hello, i'm new with GC and i've a problem... while operating tcd signal decreases but at the end of the "run" it turns back to high values! is that normal? i don't think so, so what could be the problem? thank you very much, sorry for my bad english

Your method may have two settings for the sensitivity of the detector and it may switch between these setting during the run.

Or you may have a column that bleeds greatly and you are performing a temperature program run.

As the bleed drops after cooling it may not reach equilibrium when you start increasing the oven temperature again which again increases the signal.

That would produce what you have described in your post.

Pascal,
I haven't used a TCD in awhile but if I remember correctly they are very flow sensitive. So if you have a constant head-pressure in the injector and are temerature programming the column, the actual flow through the column will decrease as the temperature increases cuasing the baseline to shift. If your GC is equipped for constant flow use this instead of constant pressure and see if this doesn't help with the baseline shift. Please let us know the results.
Regards,
Mark