Getting back to hydrogen peroxide, first let me illustrate what a reasonable ATL peroxide oxidized Atlas, assuming the decision to downsize the goal in the light of lighter H-bomb designs had been preempted.
Here are some statistics of OTL Atlas--this collated from Encyclopedia Astronautica articles on the OTL engines and
this site's description of the Mercury launcher.
Gross all up mass (no payload)
115,700 kg
Propellant load
110,300
Dry mass inferred
5400
Various anecdotal sources claim the tank portion of this was as little as 2 metric tonnes, due to being finely milled. The tank was stainless steel instead of aluminum; this is great for switching H2O2 because stainless steels are among the best, least reactive containers to put HTHP in.
Engines--2 LR89-5-- for each, vacuum thrust 822.5 kN, at vacuum ISP of 290 sec, burning for 131 sec, engine dry mass 720 kg
----------1 LR105-5, vac thrust 386.4 kN, vac ISP 316 sec burns to propellant exhaustion at 310 seconds, dry mass 460 kg
Subtracting the engine masses leaves 3500 kg left over, a lot more than 2 tonnes but after all we need mass for other auxiliary stuff, don't we? That stuff will tend to scale with the engines, being infrastructure such as thrust structures; assume we add 1500 kg to the 1900 kg of the combined engine dry weights, so we multiply each engine kg by 1.79.
The Silverbird Launch Calculator wants stage mass breakdowns, the vacuum thrust and ISP of the engines, and then if your rocket design is close enough to the data base it is based on (it is basically a glorified n-dimensional slide rule, or lookup table) it will estimate the performance to a specified orbit or escape trajectory. There is a mode whereby one can specify strap-on boosters burning in parallel with the first stage. The Atlas was essentially a single stage rocket with two liquid boosters that happened to draw their fuel from the single central tank that fed the sustainer as well--but we can define the dry mass in terms of the engine masses multiplied by that approximate 1.8 factor. One also needs to input the propellant mass each booster engine will burn, which based on vacuum thrust and ISP and a burn of 131 seconds I estimate was 37,800 kg per engine. This leaves 34,500 for the core sustainer--good for 276.8 seconds burn, not 310, but if we figure the engine throttled back later in the burn this would square.
Assigning 2850 dry mass to the core stage and thus 1275 to each of the two side booster engine sets, which mass is dropped on booster stage burnout (here actually set by a timer as there is still more propellant for them to burn when they drop) the calculator gives 2300 kg to a 200 km orbit at the inclination of Cape Canaveral's latitude of 28.5. This is considerably better than the given payload of 1.4 tonnes and so I figure that we need to be more stringent, crediting less mass to the drop-engines and more to the core. The least we can credit the drop boosters with is the dry engine weight--figuring at 850 kg each (a little more) that leaves the core massing 3700. On that assumption we still get 1600 kg to LEO. It is a low orbit to be sure!
Now then, having arrived at a model for the Atlas that seems reasonably close to the reality in Silverbird, what happens if we were to change the propellant mix over to hydrogen peroxide instead of oxygen, and scale it up on the assumption that the booster engines were meant to be five essentially identical engine cores, one of which (the central sustainer) has a high nozzle expansion ratio of 25 for better efficiency in vacuum, while the 4 boosters have a lower ratio of 8 that serves better at sea level? These were the ratios in the LR 105 and 89 respectively.
In addition to Silverbird, I have fairly recently been apprised of a rocket engine calculator. I can't afford to buy the full features version which does all kinds of nice things like factor in various types of fuel pumps, but I can compare performances of various propellant mixes, at different chamber pressures and nozzle expansion ratios. Using this tool I find that switching over to hydrogen peroxide, 100 percent pure stuff, at a chamber pressure of 41 bar (same as the LR89) one can expect essentially the same thrust. The exhaust comes out at a lower speed but with more mass flow. The ratio of oxidant by mass to a kilogram of fuel consumed is 6.2 instead of 2.25 for the OTL ker-lox engines. This combined with the fact that pure peroxide has a density of 1.45 versus LOX at 1.15 means that overall, the complete mix is 1.3 times denser. This high density offsets the lower Isp and so performance is remarkably similar. Furthermore higher density means that although the total mass flow is higher, the volume that needs to be pumped to chamber pressure is lower. Looking at pure peroxide catalyzed as a monopropellant, 41 bar pressurized peroxide could reach theoretical exhaust speeds of 1800 m/sec or more, which I believe is in the same ballpark as the gas generator mix in the more advanced later H-1 and F-1 engines, which I believe would reach well under 2500 m/sec if run as rockets. The reason the turbopump gas generators in the later 1960s ker-lox engines ran so cool is that their mixes were far from optimal, to cool down the gases to the point that turbines could run on them to pump the fuel. But already back during WWII, von Braun's engines including those used on V-2s were pumped with catalyzed peroxide instead, which would run even cooler, and in the USA designs as late as the Redstone medium range missile, used to launch suborbital Mercury missions, continued to use this method of pumping the main propellants. Meanwhile in the USSR, the R-7 engines also ran this way, and to this day the Soyuz rocket engines that today are the only method we have operational to put American astronauts at the ISS still use it! Clearly then continuing to rely on pure hydrogen peroxide monopropellant to run the pumps of this first-generation ICBM is quite reasonable. On paper higher energy gas generators would be more efficient, using less propellant to generate the power, but the scale of pumping power is low enough that the difference won't matter much, especially since if we use peroxide in the main combustion chamber as well, we have simplified three fluids down to just two, and since the power requirements are somewhat lower.
Most importantly, the peroxide fed chambers will burn considerably cooler than ker-lox, by hundreds of degrees out of 3000 and more--for the LR89, the chamber temperature is nearly 3500 K, versus 2950 for a peroxide version.
It is for this reason mainly that I believe that had a major effort been made to develop kerosene-peroxide engines early in the USA, for high priority missile projects, that it may have been possible to deliver engines of a given combination of performance and reliability several years earlier than for an equivalent ker-lox engine.
In addition to that, for the specified purpose of making an ICBM, it may have been possible to treat peroxide, along with unproblematic kerosene fuel, as essentially a "storable" propellant. I do believe that to keep the peroxide stable, it would have been important to try to keep it near its freezing point, which is below 0 Celsius. But this could be done by the simple expedient of bubbling a chilled neutral gas, such as nitrogen, through the tanks, valving it off on top and rechilling it to bubble back in on the bottom. At a sufficient rate, such gas cooling could probably keep it cold enough even on a hot desert day in direct sunlight. However, the conventional OTL Atlas requiring liquid oxygen could only be kept filled with the stuff quite briefly before it had to be allowed to be valved off to keep the tanks from bursting. Operationally this meant Atlases could not be kept fueled up, and had to be filled with oxygen immediately before a planned launch, which meant an enemy would have a window of a half hour or even many hours before the missiles could be launched. If on the other hand it proved possible for the ATL Atlas-peroxide missiels to keep a load of peroxide cool enough in normal day-night cycles for long periods of time, the missiles could be loaded up and kept on alert for weeks, with only a fraction of the force being stood down for maintenance every month or so. Thus, assuming the peroxide Atlases were functional at all, they would form a much more robust deterrent than the OTL operational one.
Now then I assume as I said that the plan was to develop a single engine core, and by attaching a longer nozzle to the central sustainer, optimize it for vacuum thrust while the otherwise identical boosters get a shorter nozzle for better sea level thrust. However these baseline engine cores would have to be bigger than the kerosene designs to achieve the same initial acceleration off the pad, since the bulk of the mass at launch, the propellant, would mass 30 percent more.
The Big Peroxide Atlas then would have these statistics versus the OTL Atlas D:
Gross all up mass (no payload)
313,740 kg
Propellant load
301,600 = 4*37,800*1.3+ 80,800*1.3; that is 49,150 per booster engine plus 105,000 for the sustainer
Dry mass inferred by extrapolating a core structural mass of 2900 by the ratio of volumes (assumes skin gauge thickness increases in proportion to length, a good assumption for pressure structures like balloons, or the Atlas tank) to 6116.4 kg, plus a heavier engine and associated auxiliary masses--was 800 kg but now we have 5 standard engine cores each 720 being kicked up by 1.3 to 940 kg; assume core engine adds 1600 kg to 6120 for 7720 all up; each booster engine plus associated mass goes up to 1100
Total dry mass is thus 12,140 kg
Engines--4 LR89-Peroxide-SL-- for each, vacuum thrust 1070 kN, at vacuum ISP of 270 sec, burning for 131 sec, engine dry mass 940 kg
----------1 LR89-Peroxide-VAC, vac thrust 1160 kN, vac ISP 294 sec burns to propellant exhaustion at 280 seconds, dry mass 960 kg
Silverbird calculator gives a payload of just over 4 tonnes with these inputs, to a 200 km orbit due east of Cape Canaveral. Note someone might worry that with the propellant load being 1.3 times as dense, it is necessary to beef up the tank balloon structure by that factor, above and beyond having expanded it for higher volume. Now if we believe that the tank itself on an Atlas massed only 2 tonnes, here we have expanded it by 2.3 volume ratio to 4.6, and another 30 percent would add 1375 kg, but our prior accounting lumped in a total of 6120 kg of tank plus other stuff, some of which might also need to be beefed up for denser peroxide and others of which might not. Worst case scenario we must raise that whole mass by 30 percent for an additional 1842. Adding all that to the core dry mass, raising total core mass to 9560, we still will have 2172 tonnes of payload. In the better case we can have a payload 467 kg heavier or 2.64 tonnes, splitting the difference we have 2.4 tonnes. Even in the worst case we come out well ahead of the OTL launcher--as indeed we should, since that was after all the downsized version and this is meant to be the big heavy version designed to lift older style big H bombs! Versus the Silverbird result for the OTL model, we only come out 35 percent ahead which is somewhat depressing; versus the reported 1.4 tonne payload that implies a peroxide payload of 1.9, for a tonnage of rocket nearly 3 times as great as the OTL Atlas!
However, having put in that 30 percent safety factor we can be pretty sure this baby is robust enough to handle dense peroxide, and it is still at least half a tonne gain over the OTL Atlas D.
And this is the rocket we might believe could be operational 5 years ahead of the OTL dates, provided everything goes right for it.
I looked briefly at what it would take to lower the temperature of the chamber down below 3000 K for a ker-lox mix; even slashing the pressure down to an insanely low 2 bars still leaves it considerably hotter, while performance at sea level goes negative! And vacuum performance takes a major hit too. If people who swallowed a compromise with peroxide wish to now forge ahead for an improved performance ker-lox version like OLT, they must accept that the engine burns hotter and requires more expensive materials and methods. Assuming that some years after the peroxide version becomes operational, ker-lox engines similar in design to the OTL versions become available, and are upsized versus OTL to match the thrusts of these rockets, we can expect the following changes:
Propellant mass drops down to a total of 232 tonnes; each booster having 37,800 and the sustainer, 80,800
Dry mass can safely shed 1500 kg or 1.5 tonne since the ker-lox is less dense; All up mass is therefore 244,460
Isp for the two engine versions rises to 290 and 315
Thrusts are the same
The payload would nearly double, to 3800 kg!
Clearly then, for purposes of launching objects into orbit, the way forward is to abandon peroxide and move on to ker-lox or even more advanced methods.
However for purposes of an ICBM, despite the big advance in throw weight, this is a step backwards, for operational, on-alert ICBMs cannot rely on LOX. As a missile system the ATL giant Atlas is pretty good; it cannot be deployed ready to launch in hardened silos unfortunately, but assuming the bases it is on have some defenses against guys holding rifles a few miles away and that launch commands come well before enemy first strikes, most can be kept at the ready for days, weeks, or longer and ready to launch within minutes or even seconds of getting an authorized command to do so.
A space launcher using oxygen instead of peroxide would be a pretty different article, with different internal bulkheads separating the propellants, different infrastructure, different engines...the deployed ICBMs cannot be converted into the more powerful and efficient oxygen burning versions.
Furthermore, I looked at what might happen if a second stage were added to the Atlas versions. Here, as with the Soviet Union not being able to catch up for years, is another long delay versus OTL to expect--namely that OTL such upper stages as the Agena were developed pretty much at the same time as the basic launchers became available, but here there is no reason to think anyone in the industry is prepared to deploy something so advanced. It is possible that in the ATL, foreseeing success soon with the basic rocket, contracts are undertaken earlier, but as with Soviet efforts, the general state of the art is backward; we might reasonably expect something almost as good as say Agena A to be available two or three years earlier than OTL, but there is still going to be a delay.
Anyway assuming something like Agena A might be available someday, on a peroxide Big Atlas, it raises payloads to around 4 tonnes--but on a LOX Big Atlas, it only raises them further to 4.6 tonnes!
This suggests to me that instead of pushing hard to develop a ker-lox Atlas, it might be smarter just to develop upper stages for the peroxide version.
And then I thought--why wait for Agena, anyway? Kerosene peroxide is inferior to hypergolic but not by a lot, and it is comparably storable, and here we have a crash program to get a powerful kerosene-peroxide engine operational. Why not consider such an engine with the lower parameters of Agena instead?
Indeed, a realistic stage in the ballpark of Agena but using kerosene-peroxide tech might be defined as massing 500 kg dry, holding 3 tonnes of propellant, having an engine capable of 70 kN thrust in vacuum; such an engine raises the Sliverbird calculation (which might be optimistic by a quarter tonne or so) to 4.7 tonnes. And then I wondered, why not a third stage? Rather, turn this thing into a third stage, and have a second stage of intermediate mass and thrust--say, 1.6 tonnes dry, 30 tonnes of propellant, thrust of 250 kN, Isp of 292? With that, Silverbird offers something like 7.6 tonnes in orbit--very near the mass of a Soyuz spacecraft!
All of this is on the base of a standard missile spec Big Atlas using peroxide; I doubt the performance with an LOX variation would be better.
These upper stages are not likely to be ready nearly as early as 1955; the most ambitious stack might not be available until 1960 or so.
I also wondered, are we overloading the basic Atlas structure with an upper stack amounting to over 40 tonnes with payload? Well, I suppose it would overload an OTL ker-lox Small Atlas. But not only have we expanded the balloon tank shell in proportion to volume, we have also beefed it up further to account for the propellant higher density, and with all that we have a shell some 3-4 times more massive than the OTL Atlas shell. Whereas, assuming we scaled in proportion to volume, the linear dimensions are just 28 percent greater, and that means that the cross-section of load bearing steel in the shell has been raised in area by more than a factor of two. OTL Atlas-Centaurs have been launched with upper stacks amounting to more than 20 tonnes mass, so with double capacity we can contemplate a 40+ tonne upper stack quite calmly.
Another metric other than raw mass is the ratio of thrust force borne by the shell to strength in the worst case, which is just before sustainer burnout. An OTL Atlas with a 20 tonne upper stack has its sustainer producing nearly 40 tonnes-weight of force on that upper stack, meaning the shell of the first stage is conveying that. The ATL Big Peroxide Atlas producing nearly 120 tonnes-weight force on our purported 45 tonne upper stack, or about three times as much. It has 4 times the mass of steel in a cylinder that is, ignoring the caps, 1.28 times as long and at 1.28 times the diameter, so the area is 1.64 as great, thus the thickness of the skin is 2.44 times as much, and the circumference is 1.28 times greater so the load-bearing area is 3.125 as great. Thus the skin, duly stabilized by pressure against buckling, ought to bear a tripled thrust quite as well.
Also note that with its thicker skin, although the diameter is raised which raises the tension force the skin must retain at a given pressure, it is thicker in proportion; the internal relative pressure can be nearly double whatever it was with OTL Atlas and remain in proportion.
That's a good thing when most of the volume is filled with high test peroxide; if a certain amount of outgassing of oxygen occurs the hull has margin before the rising pressure becomes an emergency, buying time for various countermeasures--including of course declaring a launch emergency and evacuating the launch area in anticipation of a pad blowup! Opening valves to relieve the pressure ought to be an option of course, unless the launch is far from ready and the crisis is being caused by contamination in the peroxide tank.
OK, that's my ATL Big Peroxide Atlas proposal. Maybe we can discuss if it is feasible at all, and if so, would opting for this have allowed Convair to accelerate development, and what sort of PODs might plausibly have sent the Air Force and its contractor down this path?
) the idea was to still have them built and fired from relatively the same area only the factory would be built into a mountain and the 'launchers' a distance away connected by underground railroads.