As the cyanide synth on Rhodium's page sounds slightly hazardous, and the other one is not too well scalable, SWIM has found you an alternative way of preparing KCN. This is based on SWIM's own idea, no refs this time; of course only for informational purposes. Cyanides are very poisonous, anyone attempting this illicit reaction should wear gloves and perform it in a fume hood (or outside), even if no HCN is needed to be generated.
K4[Fe(CN)6]* 2H2O (121.2 g, 0.3 mol) are dissolved in dH2O, FeCl3 soln (35.5 g FeCl3, 0.2 mol) is added in small portions with stirring. Doing this while hot will speed up filtration considerably.
(FeCl3 can be prepared the following way: iron turnings and HCl, the Cl2 bubbled through it after dissolution is complete, or alternately, calculated amt of HCl and H2O2 can be added to the green FeCl2 solution. If this procedure is used, make sure the soln contains no excess acid or HCN gas will evolve.)
The precipitated Blue of Berlin is washed with water, some of it is colloidal and will come through the filter paper. The solids are added to a solution of 66.3 g of KOH (1.2 mol) pellets in 300 mL of water (cautiously, with good stirring), again with a higher temperature particle size will be larger. The resulting brown mess is vacuum filtered (Gravity filtration will be unsuccessful), the filtrate evaporated to dryness and the KCN recrystallized from dH2O then dried over CaCl2, do NOT use H2SO4 or P2O5. Yield, ~110 g.
Blue of Berlin can also be prepared from K3[Fe(CN)6] and FeSO4.
3 K4[Fe(CN)6] + 2 Fe3+ -> Fe2[Fe(CN)6]3 + 3 K+
Fe2[Fe(CN)6]3 + 12 OH- -> 2 Fe(OH)3 + 3 Fe(OH)2 + 18 CN-
-Wraith