# How to measure temperature of a syrofoam cutter hot wire?

I would like to create a hot wire foam cutter device. The wire material is nichrome, diameter 0.9mm, resistance is about 1.7 Ohm/meter. For best cutting results, I would like to control the temperature electronically. This will be a 4 axis CNC system where the wire length between the mounting points changes while the axes move. Planned length of the wire is abount 1.5m. Another extra (non-heating) wire is added, that is able to change its length to ensure the tension of the wire while the axes move. So the length of the heated part of the wire would not change, but it would be hard to measure the contraction caused by the heat.

The main question is this: how can I measure the temperature of the wire? I was first thinking about a thermistor, but then I realized that the hot wire itself could be acting as a thermistor. When temperature goes up, then probably the resistance will also go up. However, I have no idea how to create a circuit that measures the current flowing through (it can be 10A), and how to calculate the temperature from that. Another problem is that the hot wire will be heated with a 24V 200W transformator, unregulated AC power. My idea for power control was to use a diode bridge, a high power fet for switching (e.g. IRF44N or similar) and PWM signal to control the temperature. But this would make temperature measurement impossible - PWM controlled AC cannot be used together with the other idea.

The workaround I have in mind is this: use regulated DC and a simple A/D converter to measure the resistance of the wire in every (say) 1 second. Heating would be turned off completely during resistance measurement. This would take just a few msec. Use the PWM controlled driver to heat the wire in the remaining time. Would that work? How can I separate the heating circuit from the measurement circuit? (E.g. to protect the A/D converter safely.)

The secondary question is this: do any of you have an idea about how much power and voltage I need to heat the wire? I thought that 200W would be enough. That is about 8 amps at 24V. But don't have any experience. Also, even if I can measure the resistance of the wire, I have no idea how to convert it to a temperature value. (It would be nice to display the absolute temperature value, but it is not important - all I care is to adjust and maintain a constant temperature that has the best cutting results.) I have a 24V 210W transformator for heating, but the power may not be enough.

Or maybe you have an easier solution in mind - please don't hesistate to share.

Mechanical part:

Heating goes through the green lines and the red wire. Tension is provided by the gray cable (which is not heated).

• Look at a Wheatstone Bridge, always the general solution I use when needing to measure a varying resistance. Though I am not sure how efficient it'll be with that high of a power, I've only used it in low power applications. – Jarrod Christman Jan 10 '16 at 13:23
• To keep temperature constant, the amount of power required will depend upon how fast you are cutting the styro-foam. Notice that the styro-foam is removing heat from your hot wire. – Marla Jan 10 '16 at 14:48
• It's even trickier than that. At high cutting speed you may remove a lot of heat from the portion of wire in the foam and only air cooling from the rest. Running the cutting wire hot enough may overheat the air-cooled portion. – Transistor Jan 10 '16 at 15:51
• Yes, the power required depends on the thickness of the foam and the cutting speed too. This is why I want at least try to keep the temperature of the wire constant by measuring its resistance. Heat conduction of the wire is good enough, so if I could measure the total resistance then I could interpolate its temperature with an MCU, and modify the PWM accordingly. This was the idea, just don't know how to do the heating and the resistance measurement at the same time. – nagylzs Jan 10 '16 at 17:45
• It suddenly got much more complicated with the addition of variable length. With fixed length cutting wire we could calculate temperature by using current and voltage feedback, calculating temperature rise and getting temperature from lookup table. With variable length the cold resistance value will change. Is the length changing on the fly or could you stop, change length, allow wire to cool, recalibrate and resume? Another, simpler option is to control current. A set current will give constant power per unit length of wire and automatically compensate for change in resistance. – Transistor Jan 10 '16 at 18:56

Because your supply is unregulated, you can try something like

simulate this circuit – Schematic created using CircuitLab

Using a current sense resistor (in this case, 3 0.1 ohm power resistors in parallel), you can find the current through the wire. Then, knowing the voltage across the wire you can divide the voltage by the current to get the resistance. Referring to a table of resistance for your wire, you should be able to derive the resistance and hence the temperature. Nominal current voltage will be about 0.27 volts, which is a fairly convenient voltage level to measure, and a x10 amplifier using an op amp should be simple enough if you need it.

For the setup shown, 3 5-watt resistors could be used, since the nominal dissipation for each resistor is about 2 watts. Heatsinking the resistors will help accuracy.

While this is a sort of roundabout way of measuring resistance, it is arguably simpler than interrupting the heating cycle to make more conventional resistance measurements.

• I've seen, but can't find, pictures of a power meter with two needles pivoted at bottom left and bottom right of the square case. One needle indicated current, the other voltage. The scale was a series of 2D power curves. Power reading was given by the curve under the intersection point of the two needles. Clever stuff! – Transistor Jan 10 '16 at 19:42
• This could work with DC power, but power supply is unregulated AC. How to measure "average voltage" with an A/D conveter? A/D conversion takes time, and the time it needs depends on the voltage. – nagylzs Jan 11 '16 at 8:34
• @nagylzs - So you rectify the voltage (but only at the measurement instrument). Google, for instance, "precision rectifier". How do you think DMM makers measure AC voltage, anyways? – WhatRoughBeast Jan 11 '16 at 13:57

High current AC control

simulate this circuit – Schematic created using CircuitLab

Given that your specification requires variable length cutting wire it appears that current control may be better. A quick web search shows up several devices.

Nu Wave current controller.

I have never heard of this company before but something along these lines could provide you with an off-the shelf solution. A brief look at the data sheet indicates that you need a separate SSR to switch the load and a small transformer to power the module.

Note that since $P = I^2R$ the potentiometer will give linear control of current but exponential control of power. (Confirm the linear current with the manufacturers' data sheets.)

• Yes, constant current will give constant output power. It is must better than nothing, but I would rather have constant temperature instead. (When the thickness of the foam changes in time, the power needed for constant temperature also changes.) – nagylzs Jan 10 '16 at 19:18
• Note that the constant current solution gives constant power per unit length which will be most useful for variable length cutting wire as indicated in your update. Wire temperature is only one of the variables. Foam material, thickness and cutting speed will also vary so you're going to end up with a table of values that work for you. Current control will probably be just as good as power / temperature. – Transistor Jan 10 '16 at 19:46

One of the problems with hot-wire cutting is that the wire expands with temperature and a tensioning system is required. You could use this to your advantage and control wire temperature by controlling wire length. This in turn could allow very simple control from your AC transformer through a 'phase-angle firing module'. These are glorified dimmer controllers with a DC control input.

simulate this circuit – Schematic created using CircuitLab

Your challenge with this approach is to come up with the linear transducer and some means of calibration. Setup would be by trial and error.

For my one attempt at this I used a guitar string, E, 1st, 0.010". I had a 12 V, 5 A power supply and adjusted a contact along the wire until I got the required result.

• This will be a 4 axis CNC cutter. The length of the wire is also dependent of the position of the axes. Tension system was built in, but measuring the expansion mechanically while the axes move would be difficult. I was hoping that measuring the resistance would work. – nagylzs Jan 10 '16 at 17:41
• That's fine. Can you edit your original question to explain what power supplies you have available and I don't think you mentioned what length your cutting wire will be. (Shorter is better for straightness.) – Transistor Jan 10 '16 at 17:47
• Added some more details – nagylzs Jan 10 '16 at 17:53