Sideways Engineering the SpaceX Dragon

[mrmorris]

Given the above issue with the parachutes, the question arises -- just how much faith can I put in the graphic images that SpaceX has supplied? Unfortunately, the simple answer is: HeckifIknow. The data in the images is more detailed than you'd generally get in an artist's impression, but doesn't contain <b>enough</b> details to be a 'working' CAD file from (or for) SpaceX engineers. So did this come from an engineering-quality CAD file that was more detailed -- having been 'dumbed down' enough to be publicly released? Alternately was it one that was built from scratch simply for use in the COTS proposal, with engineering accuracy a secondary concern at best?<br /><br />I have to assume the latter. There's lots of little things in the images that scream to me that this was never an engineer-grade CAD file. This is unfortunate in that I have to take everything in the images with a bigger grain of salt. I think that the capsule dimensions are a safe area -- this isn't somewhere I'd expect them to take liberties. Even in an artist's impression, using the wrong dimensions would really screw up the perspective. Beyond that -- everything is murkier. I'm of the opinion that the more detail provided on a given element, the more solid the information. The PCBM, for example, is quite detailed. This makes sense, as SpaceX can get specs, dimensions, etc. on the PCBM easily. Since the design is a NASA mandate -- SpaceX knows exactly what it will look like right now. <br /><br />What's interesting to me is that the other item in the semi-transparent views that is highly detailed is the RCS. SpaceX has said that the Dragon sitting on the floor could fly now except for the lack of an RCS and a heat shield. So... why all the detail? Personally, I think it's another clue that SpaceX is building their own RCS. The attached image shows just how much information they've provided in the graphic.<br /><br />In Apollo and Gemini, there were two independent systems in the +X/-X

 

[rocketman5000]

Whattheheck is this could be some sort of accumulator to eliminate spikes in pressure??

 

[mrmorris]

<font color="yellow">"...accumulator to eliminate spikes in pressure"</font><br /><br />Having freely admitted ignorance, I'm not likely to argue. <img src="/images/icons/smile.gif" /><br /><br /><br />While assumptions based on the graphics must be taken with skepticism as mentioned earlier, I have the following preliminary deductions. Even if they aren't what Dragon ends up having -- it at least provides a look into what they were thinking along the lines of... at one time.<br /><br />- They are looking at a bi-propellant system. No real shock there. Only very small RCS systems use hydrazine monoprop, and H2O2 was never really all that likely.<br /><br />- They are looking at regulated pressurization tanks rather than blowdown. Again -- no shocker here.<br /><br />- The graphic shows equal-sized tanks for both propellant and oxidizer. If you look at my earlier post on the volume required for the various combinations here, you can see that the *closest* the two volumes come is 2x the propellant volume to 1x the oxidizer volume. This is likely simply a case-in-point of not being able to take the graphic as gospel, however.<br /><br />- What is more telling, however, is taking the above a bit further by glancing at the minimum number of tanks required in that post to hold enough propellant to provide 300 m/s of dv for 9300 kg. That would be... ten tanks. Of course that was using my off-the-cuff volume estimate for a .5m diameter spherical tank with 20% ullage. However -- even if I eliminate the ullage, and assume a .7m diameter sphere (larger than will fit in the aft equipment hold by my measurements), it would still require almost seven tanks of propellant. There are several possibilities to reconcile this:<br /> * It's possible that either my dv assumption of 300m/s is too high. However,

 

[mrmorris]

Assumptions will get you every time -- especially when you don't realize you're making them. The external views of the Dragon capsule with the various concentric rings down the length of the capsule caused me to unconsciously divide the craft into four sections. From the picture below -- everything above 'A' was the nosecap section. A-to-B was the CBM section (and later the forward equipment section). B-to-C was the crew section. Finally C-to-D was the Aft Equipment Section (AES). I knew there was likely some usable space aft of D, but I essentially left it as 'bonus' space if things got tight in the AES. Mentally I placed bulkheads at each of those points.<br /><br />*sigh* While staring at my closeup of the RCS above, I kept getting this niggling feeling that it was strange to be seeing the heads and shoulders of the crewmembers just above the oxidizer tank. For this to happen, the crew would have to be 'behind' the tanks. I pulled up the complete transparent view and really <b>looked</b> at it. Having actually done so (looked, that is), I realized that yes indeed, the crew section extendes further back than I'd thought. The Aft Equipment Section isn't a solid cylinder, as I'd been assuming, but rather a hollow one. The crew section extends back beyond point 'C' by about .75 m in a 2m cylinder. The shape from the transparent graphic appears to follow something similar to the red lines in the picture below, with the RCS tanks positioned approximately where the red circles are.<br /><br />There are multiple ramifications from this revelation. First of all -- the AES volume has dropped by 15%. The only reason it dropped that little is because I extended it further back. Previously I assumed it went from C-to-D, but now am assuming C-to-E. Without that, the volume would have dropped by over 30%. Second -- given that I've increased my assumed propellant tank size to come up with a reasonable number of tanks for the expected Dragon configurations, I had to refine the calcu

 

[mrmorris]

I was just doing some preliminary research to add parachute calculations to the application. I plan to make a report similar to the one I made for prop/ox combinations. It will calculate things like volume, mass, and descent velocity for various chute diameters -- dynamically determined based on the current mass calculation the app has for the applicable portions of Dragon. It will also calculate velocities assuming deployment failure of one and two chutes. I think it's be interesting to see what happens to these numbers over a parachute diameter range of say 25m to 35m.<br /><br />Anyway -- I ran across an interesting paper and figured I'd take a couple minutes to post it. It's fairly dry reading unless you're a parachute fanatic, but I'm largely putting it here for me. Similar links that I put in my G-X3 thread have been on great value to me in making this one. The paper also covers some things like ballutes and deployable wings.<br /><br />NASA SP-8066 -- Deployable Aerodynamic Deceleration Systems.

 

[mrmorris]

OK -- I have the first draft of my parachute calculations section in the Dragon Builder app. Right now it has terminal velocity for 1, 2, and 3 chutes over diameters ranging from 25 to 35 meters. The document I referenced above indicates that multi-chute efficiencies are lower than that of single chute, and specifically noted that a 3-chute system is only 85% as efficient as the three chutes would be in single configutation. That is accounted for in my calcs. I didn't see an efficiency for a 2-chute system, so WAGed it at 94%. <br /><br />I have not yet been able to locate a good figure for determining the mass of a nylon parachute of a given area. I have the <b>density</b> that the Apollo parachutes were packed to, but without a mass -- I can't use it. Right now the application simply computes a 'size factor', which gives a ratio of the area of the chutes to that of the lowest size (25 meters). This will give some idea of how mass and volume requirements will scale until I have more data.

 

[mrmorris]

I was able to locate a reasonable source for parachute densities and strengths. I can't locate the tear strength for the APollo chutes, so I simply picked the highest strength chute from the following site: SPECIFICATIONS - MILITARY PARACHUTE / CANOPY FABRICS.<br /><br />PIA C 17208 -- Type 1,Class C <br />Mass: .23 kg/m2<br />Tear W/F: 27 kg/27 kg<br /><br />Following is the second draft of my parachute calculations -- updated with area, mass, and volume for a three-parachute system of the given diameter. From the earlier document on re-entry systems was a chart on the maximum packed density of chutes given various methods. I picked "Vacuum/light mechanical" to use in calculating density. From what I've read, I believe that's closest to the method used for Apollo.

 

[mrmorris]

I'm still looking for data on the Apollo parachute tear strength. I thing that 27kg is *probably* overkill, and reducing this would considerably reduce the mass of the parachues. Anyway -- while searching I located another cool link. It shows data & modelling on the water landing of the Apollo CM: The Use of LS-DYNA to Simulate the Water Landing Characteristics of Space Vehicles.

 

[mrmorris]

OK -- I finally hit the parachute jackpot. I stumbled across a study done for the EELV program using apollo-style parachutes when there was a consideration of using (and reusing) the SSMEs for the program (presumably before it was decided to make the RS-68). <br /><br />The paper gives 0.767 as the original Apollo Cd (coefficient of drag) as well as that for the system using current tech for one and three parachutes (1.12 and .965 respectively). I've updated my parachute section to use those numbers, and split the differene to 1.06 for a 2-chute splashdown. It didn't give me the mass of the original Apollo parachutes, but it *did* give me the mass of theirs, from which I was able to get a figure of 0.134 kg/m2. This is a significant mass improvement over the figure I was using before. The mass of the engines was well over what the Dragon would be, so the fabric is easily sturdy enough.<br /><br />At this point, I feel the parachute calculations are based on data as firm as I'm going to get without a press release from SpaceX, so I'm going to switch back to updating the RCS portion of the application.

 

[mrmorris]

OK... one last parachute change... <img src="/images/icons/smile.gif" /><br /><br />I compared the velocities in my parachute calculations to that of the Apollo CM at splashdown (typically ~30 ft/s or 9.1 m/s). Given that midway through my list, a 29m chute diameter would have a comparable velocity with a single parachute (i..e <b>two</b> parachute failures)... I decided that my range of parachute sizes was on the high side. I expect that capsule mass to grow somewhat, but likely not enough to account for that. The section has been altered to show parachute diameters from 20-30m instead of 25-35m.

 

[mrmorris]

Looking more at the EELV study, it does not have any drogue or pilot chutes. I have to wonder if this is feasible for Dragon. The drogues fo Apollo were to stabilize it's flight, provide a small deceleration, and orient the capsule properly for the main chute release. I have to assume this will still be required with Dragon. Accordingly, I looked up information for the drogue and pilot chutes. The Apollo D&P chutes had the following (approximate) specs:<br /><br />2 ribbon drogues -- mortar-deployed<br />Diameter: 5m<br />Surface area: ~40m2<br />Mass: 0.9 kg<br />Volume: 122 in3<br /><br />3 pilot chutes -- mortar-deployed<br />Diameter: 2m<br />Surface Area: ~6m2<br />Mass: 0.136 kg<br />Volume: 18 in3<br /><br />Looking up mortar parachute deployment systems -- it appears that Irvin is the place to go. In fact, Irvin supplied the parachutes for Falcon I, so we can be pretty sure they'll be the supplier for Dragon as well. There are no mortars with the *exact* specs for the dimension chutes above, but the closest are as follows:<br /><br />Ribbon: 5.5" x 12.19" -- 8.5 lb<br />Drogue: 2.126" x 5.35" -- 1 lb<br /><br />The ribbon mortar in particular is a bit oversized for the chute we have. Regardless, it gives me a mass and volume in the right ballpark.

 

[mrmorris]

Just for curiosity's sake, I located the most equivalent drawings of the Gemini, Apollo, and Dragon crew capsules, then 'tweaked' the sizes as close to an equivalent scale as I reasonably could. Having taken the time to do it, I figured I'd post the result here. Originally, I wanted to have the CEV in it as well, but I wasn't able to locate a good 'side-on' view. All the current graphics I was able to locate were shown from oblique angles.<br /><br />Anyway -- the result follows:

 

[mrmorris]

On Monday NASA issued an RFI on COTS resupply. I'm not really clear on where this fits into the big COTS picture. Engineering interests me -- the politics of space is a big yawn. I'm assuming it's simply a note for all the interested parties other than SpaceX and Kistler so that they can compete. FIrst one to a cheap COTS resupply solution wins. <img src="/images/icons/smile.gif" /><br /><br />At any rate -- the text from the RFI is interesting:<br /><br /><i>"NASA is soliciting information from potential sources that can provide:<br /><br />1.) A single flight in 2009 for the delivery of 2,000 kilograms of passive dry cargo to ISS with equivalent volume of trash or waste disposal removed from the complex. Dock time shall be not less than 30 days with an operational readiness date in April 2009.<br /><br />2.) Cargo Characteristics: - Unpowered, pressurized cargo and trash/waste. Average cargo density - 230 kg/m3 for launch cargo - 200 kg/m3 for trash/waste items at full capacity.<br /><br />3.) Launch cargo may include but is not limited to: - food containers - crew provisions and crew health care packed in half cargo transfer bags (CTBs) - logistics spares (all items no larger than the Russian Progress cargo ship hatch) - and other CTB packed cargo.<br /><br />4.) Late Access Cargo: provide late access/stowage for 10 percent of capability at launch minus four days. - trash/waste items may include but is not limited to: - trash packing foam (zotek, minicell or pyrell) - consumable logistics items, and waste consumables."</i><br /><br />So we have a bunch of numbers and facts here.<br /><br />-- 2000 kg of cargo<br />-- Cargo is 'dry' (one has to wonder if this means <b>no</b> water, or any water would have its own containers... i.e. 'dry' in the sense that the spacecraft doesn't have to provide native tankage).<br />-- Operational by April 2009<br />-- 30-days plus docking time required<br />-- Cargo does

 

[soyuztma]

The presentation charts of the 2nd Space Exploration Conference are now online. It includes the presentation given by Elon Musk of SpaceX and includes some details about the Dragon:<br />Standard Falcon 9 booster: 8500 kg total payload capacity to ISS orbit, allows ~3100 kg of cargo and/or crew total<br />Dragon: NTO/MMH Thrusters (RCS, OMS, and De-Orbit) <div class="Discussion_UserSignature"> </div>

 

[mrmorris]

Thanks for posting those links. Great information... plus I was starting to feel like I was talking to myself.<br /><br />The fact that Dragon will have hypergolic propellants is a bit disappointing, but not unexpected. N2O4/MMH is after all what I baselined the Dragon at. At this time, SpaceX has to be primarily interested in getting a working spacecraft out the door. Eliminating the hypergolics can wait for the next incarnation.<br /><br />The 8500kg figure is interesting... and confusing. Previously they'd indicated that the F9 would take Dragon to a 185km elliptical orbit -- from which it would have used its OMS to attain an ISS orbit. The text of the powerpoint plus the mass figure (8500kg vs. the F9 rated LEO capacity of 9300 kg) both seem to indicate that the second stage of F9 will be taking Dragon all the way to an ISS orbit. This isn't a bad thing -- although it makes recovery of the 2nd stage more problematic (if that's still the plan). It does reduce dv requirements for the capsule by a third or so.<br /><br />Changing the Falcon 9 MTO figure in my Dragon Builder application from 9300 kg to 8500 kg does interesting things to all of my calculations. 800 kilograms seriously reduces the amount of 'missing mass' that I've been agonizing over. Given a 'hard' propellant/oxidizer statement, I'll have to finish up my RCS redesign so I can finalize my mass and volume figures for it, and be able to use the calculated results rather than the simple 'RCS Mass' parameter I'm using now that's based on Apollo data.

 

[mrmorris]

Ah ha! AWST has an article explaining NASA's 2009 ISS Resupply RFI:<br /><br /><i>"NASA is seeking information from industry about the feasibility of launching an additional unmanned cargo resupply mission to the International Space Station (ISS) in 2009 to make up for an anticipated shortfall in dry cargo deliveries.<br />...<br />Although NASA is already preparing to fill the gap through its Commercial Orbital Transportation Services (COTS) program, routine deliveries to the ISS by COTS suppliers would not begin until the next decade. The anticipated shortfall in 2009 is two metric tons, and can't be made up by the space shuttle, according to NASA. "</i><br /><br />Basically NASA is telling them they <b>really</b> need a shipment sooner than they thought they would... <img src="/images/icons/smile.gif" />

 

[mrmorris]

Trolling around for Dragon tidbits, I found a photo of the prototype SpaceX has on their factory floor. It looks very little like the graphic SpaceX released. What struck me first was that it has a much larger hatch. However, this is not suprising -- they are more interested in being able to get in and out than being true to what a vehicle destined for space would actually have. <br /><br />The second thing I noted was that the shape doesn't match the released graphic. In my diagram, I performed a dimension estimate similar to what I did on the released graphic. I assumed the height of the door at 5' (although it might well be 6') and used this to estimate the height and largest diameter. Assuming a 5' door, the capsule specs would be ~4.6m x 2.3m (a 6-foot door would make it ~5.5m x ~2.75m). This contrasts to the graphic dimensions of ~4.4m x 3.8m. The ratio of the height to largest diameter is 1.16 for the graphic vs. 1.99 for the prototype -- quite a difference. <br /><br />Finally, the top comes to a point rather than a dome. I don't think this means anything significant at all other than a design change -- likely due to aerodynamic or heating considerations on re-entry.<br /><br />Mind you -- none of it really means anything. The prototype was simply something to give SpaceX experience in building a capsule structure, testing out their life-support system and placing people & equipment in a *very* claustrophobic space. The picture is interesting nevertheless...

 

[bpfeifer]

"Thanks for posting those links. Great information... plus I was starting to feel like I was talking to myself. "<br /><br />You're not talking to yourself. You're giving a fascinating lecture. Last I checked, your thread had in excess of 10,000 views. We really do appreciate what you're doing. We just didn't want to interrupt the professor... <div class="Discussion_UserSignature"> Brian J. Pfeifer http://sabletower.wordpress.com<br /> The Dogsoldier Codex http://www.lulu.com/sabletower<br /> </div>

 

[j05h]

what bpfeifer said! You're doing a very interesting analysis, keep up the good work.<br /><br />Josh <div class="Discussion_UserSignature"> <div align="center"><em>We need a first generation of pioneers.</em><br /></div> </div>

 

[mrmorris]

<font color="yellow">"You're not talking to yourself ... Last I checked, your thread had in excess of 10,000 views."</font><br /><br />Ah -- but that just means people are <b>watching</b> me talk to myself. <img src="/images/icons/smile.gif" /> <br /><br />But yes -- the thread views are encouraging. In any event, I'm having fun. I'd do this even without a public forum to distribute the results.

 

[mrmorris]

As I was driving home last night from work, I kept thinking about the photo of the Dragon prototype. My "Spidey-Sense" was tingling (which in my case would be better termed "You're Missing Something Obvious Again Ya Idiot-Sense"). About a third of the way home it struck me that the silver ring towards the top was almost assurecly the CBM. From memory, the dimensions were just about right. But... if the minor diameter of the capsule was determined by the CBM as with the Dragon-graphic, why was the major diameter not synced to the diameter of the Falcon 5/9? After a few more minutes of commuter combat, it struck me that the 'trunk' on the Dragon-graphic is of a constant diameter. However, if it were to flare outwards, this could easily account for the difference.<br /><br />Once I got home, I verified that with a 6' tall door, the "CBM" dimensions were too close to be coincidence. I then redid all my calculations using the CBM ring (83 inches) as my 'known' dimension to develop a scaled ruler. The results would be:<br /><br />Door: ~6'<br />Base: ~9' (2.74 meter)<br />Height: ~17.5' (5.33 m)<br /><br />I then CADed up an outline diagram of the capsule to scale, with the trunk/adapter (and minus heat-shield). Mind you, I have <b>no idea</b> how long the adapter was. The trunk on the Dragon graphic is roughly half the length of the capsule, so I simply assumed the same length for the original 'trunk'. It *will* be off some... but it doesn't really matter. The resulting diagram looks like... Gemini. I have to wonder just how far the resemblance stretched. The capsule on SpaceX's factory floor has no RCS. Of course if it were <b>truly</b> Gemini-oriented, the RCS/OMS would have been in the equipment section (trunk). Mind you, Gemini had a second RCS in the nose for re-entry control, so I don't know how they would have been handling that.<br /><br />So -- the middle outline diagram is the current Dragon capsule from the graphic -- to the same scale. The shape has <b></b>

 

[mrmorris]

Well I was able to reduce even further the difference between my <i>maximum</i> capsule mass that's based on Falcon- capabilities and my <i>calculated</i> capsule mass that's based on summing the elements that I've gathered together for Dragon. I wasn't taking into account anywhere the mass of the adapter/trunk structure. Right now I just put in 250kg as a placeholder figure, but I'll try to develop a number that has some research behind it. The 250 puts me in spitting distance of my calculated number. The closer I can get the two figures -- the more confidence that I'll have in the result.<br /><br />I have to make a couple more tweaks to the application to get mycalculated parachute and RCS mass figures accounted for rather than the placeholder parameters I used for my initial estimates. After that, I'm going to build a report that generates a mass breakout similar to the one I posted early in this thread, i.e.:<br /><br /><font color="orange">Mass: 5512 kg / 12151 lb <br /><br />Structural Mass: ~1800 kg <br />Heat Shield Mass: 200 kg **<br />Reaction Control System: 400 kg **<br />Recovery Equipment: 245 kg **<br />Navigation Equipment: 50 kg **<br />Telemetry Equipment: 20 kg **<br />Electrical Equipment: 70 kg **<br />Communications Systems: 25 kg **<br />Crew Seats and Provisions: 75 kg<br />Crew mass: 500 kg **<br />Miscellaneous Contingency: 200 kg <br />Environmental Control System: 125 kg </font><br /><br />Given that the <b>maximum</b> mass that I have for the Dragon capsule at the moment given the assumptions to-date is 4300kg, there's liable to be a good bit of difference from the above. All of the figures with a double-asterisk beside them will actually have a calculation or parts list based on reasonable data (or at least what *I* think is reasonable data, anyway). The single largest unknown is the structure, and even that I've based on the Apollo CM mass, albeit with a hefty fudge factor.

 

[mrmorris]

Does anyone know where I could find a copy of "AIAA-2005-0703-- Evaluation and Application of Apollo ECLS/ATCS Systems to Future Manned Missions". Mind you -- I'm aware that I could purchase it from AIAA... for $25 (not gonna happen). AIAA will let you look at the first page as a 'teaser'. Based on the abstract, I'd **really** like to get my hands on a copy.<br /><br /><i>"...the systems designed for Apollo would not be applicable to today’s spacecraft system development requirements due to more rigorous life cycle costs, design robustness, maintainability/reliability, safety and operational constraints. What would the Apollo design have looked like given today’s technology and design constraints? A brief but complete synopsis of Apollo mission ECLS and ATCS systems is presented detailing the system level design and operational limits. Multiple technology candidates and scenarios are developed and described with rationale for their adoption in the theoretical Apollo Lunar Mission of today. Technologies discussed include ECLS and ATC systems to maintain temperature and humidity control, CO2 removal/reduction, pressure control/equalization, trace contaminant control, waste management, fire suppression, and water management. Each Apollo technology is briefly described and then functionally equivalent subsystem choices are explored."</i><br /><br />The article was written by Grant Anderson and C. E. Martin of Paragon SDC. This is the best look into their ECLSS that I'm likely to find. I'd be able to refine my guesstimates on mass and volume of their ECLSS if I can locate a copy.

 

[mrmorris]

Well I did a first pass on generating a mass breakout report in my Dragon Builder application. The results actually aren't as far off as I was thinking. Of course part of that is because I goofed again. Have I mentioned that a large part of why I wrote the app was to eliminate stupid math errors? Looks like I added the structure twice in the mass total when I calculated an estimated mass from the breakout values at the beginning of the thread. The number is 1802 kg too high (dunno where the '2' came from). Anyway, that number should have been 3710kg. The mass summation from Dragon Builder comes to 4040 -- so my individual estimates weren't too far off (to be precise, they're not far off my current estimates. Obviously I can't say how close they are to the actual capsule).<br /><br />Looking at the calculated figures vs. my early estimate -- the figure furthest from the mark is the structure. This is off mainly because I put in 1800kg for my 'parameter' (read 'uncalculated guess') for the structural mass, and then to this was added 308 kg as the mass for the PCBM. In real life -- the PCBM forms part of the struture of the capsule, and so the 1800 really should have been reduced. However, as mentioned above -- the figure itself is simply a guess. I wouldn't be shocked to see the actual structural mass vary 200-300 kg on either side of the 2108kg figure. Since the Dragon is a unique structure, I have no means of getting a close estimate, and trying to 'fine-tune' the figure is pointless.<br /><br />The thermal protection system figures are extremely close -- amazing because I did *not* try to match the figures. I calculated the mass for enough AFRSI to cover the front of the capsule and enough ARA-SRAM-14 to provide a 4-cm average heat shield thickness in the rear. Mind you, the 4cm is a guess. I contacted one of the developers and asked for an off-the-cuff estimate on the thickness required. He replied that such information was proprietary and ITAR-restr

 

[mrmorris]

I figured I'd post the detailed mass breakout report to show the current components that make up the figures in the summary report above. If nothing else -- it'll be interesting to look back in a month or two and see what has changed. Just to be clear -- I <b>don't</b> expect that Dragon will have the exact makes/models of equipment that is shown here. Everything is simply a placeholder for the comparable equipment that must be in Dragon and is intended only to give me solid estimates for mass, volume, and power usage.<br /><br />I re-read the whole thread a couple of days ago to get a feel for how things have progressed. I'm pleased with how the reverse-engineered model is tightening up. I hope to keep spiraling in closer and closer on my estimates as I glean more information from data provided by SpaceX and do further research on the various subsystems.