I want to build a simple and cheap circuit to power a NiCr heating wire using 120VAC 15-amp household current. I want to be able to vary the heating power across the NiCr element from 0 to 100 watts.

Assuming I'm not cutting the voltage, my understanding is that this means I need a rheostat circuit in series with the heater capable of ranging from 120Ω up to kΩ or MΩ, and capable of dissipating 120 watts at its lowest resistance!

The only rheostat I can readily purchase is a 25Ω 3W pot from Radioshack. Does anyone have clever ideas on how I could build my circuit using household or readily salvaged components?

 

A light dimmer should let you vary the power into the nichrome wire. There are several caveats, though. Just placing 120 VAC across any old length of nichrome does not give 100 watts of heat dissipation.

The wire must be supported on non-conductive posts (usually porcelain) and be placed in a housing so accidental contact is prevented. The wire will always carry a lethal voltage. The supports must allow for the tendancy of nichrome wire to sag when hot.

Due to heat and the nature of nichrome, electrical contacts must be crimped. Uninsulated barrel connectors are best. Wire nuts are not suitable.

Do you have a 100 watt length of nichrome? Power is the product of voltage and current. The gauge of the wire and its length determine the resistance, and therefore the power the length of wire will dissipate. The current for 100 watts in a 120 volt circuit is 833 milliamps. The resistance of the wire when hot should be on the order of 144 ohms.

This site will let you figure the output of your nichrome - http://www.wiretron.com/design.html

 

Thanks for that info, I think I have the heater element squared away.

My biggest problem now is finding a rheostat and resistive elements that can handle the power of this circuit. First I want to limit the amperage in the circuit to about 1 amp. What efficient means can I use to accomplish this? A simple resistor is going to need to dissipate over 100 watts, which isn't very efficient! But I suppose it would be simple to do with household items since I could just put a 100 watt light bulb in series in the circuit, right?

The second problem is the dimmer: All modern dimmers I can find in stores apparently use triarcs, not rheostats. (It didn't take me long to discover this: I exploded a NiCr wire as soon as I turned on a dimmer at the lowest setting, because when the current is on it's full voltage and amperage!) So I still either need a true rheostat that can handle up to 100W (nothing readily available I can find) or something more clever?

 

Actually I guess I should clarify that my objective for the heater is not power but rather point temperature in contact with the NiCr wire. It will probably take substantially less than 100W but since I need to get up to point temperatures of 500C with a 20AWG wire it looks like I will need to allow up to 7amps into the circuit. So if I can reduce the circuit's voltage and preserve the ability to continuously vary current up to 7amps that would meet my objective.

I do happen to have a transformer that claims to output 10amp at 24VAC if that simplifies things.... But I still need a rheostat that can handle this amperage, and even at that voltage it's hundreds of watts of dissipation!

 

So I still either need a true rheostat that can handle up to 100W (nothing readily available I can find) or something more clever?

Power rheostats are obsolete, as there are more cost-effective ways to dissipate power (e.g., semiconductors and heat sinks).

If you tell us what you're trying to accomplish, there are numerous scientists and engineers here who might have some good ideas on getting you where you want to go.

 

I am trying to build a tiny NiCr wire heater whose temperature I can relatively precisely adjust, and with a range up to at least 500Celsius.

I can read its temperature with a thermocouple, so what I am looking for help constructing is: A circuit to connect to the NiCr wire with continuously variable amperage up to roughly 7amps. I am nothing of a solid-state engineer, so I am hoping to do this with a minimal number of parts, ideally ones I could easily purchase or scavenge locally. For power I have:

• 120VAC 15-amp household power
• A 24VAC 10-amp transformer
• 12VDC 7-amp batteries

 

What do you mean by "precise"? Your design choices are largely dictated by how close you want to control the temperature. Do you want open loop or closed loop control?

Open loop can be you monitoring the temperature, then setting a knob to control the power. Cheap and effective if there's a constant thermal load on the wire. One of the easiest ways to do this is with a Variac (and, possibly, a step-down transformer).

Are you measuring the temperature of the wire itself or is the wire going to be used to heat something else? A sketch would help a lot along with a description of what exactly you're trying to accomplish. What's the diameter, length, and resistance of the wire?

Closed loop control can be more precise and less tedious, but is more work to get to. One thing to consider are the little PID controllers sold for about $50 on ebay. If you're heating something with a reasonable amount of thermal mass, you could consider just using them with on/off control (or use a simple bimetallic controller). For better control, the proportional features can be used, but then you'll need some way to continuously vary the heating power. A simple way would be to rectify the AC to DC with a full wave diode bridge (no filtering needed), then use a MOSFET controlled by the temperature controller to vary the current. With no further constraints, this would likely be my choice, especially if the temperature controller had a DC voltage out that could directly be used for the gate-source voltage of the MOSFET. If you're a do-it-yourselfer, then you could build your own control circuit if you have the requisite linearized & amplified thermocouple signal.

If you have access to a suitable DC power supply, you can characterize the steady state temperature vs. current and perhaps use that to set the temperature you want (i.e., no closed loop control). Of course, this won't work if there's a varying thermal load.

If you have the thermocouple connected to the wire, you may have troubles with common-mode noise. Nichrome and chromel are very similar (both Ni-Cr alloys), so you might be able to use a chunk of alumel wire and spot weld it to the Nichrome and (nearly) have a type K thermocouple junction. In case you weren't aware, you can use some chromel wire in the place of Nichrome in a pinch. Oh, and the thermocouple junctions don't have to be welded -- good mechanical contact (e.g., a screw clamp) can work (I've made thermocouples for quicky projects by just twisting the wires together). But a clamp increases the response time. I use an oxy-acetylene torch to quickly weld a bead on the thermocouple wires -- no muss, no fuss.

 

I was going for a simple open-loop system. I lack both the time and experience to build a closed-loop system, although I appreciate your suggestions and will research them to expand my skills when I get a chance! (I dream of the day I am competent enough to weld a functional thermocouple joint!)

I actually rigged your first proposal last night (triac dimmer through step-down transformer to smooth the current) and was getting good results with that.

Can you clarify the potential "common-mode noise" problems I can have with a K-type thermocouple directly touching the NiCr wire for temperature measurement? With my prototype I simply wrapped the NiCr once around the tip of the thermocouple. I turn up the current little by little, waiting for the temperature reading to stabilize at each step.

FYI, the purpose of the circuit is to determine the combustion temperature of tiny samples of pyrotechnic powders, so once I put samples on the wire I'll probably give it a little longer dwell time at each step to ensure that it has had time to bring mass in contact with the wire to the wire's temperature.

 

(I dream of the day I am competent enough to weld a functional thermocouple joint!)

It honestly doesn't require much skill. You use some needlenose pliers to twist two wires together for a couple of turns (provides a mechanical joint to hold the wires together for welding), then briefly dip the wire into a torch flame near the end of the light blue cone near the tip (hottest spot in an oxyacetylene flame) and you'll get a molten bead of metal. Done. For production work, spotwelding is often used, but most homeowners don't have the necessary equipment.

Can you clarify the potential "common-mode noise" problems I can have with a K-type thermocouple directly touching the NiCr wire for temperature measurement?

This is a fancy term for the fact that both leads of the thermocouple may have a large AC voltage on them (the same amplitude and in phase, hence the term "common") due to the heating current. Our EE friends know all about this stuff and have nice tools to deal with it. If you're a DIY-er, read about op amps and common mode rejection in "Art of Electronics". Otherwise, some of the experienced folks on this board can give you guidance. The commercially available instrumentation amplifiers should be able to deal with this noise with no problem.

the purpose of the circuit is to determine the combustion temperature of tiny samples of pyrotechnic powders

This sounds like a neat application. I'm not sure I see the need for a relatively long piece of wire like it sounds you have. If I was doing the experiment, I think I would have used a short piece of wire (say, 1 or 2 cm). Then I'd use a DC power supply in constant current mode to heat the wire. I would first have made a calibration table of temperature versus DC heating current. I've attached a simple voltage/current plot I made for a small light bulb I was using for something -- this shows the utility of a good bench DC power supply: as you turn the knobs, read the numbers off the supply and type them into a spreadsheet, and plot the results.

If you continue to experiment, I think one of the best investments you can make is to get a good DC power supply for your bench. My favorite is an HP E3615A (20 V, 3 A) which I found for $100 in virtually unused condition on ebay (I consider this almost a steal as I was the only bidder). I use it constantly. It has 10 turn pots for the current and voltage controls and can work in either constant voltage or constant current mode. Agilent still sells this supply for $760 (ouch!). I'd recommend looking for used supplies on ebay -- that's what I did, as I can't afford to pay the new prices for my hobby.