Hi Kai, please sorry me for not replying your message. I have been in a "rush time". I haven´t tested your ciruit yet, but I have some doubts about it and I was wondering if you could answer them:

1) What is the function of the third 10k resistor next to the 10nF capacitor near the sensor. Y understand that the others 10k down it are used to create a virtual earth to 4,5V
2) I agree with you that the comparator is not suitable. I couldn´t find the tlc393. Instead, I bought the tlc374. I think it may be the same according to its datasheet (including the LinCMOS technology)
3) What is the function of the pair of 4,7k resistors at each input of the comapartor. Are their use to avoid oscilations on the tlc393?
4) Could you explain why you said "The threshold voltage generation with the diode at TLC393 is tolerant against supply voltage changes. " I can´t see know the diode and the 1M resistor work to do that
5) I have calculated a total gain of 1055,6. Each stage´s gain is 5,7. But why the first one has different resistor values than the others if all of them have the same gain. Has it got something to see with the offset error at the output of the first opamp? Or perhaps something with the S/R at the input of the hgain chain?
6) You have a point there: Yes I am using batteries. But no, I amn´t using a 9V one. Instead am I am using 4 AA batteries and a step-up converter that is working fine. I have two outputs: one of 12V and the other of 15V You didn´t say anything about the LM324 power. I suppose it is 9V. But I am sure that I can use 12V. Am I wrong?

Thank you for your time