ContinousTempSensor

That's a complex topic where the basic problems are the same, no matter who the manufacturer is.

Basically, a continuous temperature sampling is only useful under very specific circumstances and should not be used in most cases. An active temperature compensation works by changing the interpretation of the machine scales. When the temperature of a material changes, the material changes not simply its length, but what it really does is change its "length per length per temperature difference". That means, for every degree of temperature change, the length of the material expands or shrinks depending on the length difference of the work piece (between points taken). It's a linear change, but different for every part of the work piece.

Since every point on a scale is affected differently by the temperature compensation, there will be one point on the scale where the actual length change will be zero. It's like an origin for the temperature compensation. During a measurement, it doesn't matter, where that origin actually is, because the temperature difference per length difference is linear, but only as long as the temperature difference doesn't change throughout the measurement.

That effectively means, that the measured temperature MUST be the same throughout the whole measurement routine. If you want a continuous temperature change, you MUST be able to define where the origin of the temperature compensation is, and that is a problem. Usually, your part is clamped down to hold it in more than one point, not allowing to expand at all. Or maybe you're measuring a heavy part that isn't clamped down at all, but without a defined clamping point, the part makes contact with the granite table in many undefined places. When the part expands/shrinks during measurement, either the part expands or shrinks in unknown ways, or when it's clamped down tightly, it will not expand or shrink at all, but rather build up tensions.

The best solution is to wait until the part temperature is somewhat stable and then use only this one temperature from begin until the end of measurement. Any change in temperature compensation during the measurement will totally render the measurement bad, because it will change the whole relationship between all points taken. It will be worse, the farther away the temp compensation origin is from the actual part origin.

The only way to make a constant temperature sampling work and be useful, is to clamp your part down in ONE defined spot, and every other clamping spot must clamp the part down but still allow for lateral motion, much like a bridge suspension (a "floating" clamp, if you will).

I only know of one useful case, where this is needed: If you measure a work piece made of two different materials with two different temperature compensation coefficients. The different materials must be specified as well as their respective fixed clamping points (ONE for each material). All other clamping points must be floating clamping points, allowing for movement.

There is NO WAY that any kind of software can predict the motion of a part that is clamped down in undefined spots, and therefore a continuous change of the temperature read-out doesn't make ANY SENSE, except in the very special case that I laid out.