Coin Miner Crypto Mining Blog

Transformer & Cable Losses: The Hidden 2–4% You Can Win Back | Coin Miner Blog

Published September 14, 2025 · 4 min read

Bitcoin
Transformer & Cable Losses: The Hidden 2–4% You Can Win Back | Coin Miner Blog

TL;DR

  • Losses come from transformer core + copper and cable I²R (plus harmonics/imbalance).

  • Cut current (higher distribution voltage, good PF), cut resistance (thicker/shorter conductors), and cut constant losses (efficient transformer close to load).

  • Target ≤2% combined transformer + LV cable loss at steady state; >3–4% means money left on the table.

  • Typical quick wins: shorter LV runs, parallel smaller cables, phase balancing, k-rated/xL-low-loss transformer, busway, and EMS caps to avoid peak heating.

Where the watts go

1) Transformer losses

  • Core (no-load): constant with energization (hysteresis + eddy).

  • Copper (load): Pcu=I2×RwindingP_{cu} = I^2 \times R_{winding}, rises with current and temperature.

  • Harmonics: SMPS currents (even with PFC) add extra heating → consider k-rated or low-loss designs.

Operator targets

  • Specify low-loss pad-mount transformer sized near your steady load (not 3× oversize).

  • Place it close to the container/room to minimize low-voltage cable length.

  • Verify nameplate no-load (W) and full-load (W); add them to your PUE ledger.

2) Cable losses (I²R and voltage drop)

For any feeder:

Ploss=I2×Rac,ΔV%≈I⋅(Rcos⁡ϕ+Xsin⁡ϕ)V×100P_{loss} = I^2 \times R_{ac}, \quad \Delta V\% \approx \frac{I \cdot (R \cos\phi + X \sin\phi)}{V} \times 100

  • Lower current: distribute at higher voltage (e.g., 415 V 3φ to PDUs) and keep PF ≥ 0.97.

  • Lower resistance: short runs, larger cross-section, or parallel conductors.

  • AC effects (skin/proximity) make R_ac > R_dc for very large cables—multiple smaller parallels help.

Quick, realistic example (illustrative)

Plant: 1.0 MW IT, 10% aux → 1.1 MW total.
Assume combined transformer + LV cable losses are 3.0% → 33 kW lost.
At $0.10/kWh that’s $79/day (≈ $2.4k/month).

Improvements (typical stack):

  • Shorten LV runs & upsize conductors: –0.8%

  • Swap to low-loss transformer near right kVA: –0.5%

  • Balance phases within ±5%, verify PF 0.98+: –0.2%
    New loss ≈ 1.5% → 16.5 kW → $39/day saved (≈ $1.2k/month, $14k/yr).
    (Scale linearly with your kW and tariff.)

Design choices that pay back

Voltage & distribution

  • Keep 415 V 3-phase as far as possible; avoid long 230 V single-phase branches feeding large groups.

  • Use busway or paralleled smaller cables over a single massive conductor to limit AC resistance.

  • Aim for ≤3% voltage drop to last PDU.

Cable spec & layout

  • Short, straight runs; avoid hot trays.

  • Prefer copper where runs are long/high current; aluminium can pencil if cross-section is increased and terminations are correct.

  • Terminate cleanly; loose lugs = heat = watts.

Transformer spec

  • Select low-loss core steel, k-rating appropriate for harmonics, and tap settings to keep voltage in band.

  • Right-size: too small → overheating; too big → needless core loss.

Power factor & harmonics

  • Most miner PSUs have active PFC (≈0.95–0.99). Verify under your undervolt profile—if PF drops, current climbs and losses follow.

  • Consider isolation / k-rated units or line reactors if THDi is high and gear runs hot.

Phase balance & protection

  • Keep phase currents within ±10% (ideally ±5%).

  • Distribute rigs evenly by PDU phase; monitor neutral heating in wye systems.

Measurement plan (one-day test)

  1. Log kW at transformer primary & secondary (or upstream meter vs PDU meters).

  2. Log feeder currents and temperature (IR camera) at the same time of day.

  3. Compute implied % loss = (upstream kW − downstream kW) / upstream kW.

  4. Tag where the delta lives (transformer vs feeders) and cost it at your tariff.

  5. Prioritize fixes with ≤12-month payback.

Implementation checklist (printable)

  • Nameplate no-load & full-load transformer losses in PUE sheet

  • Map feeder lengths, gauge, material, and R_ac estimate

  • Check PF and THDi under your actual tuning profile

  • Verify phase balance (±5–10% target)

  • Cap site kW in EMS; sequence warm-starts to avoid thermal peaks

  • Plan upgrades: shorter LV, parallel runs, busway, transformer relocation/spec

  • Re-measure after changes; update KPI pack

FAQ

Is upsizing cable always worth it?
Not always. Price the kW saved × hours × tariff vs capex. Long, high-current runs usually justify bigger or parallel conductors.

Do k-rated transformers boost efficiency?
They tolerate harmonics without overheating; pair with a low-loss design for best results.

Will higher voltage damage miners?
Stay within PSU input spec and local code. Higher distribution voltage means lower current—good for losses—while PDUs supply the correct final voltage to rigs.

From the Blog