When specifying or purchasing an uninterruptible power supply, it is essential to choose the right topology, as getting it wrong can have a big impact on both performance and costs. However, there is no one topology that’s right for every application.
With only a very few exceptions, UPS systems are currently based on one of three topologies: passive standby, which is sometimes called off-line; line interactive; and double conversion, which is also referred to as on-line. The essential elements of each of these are shown in the accompanying diagram (Figure 1) and these will be examined in further detail below. Before doing so, however, it is useful to look at the functions a UPS may be required to perform.
The most obvious function of a UPS is to maintain power to connected equipment when power from the mains is interrupted. However, there are no fewer than eight other power problems that a UPS may be called upon to perform.
These are compensating for short-term low voltage conditions (power sag); eliminating supply transients (power surges); compensating for medium-term low voltage conditions (brownouts); dealing with medium-term overvoltage on the mains supply; eliminating electrical noise on the mains supply; compensating for supply frequency variations; eliminating switching transients (notches) and dealing with harmonic distortion of the supply.
Not all of these functions are required in every application, of course, and choosing a UPS that offers only those functions that are actually needed, instead of notionally playing safe by specifying a UPS that does everything, can be the key to making very useful savings. With this in mind, let’s take a close look at the three UPS topologies.
Passive standby/off-line UPS
In a passive standby UPS, when mains power is available, it is filtered to minimise power surges and supplied directly to the load, without any form of active conversion. At the same time, the battery is charged. If the mains power is interrupted, the UPS delivers stable power to the load from the battery, via the inverter.
UPS systems of this type are inexpensive and can be a good choice for protecting, for example, individual PCs in SOHO (small office/home office) and non-critical applications. They are not, however, generally considered suitable for more demanding applications.
This is partly because they provide only three of the nine possible protection functions (power failure, power sag and power surge) and partly because under brownout conditions they will use battery power, which will be exhausted after the planned back-up period.
By contrast, other topologies will compensate for the low mains voltage without calling on power from the batteries and can therefore continue to operate indefinitely during a brownout. For these reasons, the choice of UPS topology for critical applications such as data centres and other medium to large scale IT installations is invariably between line interactive and double conversion.
Line interactive UPS
At first sight, the line interactive topology looks very similar to the passive standby technology. The key difference, however, is the inclusion of circuitry that can either increase or reduce the supply voltage before delivering it to the load, without needing to call on battery power.
In modern line interactive UPSs, the incoming mains voltage is continuously monitored by a microprocessor-based system and, if this voltage moves outside of a pre-determined range, compensation is applied instantly and automatically. This type of UPS can, therefore, provide not only the three protection functions available from passive standby systems, but also protection against undervoltage and overvoltage conditions.
Line interactive UPS systems are typically priced somewhere between passive standby systems and double-conversion systems, and are certainly worth considering for most applications, provided that the frequency of the mains supply will be reasonably stable, and that the supply is not unduly contaminated with transients.
Double conversion/off-line UPS
UPS systems based on double-conversion topology are usually seen as the top-of-the-range option and, in many respects, this is true. They comprise a rectifier that takes power from the mains and converts it to DC. This is used to charge the batteries and to supply the output inverter, which converts the DC back to clean AC to supply the critical equipment. If the mains fail, the battery takes over the duty of feeding power to the inverter without any gap in the voltage during changeover.
Double-conversion topology very effectively isolates the load from anything untoward that may be going on with the mains power, and a double-conversion UPS can, therefore, provide all nine of the possible protection functions discussed earlier. In the most critical of applications, therefore, double-conversion is seen by many as a natural choice, but that is not to say that it has no drawbacks.
Double conversion is, for example, the most expensive of the three topologies and because power is converted twice on its way to the load and each stage of power conversion unavoidably involves losses, it is inherently less efficient than the other two topologies when operating with the mains supply present.
It is worth noting, however, that leading UPS suppliers have now devised ways of addressing the efficiency concerns relating to double-conversion UPS systems. Part of the solution, of course, is to improve the efficiency of the converter and inverter. Improvements in semiconductor and related technologies are helping with this, but there is nevertheless a limit to the gains that can be made in this way.
Energy saving techniques
This has led to the development of more innovative solutions, such as energy saver system (ESS) technology. This technology is based on a very simple idea – for the vast majority of the time, power from the mains is of perfectly satisfactory quality to feed direct to the load. So under these conditions why not, in effect, bypass the UPS? After all, with the UPS bypassed, there is no power conversion at all, and efficiency closely approaches 100%.
Of course, it is not quite so simple to translate this idea into a practical solution. In particular, the switching from bypass mode to double conversion mode when the supply quality deteriorates needs to be close to instantaneous so that the load sees no disturbance. This means that it is necessary to develop an algorithm that monitors supply quality and reacts very rapidly to deterioration and disturbance, and also to devise an ultra-fast mode switching mechanism.
These challenges have, however, now been addressed and the best currently available UPSs that incorporate ESS technology can switch from bypass to double-conversion mode in less than two milliseconds, guaranteeing that there is no disturbance to even the most sensitive of loads. In addition, when operating in bypass mode, these UPSs achieve efficiencies of around 99%, which is far better than can be achieved by UPS systems employing more conventional designs.
Another aspect worth noting, when operating in bypass mode, a UPS with ESS technology maintains its high efficiency even when lightly loaded. In double-conversion mode, however, efficiency falls significantly with light loads, just as it does with all other types of UPS.
UPS module management
This is an important issue that should be taken into whatever the topology under consideration, yet it is all too often neglected. The simple truth is that any power converter operating at below, say, 40% of its full load capacity will have poor efficiency. This particularly matters in applications where, for example, the UPS has been oversized either to provide a margin of reserve capacity or to make allowance for future expansion.
Once again, technologies have been developed to address this issue, such as the Variable Module Management System (VMMS) from Eaton. This uses a UPS design that splits the power conversion processes within the UPS over several modules rather than adopting the more traditional approach of using just one large power conversion unit.
The benefit of the VMMS arrangement is that, if the UPS is lightly loaded, some of the power conversion modules within it can, in effect, be taken out of service by putting them into a quiescent state where they consume very little energy.
The remaining modules are now much more fully loaded and will, therefore, operate efficiently. As might be expected, the VMMS technology manages the UPS dynamically, taking more modules out of service when the load falls, and bringing them back on line seamlessly within milliseconds when the load increases.
With a VMMS system, this management arrangement can handle not only the modules within an individual UPS, but can also manage power sharing between multiple UPS systems in an installation to ensure that they all deliver the best possible overall efficiency irrespective of the total loading.
As we have seen choosing the right topology is an essential first step in specifying a UPS installation. Making that choice is, however, not always as straightforward as it may at first appear. For this reason, there is much to be said for seeking advice from an experience UPS supplier that can offer all three technologies and that will not, therefore, be swayed by commercial bias but will offer dependable guidance based on the requirements of the application alone.
