Why large power transformers are tall: coil height, impedance and flux
On a large unit the flux limit fixes the turns, and the impedance band then decides how tall the windings must be. Get the height wrong at the start and no amount of iteration will close the design.
· 8 min read · Design
Three numbers that are not independent
For a given rating and voltage, the volts per turn decide the flux in the core: fewer turns means more volts on each one and a higher flux density. The flux density in turn decides the no-load loss, and the specification puts a ceiling on it, typically 1.65 to 1.7 T for grain-oriented steel. So the flux limit sets a minimum number of turns, and the no-load loss guarantee usually pushes that minimum higher still.
The percentage impedance depends on those same turns. Leakage reactance grows with the square of the turns, with the mean length of turn and with the radial space between and within the windings, and it falls as the windings get taller. Once the turns are fixed by the flux, the coil height is the main lever left for putting the impedance inside its band.
Why the obvious fix does not work
When a design comes out with too much impedance, the instinctive correction is to remove turns. On a small distribution transformer that often works, because there is headroom between the operating flux and its limit. On a large power transformer the headroom has usually been spent on the no-load loss, and removing turns pushes the flux straight past its maximum.
The result is a design that meets impedance and both loss guarantees on paper but runs the core at 1.8 or 1.9 T. That is not a design that can be built and tested; it is a symptom that the winding geometry was wrong from the start.
Height is the lever, and it costs something
Spreading the same turns over more discs makes each winding taller and radially thinner. The leakage path lengthens, the radial build shrinks, and the impedance falls without touching the flux. An 80 MVA, 132 kV unit that came out at 20 % impedance on 1.05 m windings landed at 12 % on 1.6 m windings, with the flux at 1.2 T and both losses inside their guarantees.
Height is not free. A taller window means longer core legs and more core steel, a taller tank and more oil, and on the largest ratings the transport height becomes the binding limit. That is why the right height is a design decision to be optimised, not a value carried over from the last unit of a similar rating.
What this means for a design tool
A design tool that only moves turns and conductor sizes will grind against this wall on large units: it will either fail to converge or hand back a design that breaks the flux limit. The search has to treat the core size and the winding height as levers in their own right, and it has to hold the flux limit as a hard constraint while it does so.
The practical test is simple. Give the tool a large rating with a tight loss guarantee and a normal impedance band, and look at the operating flux in the answer. If it is above the specification's maximum, the tool has found a number, not a design.