Adhere your sample with a carbon tape. A long strip of tape that sticks the sample to the holder and runs to the top surface of the sample. It helps a lot specially if you are using low KV.
If the samples surface is stable enough you also can try to arrange an additional lattice pattern with conducting aluminium tape. The aluminium tape itself should be well connected to a carbon tape wich sorounds the sample completly at its edges. The spaces in between the lattice are big enough for SEM images. This approach is very helpful for the analysis of non-conductive plastic materials in combination with low voltage.
Depending on what kind of resolution you need, I've had good luck with environmental SEMs: http://en.wikipedia.org/wiki/Environmental_scanning_electron_microscope
The trade off you pay for not needing to coat is resolution but I've been able to see feature about a micron in size fairly easily.
You can obtain a good quality image via Variable Pressure mode of your SEM. Usually for conductive samples, the sample chamber and the electron tunnel both operate at high vacuum. However, in variable pressure mode the sample chamber will be at a lower pressure to allow some ionized air to make non conductive sample at little or no charging level. By this strategy a research can obtain a reasonable quality image with non-conductive sample without having Au-Pd plasma coating or carbon coating.