A transformer specification is built from roughly nine decisions: type, capacity, voltage ratings, winding configuration and vector group, impedance, temperature rise, cooling class, efficiency class and physical envelope. Each affects performance, cost and which manufacturers can build it. Several are routinely inherited from previous projects without being re-examined, which is where both unnecessary cost and unnecessary delivery time tend to accumulate.
Start with type: dry or liquid-filled
This decision is usually made by the building rather than the electrical design, and it constrains everything after it.
Dry-type transformers use air and solid insulation. No liquid means no containment and no fire code implications from the fluid itself, which is what makes them the usual answer for indoor installation, upper floors and anywhere a spill would be a problem. They are generally larger for an equivalent rating, generally noisier, and generally less efficient at higher capacities.
Liquid-filled transformers immerse the core and windings in dielectric fluid that both insulates and carries heat away. Liquid is a better insulator and coolant than air, so these units are more compact and more efficient at higher capacities, and handle overload and harsh ambient conditions better. They require containment sized to the fluid volume, fire separation governed by the fluid type and local code, and a maintenance regime for the fluid.
The rough pattern: indoor and lower capacity leans dry-type, outdoor and higher capacity leans liquid-filled. The interesting cases are in the middle, where either would work electrically and the decision comes down to the fire strategy, the maintenance capability on site, and available space.
The parameters, and what each one affects
These are the fields that appear on almost every transformer specification.
| Parameter | What it is | What it affects |
|---|---|---|
| Capacity (kVA / MVA) | Continuous load the unit can carry | Everything. Size for actual load plus deliberate headroom, not nameplate totals |
| Primary voltage | Incoming voltage | Must match the utility or upstream supply exactly |
| Secondary voltage | Outgoing voltage | Determined by the downstream distribution design |
| Winding configuration | Delta or wye on each side | Neutral availability, harmonic behaviour, fault characteristics |
| Vector group | Phase relationship between windings | Paralleling compatibility. Unusual groups narrow the supplier field |
| Impedance | Opposition to current under fault | Fault current at downstream equipment, and therefore switchgear rating. Tight tolerances restrict who can bid |
| Temperature rise | Permitted rise above ambient at full load | Loading capability, insulation life, physical size |
| Cooling class | How heat is removed | Capacity under different conditions, and audible noise |
| Efficiency class | Losses at defined load points | Running cost, regulatory compliance, size and price |
| Sound level | Audible noise at rated load | Matters where the unit sits near occupied or sensitive space |
| Enclosure and rating | Protection against environment | Where the unit can be installed |
| Physical envelope | Dimensions and weight | Whether it fits, and whether it can be delivered to position |
The parameters worth challenging
Four of the fields above are frequently copied forward from a previous project without being re-examined. Each is worth an explicit decision, because each narrows the field of manufacturers who can build the unit.
- Impedance tolerance. Impedance affects downstream fault current, so it genuinely matters. But a tighter tolerance than the coordination study requires rules out otherwise suitable capacity for no benefit. Ask the engineer what tolerance the design actually needs.
- Vector group and winding configuration. Standard configurations are built by more manufacturers than unusual ones. If the configuration was chosen to match existing equipment that is being replaced anyway, it may no longer be a requirement.
- Temperature rise class. Frequently specified more conservatively than the application requires. A lower rise class means a physically larger and more expensive unit, which is worth it when loading demands it and not otherwise.
- Physical envelope. A dimensional constraint inherited from a site layout that has since changed can eliminate most of the available market before any electrical parameter does. Confirm the constraint is still real.
None of this argues for specifying loosely. It argues for knowing which requirements are genuine and which are habit.
Sizing and headroom
Two errors recur in transformer sizing, and they pull in opposite directions.
- Undersizing by ignoring growth. A transformer sized exactly to current load leaves no room for the density increases that have been consistent across the industry, and replacing a transformer is disruptive in a way that replacing downstream distribution is not.
- Oversizing by summing nameplate ratings. Adding up the nameplate figures of everything downstream produces a number well above realistic simultaneous draw. An oversized transformer costs more, takes more space, and runs less efficiently at light load.
The middle path is to size against measured or calculated actual load, apply a deliberate growth allowance based on how disruptive future replacement would be, and account for diversity across the connected load rather than assuming everything peaks together.
Any real sizing exercise should be reviewed by a qualified electrical engineer against the specific application, the applicable code and the coordination study.
What to include beyond the electrical parameters
A complete specification covers several things that are not electrical and are often left out, then negotiated awkwardly later.
- Testing and documentation. What factory tests will be performed, what test reports will be provided, and what drawings and wiring diagrams are included.
- Accessories. Temperature monitoring, tap changer type and range, surge protection, and any communications or metering integration.
- Delivery and rigging. Shipping split if the unit will not fit through the access route assembled, lifting points, and whether the unit can physically reach its final position. This is a routine source of expensive surprises.
- Warranty terms and what starts the clock — shipment, energization or commissioning.
- Spare parts and service availability, particularly for units with tap changers or active cooling.
- Storage requirements if the unit will arrive before the site is ready, which on a long lead item is a realistic prospect.
A specification checklist
Frequently asked questions
What is the difference between a dry-type and a liquid-filled transformer?
The insulating and cooling medium. Dry-type transformers use air and solid insulation, so they need no liquid containment and are commonly installed indoors and on upper floors. Liquid-filled transformers immerse the core and windings in dielectric fluid, which insulates and cools more effectively, making them more compact and more efficient at higher capacities. They require containment, fire separation governed by the fluid type and local code, and ongoing fluid maintenance.
Why does transformer impedance matter?
Impedance determines how much fault current flows to downstream equipment, which in turn sets the interrupting capacity that switchgear and protective devices must be rated for. Lower impedance means higher available fault current. It also affects voltage regulation under load. Because it links the transformer to the coordination study, impedance tolerance should be confirmed against what the design actually requires rather than copied from a previous project.
How much spare capacity should a data center transformer have?
Enough to cover realistic growth, sized against how disruptive future replacement would be, but not so much that the unit runs inefficiently at light load. Sizing by summing nameplate ratings of everything downstream typically overstates real simultaneous draw considerably. Sizing to current measured load with no allowance typically underestimates, given that rack density has risen consistently. The allowance should be a documented decision reviewed by the engineer of record.
What is a vector group on a transformer?
A vector group describes the winding configuration on each side, delta or wye, and the phase displacement between primary and secondary. It is written as a code such as Dyn11. It matters because transformers intended to operate in parallel must have compatible vector groups, and because the configuration affects neutral availability and harmonic behaviour. Unusual vector groups are built by fewer manufacturers, so an inherited requirement is worth confirming before it narrows the supplier field.
Next step
Trilogy supplies oil-filled and dry-type transformers across low and medium voltage classes, engineered to project requirements. Send us a specification and we will tell you what it implies for both cost and availability.
