Argon tanks:

From a safety stand point that would be a very poor idea as the valve and flowmeter would be exposed to damage very easy.

You would need to see a 2200psi cylinder release it's gas when the valve has been broken off to fully understand this. It's very scary to say the least.

With a Co2/Ar mix the Co2 will be at saturated vapor at times thus spitting liquid droplets into the supply line.

In ceritan temperatures, even Oxygen has a very bad problem for freezing the valve when layed horizontal then when stood up the valve will break off, shatter. They discovered this problem while building the Eisenhower tunnel in Colorado where it killed a couple guys.

Wouldn't a verticle bottle rack on the back of the machine be better? They work extreamly well.

T_Bone

Pure CO2 is a problem because it would be in a liquid state until the bottle is near empty and the regulator wants to draw the gas off the top, not the liquid from the bottom.

C25 will be the same as pure argon PROVIDED the ambient temperature is above about 40 degrees (F). At temperatures below this, with a completely full tank (2200 PSI), the CO2 can begin to condense out into a liquid. Your regulator may not appreciate a dose of liquid CO2 running into it, which is a slight possibility if the tank is mounted horizontally and the cart is on a slope.

Provided the tank pressure is 2000 PSI or below, the CO2 component will be a pure gas down to 22 degrees (F). The critical temperature gets colder and colder as the tank pressure drops.