A power supply’s output power is conditional. Long, reliable equipment life demands a more detailed analysis based on the conditions the supply will encounter in the end application. Simply stated, your design and attention to equipment cooling has a profound impact on the operational life of the power supply.
I demonstrate the constraints using the XP Power CCR550PS12 as shown in Figure 1. This is an enclosed AC-DC converter with a maximum power rating of 550 W. This particular power supply is ideal for analysis as XP Power has given us enough data to do a careful power output vs heat analysis. To better understand this specification, we will refer to the XP Power Your Essential Guide to Power Supplies. This 154-page exemplar document introduces power supplies and covers advanced topics such as cooling and derating at elevated temperatures.
Key Takeaways
- Usable power and equipment life is baked into your design because the full rated power is conditional.
- Do not assume the equipment’s sheet-metal enclosure will have sufficient cooling ability.
- Design for elevated temperatures, especially if the enclosure is sealed.
- Don’t design for ideal conditions. Equipment often has a life measured in decades. Considerable dust accumulation or poor maintenance on filters can degrade thermal performance.
- The power supply’s nameplate output is correct only under the correct conditions. Capacitor life, and consequently equipment life, will be significantly reduced if the heatsink is insufficient.
Figure 1: Image of the XP Power supply on the author’s workbench.
What is the output power of the CCR550?
The featured power supply has an output power that is entirely dependent on the cooling mechanism:
- 230 W: Convection cooled
- 450 W: Conduction cooled
- 550 W: Fan cooled
To put this into perspective, let’s consider the 450 W output specification. This can be achieved using a heatsink with a 1.0 °C/W rating. One possible solution is the Wakefield 127684 with a 1.04 °C/W natural (no fan) thermal resistance. This aluminum heatsink has 108 square inch footprint with a height of 0.9 inches.
This natural convection heatsink’s footprint is about six times larger than the power supply itself! This is a very important distinction as designers will be tempted to use the equipment chassis as the heatsink. Clearly this will not work as all but the most robust metal enclosures will have a higher thermal resistance than the massive 1.0 °C/W heatsink. Note that a cooling fan may be used to reduce the heatsink size.
Figure 2: Datasheet specifications for the CCR550 series power supply.
What is the problem with long-term sustained elevated temperature?
Elevated temperature accelerates component failure, especially electrolytic capacitors. Note that the featured supply is semi-potted, as shown in Figure 3. Consequently, all components have the same temperature. Stated another way, the capacitors have the same temperature as the power supply’s metal base.
The XP Power document clearly describes the capacitor situation:
They [data sheets] typically also provide an estimated service life curve based on the temperature of key electrolytic capacitors, which are the only parts with a wear out mechanism within the power supply.
The XP Power document also provides this gem:
All electrolytic capacitor lifetime calculations are based on the Arrhenius equation, where the rate of reaction halves and hence the lifetime doubles for every ten degree Celsius reduction in temperature, making this a critical element in the service life or service interval of the entire end application.
I wrote more about the Arrhenius equation in this article. It introduces the Rule of 10 which provides a useful rule of thumb implying that equipment life is cut approximately in half for every 10°C rise in temperature. However, the article also points out that ripple current also impacts capacitor life.
Applying the Rule to the Output Capacitors
The Gemcon-manufactured output capacitors are visible in Figure 3. These GPH Series conductive polymer aluminum solid capacitors are rated for 2000-hour endurance at the extreme operating temperature of 125°C. A simplified calculation suggests that the temperature should be held below 55°C for a three-decade service life. Note that this does not include the ripple current calculations or self-heating due to the high-frequency signal components of the switched-mode power supply. Also note that 55°C is painfully hot in human terms. It’s like scalding hot water.
Using the simplified Rule of 10, the power supply’s longevity may be expressed as:
- 2,000 hours (83 days) at 125°C
- 4,000 hours (167 days) at 115°C
- 8,000 hours (0.9 years) at 105°C
- 16,000 hours (1.8 years) at 95°C
- 32,000 hours (3.7 years) at 85°C
- 64,000 hours (7.3 years) at 75°C
- 128,000 hours (14.6 years) at 65°C
- 256,000 hours (29 years) at 55°C
Tech Tip: Treat the lifecycle as an aggregate calculation where the chain is only as strong as the weakest link. For example, the solid polymer capacitors pictured in Figure 4 may have a significantly better life than the simplistic estimate suggests. Application ANP701 by Würth Elektronik by provides this enlightening quote:
With liquid electrolytic capacitors, the expected lifetime doubles when the temperature at the component is reduced by 10 °C (2). For polymer electrolytic capacitors, the life increases tenfold when the temperature at the component is reduced by 20 °C (1).
Note that the Würth Elektronik application note refers to parts similar to this DigiKey supplied capacitor.
Figure 3: Semi-potted output filter capacitors as installed on the XP Power CCR550PS12.
What is the problem with short-term extreme temperatures?
The long-term sustained lifetime is dominated by the capacitor aging. We have already established that operation above 50°C is undesirable. Yet, in Figure 4, we see a slope starting at 50°C and ending at 80°C. This short-term operation at extreme temperatures is dominated by semiconductor limitations. This is the classic linear derating curve applied to semiconductors operating at elevated temperatures.
Actually, this is rather remarkable as many semiconductors are linearly derated by a specific Watts per degree Celsius for operation above 25°C. The fact that the XP Power supply starts at 50°C implies a wide design margin. It also reflects the fact that power supply chassis temperature will be lower than the semiconductor junction temperature due to thermal resistance.
As another example, consider the blue (convection cooled @110 V) curve and the purple (convection cooled @220 VAC) thermal derating curves. The power supply is rated at 230 W and 260 W respectively. We can assume the difference between the two lines is associated with the rectified voltage and the higher current associated with the lower supply voltage. We can interpret this as greater losses (increased heating) at higher currents or simply as less efficiency at lower input voltages.
Figure 4: Thermal derating curve from the XP Power datasheet.
Power Supply Operational Checklist
Let’s conclude with these rules of thumb for equipment design.
-
Fan noise is objectionable especially in AV and medical applications. The power supply can be derated, or a larger heatsink with a lower-speed (quieter) fan may be suitable.
-
Contamination will reduce cooling efficiency. Figure 5 presents an extreme example where accumulated dust has reduced the cooling efficiency of an old fan-cooled supply. Design this dusty assumption into your equipment.
-
For long operational life, maintain temperatures well below 50°C. This requires that you find the balance between equipment lifespan, size, fan cooling requirements, and heatsink size.
-
There may be times when intermittent operation at extreme temperatures is warranted. However, you must be careful that intermittent operation does not become sustained operation. For example, operating near the 80°C limit will reduce equipment life from decades to years.
-
Consider adding temperature sensing to the equipment. We are all guilty of installing equipment in small enclosures and inadvertently blocking the airflow. Alerting the operator to an overtemperature condition could lead to longer equipment life.
Figure 5: Dust ingress into an old fan-cooled power supply.
About This Author
Aaron Dahlen, LCDR USCG (Ret.), is a Senior Applications Engineer at DigiKey in Thief River Falls. His background in electronics and industrial automation was shaped by a 27-year military career as both technician and engineer, followed by over a decade of teaching.
Dahlen holds an MSEE from Minnesota State University, Mankato. He has taught in an ABET-accredited electrical engineering program, served as coordinator of an electronic engineering technology program, and instructed military technicians in component-level repair.
Today, he has returned to his home in northern Minnesota, completing a decades-long journey that began with a search for capacitors. Read his story here.