I spent some time already and I think I forget about something in my problem.
I have a motor, which single phase is R17 = 2 Ohm. I want to measure a current through it's circuit. I used a shunt with R11 = 5 mOhm. I measure voltages before and after the shunt, so I get the circuit:
R12 and R18 is R = 55 MOhm - those are ADC inputs of my AtMega8. I used two voltage dividers R13 & R10 and R15 & R14 to output the voltage to ADC-friendly level - in this case ~4 V. Measuring the ADC would output me almost the same result, because the voltages are different in the level of 0.01 V.
What should I do to output a nice, measureable results (like 4 V and 2 V is a good difference)?
I created another circuit with LT6106 as pipe suggested:
I think it worked like a charm. How do you think? Maybe I should add anything else?
I had a problem. I can't get fast enough any LT6106 and it's not a cheap one. I wanted to find any on my old boards(PCs, supplies) but I have none.
I have got few desoldered LM358 from ST. I managed to design a simple schematic that amplifies the differential voltage (dropout on R52).
What do you think of it now? Is something wrong or something still lacks?