CDCP Data Center Power Infrastructure 2 — Questions and Answers
Question 1: What is the purpose of a static transfer switch (STS) in a data center power system?
- To automatically transfer an IT load from one power source to another in less than a quarter cycle (4ms or less) without interruption (Correct answer)
- To gradually ramp up generator output to prevent voltage surges during startup
- To isolate a faulty UPS module while the system remains on bypass
- To convert single-phase power to three-phase power for high-density rack deployments
Correct answer: To automatically transfer an IT load from one power source to another in less than a quarter cycle (4ms or less) without interruption
An STS uses solid-state electronics to transfer critical loads between two independent power sources in less than 4 milliseconds, fast enough that IT equipment does not experience a power interruption.
A Static Transfer Switch (STS) provides seamless transfer between two power feeds — typically Source A and Source B from redundant UPS outputs or utility feeds. Using thyristors (SCRs), it monitors both sources continuously and transfers the load in less than 4ms (one quarter of a 60 Hz cycle) if the preferred source deviates from nominal voltage or frequency. This is fast enough that most IT equipment, which tolerates brief sags, does not lose operation. STS units are used in Tier III and IV designs to ensure dual-corded servers receive uninterrupted power even if one UPS fails or requires maintenance.
Question 2: What is the purpose of an isolation transformer in a data center power distribution system?
- To electrically isolate equipment from the power distribution system, reduce common-mode noise, and provide a separately derived system for grounding (Correct answer)
- To step up utility voltage from 480V to 13.8kV for transmission within a large campus
- To synchronize generator output frequency with the utility feed before transferring load
- To prevent reverse current flow from UPS batteries back into the utility grid
Correct answer: To electrically isolate equipment from the power distribution system, reduce common-mode noise, and provide a separately derived system for grounding
Isolation transformers provide galvanic isolation between primary and secondary circuits, eliminating common-mode noise, providing a new neutral-ground bond, and improving electrical safety.
Isolation transformers use inductive coupling without a direct electrical connection between primary and secondary windings. In data centers, they serve multiple purposes: they reduce common-mode noise and electromagnetic interference (EMI) that can cause bit errors in sensitive equipment; they provide a new separately derived system with a neutral-ground bond, improving safety; and they allow the facility to use different voltages on each side (e.g., 480V delta to 208Y/120V). Isolation transformers are commonly used between UPS output and PDU input in critical power distribution paths.
Question 3: What does 'kVA' measure in the context of data center power systems, and how does it differ from kW?
- kVA (kilovolt-amperes) measures apparent power including reactive components; kW measures real (active) power. The ratio kW/kVA is the power factor. (Correct answer)
- kVA measures power consumed by cooling equipment; kW measures power consumed by IT equipment.
- kVA is used for three-phase circuits; kW is used for single-phase circuits.
- kVA includes transmission line losses; kW is the net power delivered to equipment.
Correct answer: kVA (kilovolt-amperes) measures apparent power including reactive components; kW measures real (active) power. The ratio kW/kVA is the power factor.
kVA is apparent power (voltage × current), while kW is real power (the work-performing component). Power factor (PF) = kW/kVA; modern IT equipment typically has PF of 0.9–0.99.
In AC power systems, inductive and capacitive loads create a phase difference between voltage and current, resulting in reactive power (kVAR). Apparent power (kVA) is the vector sum of real power (kW) and reactive power (kVAR). Real power (kW) actually does work; reactive power (kVAR) oscillates back and forth. Power Factor (PF) = kW/kVA. UPS systems are rated in kVA; generators in kW. Modern IT equipment power supplies have active PFC (Power Factor Correction) achieving PF close to 1.0. Data center operators must understand both when sizing UPS, generators, and distribution circuits.
Question 4: In a 2N UPS configuration, how is the IT load typically distributed between the two UPS systems during normal operation?
- Each UPS carries 50% of the total IT load, so each operates at 50% capacity and either can support 100% of the load alone if the other fails (Correct answer)
- One UPS carries 100% of the load (primary) while the second is in hot standby (backup) at 0% load
- Each UPS carries its assigned half of the racks; there is no cross-connection between the two systems
- The load is dynamically balanced by a load-sharing controller to equalize battery wear
Correct answer: Each UPS carries 50% of the total IT load, so each operates at 50% capacity and either can support 100% of the load alone if the other fails
In a 2N configuration, both UPS systems are active and each carries half the load, ensuring that if one fails the other can immediately assume 100% of the load without any switching.
A true 2N UPS architecture has two fully independent UPS systems (System A and System B), each sized to carry 100% of the IT load. Under normal operation, dual-corded servers connect one power supply to each UPS, so each UPS carries approximately 50% of total load. If one UPS fails or is taken offline for maintenance, the other assumes 100% without any switching event. This is different from N+1 (where the extra unit is a spare, not active). CDCP candidates must understand that 2N requires fully redundant UPS, distribution, PDUs, and feeds to deliver true Tier IV-level power path redundancy.
Question 5: What is the function of a power distribution unit (PDU) in a data center?
- To receive high-voltage power from UPS or transformers and distribute it to individual racks at the appropriate voltage and with circuit protection (Correct answer)
- To convert DC battery power to AC for server use during utility outages
- To monitor server power consumption and report data to DCIM software
- To provide automatic voltage regulation (AVR) for sensitive laboratory equipment
Correct answer: To receive high-voltage power from UPS or transformers and distribute it to individual racks at the appropriate voltage and with circuit protection
A PDU takes incoming power from the UPS or facility transformer and distributes it to multiple branch circuits serving individual racks, providing overcurrent protection and often metering.
PDUs (Power Distribution Units) in data centers are large panels (often floor-standing or wall-mounted) that receive high-current power from UPS outputs or transformers (e.g., 480V 3-phase or 208V) and break it into individual branch circuits that run to rack-mounted power strips (also called rack PDUs or rPDUs). They include main circuit breakers, branch circuit breakers, and often include metering (per-branch or total). Intelligent PDUs can report per-outlet current via SNMP/HTTP to DCIM systems. PDUs are distinct from in-rack power strips, which plug into PDU circuits.
Question 6: What is the typical battery runtime design target for data center UPS systems, and why isn't it designed for longer runtimes?
- 10–15 minutes — long enough for generators to start, transfer to load, and stabilize; longer battery runtimes are expensive, heavy, and unnecessary if generators are on-site (Correct answer)
- 4 hours — to bridge all maintenance windows during business hours without starting generators
- 72 hours — to provide full operation during extended grid outages without generators
- 30 seconds — only needed to provide a clean shutdown signal to servers before power loss
Correct answer: 10–15 minutes — long enough for generators to start, transfer to load, and stabilize; longer battery runtimes are expensive, heavy, and unnecessary if generators are on-site
UPS batteries are typically sized for 10–15 minutes at full load — sufficient to start and load generators (90–180 seconds) with a comfortable margin. Extended battery runtimes are costly and unnecessary when generators are available.
Modern diesel generators typically start and accept load in 10–30 seconds. UPS systems are designed to bridge this gap with margin to spare. Standard data center UPS battery strings provide 10–15 minutes at full load, giving operations teams time to confirm generator operation, troubleshoot minor issues, and allow the generator to stabilize before transfer. Longer runtimes (extended battery modules) are available but add significant weight, cost, and floor loading. For extended outages without utility, generators with on-site fuel storage (typically 24–72+ hours of fuel) are the appropriate solution, not oversized battery banks.
CDCP Data Center Power Infrastructure 2
What is the purpose of a static transfer switch (STS) in a data center power system?