As posted by GI3 to techdiver

GAS MIXING LESS THE BS:

 Let's assume all bottles are correctly marked - that means MOD only,
horizontally, in three inch high letters , on either side of the upper
bottle in the orientation of the tank. All tanks turned off, all regs
parked. To deploy, we locate the bottle by MOD, we remove the reg and put it
in our mouth, we then relocate the bottle by MOD, and turn if on. If we can
breathe, we are breathing the right gas.

Now, mixing. First put two pieces of tape or one GUE split tape on the empty
bottle. Hook up to either gas, but I do helium first since I want my oxygen
addition to be more in the middle of the operating range of the gauge, but
it can be done either way. Figure the correct amount of helium for your mix
considering coefficient of expansion and heat expansion. Helium will need
about 17% overfill to get the % you seek. Let's say we want 50% and we are
filling to 3000. Fill the tank first to 1750 to get to half, but then add
another 10% or so for heat expansion, so go to a total of 1900 roughly. It
should cool back to 1750 or so. No sense getting real anal here, the heat
expansion is simply equal to the ratio of absolute temperature change, using
the Kelvin scale - just guess at it. Turn the bottle off, but do not remove
the whip until you have written down the contents of the tank and the date.
Now remove the whip .

Now, add the oxygen. Keep in mind that it will be heated and expanded, but
not too badly. Go a little over your intended amount but not too much.
Calculate this independently of the helium, and add it without regard for
the "pressure" of the helium. If you added the oxygen fist, keep in mind
that it will have expanded as well from the heat of the helium filling and
be giving you a higher overall helium pressure reading than you really have,
so add a touch there if that is the case over and above the other two
reasons. Do not remove the whip until you have written down  the new gas
added and the date. Analyze it if you please, to solve for the helium later
to satisfy yourself.

Now add the air, unless that was heliox and we are done. Immediately analyze
the gas and write the analysis on the other piece of tape or the split tape,
write the date on it as well,  and take the original tape off the tank and
cover the valve with it to indicate that the tank is full. Do not move the
tank until this is done. Obviously , the analysis should jive with the MOD.
To dive the tank, you can remove the tape so as not to litter ( and
obviously the only thing we go by in water by is MOD ) , but if you do not
use the tank, retape it and the valve, rewrite the contents and date, and
transport that way, including partially full tanks that you intend to reuse
( I reuse my deco tanks for two dives usually, so I may write some thigh
like "2000 psi 50%  10-18-00") . If you use it and are not going to reuse
it,  it is now assumed that the tank has some amount of that gas in it, but
can not be used again unless reanalyzed, so is not retagged. It can not be
stored full without a tape, and it can not be transported full without a
tape. Doubles can not be stored untagged if they have gas in them, and if
that mistake is made, dump the partial gas and remix rather than adding to a
mix. A lot of people fool themselves this way on the helium %.