Certain online UPS Systems have the ability to be connected in parallel in order to provide redundancy
. Expressed as an "n+x" system, "n" is the number of UPS needed and "x" the number of UPS over and above the actual power requirement.
This is by no means the only way to achieve redundancy (or indeed capacity as UPS can be connected in parallel for capacity rather than redundancy reasons). Instead of connecting two or more UPS together at source you could use two separate UPS Systems powering, for example, dual server inputs - in which case there is no need to have the UPS paralleled.
The issue arises with this situation in that if the output voltage waveform are not synchronised they could end up out of phase. For a single phase system operating at a nominal 230Vrms this could mean a potential between two "live" conductors in the same enclosure of 460V!
The UPS output frequency is usually synchronised to the input frequency. When a UPS enters bypass mode as a result of a fault, overload or even by just being switched off, this means that there is no loss of power as the bypass operation can be done seamlessly. This means that the power feeding the individual UPS must be in sychronisation - which means fed from the same AC power phase.
The problem then arises if the UPS systems were to drop onto battery power. In this event the UPS are disconnected from the mains power which means their UPS frequency is determined by the digital controller. If you imagine that a full waveform is generated in only 20 thousands of a second. In order for two waveforms to become anti-phase there needs to be a change of only 10 thousands of a second. In a 5 minute power outage this can happen with a variance of only 33 parts per million between the UPS.
The UPS Systems need to have synchronisation to prevent this from happening, and not all systems are capable of this. The E200 6KVA and 10KVA UPS Systems from Dale Power Solutions have this very important feature allowing the UPS to be independent but have a synchronized output.