeful approach, one that puts airflow exactly where heat density demands it, is changing the rules of system design.
There is a tempting simplicity to the way thermal management has traditionally worked in laptops and compact electronics: install a fan (or two), route a heat pipe from the processor to a fin stack near an exhaust vent, and let moving air do the rest. For the better part of two decades, this approach served the industry well enough. Processors were hot, fans were loud, and users mostly accepted both as facts of life.
That calculus is changing. AI inference workloads running locally on client devices are pushing sustained power envelopes that premium thin-and-light notebooks were never designed to handle. Chips are denser, chassis are thinner, and the acoustic expectations of the market – particularly in the premium segment – are tightening. The fan-based thermal model that got the industry here is increasingly the constraint preventing it from going further.
The problem with centralized cooling The fundamental limitation of a centralized fan-based approach is not that fans are ineffective, as they are quite good at moving air. However, they can only send air where the ducting allows, and every component that sits downstream of the heat source gets progressively warmer air to work with. By the time airflow traverses the length of a notebook chassis, much of its cooling capacity has already been spent.
Also, the physical footprint of a fan creates cascading design constraints. As shown in Figure 1, large circular fan cutouts on a motherboard consume contiguous PCB area, force routing detours that increase trace lengths, create pinch points for high-speed signals, and push components farther apart than signal integrity guidelines would prefer.
26 June 2026
Figure 1. Ubiquitous thermal design for notebooks where either one or two centrifugal fans consume a vast amount of motherboard area.
To keep noise levels acceptable, which is a key metric in the premium segment, fans must spin slowly, which means they must grow in diameter to compensate. Bigger fans mean bigger cutouts, which means less usable board space, which means longer signal paths, more PCB layers, and higher manufacturing cost. It is a compounding tax on every other aspect of system design. In a 28 W thin-and-light platform targeting sub-16 mm chassis thickness, this space pressure is acute. The motherboard footprint is already competing with battery volume, and both are losing ground to a thermal module that was designed around the limitations of rotating machinery rather than the needs of the system.
Ionic cooling: an enabling technology
A conventional fan, by virtue of its circular geometry and the inlet plenum it requires above the impeller, cannot be placed anywhere in a chassis without claiming significant volume and imposing its shape on everything around it. You design the system around the fan, not the fan around the system.
Ionic cooling eliminates most of these constraints, using electrohydrodynamic (EHD)
Components in Electronics Figure 2. Schematic representation of airflow generated by an ionic cooling device.
www.cieonline.co.uk
airflow to move air by generating a flow of ions between high-voltage electrodes, entraining neutral air molecules in the process, as shown in Figure 2. Ventiva’s ionic cooling solutions include an air blower, fin stack, and vapor chamber or heat pipe. They are solid-state, produce no acoustic noise, and the air blowers have a thin, rectangular profile that can be oriented to take air in from one side and exhaust it from the other. This side- in, side-out flow pattern requires no plenum,
no inlet clearance above the device, and no large cutout in the underlying PCB. The practical implications for system layout are significant. An ionic cooling solution can be placed at the rear of a chassis, directly adjacent to the heat exchanger, with airflow travelling in a short, direct path from inlet to fin stack to exhaust vent. The absence of fan cutouts on the main board means the PCB can be a large, contiguous rectangle as shown in Figure 3. This means it is simpler to route, easier to panelise, and better suited to hosting high-speed components with tight placement requirements. High-bandwidth memory can sit closer to the SoC, PCIe lanes can route more directly, and fewer board layers are needed to route around obstacles that no longer exist.
Multiple ionic cooling devices can also operate in parallel within the same system, each assigned to a distinct thermal zone. One or two devices can handle the processor heat exchanger through a short, high-flow, low- impedance path, while a third handles skin temperature management through a separate channel entirely. Because the heat load is distributed across these parallel paths, the total flow demand on any individual device is reduced, and exhaust temperatures from each path remain roughly equalized.
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48
