I agree upon the first part. Its not possible to identify an unknown substance via a melting point. But that is not the question.
A melting point test for pure substances, such as Cocaine.HCl can be, is absolutely reasonable if you have a proper melting point apparatus, many of which work highly reliable. I do melting point tests on incoming materials as a first measure, only then tlc if I notice something is off on the meltingnor mp, which you develop a feeling for over time.
Back to the original question:
You are correct. If you bought a proper melting point apparatus it is as easy as taking up a very small amount of material into a capillary tube that is sealed on the other side. Don't use too much pressure on the first uptake, knock it several times for the material to get to the bottom, towards the sealed side of the capillary. It is a small diameter, so you want it to not get stuck in the opening as its hard to push it down, even with a small needle.
Once about 5-10mm substance is placed into the capillary, you place it into the apparatus, set the temp to around 2-3°C below expected melting point (which is 198°C for cocaine), let the apparatus heat up fastly to there, press a button once the close-to-mp-temp has been reached (to continue heating much slower than before) and watch the melting illuminated through a magnifying glass which is part of the apparatus.
Then, the link given by G.Patton is useful when it comes to the observable phenomena. I dont like the emphasis on the bad apparatuses.
I suggest the büchi and pharmacy standard apparatuses that are electrically heated. Definitely stay away from doing a mp test with home made equipment, you waste your time.
There's old büchi ones by Dr. Tottoli, which rely on heating oil, too, which I advise against. Also, 'Kofler Bank' is a nice tool for a fast test but not for precise, reproducible, recordable results.
You do not need any chemicals of any sort for this test, it is purely physical. Only capillary tubes of the size suited for your apparatus are replaceables.
It is, to be honest, a sad situation when the knowledgeable suggest there is no pure substances out there and basic analytical skill not applicable. A sad insight into the quality of material available on the market. I just did this exact test a month ago. My apparatus gave me a result of 197.8°C. Then again, this was an analytical standard of a chemical supply house. It stayed good for decades, which is when I bought it. Hydrochlorides seldomly go bad to a significant amount if kept under proper conditions.
