Possible new directions
As the legal side catches up with the experimental chemist, it means you have to cast your net further afield in order to "catch" more useful (but as yet uncontrolled) stimulants. I reckon some of the field with really rich pickings are derivatives of phenmetrazine, fencamfamine and pipradrol.
Starting with phenmetrazine
Now phenmetrazine, is quite well known as a drug of abuse, but when people think of phendimetrazine, they tend to think of 2-phenyl-3,4-dimethylmorpholine (one of the methyl groups being attached to the nitrogen atom); this isn't too hot as a stimulant. If the second methyl group is moved to the 6 position (cmpnd 2) though, you end up with a compound that has the same potency as phenmetrazine, but isn't yet controlled. Or alternatively, you can replace the oxygen with a methylene group, to give 2-methyl-3-phenylpiperidine (cmpnd 3). This doesn't release dopamine as phenmetrazine does, but is actually a reuptake inhibitor (similar mechanism to methylphenidate). One other alternative is 2-methyl-3-phenylquinuclidine, which is also a reuptake inhibitor (no dosage info for last two, only that they are active; cmpnd 2 is same dose as phenmetrazine, namely 25mg)
Fencamfamine also lends itself to a lot of fiddling and tweaking of the basic skeleton
Both 2-phenylcyclohexylamine and 2-phenylcyclopentylamine are active CNS stimulants at 20mg, although this is a two fold drop in potency when compared with fencamfamine (10mg), although N-methylation of the amine function might give some increase in activity; cmpnd 4 is active, but alas, no dosage data for that one either. All four compounds are unusual in being both dopamine releasers and inhibitors of its reuptake (and fencamfamine is probably the smoothest stimulant I have ever tried - absolutely no peripheral activity such as increase in heart rate)
Pipradrol derivatives are the one where most hope lies.
Shulgin wrote about 2-(diphenylmethyl)piperidine as being a possible future drug of abuse, as it is estimated to have the same CNS stimulant properties as methamphetamine, but to be active at 2.5mg as opposed to 5mg for methamphetamine, and also to have less peripheral activity as well. It's a bit more complicated to synthesise than meth from pseudoephedrine, but once governments have finished trying to stop bathtub meth manufacture, it may become a possibility for some chemists. Pipradrol os active at 2mg, but is not a strong stimulant (prob because of the tertiary alcohol group - remove it to get 2-(diphenylmethyl)piperidine). The pyrrolidine analogue of pipradrol is also a CNS stimulant, similar to pipradrol and active at 5mg (it was offered by one of the research firms that were closed by Web Tryp), and as with pipradrol, almost all of the activity was present in one optical isomer only. Finally to the great hope that is 3-(diphenylmethyl)morpholine. In a paper where a whole series of diphenylmethyl derivatives of heterocyclic rings werte produced , it was the most active of all the compounds teste; should have an active dose in man of around 1mg, and similar in effect to the compound Shulgin favoured: And the best bit, it can be synthed from the synthetic amino acid diphenylalanine (which is fairly available) by reduction to diphenylalaninol, reaction with 2-iodoethanol, then ring closure to the morpholine by removal of a molecule of water by conc sulphuric acid (same way phenmetrazine is synthesised). It would definitly be worth considering for a chemist with a desire to create (and it's not controlled... yet).
One last bit for the cocaine lovers among you.
The 3-alpha-phenyltropane with an acetylated 2-hydroxy group in the axial position has been assessed as being 60x the potency of cocaine, with a fairly similar pharmacological profile (ie abusable, unlike the Win series compounds with an axial 2-methylcarboxyl group). I remember reading that the unesterified compound is less active, but still moreso than cocaine. The OH goup must be axial to the ring though, or its potency becomes a fraction of that of cocaine's.
The last two are very simple compounds that are esters of pseudotropine. The benzoate ester is also known as tropacocaine and has a potency of about 0.1x that of cocaine. replacing benzoic acid with 4-fluorobenzoic acid gives a compound of aprox 0.4x the potency of cocaine; considering how easy it is to synth (and not from monitored reagents - you can actually start from atropine and get to 3-pseudotropyl 4-fluorobenzoate in a few easy steps), I'm really surprised that it hasn't shown up in street cocaine. Maybe the market is so saturated that no one thinks it's economically viable