While customers often consider joining two or more plug-and-play UPSs, under no circumstances should the input of one UPS be attached directly or indirectly to the output of another. Commonly referred to as daisy-chaining or cascading units, this method dramatically increases the chance of UPS failure.
Daisy-chaining introduces another point-of-failure into the system – with potentially disastrous results. One threat is that the current draw can cause the upstream UPS to go into an overload condition or distort the output voltage waveform. When this occurs, it can prompt the UPS to detect a power fluctuation, revert to battery power, and continually cycle between the battery and normal operation. Plugging one UPS into another can also spark inverter compatibility issues, where the unit closest to the load doesn't recognise the other UPS's power as sufficient. This compels both UPSs go to battery, defeating the intent. Daisy-chaining can also force UPS controllers to work against each other, leading to harmonic distortion on power lines.
There are much better approaches to achieving the outcomes that customers most commonly seek from daisy-chaining UPSs, such as bolstering redundancy, gaining additional runtime and attaining surge protection.
One of the most frequently cited reasons for daisy-chaining UPSs is the desire to enhance reliability and eliminate the possibility of a single point-of-failure. However, the process of daisy-chaining actually increases the risk of UPS failure by combining failure rates for the two units. If either unit fails, the remaining UPS provides no value, an outcome demonstrated by a recent mathematically based probability study.
For customers desiring redundancy, an architecture using dual power supplies is recommended. In this type of solution, two UPSs are isolated from one another so a failure in one does not affect the other. In configurations where dual power supplies are not available, consider using an automatic transfer switch (ATS) PDU, which reallocates power between sources without interruption if the main source fails. Find out more about redundancy and system-level design in Eaton's network cupboard white paper.