Power amplifiers – Playing it safe
Here’s a guest blog this week from Apex Microtechnology
When your power amplifier mimics a steam engine, you know that something is wrong – terribly wrong. Using power amplifiers for high voltage or high current applications has its challenges, in particular when the output of the amplifier is used for equipment that is handled by others or that is used in harsh environments. The impact for mishandling can go way beyond just a straight damage to the power amplifier. It can shut down a production line, damage other portions of the equipment, or can even pose the threat of harming an individual.I didn’t do anything wrong!?
It’s common that you might burn an amplifier or two when you design and test a new circuit. But once you worked out the kinks, protected the inputs against voltage transients, put in the right compensation to get the desired stability, you expect a design to work reliably.
When we see returns of our power amplifiers from the field, it is very likely that the failure analysis reveals an operation of the power amplifier outside its safe operating area (SOA). This can be extensive stress due to operating the device outside its input or supply voltage boundaries or beyond its temperature limits, running higher currents than allowed for a certain period of time, or by the end user causing a short in the output circuitry. And stress at that time is contagious – once it did its damage to the device and the system, it does try to infect the operator of the equipment and other people involved.Aim for the Early Bird Special
While you often can’t control the handling of your equipment, various options are available to protect your system. The best time to start thinking about system protection is right when you write your system specification. It’s a good idea at that time to analyze potential fault and mishandling scenarios and their impact on the overall system. Once that list is complete, you can then plan for measures in your design to limit the impact of the faults you identified as critical. Pushing out the fault analysis to a later time will be costly – the later you come up with a change or a new requirement, the higher the impact on the cost for the design and the schedule. A diode that protects the input pin of an amplifier or a filter capacitor is added easily when you start the schematics and the layout of a board, but once the board is complete, it is a different story.Let’s talk Protection
We know that high current or high voltage analog design is tricky, and our components often play an integral part for the safe operation of the equipment. Through our long experience in analog power amplifier design we developed a menu of options for protecting the amplifier and the system. In the following I would like to introduce and discuss some of the measures available.
As mentioned earlier, one of the causes for field failures is operating a unit outside its temperature limit. Putting a temperature sensor close to the components that cause the most heat can limit the operation of the power amplifier to a safe temperature range. In a discrete design, one needs to analyze the system layout carefully to determine the best position for the temperature sensor. Often enough, though, the best place for the sensor is already reserved for other components to eliminate long routes and noise, so it’s a trade-off. Hybrid or multi-chip technology allows for putting the sensor right inside the power amplifier package, eliminating the hassle of defining the best spot on your system. Depending on the device, an external controller will read the temperature of the power amplifier and control the health of the system, or the power amplifier internally shuts down its operation once the temperature exceeds a certain threshold. Integrating a temperature hysteresis between system shutdown and system restart, either by the external control or by built-in circuitry, is key. It will ensure that the device and the system can cool down enough before getting back to work again.
Limiting the maximum output current of the power amplifier provides another level of protection. Most power amplifiers provide pins for a current limit resistor. The amplifier will measure the voltage across the current limit resistor, adjusting the current if the voltage exceeds the set threshold. However, that current limit might not be fully comprehensive, as the power dissipation of the device needs to be taken into consideration as well. As an example, if we output 40V/5A out of a class AB power amplifier with ±50V supply voltage and 100W heat sinking capability, we have a voltage drop of 10V between the positive supply and the output voltage, resulting in power dissipation within the amplifier of 50W (10V x 5A), meaning within its specified range. At the same time, if we output 10V/5A under the supply voltage same conditions, we have a 40V voltage drop between the supply voltage and the output voltage. The power dissipation within the amplifier is now 200W (40V x 5A), exceeding the 100W heat sinking capabilities by a factor of two. So although we limited the current to 5A, the power dissipation within the amplifier of 50W might be in the safe operating area, while the 200W dissipated power is beyond what the power amplifier can handle over an extended period of time, resulting in over-heating.
SOA charts are available that indicate what current the device can handle relative to its supply and output voltage. In the example the SOA of the MP108 is shown, an operational class AB amplifier that can drive 12A peak, has a power dissipation capability of 100W and a maximum supply voltage of 200V. Assuming we operate the MP108 amplifier at room temperature (25°C), it shows that even with the current limit set to 8A, if the voltage drop from the supply voltage to the output voltage is around 30V, the part will get damaged if drive 8A output current for more than 10ms.
Implementing current limits can be done in various ways. One might provide a current power supply that has a built-in current limit. However, when planning to use that option, capacitances on the system have to be considered, as well how as quickly the current limit kicks in. Capacitances on the board that are used to stabilize the power supply can also turn into current sources under some fault conditions, allowing the system to exceed the maximum current for a limited period of time. An alternative is to place a fuse in the supply line of the amplifier. It is better to replace a blown fuse than replacing major components of your system, but board space and accessibility to the fuse need to be considered. However, fuses have some response time. That can be useful if you want to allow an overcurrent for a short period of time while protecting the system from a long-term overload. But it can also be of disadvantage as the delay could be too long to provide full protection. Putting the current limit inside the power amplifier usually is the most straight-forward solution. High power amplifiers typically provide such current limit option, either with a built-in fixed current limit or by offering two pins for setting the current limit through an external resistor.
But as explained earlier, current limits don’t necessarily provide the proper protection against short circuits, as a component might be operated outside the SOA. One potential answer is the use of power amplifiers that offer internal short circuit detection and protection. Once a short is detected, the amplifier will shut down immediately – immediately meaning within nanoseconds or microseconds. A reset pulse or a power cycle will bring the amplifier back to life. All of the options to protect against shorts have their advantages and disadvantages. Before settling for any of them, the impact on the system needs to be analyzed and weighted.Weigh the Options
Designing a circuit with power amplifiers is a challenge, and while settling on power amplifier architecture you need to look beyond the pure performance characteristics offered. A good design also includes appropriate measures to protect against malfunction.
Power amplifiers have become more sophisticated and provide increased protection functionality, like input voltage protection, temperature monitoring or short circuit detection. However, they can’t cover all the faults that can damage a system. The importance of an early fault analysis can’t be stressed enough; it will save you money at the end. A good source of information is typically the application note library that is offered by most power amplifier manufacturers. But if that also leaves questions or concerns unanswered, go a step further – drop an email to the technical support team and ask for help!