Is the attic always colder than the ground floor?

Surface temperatures are not quite trivial to measure, the values determined here are probably worthless. Flow and return temperatures should be easier to measure. Flow rates at least relatively based on the floats in the heating circuit distributor.

Obviously, one should measure steady states, in this case that means waiting a long time after a change in the setting (and turning off outdoor temperature control, instead operating "manually").

Surface temperatures are not quite trivial to measure, the values determined here are probably useless. Supply and return temperatures should be easier to measure. Flows at least relatively based on the floats in the heating circuit distributor.

Of course, one should measure stationary states, in this case that means waiting a long time after a change in the setting (and turning off the outdoor temperature control, instead operating "manually").

Why is the surface temperature not so easy to measure?

In non-contact measurement (radiation methods), one must know the (temperature-dependent) emissivity of the surface. If this is low, there is a significant influence from the ambient radiation.

With attached, glued-on thermometers, there is always a feedback between the body to be measured and the thermometer (heat flow).

If one now wants to balance a underfloor heating system, where there are only a few degrees difference between the supply and return flow, and the measurement is carried out by a layperson (or craftsman) in the absence of experienced physicists with sinfully expensive measuring equipment, then I tend to consider the measurement results (with regard to the intended use) as worthless.

It is better to measure directly in the medium itself. That is trivial as long as one can access it, and the measured temperatures are then very accurate. The remaining uncertainty is the state of the system (stationary?).

Ok, the evaluation or interpretation is then something different. But with the thermal imaging camera, one should at least be able to measure a delta of the surface temperature, which increases the supply temperature by 5° or more. If, then, even when measured in a layman's way, with the same "measurement method" the surface temperature does not change, that should already make one skeptical. Of course, I also do not know which value, for example, is good for window frames, roofs, etc., but even a layman can possibly detect leaks.

For non-contact measurement (radiation methods), you have to know the (temperature-dependent) emissivity coefficient of the surface. If this is low, there is a significant influence from ambient radiation.

With attached, glued-on thermometers, there is always feedback between the body to be measured and the thermometer (heat flow).

If you now want to balance a floor heating system, where there is only a few degrees difference between supply and return, and the measurement is carried out by a layperson (or craftsman) in the absence of experienced physicists with outrageously expensive measuring equipment, then I tend to consider the measurement results (with regard to the intended use) as worthless.

It is better to measure in the medium itself. This is trivial as long as you can access it, and the determined temperatures are then very accurate. The only remaining uncertainty is the state of the system (stationary?).

Let's keep things in perspective. It's about measuring whether the surface temperature changes with the increase of the supply temperature. For this, the emissivity coefficient doesn't matter, and especially not its temperature dependence. You just stick a piece of black insulating tape on the surface, and that's it. And with contact measurement, you worry about accuracy because of heat flow? Between screed and, for example, a Pt100? Seriously? If I were 15 years older, I would ask: What’s wrong with engineers nowadays, do they only see obstacles??? Do you also call an "experienced physicist with outrageously expensive measuring equipment" to work when there is a small thing to measure?

Exactly, the good camera still registers for a "long" time when the hand has only touched the ground for a short moment. More than 5 degrees difference in the pre-run temperature should therefore be noticeable to anyone. To what extent the surface temperature should then change is not so easy to predict, but even with bare feet you can clearly feel the difference...