How to Size a Servo Voltage Stabilizer: Complete Engineering Guide

Sizing a servo stabilizer wrong is expensive. Undersized units run at continuous overload and burn out wiper contacts in months; oversized units waste 30–40% of capital and run inefficiently. This guide walks through the proper engineering calculation and gives worked examples for the most common load types.

The sizing formula

Stabilizer rating (kVA) = Total connected load (kVA) × Diversity factor × Headroom factor (1.25)

Total connected load — sum of all equipment kVA that will run on the stabilizer, at nameplate rating.

Diversity factor — the fraction of connected load that runs simultaneously. For a machine shop where all machines rarely run at once, this may be 0.7–0.8. For continuous process, 1.0.

Headroom factor — 1.25 (25%) accounts for inrush currents, ambient derating, and long-term reliability margin.

Step 1: Inventory the connected load

Load inventory worksheet

Load type	How to measure	Note
Motors	Nameplate kW ÷ (efficiency × PF)	Typically kW / 0.72 for 3-phase induction
Heaters, resistive loads	Nameplate kW = kVA	PF = 1
SMPS, IT equipment	Nameplate VA (or W × 1.2)	Assume PF 0.85 if only W given
Lighting (LED, fluorescent)	Sum of fixture VA	LED PF often < 0.9
Medical imaging	OEM installation kVA spec	Use OEM value, not measured
HVAC compressors	Nameplate LRA × V × √3 for inrush	Compressor cycling: use running kVA + 20%
Welding machines	Nameplate primary VA × duty cycle	Segregate from stabilizer if possible

Step 2: Apply diversity factor

Diversity factor is the ratio of actually simultaneous load to connected load. Look at a typical operating hour and estimate what fraction of loads run at the same time:

Typical diversity factors by facility type

Facility type	Diversity factor	Reason
Machine shop	0.75–0.85	Not all machines run continuously
Continuous process (pharma, food)	0.95–1.0	All process equipment runs together
Hospital	0.60–0.75	Diagnostic loads intermittent; only critical care continuous
Data centre	0.85–0.95	IT load nearly constant; HVAC follows heat
Office building	0.55–0.70	Lighting/AC intermittent by zone
Textile mill	0.90–0.95	Weaving/spinning runs 3-shift
Cold storage	0.65–0.75	Compressor cycling; not simultaneous

Step 3: Apply headroom (1.25×)

The 1.25 headroom covers:

1. Motor inrush — starting current up to 6× rated for a few seconds
2. Ambient derating — stabilizer nameplate is at 25 °C; derate ~10% per 15 °C above
3. Voltage input at low end — at 150 V input, the stabilizer needs to source 30% more current than at 230 V for the same output kVA
4. Long-term reliability margin — running at 95% of rating drops MTBF significantly

View data table

Source: Ozmist reliability data

Running at 80% average load (which is what a 1.25× headroom produces at nameplate demand) gives 3–4× the wiper life vs running at 100% of rating.

Step 4: Choose input voltage variation range

Input variation range determines the size and cost of the buck-boost transformer. Options:

Input voltage variation range selection

Input range	Input voltage (1-phase)	Input voltage (3-phase)	Best for
±30%	160–300 V	280–520 V	Urban feeders with modest swing
±40%	140–320 V	240–560 V	Peri-urban feeders; standard industrial
±50%	115–345 V	200–600 V	Rural feeders; heavy voltage swing

Always measure before speccing. Install a voltage logger at the connection point for 7 days and record min/max. Choose the range that covers the observed extreme with 10% margin.

Step 5: Choose single-phase or three-phase

Single-phase vs three-phase stabilizer decision

Criterion	Single-phase	Three-phase
Load type	Small IT, residential, single motor	Industrial, HVAC, machine shop
Rating range	0.5 – 20 kVA typical	10 kVA – 3 MVA
Balancing	Simple	Balance loads across 3 phases
Correction independence	N/A	Ozmist supports independent phase correction (unbalanced input)

For three-phase, ensure the stabilizer supports independent per-phase correction. Cheaper stabilizers correct all three phases based on average, which is inadequate when input phases are unbalanced (common on Indian feeders).

Worked examples

Example 1: Small machine shop

• 3× VMCs at 30 kVA each = 90 kVA connected
• Diversity factor 0.85 (not all running simultaneously): 76.5 kVA effective
• Headroom 1.25: 95.6 kVA
• Round up to next standard: 100 kVA 3-phase, ±40% range
• Ozmist SVS-100 oil-cooled recommended

Example 2: Hospital OT complex

• 1× CT scanner 40 kVA + 4× OT (lights, anaesthesia, monitoring) 12 kVA each + HVAC 25 kVA = 113 kVA connected
• Diversity factor 0.70 (CT and OTs not always running together): 79.1 kVA effective
• Headroom 1.25: 98.9 kVA
• Round up: 100 kVA 3-phase, ±40% range with medical-grade features
• Ozmist SVS-100M medical series recommended

Example 3: Injection moulding factory

• 10× injection moulding machines at 25 kVA each = 250 kVA connected
• Diversity factor 0.90 (continuous 3-shift operation): 225 kVA effective
• Headroom 1.25: 281 kVA
• Round up: 300 kVA 3-phase, ±40% range, oil-cooled
• Ozmist SVS-300 oil-cooled recommended

Example 4: Data centre

• 100 kW IT load (PF 0.95) = 105 kVA + 40 kW HVAC (PF 0.85) = 47 kVA = 152 kVA connected
• Diversity factor 0.90 (nearly constant): 137 kVA effective
• Headroom 1.25: 171 kVA
• Round up: 200 kVA 3-phase, ±30% range (assume urban feeder)
• Note: Data centre often prefers online UPS; stabilizer only if UPS is line-interactive

Example 5: Textile mill

• 30 looms at 8 kVA each + auxiliaries 40 kVA = 280 kVA connected
• Diversity factor 0.90 (3-shift, most looms running): 252 kVA effective
• Headroom 1.25: 315 kVA
• Round up: 400 kVA 3-phase, ±40% range (peri-urban feeder), oil-cooled
• Ozmist SVS-400 oil-cooled recommended

Common sizing mistakes

Common stabilizer sizing mistakes

Mistake	Consequence
Sizing to running load ignoring inrush	Overload trip on every startup
No diversity factor for machine shop	Overspec by 20–30%
Ignoring ambient temperature	Derating overlooked; runs hot
Wrong input range for feeder	Under- or over-voltage trips; wrong size buck-boost
Not measuring actual feeder for 7 days	Design based on assumption; misses peaks
Ignoring future load growth	Full stabilizer replacement in 3–5 years
Sizing on kW instead of kVA	Undersize by 15–20% depending on PF

Verifying sizing after installation

Post-install, verify:

1. Voltage at load terminal — logged for 7 days, should hold within ±1% of setpoint
2. Current at stabilizer output — should not exceed 80% of nameplate on any phase during peak hour
3. Stabilizer body temperature — top surface should not exceed 45 °C in a 30 °C ambient
4. Wiper motor activity — should update within seconds, not hunt continuously

If any of these fail, revisit sizing.

Frequently asked questions

Can I use a smaller stabilizer than sized because my load rarely peaks?

No. The stabilizer must handle worst-case peak, not average. Cheaper alternative is a bypass switch to route peaks around a smaller stabilizer, but this defeats the purpose.

How do I handle unknown future loads?

Add 20–30% growth headroom on top of the 1.25 factor. It is cheaper to buy one 200 kVA unit now than 100 kVA now plus 100 kVA in 3 years.

What about power factor — do I need to correct it separately?

Yes. Stabilizers do not correct PF. Install PF correction capacitors separately, before or after the stabilizer as appropriate.

Can I use a 3-phase stabilizer for a single-phase load?

Not economically. The unused phases waste capital. Use single-phase stabilizer for single-phase loads.

How does harmonics affect sizing?

Non-linear loads (VFDs, SMPS) draw harmonic current that raises stabilizer heating. Add 10–15% to sizing for harmonic-heavy loads. Or use a stabilizer rated for K-factor loads.

Does site altitude affect sizing?

Yes for sites above 1000 m. Derate 1% per 100 m above 1000 m ambient reference. Ozmist provides altitude-derated versions for high-altitude installations.

What about outdoor installation?

Choose IP-55 or higher enclosure for outdoor. Add sunshade to reduce solar heating. Sizing headroom increases to 1.35–1.40 for outdoor installations exposed to summer sun.

Can I parallel two stabilizers for redundancy?

Yes with proper synchronising panel. Parallel operation halves the load per unit and enables one to be serviced while the other carries the load. Common in critical installations.

References

1. BIS IS 9815:1997. Auto-transformer type servo controlled voltage stabilizers.
2. IEEE 141 Red Book. Recommended Practice for Electric Power Distribution for Industrial Plants.
3. IEC 61558. Safety of transformers, reactors, power supply units.
4. Central Electricity Authority. Report on Voltage Quality in LT Distribution Feeders. CEA, 2024.
5. Bureau of Energy Efficiency. Voltage-Sensitive Equipment Loss Assessment. BEE, 2023.
6. Ozmist Sizing Tool. Internal engineering worksheet, 2026.
7. IEEE 519. Recommended Practices and Requirements for Harmonic Control in Electric Power Systems.
8. IEEE 1100 Emerald Book. Recommended Practice for Powering and Grounding Electronic Equipment.