When you purchase a UPS system or any other type of capital goods like a precision cooling system, it is important to consider what is referred to in engineering as the bathtub curve. Why is this important? Well at some point during the working life of your new system you may have an alarm condition that requires technical support and an emergency call out and if you are not covered by a UPS maintenance contract then this could be a chargeable service.
The bathtub curve is used to describe three failure periods in reliability engineering (more information at: https://en.wikipedia.org/wiki/Bathtub_curve). What the curve looks at the life cycle of a product and predicts most failures occur during their first few days of installation and operation and then towards the end of their working life.
The curve is typical for almost all electronic, electrical and electro-mechanical devices. For a UPS system or cooling solution the curve shows why it is important to have a comprehensive warranty included when you purchase a new system and to ensure that the system is properly installed and commission by factory / manufacturer trained engineers. If you have a plug-in system that you are installing and connecting yourself then access to good 24/7 technical support is as important, as are clear instructions.
So what happens if there is a ‘day one’ issue? Well you should look for a warranty service that offers 24/7 support and potentially an engineer on site or a swap-out unit next working day. Sometimes a technical issue can hold-up not just the installation and commissioning of a new device but also the handover of a complete system, server room or datacentre facility.
Warranty periods typically run from 12-36 months for most UPS suppliers. 60month warranties may be available as an ‘extension period’ or promotion. What’s covered in the warranty period is also important and especially in terms of exclusions. For a UPS system or cooling system consumable items may not be covered. In an uninterruptible power supply this could exclude batteries (whether lead acid or lithium-ion), fans and capacitors. During the standard warranty period these are typically covered by a manufacturer’s warranty and may be on a pro-rata basis meaning 100% year 1, 90% year 2 etc.
There are several important components within a UPS maintenance contract and perhaps the most important is a guaranteed call-out time. This can be measured in either clock hours or working hours with a typical high-service level contract offering 4 clock hours on a 7day period, and a lower level of service offering 12 working hours over a 5 working day period. Some UPS maintenance contracts may also include a service level agreement and guaranteed 8hour fix. If the UPS problem cannot be rectified, the entire UPS system is replaced with the UPS manufacturer covering all the labour, parts, installation, commissioning and logistics costs.
Preventative maintenance visits are different to emergency call outs. This type of site visit is planned (during normal or outside working hours) and may be annual or bi-annual. The preventative maintenance visit provides an opportunity to health check a UPS or cooling installation and to carry out any necessary upgrades, advise on ‘wear and tear’ issues and to replace consumable parts.
In terms of reliability engineering, preventative maintenance also helps to extend the working life of a system and the middle part of the bathtub curve.
A regularly maintained UPS system should give up to 10 years or more of consistent high reliability and service operation. Without regular UPS maintenance, the probability of a fault or alarm condition becoming critical is almost inevitable. Why? Well any UPS system has consumable parts and the major ones are the batteries used for energy storage, fans and capacitors.
Most UPS systems rely on valve regulated lead acid batteries and either 5-year or 10-year design life battery blocks. Within a server room or datacentre the ambient temperature is typically within the required 20-25˚C range to ensure battery performance. In terms of replacement this means that a 5-year battery set will be replaced around years 3-4 and a 10-year within years 7-8. Lithium-ion batteries typically have a 10-year or more design life and though having a higher initial capital cost are more temperature tolerant as well as being more suited to a higher number of charge/discharge cycles. It is important to remember that VRLA UPS batteries are best suited to a standby power operation and over 80% of UPS failures are linked directly to a poorly monitored or aged battery set.
Fans and capacitors have long working and design lives and will typically be replaced as part of a refurbishment project around years 8-10. Fans may require earlier replacement if the environment is dusty or dirty and especially if fan filters and covers are not used. It is important to consider capacitor replacement as electronic components like these can dry-out over time and develop ‘hot-spots’ with the potential for thermal runaway.
A final point should be noted in terms of whether the UPS uses a closed or open protocol. This refers to the communications used by UPS maintenance and service engineers to access alarm logs and set points within the uninterruptible power supply. The issue is that an ‘closed’ protocol ties service and maintenance into the manufacturer and service partner. This can lead to a higher cost for both labour and manufacturer certified parts. This higher cost should however be weighed against the fact that the work and any replacement parts should carry a manufacturer or supplier warranty. Sometimes you can get a lower cost service engineer, but they do not have the right software or access to certified parts. Your system is ‘fixed’, but the overall long-term reliability is degraded.
Total Cost of Ownership (TCO) is an important calculation when considering the purchase of any system for a critical environment such as a server room or datacentre. The TCO calculation should consider costs and details of any warranties (both standard and extended) and UPS maintenance contract provisions and the levels of service provided. Without regular maintenance the longer-term reliability and performance of a critical infrastructure component such as a UPS or cooling system cannot be guaranteed, and the overall bathtub curve will inevitably be far shorter.
Server rooms should be designed to provide a secure and managed environment in which to run critical IT servers and network infrastructures. Facilities vary in size from IT closets to computer rooms, Edge to regional and even hybrid datacentres. Whatever the size of the setup, each will have a core set of systems that require regular inspection and maintenance in order to ensure their uptime and resilience.