Problem with High Temperatures Surrounding Coil.

[chemelec]

One person I am dealing with needs to detect large pieces of metal in a VERY High Temperature Area, using a special designed PI I designed for his application.

He Wrote to me:

I have been working with the pulse circuits for many months I am focusing on temperature compensating the coil when its at high temp (450F) I have managed to encapsulate it in ceramic which is no problem in this temp but the sensitivity on the coil goes down with an increase in temp.

I suspect that as Temperature goes Up, So does the DC Resistance of the Coil.
Causing a Lower Magnetic Field.

Also I know magnets LOSE Magnetism at higher temperatures.
Not sure if this also applies to Electromagnets.

ANY OTHER SUGGESTIONS?
Or Possible Cures?

[Skippy]

I'm sure if you look online, you can find out data on resistivity of copper, and how it varies with temperature. From this you can calculate how much variation in coil resistance will result. I'm no PI expert, but I assume the "inverse 6th power" (perhaps inverse 5th power might be more appropriate at the distances hobbyists use) applies, so if your hot coil is running at half the normal (cold) current, you might expect to lose about 15% depth?. Maybe if you added some series resistance inline with the coil, that would reduce the variations caused by coil changes? Another random idea - could the coil be made from a low tempco alloy? These are usually higher resistance, so probably not suitable, but without looking at the specs, I couldn't say. I'm sure others will have ideas.
Here's nichrome data:
http://en.wikipedia.org/wiki/Nichrome

[Tepco]

At 450F, or some 230 deg C coil resistance will double, reducing peak current twice, so longer pulse, or higher drive voltage can fix this, depending on coil design. But if metal object to be detected is also on high temperature, TC and amplitude will change too. There is no critical temperature for electromagnet, will work in molten state if needed.

[techo_bob]

Hi guy's

If this is an industrial situation
you could try small copper tube and water cool the coil as they in induction heating
systems..

Bob..

[Skippy]

Here's some resistivity data, including tempco's. It looks like manganin or constantan alloy would be the most obvious choice, taking into account the need to keep resistance low. But you would need a significantly thicker conductor (ie. paralleling up several strands, for example) to get the low resistance.
http://en.wikipedia.org/wiki/Resistivity

[green]

Is the depth lose about 15 percent mentioned in reply #2? Maybe control coil driver volts with temperature feedback

[Qiaozhi]

The temperature coefficient of copper is

If we say the coil resistance is 4 ohms at 20 degrees C, then from:





which is almost double the original resistance value, as Tepco said in a previous post.

Please be aware that this is not an accurate value by any means. The equation is a linear approximation that is only valid over a small range around the reference temperature, and R is actually the resistivity and not the coil resistance. But it's probably close enough to prove the point. As a practical experiment you could try monitoring the current with the coil in the oven, and then plot a graph to seen how it varies with temperature.

[Skippy]

...or you could measure the DC resistance of your coil at ambient and at elevated temperature(s). Presumably you have some cable length between the hot coil and the circuitry - this may have an ohm or two resistance, which doesn't vary so much. Rewinding your coil with thicker wire so it's 2 ohms instead of 4 ohms, for example, and then adding 2 ohms fixed resistance would reduce the temperature variation: Ambient R=2+2=4, Hot R=2+4=6, as an example.

[Tepco]

Or even better, use constant peak current control circuit, it will automatically adjust pulse width to compensate for resistance change.

[chemelec]

Thanks for all your Replies!
Take care....Gary