Title: Thermal design for Six Sigma of an imaging device
Thermal capability of an imaging device with smart temperature control was estimated using Monte Carlo simulations with a transfer function derived from a Design-of-Experiments scenario, executed in Flotherms Command centre.
The results show that typical random errors +/- 6 C (3σ) in the embedded temperature sensors of the imaging IC-set can lead to significant differences in performance of devices equipped with smart temperature control i.e. where the performance is regulated down to cope with high ambient temperatures. Hence, a strategy of ‘smart temperature control’ on component level can backfire on system level.
Using CFD simulation in architecture and implementation design phases shows the risks and enables early mitigation through an optimized thermal design and/or factory calibration. Using an external housing as external heatsink shifts the temperature curves to higher working regions, but has less effect compared to an optimized thermal architecture and does not mitigate the variation between devices.
Figure 1: Original design without (blue curve) and with (red curve) external housing used as additional heatsink, without ( Left ) and with ( Right ) calibration.
Figure 1 Shows the original design (blue curve), without (Right) and with (Left) factory calibration. The design shows 100% performance @ 20C in the lab, but very significant spread in performance – some devices start dimming at 23 C ambient and ~80% performance @ 25C ambient. Factory calibration shifts this to running at ~95% performance at 25 C.
Figure 2: Optimised design, without (blue curve) and with (red curve) external housing, without (Left) and with (right) factory calibration
Figure 2 shows the same curves for a design optimised using aan Design-of-Experimentsin archtecture phase . With the same electronics and this design has 100% performance well up to 40C, and 55 C with an external housing utilized as heatsink.