DMoz liked the idea of a pivoted launch rig...
...and here's Rob's variation on the theme:
He explains: "Rather than using a counter weight, it just uses the width of the balloon to get your separation. It has the neat property that it throws away your assumption that the balloon will float upright.
"It does require a top attachment to the balloon. I'm not sure if you can glue an attachment to the latex, but if not, a lightweight net around the balloon would do the job of providing a top attachment point.
"You'd have to experiment a bit with the geometry once you knew the weight of the payload and craft, but by adjusting lengths A, B, C, D and the position of attachment point E, you can achieve pretty much any launch angle with minimal additional weight compared to attaching LOHAN to the payload directly."
Gary Keall, meanwhile, suggests putting the launch platform "on the end of a long piece of (say kite) string to clear the balloon (I'll claim the original suggestion of that in the comments) – I'd say 50 if not 100m if the balloon expands to 10s of m across".
He continues: "My launch platform would simply be a long lightweight carbon fibre pole (fishing pole?) – about 3-4m I suggest. This is suspended from each end by a string yoke in an asymmetric 'Y' shape to hold the pole at the required launch angle, as shown. I don't think the launch needs to be that near vertical – the ~30° angle show still has plenty of vertical component, and any more vertical and the initial roll orientation of the aircraft will be less defined. The exact angle of dangle depends on where the centre of gravity of the assembly is – I've drawn it assuming most of the mass is in the aircraft at the bottom end.
"As I've drawn it, I envisaged the aircraft attached to a lightweight launch shuttle which slides up the pole on a pair of ring guides or something similar. The aircraft is released from the shuttle by a lever trigger hitting a trip 'noggin' at the end of the pole. This yanks a piece of string which pulls release pins or something like that."
While Gary's design features swing damping vanes, as you can see, Angus Wood suggested using the wind itself to control the relative positions of balloon and payload. He says: "The basic idea is to take advantage of the high wind conditions to create a situation where one is (almost) guaranteed to be able to know where the balloon is in relation to the launcher."
He continues: "Given that the balloon will have the highest surface area it will catch the most wind and so will always be 'ahead' of the payload and LOHAN. The next step is to orient the rotation of payload in respect to the balloon. This is done by attaching a rudder-like flap of cloth to some aluminium spars which are in turn rigidly attached to the launcher.
"The net effect should be to always aim the LOHAN launcher in to the wind (which is beneficial) and guaranteeing that it will always be pointed at maximal angle away from the helium balloon."
if your going to go with the triple orb launcher it should be named "Eccentrica Galumbits"
That is all
If the crowd was predominantly made up of Wile E Coyote and his family, then yes.
Too much is never enough...
I can't help but agree with lawndart (hangie pilot, perhaps?) in believing any idea of aerodynamic steering is pointless at that sort of altitude. Since the aim is to get the rocket aimed generally upwards as soon as things start happening in the balloon bursting department, it seems to me you have only two real options:
1) have it pointing in the right direction to start with, or
2) vectored thrust.
Good luck writing the flight control software for (2) - starting with an unknown position, attitude, and vector, and not very long to sort it all out before you lose all the thin-air friction advantage that the balloon's altitude gives you.
I'm not happy with triplet balloons, but even less happy with long carefully balanced struts, rails, and pointers: no-one seems to have remembered the good Doctor Newton and his 'equal and opposite reaction'. A passing fad, no doubt, but I can't help feeling that rocket going rapidly forwards is going to result in a certain amount of balloon going backwards - at the very least, it's going to tip in the reaction and that's going to cause the end of a launch ramp to tilt down. And that's ignoring the issue of a guide/release mechanism that is both light and able to force a direction change of better than sixty degrees without sticking.
Which leaves me with the launching upright approach. This has the advantage that it's gravity stabilised, assuming the mass of the payload is suspended somehow below the balloon(s).
What I would propose is not a three- but a six-balloon system. The balloons would be constrained in a light mesh into an annular shape - a poor-man's doughnut balloon.
This has the advantage that the balloons will automatically assume a hexagonal shape - with a space in the middle the size of the balloons, very suitable for a lightweight horizontal platform to be used for a vertical launch. It also has the advantage that a single balloon bursting doesn't shift the balance too much; the platform will still be vaguely horizontal and the hole won't fill.
BOTE calculations indicate that a 10 metre diameter balloon has a volume of about 105 cubic metres; a four metre balloon just over one sixth of that - very convenient. With six four metre balloons you have a four meter platform from which to launch vertically.
The question becomes then one of *when* to launch: ideally just before the first balloon pops. Presumably there is some sort of specification as to differential pressure for the balloons at bursting point; perhaps some sort of pressure sensor, and trigger just before the expected bursting point? Or, if the rocket motor can be ignited quickly enough, wait for the burst (a gyro will tell you you're tipping over) and then go.
You might want to arrange some automated bursting after launch; the balloon will of course rise faster once the load of the rocket is removed and might even catch up with it in the short term.
p.s. No hydrogen involved!