1
\$\begingroup\$

I am rather new in this field.

On 220V AC industrial power supply, a 6 ohm resistor heater (nichrome wire, 0.40 ohm/m, 15 m) will draw about 8067 watt of power. About 2400 W are needed.

What are the common ways (if there are any) to bring the power being drawn down beside extending the wire?

Looking for other options because the wire is a bit expensive and I would like to bring the cost down even just a little.

Context:

  • What are you actually trying to do? Create electric furnace
  • Are you trying to heat something? Yes, it will heat air and metal
  • Do you already have the wire so you need to use it? Yes, some of them and will buy more if needed
\$\endgroup\$
11
  • 2
    \$\begingroup\$ More context needed. What are you actually trying to do? Are you trying to heat sometyhing, or load the power supply, or something else? Do you already have the wire so you need to use it, or what? PWM or burst on/off are common ways of bringing the power down, but they may not be appropriate depending on what you're trying to do. \$\endgroup\$
    – Neil_UK
    May 13, 2022 at 13:22
  • 1
    \$\begingroup\$ context added, its 220V AC \$\endgroup\$
    – Bramble
    May 13, 2022 at 13:27
  • 4
    \$\begingroup\$ If you don't want to change the nichrome wire, then you can either reduce the voltage you're driving it with, or use a simple 'simmerstat' type of heating element controller which reduces the duty-cycle at which power is applied to the heater (you'd need about 30% to get 2.4kW from your 8kW heater). \$\endgroup\$
    – brhans
    May 13, 2022 at 13:56
  • 2
    \$\begingroup\$ You don't really have complete freedom in changing wire gauge and length- the wire surface temperature and thus element life will be impacted greatly by changes. A proper heater design is needed. With a controller you will want somewhat more power when the element is at operating temperature than is required to maintain the temperature in steady-state, perhaps +50% (not 300%). \$\endgroup\$ May 13, 2022 at 14:58
  • 2
    \$\begingroup\$ Industrial heaters are often controlled by SCR or Triacs using burst mode instead of phase angle. The thermal time constant of the system is ususally long enough to allow zero-crossing burst mode. This mode will generate much less EMI compared to a phase controller. What type of controller have you planned? Unless you are running open-loop, you need something to switch the load anyway. Do it right. \$\endgroup\$
    – Mattman944
    May 13, 2022 at 16:37

6 Answers 6

3
\$\begingroup\$

If you need 8067 watts of power to heat whatever it is you're heating, then... you need the watts, can't reduce the power.

If you do not need the watts, but find you need less watts, then you can use a higher resistor value... for example by extending the wire, or if the cost of the wire bothers you, use thinner wire, which will have higher resistance and be cheaper than thicker, lower resistance wire.

Note nichrome wire can't heat metal because it's not insulated. So if it touches the metal, current will flow in the metal instead. It's nice for heating air, but make sure there's enough airflow, otherwise it'll melt.

\$\endgroup\$
7
  • \$\begingroup\$ It doesnt actually need 8067 watts. that number is just for case example and from pure calculation only, 220*220/6 = 8067. im curious what my options are to bring the power consumption down. \$\endgroup\$
    – Bramble
    May 13, 2022 at 13:33
  • 1
    \$\begingroup\$ Well, first you need to figure out how little power your oven can use while still reaching the desired temperature in the desired time... \$\endgroup\$
    – bobflux
    May 13, 2022 at 13:34
  • 1
    \$\begingroup\$ If you wire it in triangle instead of delta, it'll get lower voltage, so power will be lower. Besides that, it's about wire length and thickness... \$\endgroup\$
    – bobflux
    May 13, 2022 at 13:54
  • 1
    \$\begingroup\$ @bobflux, "If you wire it in triangle instead of delta ...". Do you mean "If you wire it in 'Y' instead of delta ..."? \$\endgroup\$
    – Transistor
    May 13, 2022 at 14:28
  • 1
    \$\begingroup\$ @Transistor Yeah, that's what I meant, thanks for correcting! In French we say "triangle" and "étoile" (star) and it looks like I had a brain fart while translating... \$\endgroup\$
    – bobflux
    May 13, 2022 at 14:36
3
\$\begingroup\$

220 Volt / 6 Ohms = 36.6 Amps * 220 Volt = 8066 Watts

120 Volt / 6 Ohms = 20 Amps * 120 Volt = 2400 Watts

Reduce the voltage. At these currents you likely want a transformer. Or get a 120V power supply in the first place. That's the practical approach if you can't extend the wire or use a suitable replacement.

Or use a constant current regulator.

Alternatively you could use a PID controller to cycle the wire on/off until you get the required temperature. You'd still have to provide full power but then it will cycle on/off to maintain that temperature.

\$\endgroup\$
2
\$\begingroup\$

Since your application (electric furnace for fused metal) is not going to need quick reaction times, one flexible way to control the heating power is to use PWM (Pulse Width Modulation), as mentioned by Neil_UK in a comment.

PWM can be generated by a small microcontroller and then used to drive the heating wire through a driver circuit using MOSFETs. Given the thermal inertia of the wire-furnace system, the PWM signal could have a rather low frequency (in the hundreds of Hz range), easily generated by even the most basic MCU using just bit-banging, although using a higher frequency (above 20kHz) could increase efficiency of the switching MOSFET. There are tradeoffs in this choice.

Even if the MCU is just used as a PWM generator and not in a full closed-loop control system, it will allow you to set the power in a very simple way.

In fact you can hook up a simple potentiometer read by the internal ADC of the MCU and then convert the pot setting to a duty-cycle setting in order to set the desired power into the wire.

If you want to be more sophisticated, the MCU could be connected to a thermocouple that monitors the temperature of your furnace and the reading could be displayed, together with the preset duty-cycle, on a simple LCD module, also controlled by the MCU.

At this point, if you want even more sophistication, the system is easily modified (in software) to become a closed-control system, with the MCU automatically setting the duty-cycle of the PWM signal to obtain a preset temperature.

Since your application uses AC, this simple approach would need to convert the AC power to DC. This is easily done with just a rectifier diode and a cap (ripple is not an issue in such a circuit).

If DC conversion is not an option, you could use a TRIAC-based power controller using phase-control. These analog circuits are simple and reliable for powering resistive loads, and have been around for decades (the classic incandescent lamp dimmer circuit).

This latter solution doesn't involve an MCU. However, you could use an MCU to control the TRIAC and have all the benefits I mentioned before (LCD display, etc.).

Here is a quite comprehensive application note from ST about TRIAC control using an MCU in AC applications.

\$\endgroup\$
5
  • 1
    \$\begingroup\$ This answer would make sense for a few watts. But 8 kilowatts? That's just insane. Can't rectify that much power with "a diode and a cap". I can't imagine how to even switch that much power with MOSFETs (at hundreds of Hz no less) without burning them down. \$\endgroup\$
    – akwky
    May 13, 2022 at 14:47
  • \$\begingroup\$ @akwky My answer doesn't specify a complete design. My sentence didn't necessarily mean "a single physical cap". From a theoretical circuit POV it's just a diode and a cap. You definitely can have rectifiers for that kind of power range. Of course caps and rectifiers should be designed for the purpose. This doesn't mean the circuit is the most efficient you can devise. My answers focused on circuital simplicity. \$\endgroup\$ May 13, 2022 at 14:54
  • \$\begingroup\$ @akwky Look at this beast, for example: smc-diodes.com/propdf/S4D40120H%20N2409%20REV.D.pdf. It's rated at 120 odd amps (at 25°C) and 40A at 155°C! \$\endgroup\$ May 13, 2022 at 15:02
  • \$\begingroup\$ In this instance, "hundreds of Hz" duty cycle is far too high. One should switch on and off at zero-crossing of the AC waveform (two per cycle at 50/60Hz depending on the country). So a duty cycle likely around 1Hz, with steps of one half-cycle, so control steps of 1/120 to 1/100. Even better, make a one-bit DAC with dithering, instead of PWM. Same ratio of "on" to "off" half-cycles, but they will be more evenly distributed, reducing variation in the heater temperature. \$\endgroup\$
    – Ben Voigt
    May 13, 2022 at 15:45
  • \$\begingroup\$ @BenVoigt Well, it depends on the exact thermal time constant of the system and the allowed residual "ripple" in temperature. Anyway you are right in that I should have said "hundreds of Hz or lower". About synchronizing with zero-crossings, you need some sort of zero crossing detector for that (if the MCU has an internal analog comparator that could do), but the circuit and the firmware would be more complex. \$\endgroup\$ May 13, 2022 at 16:15
2
\$\begingroup\$

In addition to other answers, the simplest approach I can think of is to reduce the diameter of the wire. That will increase the resistance and drop the power. In addition, the smaller-diameter wire will probably be cheaper than what you're using now.

At 220 V, 6 ohms (15 meters at 0.4 ohms/meter) gives you 8067 watts. You want 2400, so your resistance needs to increase by (8067/2400) or 3.36, for a new resistivity of 1.34 ohms/m. You haven't told us what Nichrome alloy you're using (there are several), nor what gauge you're using. Assuming you're using about 9 ga, you can replace it with about 23 ga. I've never dealt with nichrome at 9 ga, but I expect the cost savings would be considerable, at least 50%.

\$\endgroup\$
1
\$\begingroup\$

Typically, industrial power supply is 3-phase. You can split power between 3 phases (3 heaters, star or triangle, depends of voltage you wish) so reduce current through every single phase.

Another approach is to improve thermal insulation of the furnace. Temperature comes to equilibrium when heat income from heater (constant) gets equal to heat losses (increase with temperature increase). So reducing losses will reduce power consumption while keeping temperature.

\$\endgroup\$
-2
\$\begingroup\$

You can put a resistor in series with the 6 ohm heater , in order to reach 2400W of power you will need a smaller than 6 Ohm resistor.

\$\endgroup\$
9
  • \$\begingroup\$ does it need to be special kind of resistor since the power flowing is that much? \$\endgroup\$
    – Bramble
    May 13, 2022 at 13:52
  • 5
    \$\begingroup\$ This is really just another way of saying "make the wire longer". You're increasing the overall resistance of the circuit, and this new resistor yo insert is also going to be dissipating a lot of power as heat. \$\endgroup\$
    – brhans
    May 13, 2022 at 13:54
  • 5
    \$\begingroup\$ Yeah. A longer piece of nichrome wire. The resistor would be big and majorly expensive - it'll have to be rated for thousands of watts. Oh, wait. So's the nichrome wire, and we can actually use the heat rather than having to dipose of it. \$\endgroup\$
    – JRE
    May 13, 2022 at 13:56
  • 2
    \$\begingroup\$ Thinner wire = higher resistance. Less power for the same length. \$\endgroup\$
    – JRE
    May 13, 2022 at 14:31
  • 1
    \$\begingroup\$ This is a terrible, terrible idea in this case because the resistor will waste so much power. It will end up wasting almost as much as the heater uses. \$\endgroup\$
    – user57037
    May 14, 2022 at 20:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.