Autotransformer advantages — smaller size, lower cost, higher efficiency — are real, but conditional. All of them come from one design fact: the input and output share part of the same winding, so only a fraction of the power is actually transformed. That fraction is set by the voltage ratio. Move the ratio away from unity, or add a requirement for isolation, and the advantages shrink or turn into liabilities.
This guide quantifies where those advantages hold, shows the ratio at which they stop paying for themselves, and lists five situations where an autotransformer should not be used at all. For wider product context, browse the step-up and step-down transformer range.
In a two-winding transformer, every watt crosses the core by induction, so both windings are built for the full throughput. In an autotransformer, the common winding carries only the difference between input and output current; the series portion is sized for the voltage difference.
Winding fraction = |1 − Vraus / Vin|
Take a 10 kVA unit stepping 230 V down to 200 V. Output current is 10,000 ÷ 200 = 50 A and input current is 10,000 ÷ 230 = 43.5 A. The common winding carries the 6.5 A difference at 200 V, the series winding carries 43.5 A across the 30 V difference. Both come to roughly 1.3 kVA of winding duty for 10 kVA of through power — about 13% of the copper and core a two-winding unit needs.
| Conversion | Spannungsverhältnis | Winding fraction | Winding kVA for 10 kVA through |
|---|---|---|---|
| 230 V → 200 V | 1.15 : 1 | 13% | 1.3 kVA |
| 230 V → 190 V | 1.21 : 1 | 17% | 1.7 kVA |
| 230 V → 115 V | 2 : 1 | 50% | 5.0 kVA |
| 230 V → 100 V | 2.3 : 1 | 57% | 5.7 kVA |
| 230 V → 76 V | 3 : 1 | 67% | 6.7 kVA |
| 230 V → 46 V | 5 : 1 | 80% | 8.0 kVA |
The saving tracks the winding fraction, which tracks how close the two voltages are. A 230 V to 200 V unit delivers the full benefit; a 230 V to 115 V unit delivers half of it.
Copper loss scales with winding duty, so a 13% winding fraction loses far less than a two-winding equivalent. Core loss does not scale down with the ratio, though, so at no load both designs draw a similar magnetizing current. On a load that runs a few minutes a day the gain is theoretical; on continuous duty it compounds — about 600 kWh a year on a 10 kVA unit with a 1.5-point advantage at 4,000 hours.
Tighter coupling and lower series impedance keep the output up from no load to full load. That same low impedance lets fault current climb, which is why the protection review below matters more here.
Taps on a shared winding are cheap to add, which is why the variable autotransformer became the default bench instrument for voltage testing — the TDGC2J variable transformer is a familiar example.
| Vorteil | Physical reason | Condition for it to hold | Where it breaks down |
|---|---|---|---|
| Lower cost and weight | Only the series section is built for the voltage difference | Ratio below about 1.5 : 1 | Above 2 : 1 the saving halves, then fades |
| Höhere Effizienz | Copper loss tracks the winding fraction, not throughput | Continuous or long-duty loading | Short intermittent duty, where no-load loss dominates |
| Better regulation | Lower series impedance, tighter coupling | Supply already has adequate fault protection | Weak supply or long cable runs |
| Adjustability | Cheap taps on a single winding | Adjustment only, isolation provided elsewhere | Any circuit that must float |
No switch flips at a particular ratio, but the trade stops making sense somewhere. At 2 : 1 you have given up half the material saving while still carrying the full isolation penalty. At 3 : 1 the winding is two-thirds of a full transformer, at 5 : 1 it is 80%. You pay nearly the full cost of a two-winding unit and get none of its separation.
One further boundary matters: an autotransformer changes voltage only, never frequency. A 60 Hz unit asked to run a 50 Hz load will saturate, because flux rises with the voltage-to-frequency ratio — the mechanism behind transformer core saturation.
This is the one that causes real damage. In a 230 V to 200 V unit the output sits on a tap 30 V below the top of the winding. If the series section fails open, the load is left connected to the full 230 V through the common winding; if the common section fails short, the output rises toward the input voltage. No second winding limits either event, and neither is visible at the output terminals. Where a standard or risk assessment calls for separation, the answer is an isolation transformer.
An autotransformer cannot create a neutral the supply does not have, nor establish a new grounded system. A three-wire supply cannot become a four-wire output, because the shared winding offers no independent point to bond. The bonding rules are covered in how to ground a transformer secondary.
These rely on the load side being isolated so a single fault cannot create a shock path. Medical equipment adds leakage-current limits a shared winding cannot meet, set out in medical-grade isolation transformers.
Rectifier front ends, VFDs, and switch-mode supplies inject harmonics that the low-impedance shared winding passes rather than blocks. On three-phase units, triplen harmonics circulate through the common winding as heat.
Beyond 2 : 1 the economics fail as shown above. A floating load — test benches, instrumentation front ends, anything that must not reference supply ground — fails on the same isolation point as item 1.
| Situation | Why the autotransformer fails | Use instead |
|---|---|---|
| Isolation required by standard or risk assessment | Shared winding transfers faults and surges to the load | Two-winding isolation transformer |
| Neutral or separately derived system needed | No independent winding point to bond | Two-winding transformer with grounded secondary |
| Patient contact, wet area, portable tool | No separation, leakage current too high | Medical-grade isolation unit |
| Rectifier, VFD, or sensitive electronics | Low impedance passes harmonics and transients | Isolation transformer, add filtering if needed |
| Ratio above 2 : 1, or floating load | Saving gone, separation still absent | Two-winding step-up or step-down transformer |
Nameplate kVA basis. Ask whether the rating is through-power or winding power. A 10 kVA winding rating on a close-ratio unit can pass far more than 10 kVA of load, so reading it the wrong way leaves you oversized or quietly overloaded.
Fault current. Lower impedance means higher available fault current at the load terminals than a two-winding unit of the same rating. Downstream breakers must be rated for the real value.
Neutral-ground bonding. Never create a second neutral-to-ground bond downstream — two bonds on one system put operating current on the grounding conductor.
Tap position. The winding fraction is valid only at the design tap, so a unit set elsewhere is no longer the same kVA rating and the protection settings no longer match.
The advantages are genuine in the cases they were designed for: matching two systems that sit close in voltage, correcting a persistent 5% to 10% mismatch, reduced-voltage motor starting, and giving a test bench an adjustable supply. In each, the load either tolerates the shared reference or isolation exists upstream.
When the checks point to separation instead, the DT-10000VA toroidal step-up and step-down transformer is the two-winding alternative: separate primary and secondary, 110–120 V and 220–240 V windings, fuse and MCB protection on the input. It costs more and weighs more than an autotransformer of equal rating, and that difference is what you are buying.
For the winding-level comparison between the two families, see step-up and step-down transformer vs autotransformer.
Autotransformer advantages are a function of voltage ratio, not a fixed property of the product. Below 1.5 : 1, with no isolation requirement in the circuit, the shared winding gives you a smaller, cheaper, more efficient transformer. Above 2 : 1, or as soon as a neutral, a grounded system, or a floating load enters the picture, the same winding becomes the reason to choose something else. Confirm the kVA basis and the fault current in writing; those cause more field problems than the ratio.
No. The input and output share part of the same winding, so a conductive path exists between them. A fault on one side can reach the other, and the load cannot float relative to the supply.
When isolation is required, when the load needs a neutral or a separately derived system, in patient-contact or wet locations, with harmonic-rich loads, or when the ratio exceeds about 2 : 1.
No. A shared winding has no independent point to bring out as a neutral, so a three-wire supply cannot become a four-wire output. Use a two-winding unit with a grounded secondary.
There is no hard electrical limit, but the economics set a practical one. Above roughly 2 : 1 you pay most of a two-winding transformer’s cost with none of its separation.
Usually yes, because losses track the winding fraction rather than full throughput. The gap narrows at light load, where core loss is the same for both.
It can, and the winding fraction at that ratio is 50%, so the material saving is half of what a 230 V to 200 V unit gives. If the load also needs isolation or a neutral, use a two-winding transformer.
