Ok, so I'm in the process of designing a sort of miniature amp that uses just 12ax7's. Inspiration sort of comes from here: http://www.allaboutcircuits.com/vol_6/chpt_5/19.html, except

I'm going to try just 1 valve to begin with (so 2 triodes), have one of the triode's as a voltage amplifier (so as a standard pre-amp valve), and the second triode in a similar arrangement to the above link, except a proper O/T, and it should output about 1W (which I'm going to dummy load and tap off for a line output, the idea being this baby will allow me to achieve power stage saturation.

So, ignoring things such as gain, the fact that 12ax7's aren't really made to drive a load (yes I've seen Doug Hammond's Firefly), my main question is is it a good idea to use different B+'s on each triode? The idea here being a lower B+ on the pre-amp triode so it will saturate, and then the higher B+ on the power stage triode.

Something tells me that there could be some issue between the difference in voltages on each plate, but having said that a cascode arrangement would result in something similar and that's OK....

Anyways, I realise this might be a bit too technical even for this forum, but if anyone wants to discuss my overall plan I would be more than happy to, and would welcome any advice if anyone has made anything similar.
Different B+ on different triodes is fine. It's a great way to voice the amp. You only run into problems if you try and control the B+ with the plate load resistor. 12AX7's tend to need a 100K plate load resistor so leave those alone. Control your voltage by changing the value of the resistors that decouple the capacitors in your power supply.
Not taking any online orders.
Theoretically the different voltages on the different triode sections of the tube should work. In practice there could be some problems, however I believe that it should still work fine for what you want to try. Just make sure its properly biased etc.

The schematic you have provided seems to just be something like a power amp made out of 2 12ax7s. But you are looking to have one gain stage from a triode then an output stage if i understand you correct?
^Yeah, just use a single triode for each stage instead of both.

12ax7's are relatively cheap though, so the worst that can happen is that I blow one.... I think that the final design will end up having more valves, because I don't think a single triode will result in enough for my liking, but in the event something so simple can sound great i figured it's a good idea to give it a try...
To start with make sure to change the grid resistor to 1Mohm to use a guitar, this circuit is for amplifying radio's etc and 100k is too low. There's no need for an input cap for guitar use but a grid stopper of 10k plus is a good idea as it will stop you picking up radio stations.
No problem with different volatges on the plates. Try 300 on the "power" triode and 270 on the preamp triode. You'll need a suitable resistor and decoupling cap for this.

You do realise that this circuit is not complete, right? There is no bias circuit here.

You'll need to set that up otherwise it'll sound really nasty. Try a 1.5k on the first stage, the second stage value will depend on the OT you select.
You'll also need a 100k plate resistor on the first stage plus a decoupling cap before the "power" stage.

Could I suggest building it with two valves, one 12AX7 pre amp and then use a 12AT7 like the power amp shown in you link. I think that'll give you much better results.