Filter by:  

 

How to select electrolytic capacitors for power applications

  • 01-Aug-2018
  • Articles|Technical Articles


Rudy Ramos explores the relative strengths of film and electrolytic capacitors in power-oriented applications and offers tips to select the best type for any application.
 
One important role of capacitors is to diminish ripple and noise in power conversion applications by using the energy they store to provide a hold-up (or ride-through) function. However, making the right (or wrong!) choice can have a significant impact on the size, performance and cost of the application. In this article, we will look at the relative strengths of film and electrolytic capacitors in popular applications.
 
The basics
Film capacitors are a good choice for power conditioning in applications where reliability is important, such as electric vehicles, renewable energy, and industrial drives. This is mainly due to the low ESR rating which makes them good at dealing with ripple current, although their ability to self-heal and deal with high levels of surge voltage are also important in these applications.
 
If there is no hold-up (or ride through) requirement, such as during a power outage or to manage line frequency ripple peaks, then film capacitors are ideal. Similarly, they are good at reliably sourcing or sinking high frequency ripple currents with low losses.
 
Due to their high voltage ratings, film capacitors are suited to high(er) voltage DC busses that are in common use to minimise I2R losses. Aluminum electrolytic capacitors are rarely rated above 550V, so to attain higher voltages they have to be series-connected which can be a challenge for designers. One approach is to specify and select voltage-matched components, which increases cost and takes time. Alternatively, voltage-balancing resistors can be used but this dissipates energy and increases BOM cost and PCB space.
 
In other applications, storing as much energy in as smaller space as possible is a primary concern – this would include applications such as offline power supplies where the stored energy is used to maintain the DC rail in the case of an outage. In these applications, aluminum electrolytic capacitors are the preferred choice due to their energy storage density (measured in joules/cm3). While some have concerns about lifetime and reliability with electrolytics, proper derating can address this issue.
 
Resilience is a difference between the two types, especially when related to over-voltage conditions. Over-voltage above 20% of the rated value will normally damage electrolytics, while film capacitors can withstand double the rated voltage for short durations. Film capacitors also have the ability to self-heal, allowing them to withstand real-world conditions and occasional stresses. 
 
Generally speaking, film capacitors are more convenient to use. They are non-polarised and therefore cannot be mis-inserted. They offer a wider variety of mounting and connection options including screw terminals, lugs, “fastons” or bus bars and are often packaged in easy to handle (and volumetrically efficient) ‘box’ packages.
 
Common film capacitor types are shown in Table 1, with an overview of key parameters. PET devices are generally popular in lower voltage applications, while polypropylene capacitors show the greatest reliability during voltage stress due to the high dielectric breakdown and low dissipation factor (DF). Polypropylene types have a DF that is quite stable with frequency and temperature and also exhibit the lowest losses of the types shown. For high-performance applications, segmented high-crystalline metallised polypropylene offer energy density levels close to that found in aluminum electrolytics.
 
Selecting the correct capacitor
To better understand how the usage of a capacitor (storing energy or suppressing ripple and noise) and technology selection can affect size, weight and cost, it is instructive to analyse common power-conversion topologies. One good example is the bulk capacitance in an offline 1kW converter as shown in Figure 1 as this shows the differences between film and electrolytic. In this example, the converter has power factor correction and the nominal DC bus voltage (Vn) is 400V.
 
If we assume that regulation is lost when Vn is below 300V, then this defines the dropout voltage (Vd). In this example, we will assume that the converter has an efficiency of 90 per cent. If the supply is removed as the result of an outage, then capacitor C1 provides the energy to power the output stage as Vn drops to Vd. To calculate the required value of C1 to give 20ms hold-up / ride through we would use:
 
A suitable electrolytic capacitor would be a TDK-EPCOS B43508 680µF 450V device. This meets the electric requirement and is 35mm diameter x 55mm tall, giving a volume of 53cm3 or around 3 cubic inches. 
 
If we were to select a film capacitor, then the TDK-EPCOS B32678 would be electrically suitable, but would require up to 15 devices connected in parallel, giving a volume of 1500 cubic cm (91 cubic inches) which is not viable in real-world applications.
 
It the requirement was simply to manage voltage ripple on a DC rail (for example in the powertrain of an electric vehicle) then the situation is very different. As the bus voltage is driven by a battery, there is no need for hold-up as the supply is always present. With a 400V rail, 4V of ripple (1 per cent) would be a realistic goal. In this example, we will assume that the converter is needed to supply 80A (rms) pulse-current at a frequency of 20 kHz. The capacitance can be calculated as:
 
If we chose a 180µF 450V electrolytic capacitor from the TDK-EPCOS B43508 series with a ripple current rating of 3.5A (rms) then, to handle the 80A requirement, we would need 23 capacitors arranged in a parallel configuration. This is, however, not a viable solution as the capacitance would be 4.14mF and occupy a space of 1200 cubic cm (73 cubic inches). Interestingly, this solution aligns with the 20mA/µF rule of thumb that is often used for electrolytic-capacitor ripple-current ratings.
 
However, if we were to look to film capacitors such as the TDK-EPCOS B32678 series, we could deliver a solution with only four capacitors in parallel. This would give 132A(rms) ripple current rating in approximately one-third of the size of the electrolytic solution (402 cubic cm or 24.5 cubic inches). In fact, if the operating temperature is below 70°C, then smaller devices could be selected, reducing the size of the solution even further.
 
There are other issues to be considered other than just the size, such as the inrush current due to the extremely high capacitance levels of the electrolytic solution. Also, light traction applications such as electric vehicles commonly have transient voltage spikes, so film types will deliver greater reliability due to their robustness.
 
Similar configurations of power conversion are also found in applications such as UPS systems, power conditioning in wind or solar generators, general grid-tied inverters, and welders so this analysis is equally valid there.
 
Should film be the first choice?
The costs of electrolytic and film capacitors vary depending upon their usage in the application, whether the need is for energy storage or ripple reduction. Figures from 2013 (shown in Table 2) provide a useful guide for a DC bus powered by a 440VAC supply.
 
Looking at the costs above, film capacitors are ideally suited to decoupling, switch snubbing, and filtering applications such as EMI suppression or inverter-output filtering. When used for decoupling and placed across the DC bus of an inverter or converter, a film capacitor provides a low-inductance path for circulating high-frequency currents. To size the capacitor, the rule of thumb stating 1µF per 100A switched is quite useful and accurate.
 
Any connections to the capacitor should be short to avoid parasitic inductance inducing voltage transients. With high currents and rapid switching, slopes of 1000A/µs often occur: considering that PCB traces can have inductance of about 1nH/mm, each millimeter can create a transient calculated by:
 
In a typical snubbing circuit for a MOSFET or IGBT, the capacitor is fitted in series with a parallel resistor/diode combination, and this is connected across the switch to manage dV/dt, as shown in Figure 2.
 
The snubber will reduce the speed of the ringing and thereby control the EMI. It also prevents spurious turn-on/turn-off. A rule-of-thumb to calculate the snubber capacitance is to use approx. double the sum of the switch output capacitance and mounting capacitance. The resistance value is then calculated to critically damp any ringing.
 
Suppressing EMI
X and Y capacitors reduce differential- and common-mode noise, as shown in figure 3. Film capacitors are ideal for use in this application, benefitting from their self-healing and transient-overvoltage capabilities. Safety-rated X1 (4kV) or X2 (2.5kV) capacitors are regularly connected across the mains feed. These are often polypropylene capacitors with values of several microfarads – the exact value needed is defined by the requirement to comply with applicable EMC standards. 
 
Y capacitors with low inductance are used for line-to-earth positions (see Figure 3) and are known as Y1 or Y2 capacitors. These are respectively rated for 8kV and 5kV transients in typical mains applications. The amount of capacitance that can be used is limited by leakage current restrictions. External connections to the ground rails should be kept short even though the low connection inductance of film capacitors tends to keep self-resonances high. 
 
Filtering the output of inverters
Applications such as inverters and motor drives often require filtering of the high-frequency harmonics in the AC output. Non-polarised film capacitors are commonly combined with series inductors in a single module to create low-pass filters for this purpose (see Figure 4). In many applications, the load is not close to the power source so this approach reduces dV/dt-related stress on cabling and motors as well as facilitating compliance with system-wide EMC requirements.
 
Summary
In this article, we have explored the relative strengths of film and electrolytic capacitors in power-oriented applications and are now in the position, as designers, to make informed choices about the best type for any application based upon size, reliability and cost. As a summary, the relative benefits are:
 
Electrolytic capacitors: Increased stored energy density (joules/cm3); lower cost bulk capacitance for “hold-up” applications; and ripple current rating is maintained at elevated temperatures
 
Film capacitors: Lower ESR for better ripple performance; increased surge-voltage ratings; and self-healing characteristic improves system lifetime and reliability
 
About the author:
Rudy Ramos is the Project Manager for the Technical Content Marketing team at Mouser Electronics and holds an MBA from Keller Graduate School of Management. He has over 30 years of professional, technical and managerial experience managing complex, time critical projects and programs in various industries including semiconductor, marketing, manufacturing, and military. Previously, Rudy worked for National Semiconductor, Texas Instruments, and his entrepreneur silk screening business. For more details, contact Helen Chung (Asia PR Specialist, Publitek) on email: helen.chung@publitek.com


Advertise Here