Bottom line first: MoSi2 heating elements can run on DC. A MoSi2 element is a resistive heater: heat output is set by RMS current and resistance, so DC and AC do the same work. In a KULUN test, a 60 V / 1000 A IGBT DC supply drove a two-series two-parallel MoSi2 bank from cold state to 858 °C: output voltage rose from 1.33 V to 22.41 V, input power factor from 0.449 to 0.939, and grid-side voltage THD stayed within 1.78%–2.6% throughout.
This article answers four questions in order: why DC works on MoSi2, where the conventional AC route struggles, how an IGBT DC supply removes that constraint, and what the measured data actually shows.
Why can MoSi2 elements be heated with DC?
Bottom line: a MoSi2 element is a resistive heater, and its heat output is P = I²R. With the same RMS current, DC and AC produce the same heat in the same element.
MoSi2 involves no electrolysis or polarisation process that depends on current direction, so long-term DC operation raises no direction-related problem. The real question is not whether DC works, but whether a DC supply handles cold start and grid-side behaviour better than the conventional AC route.
Where does the conventional AC route struggle?
Bottom line: it struggles with the cold-state resistance of MoSi2. The conventional chain is 380 V AC into an SCR power controller, then a step-down transformer to about 30 V–50 V, then the MoSi2 elements. MoSi2 resistance is very low when cold and rises markedly with temperature.
The test data confirms this. At 61.6 °C the two-series two-parallel bank drew 147.3 A at only 1.33 V, an equivalent resistance of about 9.0 mΩ. At 858 °C it drew 628 A at 22.41 V, about 35.7 mΩ, roughly four times the cold value.
| Condition | Load temperature | Output voltage | Output current | Equivalent resistance |
|---|---|---|---|---|
| Cold state | 61.6 °C | 1.33 V | 147.3 A | approx. 9.0 mΩ |
| Heating up | 374.2 °C | 10.36 V | 461 A | approx. 22.5 mΩ |
| High temperature | 858 °C | 22.41 V | 628 A | approx. 35.7 mΩ |
Phase-angle control on such a load brings three problems during heat-up:
- The conduction angle is squeezed very small. With an equivalent resistance of only tens of milliohms, the thyristor conducts in a narrow window each half cycle to hold the output at a few volts and limit current.
- Power factor drops. Current becomes a narrow pulse, so both displacement factor and distortion factor worsen, and reactive power occupies transformer capacity.
- Harmonics and interference rise. Narrow-pulse current is rich in harmonics and can disturb other equipment on the same transformer.
Do transformer taps solve it?
Bottom line: they relieve the problem but add cost and moving parts. Adding secondary taps lets the heat-up stage run on a lower-voltage, higher-current tap, keeping the thyristor conduction window wide and harmonics lower.
The cost is the changeover itself: manual switching, or an automatic switching circuit with its own components, plus current transients at the instant of change. It suppresses the problem rather than removing it.
How does an IGBT DC supply remove the constraint?
Bottom line: output voltage is set by chopper duty cycle, not by conduction angle. The following points describe the operating principle, not measured results.
- Constant current from cold state. The supply regulates output current, so the cold load receives the target current while voltage follows the element resistance; no need to squeeze a conduction angle.
- Low voltage, high current, directly chopped. The DC link is chopped by IGBTs at high frequency, so a few volts to tens of volts are produced without a transformer secondary plus phase control.
- Input side decoupled from output voltage. The input stage is a three-phase rectifier with a DC link; power factor and harmonics depend on rectifier topology, not on how low the output voltage is.
What did the 60 V / 1000 A IGBT supply measure on MoSi2?
Bottom line: heating was stable from cold state to 858 °C, output power rose from 0.196 kW to 14.073 kW, and both input power factor and efficiency rose with output power. Test conditions are listed below.
| Item | Value |
|---|---|
| Supply rating | 60 V / 1000 A, water-cooled module |
| Control mode | Constant current |
| Load | MoSi2 elements, two in series and two in parallel |
| Hardware / software version | V4.01 / V1.02 |
| Instruments | FLUKE 15B multimeter and FLUKE 353 clamp meter (power); HIOKI PW3360-31 (harmonics) |
| Temperature range | 61.6 °C – 858 °C |
Measured power and efficiency (setpoint 1.00 corresponds to the rated 1000 A):
| Setpoint | Output voltage | Output current | Output power | Input active power | Apparent power | Power factor | Output / input | Temperature |
|---|---|---|---|---|---|---|---|---|
| 0.148 | 1.33 V | 147.3 A | 0.196 kW | — | — | — | — | 61.6 °C |
| 0.185 | 1.78 V | 183.2 A | 0.326 kW | 0.71 kW | 1.57 kVA | 0.449 | 45.9% | 78.2 °C |
| 0.227 | 2.42 V | 224.5 A | 0.543 kW | 1.02 kW | 1.92 kVA | 0.533 | 53.3% | 105.1 °C |
| 0.278 | 3.51 V | 276.5 A | 0.971 kW | 1.56 kW | 2.63 kVA | 0.593 | 62.2% | 173.8 °C |
| 0.464 | 10.36 V | 461 A | 4.776 kW | 5.71 kW | 6.55 kVA | 0.871 | 83.6% | 374.2 °C |
| 0.492 | 12.06 V | 490 A | 5.909 kW | 6.89 kW | 7.71 kVA | 0.894 | 85.8% | 464.3 °C |
| 0.510 | 13.24 V | 504 A | 6.673 kW | 7.75 kW | 8.52 kVA | 0.909 | 86.1% | 537 °C |
| 0.547 | 15.52 V | 540 A | 8.381 kW | 9.62 kW | 10.45 kVA | 0.920 | 87.1% | 551 °C |
| 0.599 | 19.55 V | 593 A | 11.593 kW | 12.95 kW | 13.86 kVA | 0.935 | 89.5% | 805 °C |
| 0.618 | 21.16 V | 613 A | 12.971 kW | 14.39 kW | 15.33 kVA | 0.938 | 90.1% | 842 °C |
| 0.632 | 22.41 V | 628 A | 14.073 kW | 15.56 kW | 16.54 kVA | 0.939 | 90.4% | 858 °C |
Logging stopped at 858 °C, 628 A and 14.073 kW. The supply is rated 60 V / 1000 A, so output headroom remained.
What do the harmonic readings look like?
Bottom line: grid-side voltage THD stayed between 1.78% and 2.6% throughout, and current harmonic content fell as the load heated up. Dominant orders were the 5th, 7th, 11th and 13th, the characteristic harmonics of a three-phase six-pulse rectifier.
| Load temperature | Phase A current harmonic content | Phase A voltage THD | Phase B voltage THD | Phase C voltage THD |
|---|---|---|---|---|
| 78.2 °C | 85.97% | 1.92% | 1.86% | 2.03% |
| 173.8 °C | 78.51% | 1.79% | 1.80% | 2.05% |
| 374.2 °C | 47.52% | 1.91% | 1.91% | 2.10% |
| 551 °C | 37.01% | 1.84% | 1.89% | 2.07% |
| 805 °C | 34.04% | 2.15% | 2.16% | 2.39% |
| 858 °C | 33.03% | 2.22% | 2.41% | 2.55% |
Individual orders and total content are HIOKI PW3360-31 readings; the definitions follow the instrument.
How should these numbers be read?
- Cold start works. At 61.6 °C the bank took 147.3 A at only 1.33 V, which shows that constant-current control delivers current stably without squeezing a conduction angle.
- Power factor and efficiency rise with output power. Above 4.776 kW output, power factor was at least 0.871 and output/input at least 83.6%; at 858 °C the figures were 0.939 and 90.4%. The lower values at low temperature reflect the small absolute output power, where the supply's own losses dominate.
- Voltage distortion stays low. Voltage THD never exceeded 2.6%, below the 5.0% limit for 0.38 kV systems in GB/T 14549. The high current harmonic readings appear at light load, where the value is a ratio to a small fundamental current.
When is a DC supply worth considering?
- MoSi2 or SiC loads with very low cold resistance where start-up current stability matters.
- Sites sensitive to harmonics and power factor, for example where one transformer also feeds precision instruments.
- New installations that want to avoid transformer taps and changeover hardware.
Conversely, where an existing AC system runs stably and grid-side indices already comply, there is no reason to change for its own sake. Before any retrofit, run a harmonic check against the actual transformer capacity and short-circuit capacity.
FAQ
Does DC heating produce the same heat as AC in a MoSi2 element?
Yes. A MoSi2 element is a resistive heater, so heat output is P = I²R and depends only on RMS current and resistance. With equal RMS current, DC and AC produce equal heat in the same element, and MoSi2 involves no electrolysis or polarisation process tied to current direction. In the KULUN test, a 60 V / 1000 A IGBT DC supply heated a two-series two-parallel MoSi2 bank from 61.6 °C to 858 °C under stable control.
Why is a MoSi2 element close to a short circuit when cold?
MoSi2 has a positive and pronounced temperature coefficient of resistance. In this test the two-series two-parallel bank needed only 1.33 V for 147.3 A at 61.6 °C, an equivalent resistance of about 9.0 mΩ; at 858 °C it took 22.41 V for 628 A, about 35.7 mΩ, roughly four times the cold value. A supply for this load must therefore be specified for low voltage, high current and current limiting, otherwise switch-on is effectively a short-circuit condition.
Why does phase-angle control give low power factor during MoSi2 heat-up?
Phase-angle control regulates output voltage by changing the thyristor firing angle. During heat-up the MoSi2 equivalent resistance is only tens of milliohms and only a few volts are needed, so the thyristor conducts in a narrow window each half cycle and current becomes a narrow pulse. Displacement factor and distortion factor both worsen, power factor falls, harmonic content rises, and other equipment on the same transformer may be disturbed. That is the direct reason sites move to transformer taps or a DC supply.
The current harmonic readings are high. Will the supply pollute the grid?
Judge grid impact by voltage THD, not by the relative current harmonic reading. Current harmonic content is a ratio to the fundamental, and at light load the fundamental is small, so the ratio looks large. In this test the grid-side voltage THD stayed between 1.78% and 2.6%, below the 5.0% limit for 0.38 kV systems in GB/T 14549. Projects should still be checked against transformer capacity and short-circuit capacity, and sensitive sites may add filtering or multi-pulse rectification.
Can an existing AC MoSi2 furnace be converted to DC supply?
Technically yes, but check three items first. Supply rating: rated voltage must cover the highest voltage required by the hot-state resistance, and rated current must carry the maximum working current with margin. Load wiring and insulation: DC has no polarity requirement, but contact heating, busbar drop and insulation ageing in a low-voltage high-current circuit need rechecking. Control interface: the existing temperature controller and host system must match on 4-20 mA or 0-10 V setpoint and communication protocol. A trial run at reduced power is advisable before converting the whole furnace.
Timeliness note
This article is based on a KULUN test of November 2026 in which a 60 V / 1000 A IGBT DC supply drove a two-series two-parallel MoSi2 bank under constant-current control from cold state to 858 °C. Whether a DC supply suits a specific project depends on the actual load parameters, transformer capacity and an on-site harmonic check.
