So the title says it all really...

In swims mis-spent youth she hammered the N2O, and it doesn't seem to be available these days unless you know someone in the medical profession or similar. Looking at the synth on wiki it doesn't seem particularly difficult..

"Nitrous oxide is most commonly prepared by careful heating of ammonium nitrate, which decomposes into nitrous oxide and water vapor.[12] The addition of various phosphates favors formation of a purer gas at slightly lower temperatures."

NH4NO3 (s) ? 2 H2O (g) + N2O (g)

"This reaction occurs between 170 and 240 °C, temperatures where ammonium nitrate is a moderately sensitive explosive and a very powerful oxidizer. Above 240 °C the exothermic reaction may accelerate to the point of detonation, so the mixture must be cooled to avoid such a disaster."

So...maybe not for the faint hearted? Assuming the reaction occurs between 170 - 240 and this may not be incorporating the "various" phosphates, it looks pretty easy?

Has anyone experimented here? I guess you only get one set of lungs and swim isn't too keen to be a guinea pig in this instance. There's a good graphical display here: http://en.wikipedia.org/wiki/File:Nitrous_oxide_production.png

Several interesting patents pertaining to the topic are: 3360336, 2425528 & 4154806..

I would  look into the urea / nitric acid reaction if they are available to you. It works quite well, there is a write up on Rhodium archive. I gave this a go with a bollon on the end of the apparatus. I heated the HNO3 too much and it started boiling due to the exotermic reaction. I popped of the balloon and the hno3 started boiling out of the flask due to the drop in pressure! Oops!

It seemed to be fine once the generated gas was bubbled through a dilute NaOH solution, it had a nice sweet smell and definatly had the expected physiological effects.

http://127.0.0.1/rhodium/Rhodium/chemistry/nitrous.html

Yea I have read and seen instructions before for a sodium hydroxide column IIRC ~ metre in length, through which SO2-laden NO2 bubbles are bubbled, producing a purified stream through an outlet at the top

I always thought it might be possible to build some sort of scrubbing system to clean up racing grade N2O.  Seems like they use SO2 or some other nasty to render it unfit for consumption, but adapting surplus gas mask canisters might be the ticket to cleaning it. 

If you were to make your own, you would need to do some gas washing/scrubbing anyways to get rid of the other nasties that get produced along with the goods. 

SO2 and H2S are both reported in use.

Holly claims the contaminant at NOS-affiliated refill stations (branded as Ny-trous Plus) is SO2 @ 100ppm.

I doubt that any reputable. manufacturer is contaminating automotive N2O with H2S. I suspect a gas misidentification has led to the propagation of rumors. Still, H2S is nasty, so I would assume its presence until proven otherwise.

Recirculation through H2O2 would clean up both the SO2 and suspected H2S.

Dry ice cooled acetone should be able to condense both SO2 and H2S.

The biggest problem with purification of automotive grade N2O is the inconvenience. Pure N2O might bubble out at the end of a few columns, but how will you get it in the balloons? Using a sealed scrubber to use the tank's pressure to fill the balloon has to be done in the lab, which means hitting N2O alone in a lab instead of with others on something comfortable.

Whippits are as easy to source as NaCl.

Which brings up the conundrum: if you're considering the auto grade, for reasons of cost and bulk, then you are likely with others and need a big portable tank instead of a drug lab apparatus, but if you're alone, the price of whippits is more manageable.

N2O can be made by mixing Hydroxylamine HCl and sodium nitrite. My favorite method.