how do i determine the minimum toxic dose of antibiotics against germs?

First you have to determine the germs strain.
The lab will do this. Its called a "Culture and Sensitivity"
You can't do it yourself in need access to culture media, an incubator and innoculation discs.
That's a difficult point to do. In a given population of bacteria, nearby are always germs dying and new ones coming to go; the minimum lethal dose would be firm to determine since you couldn't tell for sure if it be the antibiotic that killed them or age. What I believe is the prevailing method is to index the LDX where X = some number, usually 50. The LD50 is the point at which partially of the target population die. This number is much more useful because it provides an average. If we be going for LD1, it would be a very low number since plentifully of bacteria are probably questionable and dying already. LD100 is extremely high because the antibiotic doesn't affect everything like, and killing everything take a very dignified dose.
Also, although I strongly encourage experimentation, if you plan on carrying out an experiment approaching this, you need to look up and exercise proper containment measures. We have a extremely big problem currently with germs becoming resistant to some of the best antibiotics we have, and treating microbes with low doses of antibiotic is the foolproof way to breed resistant strains, so don't do it unless you know they won't achieve out.
You certainly don't own to determine the strain of bacteria first. There are several microbiological methods that may be of interest to you: determination of MIC (minimum inhibitory concentration) and determination of MBC (minimum bactericidal concentration). You can find a breakneck review of these methods here: http://www.bmb.leeds.ac.uk/mbiology/ug/u.