# Solar Thermal Rocket

## Recommended Posts

While the sunheat engine converts efficiently sunlight into kinetic energy, the Vasimr obtains electricity first, which is inefficient and costly, and it needs also more power (but less propellant) for the same thrust. Other electric engines tend to be less efficient and strong than the Vasimr.
https://en.wikipedia.org/wiki/Variable_Specific_Impulse_Magnetoplasma_Rocket

The collecting area is a thin mirror for the sunheat engine, but expensive and heavier Solar cells for the Vasimr. 20% light to electricity is a rough figure; better cells cost more per watt. The Vasimr converts well electricity into kinetic energy of the expelled propellant, but for Isp~5000s, it needs 5* more power and 1/4 as much propellant as the sunheat engine with Isp~1267s.

The Vasimr can also use less electricity and more propellant for the same thrust, but it first spends ~100eV to ionize the argon, so tuning near Isp~1200s would waste most power.

Near Earth, the sunheat engine would commonly use an elliptic transfer to save propellant and the Vasimr a spiral one to limit the collecting area. This reduces the Vasimr's advantage on propellant consumption from 1267s/5000s to 2.4t/5.3t in the previous freight transfer from Leo to low Moon orbit - and the heavy Solar panels reduce the payload.

To illustrate the sunheat engine versus the Vasimr:

Not for the same thrust, which would be unfair, but for the same time to bring freight from Leo to low Moon orbit, as described here on Aug 13, 2017. The Solar panels are 3/4 as huge as on the International Space Station, and the 16 Sunheat concentrators for 9 months are at the same scale. Here a picture from ISS; the central modules have D=4m.

Solar panels have become cheaper than on the ISS, hopefully. For smaller areas, ATK claim their UltraFlex weigh 150W/kg, but on the ISS, the long beams holding the panels weigh tens of tons, excluded for a Moon freighter.

Marc Schaefer, aka Enthalpy

• Replies 84
• Created

#### Posted Images

Some second-hand data about ISS' solar panels:
http://news.bbc.co.uk/1/hi/sci/tech/1053725.stm
http://www.nieworld.com/special/hotcold/qtoz/space.htm
1/4 ISS' area cost 600M$and weighed 17t (without the trusses). It improved meanwhile, hopefully. As opposed, the sunheat engine can transport heavier freight to the Moon, matching the capability of heavy launchers supposed to be cheaper per kg. • 23t satellized by Delta IV Heavy or Falcon 9 Full Thrust need 24 D=4.57m engines to reach the Moon in 9 months. • 64t by Falcon 9 Heavy need 16 months plus one eclipse season, and 36 engines. A wider fairing would be useful. • The Space Launch System foresees a D=10m fairing, so 24 D=8.8m engines need 13 months plus one eclipse season. The engines' arrangement is nothing definitive. The landed mass results from O2 and Pmdpta. Instead, a few RL10 to burn hydrogen are very attractive. Marc Schaefer, aka Enthalpy ##### Link to comment ##### Share on other sites The regenerator's coating was to reflect 98% of the light emitted by the envelope of the chamber described on Jun 23, 2013 http://www.scienceforums.net/topic/76627-solar-thermal-rocket/?do=findComment&comment=753432 but tungsten that sublimes from the envelope would deposit on the regenerator, spoiling its reflectivity. And nobody tells me a word. Fortunately, hot tungsten can make a multilayer insulation. Imagine a D=56mm L=150mm eps=0.32 envelope (for Ariane-class D=4.57m concentrator that catches 22kW at 1AU) at 2800K: it would radiate 29kW, but we prefer 1kW transferred to hydrogen in the regenerator. The 1+16+1 layers need only these temperatures: 2800 2759 2716 2672 2624 2574 2521 2465 2404 2338 2266 2187 2098 1995 1874 1724 1519 1161K where the T<sup>4</sup> decrease uniformly - but the emissivity improves at cold. To separate the sheets, a few stamped bumpers limit the contact area. D=2mm contact and 500µm thickness conduct 8W to D=10mm with 50K drop: as little as D=15mm radiate. 3+3 bumpers should suffice, offset between the layers. Welding at the bumpers seems possible. I don't prefer the usual fabrics because thin wires sublime quickly. 2800K sublime 115µm of the tungsten envelope per month (years-long thrust needs rather 2620K) and this halfs at successive insulation layers. Condensation lets the sheets gain thickness as a mean, so they can be thin, especially the colder ones. From some 2000K down, fewer tantalum sheets with lower emissivity can replace tungsten. Marc Schaefer, aka Enthalpy ##### Link to comment ##### Share on other sites The following script goes from low Earth orbit to remote celestial bodies using only the sunheat engine and eccentric orbits. The script starts from a tilted 400km Earth orbit with 18500kg, the capacity of an Atlas V 551. Ariane 5 carries a bit more. • The probe raises its apogee with many pushes around the perigee, long hence using 90% of the sunheat engine's performance. Symmetric operation at the remote celestial body. • I neglect the small Oberth advantage at escape and capture, but don't include other small costs like tilting the orbital plane. • Acceleration and deceleration to Hohmann speeds are done at full efficiency. The script is for conventional situations. • Mercury's orbit is elliptic and tilted, neglected here. Reasonable planning would make one pass by Venus to gain much mass without wasting time. And a hectare Solar sail is already good at Mercury http://www.scienceforums.net/topic/78265-solar-sails-bits-and-pieces/?do=findComment&comment=763027 Aerobraking is an option at Mars and also Venus. • The asteroid is small and at arbitrary 2.6 AU. • Capture at Jupiter or Saturn is excellent with the sunheat engine. I don't detail it because missions want too varied orbits. Venus, Venus, Earth (and Jupiter) flybys would gain much. • Capture by Uranus and Neptune demands a chemical engine, which can then serve to leave Earth, so they would use the previous script. • The chemical engine gives 18% (asteroid) or 24-30% (others) more mass but the present script simplifies the craft. ---------- Erratum to the previous script on Jul 27, 2014 http://www.scienceforums.net/topic/76627-solar-thermal-rocket/?do=findComment&comment=818683 en route to Mercury, the mass in Hohmann transit is 8779kg not 6244kg, and the follow-ups increase by 1.41 too, including the arrived mass: 3835kg not 2719kg. Marc Schaefer, aka Enthalpy ##### Link to comment ##### Share on other sites The coming Vega-C would benefit from a sunheat stage: from estimated 3.1t in 400km 5.4° Leo, the sunheat engine would bring 1.7t directly in Gso, 1.6t in 300km Lunar orbit, 1.4t in transfer to Mars. As much as Soyouz from Kourou: game changer. Just 6 sunheat engines, easier to deploy, would spiral to Gso in 26 days plus eclipses - a Gto step would gain 90kg payload but take 130 days, while 2 spiralling sunheat engines gain as much in 78 days plus eclipses. 8 sunheat engines can raise the apogee in approximately 6 months if making less efficient long burns in the very last orbits, and eventually put 1.6t on a 300km Lunar orbit, or 1.4t in transit to Mars where the probe can aerobrake. That's more than Spirit and Opportunity (1063kg) got from the Delta II Heavy: cheaper science! While the launch company could operate the sunheat transfer to Gso, this needs long-term navigation sensors available at the payload, and space probes' operations are longer, so the spacecraft team will more likely operate the sunheat stage. The stage is nearly identical (1274kg LH2, 4 to 8 engines) for these three goals and may be supplied by a third player. 1973kg LH2 would transfer 0.7t directly to Jupiter and 0.8t to a low Martian orbit (figures suggest the tank insulation avoids a cryocooler). Bigger tank bigger for all, or only for them, or scale the payload down. This tank doesn't fit in the Vega-C fairing, hence is a separate stage. An interstage, guided by rolls at separation like for Zenit, gives room to deploy the concentrators. The stage needs a different fairing and trickle hydrogen before the launch. Vega has none up to now, but Ariane 1 had it on the reused ELA-1 launch pad. Here's an estimation of the stage's dry mass, built as described there http://www.scienceforums.net/topic/60359-extruded-rocket-structure/?do=findComment&comment=761740 and followings. 40kg Ballon of steel sheet 92kg 50mm foam for 15min without trickle hydrogen 52kg 20 plies MLI for months in vacuum between uses 4kg Polymer belts holding the tank 70kg Truss of welded AA6005 tubes 96kg Eight engines D=2.8m 30kg Engines' deployment and orientation 50kg Undetailed 0kg Shell already thrown away, interstage too 0kg Equipment bay is in the spacecraft -------- 434kg Marc Schaefer, aka Enthalpy ##### Link to comment ##### Share on other sites • 3 weeks later... At least two companies want to mine the asteroids for precious metals. While I have no knowledge hence opinion about the wealth available there, I claim that the sunheat engine can bring a heavy load there and back to Earth. Near-Earth-Objects would be more accessible, even to chemical propulsion, but may not be varied enough to offer useful resources, so the goal shall be in the main belt. I take an arbitrary asteroid at 2.8AU from the Sun in the ecliptic plane. As we ignore if water is available near the wanted ore, the script I propose uses no propellant gained there. ----- Escape ----- An Ariane 5 or Atlas V 551 shall put 18 500 kg on a naturally tilted 400km orbit, beginning as described here on Jul 27, 2014. Eight D=4.572m sunheat engines raise the apogee to 60 000 km and detilt the orbit as needed; this takes approximately 1 year and, the long pushes at perigee being 90% efficient, leaves 14 640 kg. Then, an oxygen-hydrogene engine adds 565+1096m/s at perigee to escape Earth and provide 5000m/s asymptotic speed to the remaining 10 190 kg. Escaping with the launcher's upper stage looks seducing but its tanks aren't insulated for one year, so the craft needs an insulated oxygen tank with cryocooler and a chemical engine. A car fuel cell or a lithium battery rotating electric pumps can make a small engine very efficient, but I've taken Isp=4590m/s only. After escape, the chemical engine and the cooled oxygen tank are discarded, leaving 10 000 kg in transit. ----- Transit ----- The sunheat engines add in 8 days 1373m/s near Earth for a Hohmann transfer and 4887m/s in 168 days near the asteroid. A spiral transfer is nearly as good, so even unfavourable inclination changes at mid-course are affordable. 500m/s more account the path corrections. This leaves 5804kg near the asteroid. Here the craft could discard a hydrogen tank and, if not done earlier, the external truss. Not vital for the mass, but it can ease the landing. ----- At the asteroid ----- The sunheat engines operated normally would land and lift the craft from a dense D<600m asteroid only, but the craft can also eject luke-warm hydrogen, and even better, it can have a spring. One small nitrogen spring lifts the craft from an asteroid of several km size. The hydrogen consumption is negligible if any. A first mission ends there, having brought 4.5t of mining equipment that begin operations. What and how, ask someone else. The following missions load 1t of precious metal purified in situ and lift off with 6804kg. ----- Return leg ----- The sunheat engines brake 4887m/s in 134 days near the asteroid. Later, they apply a 500m/s correction. This leaves 4410kg heading to Earth. One or several reentry capsule carry the precious metal and total 2700kg. Their structure makes 29% of the mass, the heat shield 21% for a reentry at 12 877 m/s (slightly more than Apollo), the parachutes 13%, and the cargo 37%. The 1710kg ferry can burn in the atmosphere or change its path. Refilling it seems more complicated than useful. ----- Ferry ----- 10 450 kg hydrogen fit in 155m3, too much for the usual fairings. Instead, a truss at D=5.4m can hold a D=4.8m H=12.5m balloon, plus the oxygen, all engines, the capsules, the equipment, making it somewhat taller than usual - or broader. Made of graphite, the truss weighs 250kg, the tank 260kg, it foam 64kg and the 20-ply superinsulation 108kg, while the aerodynamic shell elements are discarded when exiting the atmosphere. The chemical engine and oxygen tank are accounted elsewhere, so 1710-682=1028kg include the cryocooler, the equipment and undetailed items. Maybe the ton of cargo can increase a bit. The hydrogen tank would leak 60W from 300K to 20K. The cryocooler may need 3kW electricity when the sunheat engines idle. As a cheap source, the 8 concentrators can feed 180kW light on small solar cells. This drops at the asteroid, but the heat leak too. Alternately, an uncooled hydrogen tank may perhaps suffice, if its external skin is colder and its insulation better - to be thought with calm. ----- Economics ----- Chemical engines bring nothing back from the main belt, electric thrusters only gram samples, but the sunheat engine one ton cargo. Gold and other precious metals sell for 40M$, half the launch cost, and the craft costs too. Still imperfect, but that's a first try. Venus flybys improve the masses. The main belt begins at 2.2AU. Falcon 9 Heavy claims to launch for 90M$3.4* the payload, which would bring 140M$ gold back.

Marc Schaefer, aka Enthalpy

##### Share on other sites

On 9/24/2017 at 0:12 PM, Enthalpy said:

At least two companies want to mine the asteroids for precious metals. While I have no knowledge hence opinion about the wealth available there, I claim that the sunheat engine can bring a heavy load there and back to Earth.

Near-Earth-Objects would be more accessible, even to chemical propulsion, but may not be varied enough to offer useful resources, so the goal shall be in the main belt. I take an arbitrary asteroid at 2.8AU from the Sun in the ecliptic plane. As we ignore if water is available near the wanted ore, the script I propose uses no propellant gained there.

----- Escape -----

An Ariane 5 or Atlas V 551 shall put 18 500 kg on a naturally tilted 400km orbit, beginning as described here on Jul 27, 2014. Eight D=4.572m sunheat engines raise the apogee to 60 000 km and detilt the orbit as needed; this takes approximately 1 year and, the long pushes at perigee being 90% efficient, leaves 14 640 kg. Then, an oxygen-hydrogene engine adds 565+1096m/s at perigee to escape Earth and provide 5000m/s asymptotic speed to the remaining 10 190 kg.

Escaping with the launcher's upper stage looks seducing but its tanks aren't insulated for one year, so the craft needs an insulated oxygen tank with cryocooler and a chemical engine. A car fuel cell or a lithium battery rotating electric pumps can make a small engine very efficient, but I've taken Isp=4590m/s only.

After escape, the chemical engine and the cooled oxygen tank are discarded, leaving 10 000 kg in transit.

----- Transit -----

The sunheat engines add in 8 days 1373m/s near Earth for a Hohmann transfer and 4887m/s in 168 days near the asteroid. A spiral transfer is nearly as good, so even unfavourable inclination changes at mid-course are affordable. 500m/s more account the path corrections. This leaves 5804kg near the asteroid.

Here the craft could discard a hydrogen tank and, if not done earlier, the external truss. Not vital for the mass, but it can ease the landing.

----- At the asteroid -----

The sunheat engines operated normally would land and lift the craft from a dense D<600m asteroid only, but the craft can also eject luke-warm hydrogen, and even better, it can have a spring. One small nitrogen spring lifts the craft from an asteroid of several km size. The hydrogen consumption is negligible if any.

A first mission ends there, having brought 4.5t of mining equipment that begin operations. What and how, ask someone else.

The following missions load 1t of precious metal purified in situ and lift off with 6804kg.

----- Return leg -----

The sunheat engines brake 4887m/s in 134 days near the asteroid. Later, they apply a 500m/s correction. This leaves 4410kg heading to Earth.

One or several reentry capsule carry the precious metal and total 2700kg. Their structure makes 29% of the mass, the heat shield 21% for a reentry at 12 877 m/s (slightly more than Apollo), the parachutes 13%, and the cargo 37%.

The 1710kg ferry can burn in the atmosphere or change its path. Refilling it seems more complicated than useful.

----- Ferry -----

10 450 kg hydrogen fit in 155m3, too much for the usual fairings. Instead, a truss at D=5.4m can hold a D=4.8m H=12.5m balloon, plus the oxygen, all engines, the capsules, the equipment, making it somewhat taller than usual - or broader. Made of graphite, the truss weighs 250kg, the tank 260kg, it foam 64kg and the 20-ply superinsulation 108kg, while the aerodynamic shell elements are discarded when exiting the atmosphere. The chemical engine and oxygen tank are accounted elsewhere, so 1710-682=1028kg include the cryocooler, the equipment and undetailed items. Maybe the ton of cargo can increase a bit.

The hydrogen tank would leak 60W from 300K to 20K. The cryocooler may need 3kW electricity when the sunheat engines idle. As a cheap source, the 8 concentrators can feed 180kW light on small solar cells. This drops at the asteroid, but the heat leak too. Alternately, an uncooled hydrogen tank may perhaps suffice, if its external skin is colder and its insulation better - to be thought with calm.

----- Economics -----

Chemical engines bring nothing back from the main belt, electric thrusters only gram samples, but the sunheat engine one ton cargo. Gold and other precious metals sell for 40M$, half the launch cost, and the craft costs too. Still imperfect, but that's a first try. Venus flybys improve the masses. The main belt begins at 2.2AU. Falcon 9 Heavy claims to launch for 90M$ 3.4* the payload, which would bring 140M$gold back. Marc Schaefer, aka Enthalpy But can it beat that? Lots of fuel, and basically brute force. But possibly one of the most effective rockets to date once they get it started. ##### Link to comment ##### Share on other sites At the main belt asteroids, Dawn observes the surface of Ceres and Vesta, other probes passing by have imaged half a dozen small bodies more, and telescopes give optical spectra from the surface of whole bodies. https://en.wikipedia.org/wiki/Asteroid_belt A mission to analyze them in situ and bring samples back would benefit to science, potentially to the extraction or production of propellants for farther trips, and hypothetically to the extraction of precious metals brought back to Earth. A return trip passing by many main belt objects needs a huge delta-V and a decent thrust made possible by the sunheat engine. It begins like the asteroid mining mission here above: an Ariane 5 or Atlas V 551 delivers the oversized 18.5t craft in orbit. Solar, chemical and solar pushes put the probe in transit to the main belt - 10.5t because the first target orbit in the Vesta family is at 2.35AU only but inclined 6.7°. Pushing 4560m/s there, in 132 days with 8 engines, leaves 7.3t at the first asteroid. Venus flybys may improve. The asteroids' actual "osculating" orbital elements look pretty random, but Kiyotsugu Hirayama saw that Jupiter lets them fluctuate. He computed "proper orbital elements" as mean values, and then the asteroids make clearer families supposed to result from shattered bigger bodies. https://en.wikipedia.org/wiki/Asteroid_family Sampling a few bodies in each family, plus some outside all families, seems a good plan. I estimate the mission delta-V based on the mean "proper" orbital elements, but the actual "osculating" elements differ much, so among hundreds of family members, a wise and well equipped mission planner would pick more favourable ones. To evaluate individual transfers, which can't await the best date to change the orbit inclination, I combine quadratically the polar component of the speed difference with only half of the ecliptic component and leave the other half untouched. Stopping at intermediate objects is for free, and because I've been pessimistic above, I neglect the orbit fine-tuning and the operations near each object. A well chosen set of target objects would change radically the mission duration and delta-V. The unoptimized sketched tour costs 10 693 m/s from Vesta to Eumonia families, leaving 3.1t there. Landing on bodies up to km size is for free and a spring lets lift off. Ask someone else how to analyze, scoop, dig, bore samples there. A Yag pumped by concentrated sunlight and a hydrogen gun shall perform remote analyses at bigger objects. I already described light samples boxes and sealing apparatus there http://www.scienceforums.net/topic/85103-mission-to-bring-back-moon-samples/?do=findComment&comment=823276 The science instruments can be discarded before heading to Earth, one optional hydrogen tank there too or a bit earlier, the structural truss early if possible. The craft brakes within the asteroid's orbital plane by 4724m/s in two kicks well spread over two revolutions in order to synchronize itself with the Earth and makes a 200m/s fine tune. Other transfers would be faster. If nothing has been discarded, this leaves 2.1t heading to Earth for aerobraking, say over Antarctica. The insulated cooled tanks, the structure and the engines shall weigh 0.9t, equipments 0.3t, experiments 0.3t, leaving 0.6t for the reentry capsule(s); dropping unneeded mass early would increase that. The 600kg comprise 29% structure, 21% heat shield, 13% parachutes, 7% boxes and 30% = 180kg samples from dozens of asteroids all over the main belt. Marc Schaefer, aka Enthalpy ##### Link to comment ##### Share on other sites • 2 weeks later... Electric thrusters need so much power that solar panels feed only a faint push. Operations near a planet, which takes stronger engines, lead to use a nuclear reactor which can't be very light and needs a big radiator. The cancelled Jupiter Icy Moons Orbiter Jimo (not the Explorer Jime) was such a project, and this drawing (thanks to Nasa) illustrates the oversized feeder for the electric engine: The reactor is at the tip, the "wing" is the 422m2 heat radiator, and from https://en.wikipedia.org/wiki/Jupiter_Icy_Moons_Orbiter the science payload orbiting Europa, Ganymede and Callisto would have been 1500kg, similar to what the sunheat engine promises - but Jimo would have needed chemical engines to leave Earth, so despite the electric thruster's higher Isp, it would have weighed 36t in Earth orbit launched in three heavy flights and cost 16G$.

Edited by Enthalpy
##### Share on other sites

To accelerate a craft for weeks, the sunheat engine uses an energy amount impossible to store. But to raise or lower an apoapsis, escape a celestial body or get captured, the smaller energy for short kicks can be obtained during idle time and stored, as already noted on Jun 10, 2014:
http://www.scienceforums.net/topic/76627-solar-thermal-rocket/?do=findComment&comment=810218
More details now.

==========

Lithium-polymer batteries can store 475kJ/kg more or less. Before a kick, the concentrated light can be split by wavelength and sent to small solar cells of varied bandgaps for >40% conversion.
http://www.scienceforums.net/topic/76627-solar-thermal-rocket/?do=findComment&comment=826983
Other uses welcome the efficiency and collecting area far from the Sun.

To raise the apoapsis in Earth orbit, each sunheat engine without storage consumes 20kW during 1200s. A battery providing as much energy to double the force weighs 50kg. That's worse than adding one ~30kg engine. But:

• Fewer engines are easier to deploy.
• Operation during an eclipse season gains time and provides flexibility.
• Electric pumps for the propellants making the escape chemical kick justify 100-150kg batteries if starting with 18.5t at Leo, so 8 sunheat engines get 30% more energy per kick at zero extra mass, and the added thrust can concentrate on the best 1000s.
• At Mars (1.52AU) the battery is already lighter than the additional engine, at Saturn (9.58AU) the battery advantage would be 50*, if 1200s kicks were meaningful there.
• Other uses cherish the good electricity production at the outer planets. 700W from 8 concentrators at Saturn.

The combination isn't a resistojet, because direct heating by sunlight is kept, as it avoids the wasteful conversion to electricity. Far better for the long pushes. The engine can use the same chamber for both modes or not. The described regenerative insulation methods apply.

==========

Melting a metal stores heat too. Simpler than making electricity first, but heavier, and they probably dissolve their tungsten container.

Tmelt        Hmelt
K          kJ/kg
-------------------
3290   Ta     159
2896   Mo     286 ?
2750   Nb     288
2506   Hf     143
-------------------

One hope is to find some eutectic of tungsten with one or several other elements giving the desired melting point, as inspired by Sn-Pb-Ag solder not dissolving Ag. I didn't find credible data about Hf-W nor Nb-W eutectics, only calculated data; Mo-W and Ta-W seem to make no eutectic and be fully miscible.

==========

Melting a ceramic stores more heat. Beware data is inconsistent. BeO is toxic, and B is expected to corrode W. More complex formulas are possible. Mixes shall bring missing melting points.

Tmelt        Hmelt   Hform
K          kJ/kg   kJ/mol
---------------------------
WO2           -285
Ta2O5          -409
NbO           -406
---------------------------
2988   ZrO2   706    -550
2915   MgO   1920    -602
2703   SrO    674    -592
2500   Y2O3   463    -635
2318  Al2O3  1071    -559
2247   ZnO    860    -351
2218   MnO    767    -385
---------------------------

The last column shows the heat of formation per mole of oxygen atoms. If the melt's metal binds more strongly with oxygen, it could be compatible with W; computing the chemical equilibrium would be better if data exists. For water propellant at a lower temperature, container candidates are Nb or better Ta, possibly with W coating on the melt side, and some ceramics.

As heat storage keeps the engines hot for months, the temperature should rather be 2700K or 2600K. With the ceramic fully molten, 2800K remains possible during week-long accelerations.

MgO stores 4* more energy than a battery and is 2* lighter than an added engine for perigee kicks. The advantage increases at the outer planets.

==========

The freight transport to the Moon took 10 months of perigee kicks with 10 sunheat engines
http://www.scienceforums.net/topic/76627-solar-thermal-rocket/?do=findComment&comment=1007124
Let's replace 4 sunheat engines by 120kg of MgO heat storage: the 1.55* stronger kicks gain 3.5 months, and thanks to the shorter kicks' efficiency, we land 2% more mass.

Marc Schaefer, aka Enthalpy

##### Share on other sites

If we get 40% conversion efficiency from a concentrating solar array, is pumping this energy into an ion engine more efficient when its high Isp is considered?

All this technology is also known and working so the risk of development isn't there (or as high).  Also, electricity is a generally useful form of energy (compared to heat) and may have use along the coast phase or even when parked waiting for the next mission.

I believe there is some cooling needed for those concentrating collectors, but no mirrors needed, which may end up adding to the mass of the array in the balance - hard to tell.

##### Share on other sites

2 hours ago, Frank said:

If we get 40% conversion efficiency from a concentrating solar array, is pumping this energy into an ion engine more efficient when its high Isp is considered?

2 hours ago, Frank said:

All this technology is also known and working

Where did you see such a thing? A strong electric engine?

2 hours ago, Frank said:

I believe there is some cooling needed for those concentrating collectors

Why?

##### Share on other sites

I haven't looked closely at the numbers, but 7x more efficient seems optimistic to me.  Here on Earth, it's VERY difficult to make a case for solar thermal based on cost, even where thermal should obviously be better.  When it comes to mass as the currency, I don't know yet.

The working example of a concentrating solar collector I found is SCARLET https://pdssbn.astro.umd.edu/holdings/ds1-c-micas-3-rdr-visccd-borrelly-v1.0/document/doc_Apr04/int_reports/Scarlet_Integrated_Report.pdf

It's a fresnel concentrator and thermal management is on page 7 of the report.  In 1998, 20 years ago, they managed 45W/kg for the array and DS1 with its ion engine worked.

Using a 4 m diameter dish and concentrating on a PV module will generate some heat.  Quite a bit of it, because up to 60% of the energy will dissipate in heat if 40% efficient.

##### Share on other sites

Here are a couple of link to STP that might be of interest:

"Reduce tank volume with increased STP specific impulse approaching 1200 seconds

- Higher temperatures above 3000K increase Isp

- At higher temperatures above 3000K and low pressures, hydrogen starts dissociating and lowers the average molecular weight, which also increases Isp"

I guess dissociation of hydrogen will allow more Isp for a given temperature (which is limited by material properties), at the expense of efficiency, requiring more energy input.

##### Share on other sites

I stand by my figures, especially the thrust versus concentrator size. Whether they fit the decisions and policies of an administration is no proof for anything. Nor do I need nor seek bad designs elsewhere to consider them as limitations for other attempts.

I'm not interested by a refractor concentrator. No perceivable usefulness, so its mass and losses are no hint to anything.

Yes, more Isp needs more power. The sunheat engine allows to adjust the trade-off in flight, but I've seen no clear usefulness.

From the trials cited in 20160003173.pdf:
An inflatable concentrator is bad for high focus temperatures. I take a rigid one.
Several concentrators per engine are dangerous for the craft or vessel and difficult to test.
Sunlight can't make the proper temperature on Earth. How did they figure such a thing? Their results reflect that.
Their nozzle is too small.
They have no ruminator. It remains to see if their colder faceplate radiates little, and how much heat is lost through conduction to the faceplate.
I don't understand the alleged advantages of the refractive secondary concentrator. But its obvious drawback is to limit the temperature.
The Isp improvement through dissociation results from absorbed heat primarily, not from the average molecular mass.
Their mass estimates are way off. A Gso launch does save a lot through sunheat engines.
Yes, fairings must increase. Or better, make a separate sunheat stage, as I suggested here.
They only consider Leo-to-Gso missions, too bad.
=> Many bizarre choices, misconceptions and half-thought designs. This led to wrong conclusions.
=> This is why I don't seek figures elsewhere to decide if a potential technology is interesting.

Concentrate light on solar cells:
The light is filtered by wavelength, and then the efficiency of the solar cell is hugely better. 40% refers to the power intercepted by the concentrators. Light with bad wavelength is rejected.
When used at Saturn or even Neptune, no overheating to fear.

Not only is the sunheat engine extremely performant, it's also the sole and only solution beyond chemical propulsion.

##### Share on other sites

I found it encouraging that money was spent on STP and that 1200s Isp is thought to be achievable based on prototype results (even with their design).  I'm still unsure how the ruminator works, but it sure would be neat to get a (small?) grant to build a one and test it at that facility.  Most of the infrastructure is already in place!

Most high efficiency CPV works on multiple wavelengths, but light generally isn't selectively filtered and heat IS an issue.  It's surprising that there is a way around that but it isn't being used (that I know of).  I must admit that I don't understand how the selective wavelength filter works either.

##### Share on other sites

• 3 months later...
On 7/20/2013 at 10:23 PM, Enthalpy said:

Hydrogen is the propellant that improves the exhaust speed over chemical reactions, but the Solar thermal engine accepts other propellants. Water at 2400K can give some 3000-4000m/s ejection speed, depending on the dissociation allowed by the chamber pressure - and if available in space, it enables big scenarios where the in-situ propellant needs no lengthy preparation.

Corrosion is a serious worry, hence the 2400K. If metals don't survive hot vapour, ceramics may: MgO and ZrO2, with 100K less? Tantalum hafnium carbide?

If water can be used as a propellant, can aqueous ammonia work as well?  How would dissociation due to ionization (NH4OH) affect Isp?  Would the increase in hydrogen increase exit velocity and Isp?

One advantage is a lowered freezing point.  Plus many of the advantages of water - smaller tanks, ISRU, non-cryogenic...

##### Share on other sites

Perhaps anhydrous ammonia makes more sense than aqueous.  It can be dissociated in the regenerator using a nickel catalyst at a  temperature of 920 C :  2 NH3  N2 + 3 H2  then the Nitrogen and Hydrogen can be heated to high ejection speeds.

Edited by Frank
##### Share on other sites

Hi Frank,

At the temperature of the sunheat engine, you don't have to care about ionization, catalysts nor active dissociation. The state in the chamber results from the heat whatever you do before.

You should sniff a milligram of ammonia (diluted in the air!) before considering tons. You'll change your mind.

If really you want an antifreeze in water, consider guanidine and pimagedine (made on Earth), which produces only gas upon decomposition.

==========

Test facility: concentrated sunlight on Earth can't reproduce simultaneously the power density and the convergence angle attained in space, but both are important to test the heater.

As opposed, a test setup in a lab is small, decently cheap and quickly purchased, and works in cloudy Europe too. I didn't detail it here but have clear ideas and figures about it. Easy choice.

##### Share on other sites

• 2 months later...

Many telecom satellites use presently electric thrusters to raise the perigee from the transfer to the geosynchronous orbit. To achieve it in 4 to 7 months but save solar panels, the thrusters eject more propellant but slower than the Vasimr cited on Aug 19, 2017. For instance the Hall thruster PPS-1350-G in the table there
https://en.wikipedia.org/wiki/Ion_thruster
achieves 90mN at 1660s=16.3km/s from 1.5kW. This makes the solar panels 0.35* as huge as previously. The visual comparison with the sunheat engine becomes :

##### Share on other sites

It might be an even better ratio if relative mass is considered.  At 150w/kg,  ~113 kg of mass for the solar PV panels must be accelerated, so the heat engine thruster might be even smaller.  How heavy are telecom satellites?

##### Share on other sites

The first user's manual for Ariane 6 is published and it details preliminary performance to varied orbits, nice. Once put on low-Earth-orbit by Ariane 6, heavier satellites can attain the geosynchronous orbit using sunheat engine.

The sunheat stage would spiral from 400km 6° to 35800km 0°. This costs 4700m/s, 500m/s more than a Hohmann transfer, but saves a year.

With Ariane 62, the fairing can house the 3.2t hydrogen and the payload. This brings no higher stresses on the previous stages, but needs trickle hydrogen under the fairing at pre-launch. The satellite owner can operate the sunheat stage, use the existing sensors and electronics. I take 120kg of structural tank per ton of hydrogen, 4 engines of 17kg for 2 months transfer, 100kg equipment if the launch company operates this stage.

With Ariane 64 (or for missions of higher energy), 6.8t hydrogen need a tank outside the fairing and probably common to both customers. This suggests operation by the launch company. The stresses on the previous stages increase. I take 120kg of structural tank per ton of hydrogen, 6 engines of 17kg for 3 months transfer, 100kg equipment.

Marc Schaefer, aka Enthalpy

##### Share on other sites

On 4/11/2018 at 9:54 PM, Frank said:

It might be an even better ratio if relative mass is considered.  At 150w/kg,  ~113 kg of mass for the solar PV panels must be accelerated, so the heat engine thruster might be even smaller.  How heavy are telecom satellites?

An ion thruster would accelerate itself and its solar panels enough. Geosynchronous telecom satellites weigh about 4 to 7 tons in Gto, or 2.5 to 5 tons in Gso. Their 15kW take a few 100kg panels, but the claimed 150W/kg would take just 100kg, so there is margin. Fakel's Spt140D takes 15kW/N so the absolute limit would be 10mm/s2,  enough for most missions. For the sunheat engine, this limit or factor-of-merit is 2.4N/30kg or 80mm/s2. This comparison is more important farther from the Sun.

The hardest limit to ion propulsion is the cost of the solar panels.

##### Share on other sites

Here's a more detailed mass estimate of the sunheat stage for Ariane 6. Computed to break at 2.2MN*m (6.4t on A62) or 8.6MN*m (13.7t on  A64) and to work at 1.5atm (boiling 2K warmer than 1atm). The truss is cold at A64.

 kg   kg
190   360   290µm or 370µm steel balloon with brazed seams
74    94   22mm or 17mm foam for 1000s after trickle hydrogen removed
61    76   25 or 19 plies of 13µm MLI for additional 7 idle days in vacuum
10    20   Polymer straps holding the balloon
250   700   Welded truss of Di=80mm e=1,4mm AA6082 or Di=98mm e=2,6mm AA7022
no    83   Upper short insulating truss of Di=116mm e=4,4mm glass fibre epoxy
no     0   Lower insulating truss already thrown away
120   180   4 or 6 engines
20    30   Engines' deployment and orientation
no     0   Shell already thrown away
100   100   Equipment if not in the spacecraft
30    30   1.5 remaining separation belts
50    50   Undetailed
----------
905  1723   kg dry stage

282 and 254kg per ton of propellant is heavy but would improve at a lighter space probe. The 5* denser couple with oxygen would scale it as 56 and 51kg/t, outperforming the projected Esc-B. I wouldn't be surprised if Ariane 6, whose manual tells masses landed on the Moon, uses the already suggested insulation stack
http://www.scienceforums.net/topic/60359-extruded-rocket-structure/?do=findComment&comment=761740

At the A62's truss, tubes of isogrid AA7022 would improve over smooth AA6082. At the A64's truss, tubes of isogrid Ti-Al6V4 would improve over AA7022, be weldable, and could replace the glass fibre trusses too. Everywhere, balloons and trusses made of graphite composite would be half as heavy, and the walls of graphite tubes could be a balsa sandwich to save more mass. These sandwich tubes could replace the glass fibre trusses too.

The throwaway windshield shells are already described
http://www.scienceforums.net/topic/60359-extruded-rocket-structure/?do=findComment&comment=764231

Marc Schaefer, aka Enthalpy

##### Share on other sites

Here I compare the payload mass in geosynchronous orbit with electric versus sunheat engines.

----------- Ariane 64 and Hall -----------

Fakel's Spt140D has been used on Eutelsat 172b
http://www.fakel-russia.com/images/content/products/fakel_spd_en_print.pdf
and Ariane VA237 put it on the usual GTO: 250km*35706km*6°, from where a 1465m/s short kick would achieve Gso.

This Hall thruster offers Isp=1770s=17360m/s and 0.29N from 4500W, so the 3550kg satellite's 13kW (end-of-life?) solar panels feed 3 engines for 0.87N or 245µm/s2. 4714m/s spiralling from Leo would need 8.5 months, and even longer for heavier satellites, but in 4 months the Hall thrusters achieved Gso from Gto.

To save delays, the apogee pushes are very long hence inefficent. Let's say they must provide 1.2*1465=1758m/s: the payload is 0.90* as heavy once in Gso. Scaling up to 11500kg in Gto for Ariane 64, 1108kg xenon are consumed, plus 83kg for a bigger tank, and 85kg for 10 Hall thrusters. This leaves 10220kg in Gso for the satellites and a dual launch adapter.

Other transfer orbits would improve. The long pushes raise the apogee too, so a lower one is better. A perigee higher than 1000km would reduce the risk of collisions. A mid-high circular orbit would let the electric satellite spiral to Gso. Ariane 6's user's manual gives masses to these orbits. But as long as chemical satellites buy half-launches, Ariane will target the usual Gto, maybe Gso (5000kg there).

------------ Falcon 9 and Hall -----------

The outdated but documented Falcon 9 v1.0 put 4536kg to Gto 185km*35786km*28.5°, from where a short 1826m/s kick achieves Gso. Let's take 2191m/s with Hall thrusters: the payload weighs 3998kg in Gso, minus 40kg tank and 34kg thrusters, that's 3924kg in Gso.

---------- Ariane 6 and sunheat ----------

As estimated on Apr 15 & 19, 2018, Ariane 64 and a sunheat stage would put 13.1t at Gso in 3 months or less. This 2880kg or 28% improvement over Hall thrusters starting from Gto is worth 25M$, plus the saved delay. Ariane 62 and a sunheat stage would put 6.1t at Gso in 2 months, that's 1251kg or 26% better than Hall thrusters. Worth 18M$ or 20M$, the price of upgrading to Ariane 64. The transfer time is one month for 360kg or 120kg less payload. ---------- Falcon 9 and sunheat ---------- Falcon 9 v1.0 put 10000kg at 400km*28.5°, from where I estimate imprecisely to 5552m/s the cost to spiral to Gso and suppress the inclination simultaneously. This leaves 6396kg at Gso, of which 1016kg are the dry sunheat stage (282kg/t as for Ariane) and 5380kg the payload at Gso in 2 months. That's 37% heavier than Hall thrusters, a bigger advantage because Gto in two stages is demanding. At 56M$ per launch (?) the gain is 21M\$.

Marc Schaefer, aka Enthalpy

## Create an account

Register a new account