If you have done whole cell or intracellular patch before, loose patch is very easy. You just have to touch the cell membrane, push the electrode tip a bit more so you get a 'dimple' in the cell, and then apply suction - till you get the membrane in the electrode. Of course, your electrode tip for loose patch has to be a bit wider than in other methods so that it doesnt poke the cell. Also, the suction has to applied very slowly and gradually.. You could also look at this video http://www.jove.com/video/50312/recording-electrical-activity-from-identified-neurons-intact-brain
Barbour and Isope: Combining loose cell-attached stimulation and recording. Journal of Neuroscience Methods, Volume 103, Issue 2, 30 November 2000, Pages 199–208 ( http://www.sciencedirect.com/science/article/pii/S0165027000003186 )
In practical terms, it's just like normal patching except you just need to obtain a partial seal rather than a tight seal (e.g. using not a perfectly clean electrode). If you want to stimulate action potentials from single cells and see the evoked spikes, you might benefit from an amplifier specifically designed for extracellular loose cell attached recordings (e.g. ELC-03XS).
As a technique, loose patch is rather similar to cell-attached patch clamping. Nevertheless, there is a big difference: you need "Megaseal" (tens or hundred MOhm of seal resistance) instead of "Gigaseal" (GigaOhm of seal resistance). Gigaseal requires very clean glass surface at the rim of the electrode, therefore I make the rim a bit dirty before attaching the cell membrane to achieve Megaseal. It means, I release the positive pressure when I am near to the cell, then move the electrode a bit in the tissue using the micromanipulator (to make the rim dirty), then I apply the positive pressure again and touch the membrane. After touching it, apply negative pressure to achieve Megaseal. And that is all....