In spacecraft and other radiation-exposed systems, a power converter must continue regulating its output as its semiconductor devices accumulate radiation effects. A radiation-hardened Buck converter design described in a published circuit study addresses one specific vulnerability: leakage current in the adaptive on-time (AOT) timing circuit. By compensating for that leakage, the design maintained stable operation during testing after a total ionizing dose of 350 krad(Si).
The work is relevant to engineers selecting or designing point-of-load power stages for space applications. It illustrates how radiation tolerance can be addressed not only through device selection or layout, but also through the control circuit that sets switching behavior.
Why total ionizing dose can affect a Buck converter
Total ionizing dose (TID) exposure can alter CMOS device characteristics. The effects described in the study include threshold-voltage shifts, reduced transconductance and carrier mobility, and increased leakage current. As these changes accumulate, individual devices and circuit blocks may degrade, potentially affecting the operation of the complete converter.
In an AOT Buck, the adaptive on-time generator is particularly important because it sets the duration of the switching pulse and influences the converter’s switching frequency. If radiation-induced leakage changes the timing circuit’s behavior, the converter may no longer operate as intended. The study identifies NMOS leakage as a key concern in its design and focuses compensation on errors caused by that leakage during on-time measurement.
Radiation-hardened Buck converter timing design
The proposed circuit combines an adaptive on-time timing block with a charging-current generator. In an AOT control scheme, the timing block determines how long the power switch remains on. The charging-current generator supplies current to timing capacitors, allowing the on-time to respond to operating conditions.
The design’s central technique is to create a matching leakage condition for two NMOS devices in the timing block. The circuit uses a 2:1 current-mirror relationship and equal capacitances, with the charging current for one timing node set to twice that for the other. Under the described operating conditions, this produces corresponding voltages across the relevant devices. Their gate-to-source and drain-to-source voltages are made equal, so their leakage currents also match.
That matching allows the leakage associated with one device to compensate for the timing error introduced by leakage in the other. In practical terms, the circuit is designed so that radiation-induced leakage does not simply appear as an uncontrolled disturbance to the on-time interval.
Why compensation is applied to the timing phase
The analysis distinguishes between leakage during on-time measurement and leakage during the discharge interval. The compensation targets leakage from the NMOS device that would otherwise affect on-time timing. The paper states that leakage during discharge does not influence the Buck converter’s on-time. It also explains that leakage from another NMOS device in the timing circuit does not affect the interval: during the top PMOS on-time, the relevant PMOS is off, while leakage during discharge is likewise not consequential to the timing result.
This phase-specific approach is important for circuit design. Rather than treating all leakage as equally disruptive, the circuit addresses the path and operating interval that directly influence the controlled timing variable.
Maintaining switching behavior as operating conditions change
The charging-current generator and timing capacitors together establish the on-time. The study describes the on-time as proportional to the converter duty ratio through a circuit-dependent constant. This relationship is intended to keep the switching frequency near a fixed value as input and output conditions change.
Simulation compared the proposed compensation approach with a conventional AOT Buck without leakage-current compensation. Leakage was represented as a resistance in parallel with an NMOS device, with the resistance varied to represent different leakage conditions. The comparison covered duty ratios of 20%, 40%, 60% and 80%. These simulations evaluated how the timing behavior responded to the modeled leakage; they were distinct from the subsequent irradiation tests on fabricated hardware.
What the irradiation tests demonstrated
The converter was fabricated in a standard 0.18 μm BCD process. Packaged chips mounted on a PCB were exposed to Co-60 gamma radiation at a dose rate of 50 rad/s, with the total dose increased in steps up to 350 krad(Si). Output measurements were reported at an input voltage of 6 V and an output voltage of 1.2 V.
At 350 krad(Si), the conventional AOT Buck used for comparison was reported to be unstable, while the proposed design remained stable. The measured output-voltage ripple for the proposed converter was 30 mV at that dose, compared with 66 mV for the conventional design. Across the reported dose points of 0, 100, 200, 300 and 350 krad(Si), ripple for the proposed design was 24, 26, 30, 24 and 30 mV, respectively. The corresponding values for the conventional design were 25, 52, 53, 56 and 66 mV.
The study also reports that the proposed converter maintained a switching frequency of 1 MHz at 350 krad(Si) under two tested input-output conditions: 6 V to 1.2 V and 15 V to 6 V. These results describe the tested circuit and conditions; they should not be taken as a universal radiation rating for other converter designs or applications.
Practical implications for space power design
For engineers evaluating power conversion in radiation environments, the work highlights the value of examining control-loop timing circuitry alongside power devices. A converter may use suitable power components yet remain vulnerable if radiation-induced leakage shifts the timing behavior that governs switching.
- Identify sensitive control blocks: Assess how leakage in timing, sensing and current-generation circuits could affect converter operation.
- Design for the relevant operating phase: Determine whether leakage affects on-time generation, discharge behavior or another part of the switching cycle.
- Verify under irradiation: Simulation can help compare circuit responses to modeled leakage, while radiation testing checks behavior in fabricated hardware under defined conditions.
The reported results show a circuit-level method for improving AOT Buck stability under TID exposure: match device conditions so leakage errors in the timing block can be compensated. For space power systems, the practical lesson is to treat radiation tolerance as a control-circuit design concern as well as a component-selection issue, and to validate performance against the dose and operating conditions required by the application.