Sideways Engineering the SpaceX Dragon

[josh_simonson]

The falcon 5/9 avionics will traverse the van allen belts for GEO launches, so either the avionics will be rad hard or the 3x redundancy is enough to push radiation failure odds down into the noise.

 

[mrmorris]

<font color="yellow">"...falcon 5/9 avionics will traverse the van allen belts for GEO launches..."</font><br /><br />There's a timeframe involved in hardening high-radiation environments. If I have a digital camera that is not waterproof, and I take it with me to photograph coral reefs on a diving expidition -- it will almost assuredly fail within seconds of being in the water. This type of 'instant' failure isn't true of non rad-hardened processors in high-radiation environments.<br /><br />For GEO launches -- the actual time in which the SpaceX avionics will be running until the payload is deposited into GTO is on the order of minutes. The chances of the avionics failing are (as you said) so unlikely as to be relegated to noise. The LEO missions will have longer durations, but a more benign environment.

 

[mrmorris]

I've been out of town for a week -- haven't had a chance to post or do much on this thread. I'm currently developing a database application to help me keep track of all the specs/hardware/etc. I'm gathering for the sideways-engineering project. Ideally it will perform a lot of the calculations that I've been doing manually to this point. That way if I add some equipment or change a spec, everything will be recalculated automagically. Anyway -- until it's done (or I give up on it) -- posts may well be random and infrequent.<br /><br />Anyway -- I ran across the specs for the various docking/berthing elements to ISS. This is something I **really** wanted for G-X3 but never came across. I wanted to post them here to keep them handy:<br /><br /><b>Passive Common Berthing Mechanism (PCBM)</b><br />Weight: 680 lbs2 (440 lbs PCBM + 240 lbs hatch)<br />Max OD: 86.3"” dia<br />Hatch Pass Through: 50"” square<br /><br /><b>Androgynous Peripheral Docking System (APAS)</b><br />Weight: ~950 lbs (660 lbs APDA-6001 + 276 lbs avionics) (hatch not incl.)<br />Max OD: 69" dia<br />Hatch Pass Through: 31.38"” dia<br /><br /><b>Russian Probe</b><br />Weight: 700 lbs (550 lbs cone + 150 lbs avionics)<br />Max OD: 61"” dia<br />Hatch Pass Through: 31.5"” dia (approximate)<br />

 

[mrmorris]

Well I had wondered if there was possibly room in the section of the Dragon containing the CBM to store the descent parachutes (ala Apollo). This is the main reason I spent the time looking for CBM specs. However, at this time I can state with considerable certainty on this subject: <b>no flipping way</b>! <br /><br />The diagram below is one I created a couple of weeks ago to get a better set of dimensions on the Dragon. As I mentioned earlier in the thread -- my first go-round was based on the tried-and-true (albeit very low precision) 'ruler-to-the-screen' method. For the diagram below, I still used the Dragon image that SpaceX has released, but I pulled it into Paint -- created a meter-stick based on the one known dimension (i.e. the max diameter should match the Falcon 9 diameter) -- then used that meterstick to measure the remaining dimensions. The accuracy is still dependant on the image being to scale, of course.<br /><br />From the diagram (and assuming my dimensions are reasonably accurate), the exterior diameter of the CBM-section of the Dragon varies from 2.4m to 2.2m. Given an outer diameter of 2.2m for the CBM -- there's simply no way anything else will fit.

 

[mrmorris]

I started working on the dv & propellant estimates a couple of weeks ago. I quickly got bogged down in a couple of problems. First -- to calculate the propellant needed to generate a given dv -- you use the mass of the craft. Once you have calculated the amount of propellant -- the mass of the propellant you just added throws off the calcs you just made. So you have to recalculate to add propellant for the propellant mass you just added. Then you have to recalculate to add propellant for the propellant mass you just added...<br /><br />Second -- I still wasn't happy with my mass estimates -- and the prospect of spending a bunch of time re-iteratively calculating propellant needs -- just to have it all invalidated when I changed either the mass or the dv estimates was not appealing.<br /><br />This was what made me decide to write a database application to automate these calculations. It's more work up front, but if I change the parameters in the future -- everything will be recalculated automagically. The app is coming along reasonably well -- if more slowly than I'd like as it's getting written in my spare time. However, it's reached the point where I'm ready to start calculating dv. In looking back at this doc to program in the calcs, I figured I'd update this a bit and post it. The numbers will almost certainly change, but the logic behind my dv calculations should remain reasonably constant... I hope.<br /><br />I used the following as my starting base dv budget -- largely pulled from the Apollo MODAP document and tweaked a bit:<br /><br /><b>dv budget:</b><br />Each 1-degree Plane Change: 100 m/s<br />Hohman Transfer: 200 m/s (300km elliptical - /> 400km circular orbit)<br />Rendezvous & Docking: 150 m/s<br />Contingency: 100 m/s<br />Deboost dv: 250 m/s<br /><br />I know that I said the LES would take a back-seat in this doc -- but it seemed contraindicated to design the RCS/OMS for Dragon without taking it into account. Once I added it to the mix -- co

 

[josh_simonson]

The main requirement for the LAS is that it be able to fire from the pad and get the capsule high enough to engage it's parachute. <br /><br />Typically a LAS is jettisoned after the second stage fires, which saves considerable payload because it departs at a low delta-v. I wouldn't be too surprised if they found that carrying the LAS equipment all the way to orbit was a losing proposition and switched to a jettisonable version.<br /><br />It seems to me that a simple hybrid booster would make a good LAS, and it could be recovered in the ocean and re-used much easier than the first or second stage could.

 

[mrmorris]

<font color="yellow">"I wouldn't be too surprised if they found that carrying the LAS equipment all the way to orbit was a losing proposition "</font><br /><br />Either I don't understand what you're saying or you don't understand what I am. As described -- this isn't an LAS/LES system once in orbit. At that point, the solid rocket motors that *would* have provided for an abort are instead tasked to provide dv for the Hohman transfer and the de-orbit burn. If the SRMs had been jettisoned at SS ignition, then Dragon would instead need to carry additional propellent/oxidizer for its conventional RCS/OMS to perform these burns.<br /><br />There is still a mass penalty. The SRMs will have cases and nozzles -- generally only about 90% of the mass of an SRM is propellant. Also, the iSP is likely to be somewhat lower than the liquid thrusters of the RCS. However, with the method you detailed, the booster will have to carry the additional mass of the propellant/oxidizer for the RCS/OMS <b>plus</b> the mass of a dedicated abort system through the first stage burn. Of the two options -- it's almost certain that the dual-purpose SRMs will prove to have a lower mass. I may try modelling both when (if) my application gets to the stage that this is reasonable.<br /><br /><font color="yellow">"... it could be recovered in the ocean and re-used..."</font><br /><br />I haven't studied hybrids enough to make statements with any assurance of being correct. However, I expect that they are likely manufactured somewhat similar to solids in that the propellant material will have a grain pattern designed to help control the speed of the burn. After a dunk in the ocean, this pattern would be filled with seawater and multiple impurites. I doubt that reusing it would be an option. Even the Shuttle SRBs are only marginally less expensive to refurbish than expend.

 

[mrmorris]

An article at TheSpaceReview recounted some information that Tim Hughes od SpaceX had given on the Dragon. Some of the key points to add to my store of specs:<br /><br /><font color="orange">Upmass:<br />1400 kg of pressurized cargo<br />1700 kg of unpressurized cargo</font><br /><br />Given a Falcon 9 payload limit of 9300kg -- this would imply the Dragon capsule plus 'trunk' tops out at 6200kg. I have difficulty believing that they'd use the <b>absolute</b> maximum in presenting figures -- after all, telling NASA in a year or two that the payload figures have gone up is a heck of a lot more palatable than having to tell them that they've gone down. Given that -- I can't see putting the max mass being over 5500kg.<br /><br /><font color="orange">Downmass <br />~1400 kg of pressurized cargo (actually he said 'similar' but the cargo section is gone, so presumably he was referring only to the pressurized portion.</font><br /><br />This should help me in sizing the parachutes. Speaking of which -- one of the two attached photos shows a landing with parachutes. The graphic does indeed imply Apolloesque triple-chutes as I assumed. It also shows them all originating from a single attach point (also as expected). What is completely <b>unexpected</b> is where the attach point appears in the graphic. If that photo is accurate -- then the attach point is near the intersection of the crew and CBM portions of the craft. Either the chutes are stowed there (where they found space I haven't a clue), or the chutes have an attach-point channel similar to that used with Gemini. As I mentioned earlier in the thread -- the thought of that channel always bothered me. Besides that -- the attach point is immediately forward of the top hatch... which would preclude such a channel. Even further -- there's an RCS cluster in the equipment section directly 'behind' that mount point. The only reasonable answer is that they've found space in the CBM section... I'll have t

 

[mrmorris]

OK -- I'm reasonably sure I know where they hid the parachutes. I'm going to try to draft it out shortly. For the life of me I can't locate a <b>complete</b> set of dimensions for the PCBM, but I took the one picture I have of it and did a dimensioning excercise similar to that I did on the Dragon. Unfortunately, the image I found is angled, so my approximations are more approximate than usual. <img src="/images/icons/smile.gif" /> I had to assume that the angle of the vertical tilt affected the horizontal measurements in proportion. If reasonably close -- the PCBM is only about 11 inches in length, whereas the CBM section of the Dragon is ~20" long. The outer-dimensions of the PCBM are 87.6", but the inner hatch is only a 50" square. There should be room in a doughnut-shaped area on the inside facing section of the PCBM for the parachutes. My CAD drawing should give a better feel for how large this space is.

 

[josh_simonson]

Everyone flying hybrids these days are using re-useable ones (except when they crash) nobody has fired them high enough yet that they burn up on the way down. The plastic or rubber fuel grain should be pretty resistant to the effects of water, but it could be protected with a water-tight nozzle plug if needed. W/C a new fuel grain is relatively cheap.<br /><br />The most efficient use of an LAS would probably be to fire it shortly after SS ignition but not separate the capsule, instead adding it's delta-V kick to the SS's, then jettisoning the spent LAS motors. That's bad for reliability though, but so is relying on the LAS for OMS purposes - only volunteers to light off a bunch of solids 1 meter from your body and adjacent to your heat shield while in orbit? <br /><br />It's definitely more weight efficient to add more fuel to the existing liquid OMS/RCS system than to add a completely separate solid LAS/OMS system in addition to the liquid OMS/RCS.

 

[mrmorris]

<font color="yellow">"Everyone flying hybrids these days ..."</font><br /><br />What specifically do you think that hybrid buy in this instance over solids? I understand the tech -- solid propellant, liquid (generally) oxidizer. It's therefore possible cut off the oxidizer to stop the reaction, as is not possible with traditional solids. However -- for a LAS -- the 'reaction' from start to finish is about four seconds in duration... so what is being gained?<br /><br /><font color="yellow">"...volunteers to light off a bunch of solids 1 meter from your body and adjacent to your heat shield while in orbit?"</font><br /><br />Why do you say this? What are the problems that you envision?<br /><br /><font color="yellow">"It's definitely more weight efficient..."</font><br /><br />Nothing's definite except death and taxes. If you have figures -- please provide them. Otherwise, you're providing a personal opinion, not making a statement of fact.

 

[mdodson]

Why a database application? Would things go smoother with a spreadsheet? I'm not critical, I love watching what you're doing!

 

[mdodson]

"-- to calculate the propellant needed to generate a given dv -- you use the mass of the craft. Once you have calculated the amount of propellant -- the mass of the propellant you just added throws off the calcs you just made. So you have to recalculate to add propellant for the propellant mass you just added. Then you have to recalculate to add propellant for the propellant mass you just added... "<br /><br />Wouldn't the "rocket equation" http://en.wikipedia.org/wiki/Rocket_equation take care of that for you?<br />

 

[mrmorris]

<font color="yellow">"Why a database application?"</font><br /><br />Mangling a catchphrase horribly -- When you're a carpenter, you use a hammer to solve all your problems.<br /><br />I'm a database application developer -- ergo if I need something to store a bunch of information and perform calculations, I use a database application. I could do most of this in a spreadsheet -- certainly the basics. It wouldn't be faster both because I'm a DB programmer by trade, and because the primary holdup isn't the programming per se, but the fact that I'm building without a blueprint. I'm designing this application as I build it. I have to decide what I want it to store/calculate/display -- then locate (or estimate) the data, dig up the calculations I need and then find some way to integrate them harmoniously into what I already have.<br /><br />I'm attaching a graphic of what I have so far. It's pre-Alpha, but is starting to come together... a bit.

 

[mrmorris]

<font color="yellow">"Wouldn't the "rocket equation"..."</font><br /><br />Hadn't considered using the RE, quite honestly. I tend to think of that equation in terms of designing a booster that will start out sitting on a pad with the <b>intent</b> of heading into orbit -- not for something there already.

 

[mrmorris]

OK -- the CAD file demonstrates that while the width of PCBM is such that it extends completely to the outer dimension's of the Dragon's outer surface (the reason I thought there could be no room for the parachutes in the CBM section initially), its length is small enough that there is still plenty of space for the parachutes in the CBM section of the Dragon. From this, it's also obvious that the PCBM has much to do with the specific shape of the Dragon. The <i>maximum</i> diameter of Dragon was Picked as the diameter of the Falcon 9, of course. With this, we see that the <i>minimum</i> diameter of Dragon (discounting the nose dome) was dictated by the outer dimension of the CBM. The remaining question was simply how much distance to have between those two.<br /><br />An image is attached (everything is rounded to 5" increments -- I'm not not *that* worried about tolerances). Clearly there is a gap in a ring behind the CBM analogous to the one where the parachutes were stowed on the Apollo CM. The Dragon will have a 'tunnel' from the crew section to the CBM hatch. The hatch is a 50" square, but I expect that the tunnel will be circular, which would mean at least a 71" diameter to clear the corners of the square hatch (I'll use 72" just to make it an even six feet). This will give us a toroidal space around the tunnel with an inner radius of 72" and an outer radius ranging from ~91" to ~95". For volume considerations I'll split the difference and use 93". The volume of this would be ~1.8 m3 (~63.0 ft3). This is ample for our needs. The equivalent "Apollo Toroid" would have been ~15" in length, with an inner ring of ~28" diameter and an outer one ranging from ~31" to ~55" (split to an average of 43" for this calculation). This resulted in an area in the realm of 0.82 m3 (29.0 ft3). This is only 46% of the volume available to the Dragon. Given this -- it's obvious that the parachutes won't be the only items stored in this space. I'm changing my terminolo

 

[mrmorris]

I figured I'd upload the diagram I used/modified to determine the volume of the compartment where the Apollo CM parachutes were stored...

 

[mrmorris]

<b>D'oh!!</b>. I've speculated in two separate posts now what the difference between the Crew and Cargo versions of the Dragon would be, and specifically whether the cargo one would still have ECLSS capability built in to make the same capsule functional for both missions. In the two posts I decided (for different reasons) that the ECLSS likely was *not* included... but it turns out there was never a need for speculation. SpaceX has long since declared that the cargo version will have neither ECLSS nor manual flight controls.<br /><br /><i>"To ensure a rapid transition from cargo to crew capability, the cargo Dragon and crew Dragon are almost identical, with the exception of the crew escape system, the life support system and onboard controls that allow the crew to take over control from the flight computer when needed."</i><br /><br />I suppose it's nice to have my guesses confirmed... <img src="/images/icons/smile.gif" />

 

[josh_simonson]

>"Everyone flying hybrids these days ..."<br /><br /> />What specifically do you think that hybrid buy in this instance over solids? I understand the tech -- solid propellant, liquid (generally) oxidizer. It's therefore possible cut off the oxidizer to stop the reaction, as is not possible with traditional solids. However -- for a LAS -- the 'reaction' from start to finish is about four seconds in duration... so what is being gained?<br /><br />Hybrids are safer. Especially if you're firing them every flight. They're also much cheaper to re-use, especially if you're firing them every flight.<br /><br /> />"...volunteers to light off a bunch of solids 1 meter from your body and adjacent to your heat shield while in orbit?"<br /><br /> />Why do you say this? What are the problems that you envision?<br /><br />If one fails it's very likely to be a LOC event. If they fire every flight, not just in an emergecy, the odds of one failing goes up. At some point the risk due to the LAS would outweigh the risk of flying without one.<br /><br /> />"It's definitely more weight efficient..."<br /><br /> />Nothing's definite except death and taxes. If you have figures -- please provide them. Otherwise, you're providing a personal opinion, not making a statement of fact.<br /><br />Liquid or gas propulsion systems have higher ISP than solids, so the amount of extra fuel would be lighter than the fuel in the solids, plus you're just adding the tankage to hold that extra fuel, rather than the casings, nozzles, RCS, attchment hardware, electronics, ect that are all required for the solid system but already exist in the liquid one.

 

[mrmorris]

<font color="yellow">"Hybrids are safer."</font><br /><br />What are you basing this statement on? Solids have been in production use for decades. Hybrids aren't in mainstream use yet. If you feel the basic technology is safer and this is where your statement comes from -- then what aspect of hybrids is it that you feel makes them safer than pure solids?<br /><br /><font color="yellow">"If one fails it's very likely to be a LOC event."</font><br /><br />1. If a hybrid of comparable thrust fails... what exactly happens?<br /><br />2. What failure do you envision happening? These SRMs would be comparable to those used by Gemini -- namely a ~15" sphere & external nozzle with (IIRC )<100 kg of fuel. There's no seals to cause a failure as with the Challenger disaster. The fuel burns for 4-5 seconds, so the chance of a nozzle overheating in that brief of time are almost nil. Likewise, if there's a defect in the propellant that causes the burn to be uneven, too fast, etc. -- there's not enough time for forces to build up which would cause a catastrophic failure. What exactly is the event that you believe is likely to cause a loss of control?<br /><br /><font color="yellow">"If they fire every flight, not just in an emergecy, the odds of one failing goes up."</font><br /><br />Agreed -- whatever the failure percentage is on any given system (that would include hybrids and liquid thrusters, of course) -- the more times you use a particular system, the more likely a failure may occur. Solids are about as simple as it gets... so the chances of them failing would actually be *lower* than that of liquids or hybrids -- both of which have several times the number of components and failure modes.<br /><br /><font color="yellow">"Liquid or gas propulsion systems ..."</font><br /><br />Gas? Are you talking cold-gas thrusters? If not -- what?<br /><br /><font color="yellow">"...the amount of extra fuel ... but already exist in the liquid one.</font>

 

[mdodson]

"Hadn't considered using the RE, quite honestly. I tend to think of that equation in terms of designing a booster that will start out sitting on a pad with the intent of heading into orbit -- not for something there already."<br /><br />But that usage will give you an answer that's 35% low since it doesn't account for atmospheric drag.

 

[mrmorris]

<font color="yellow">"...an answer that's 35% low..."</font><br /><br />I don't believe that drag constitutes 35% of the propellant/oxidizer usage. Anyway -- I still haven't gotten around to the dv & propellant calculations at that level yet. I just finished working them from another direction. I like coming at my numbers using two or more unrelated methods to act as a sanity check. Right now I'm looking at the dv of the Soyuz and the orbiters and using that as a basis for what the Dragon needs. I'l have a post on that shortly.<br /><br />Most recently, I wanted an estimate of the mass of the Dragon capsule that was unrelated to my current method (namely summing up estimates of the structure, plus my chosen equipment, TPS, etc.). I decided to work backwards from the Falcon IX uplift capacity of 9300kg. SpaceX has stated Dragon's capacity at 1400kg of pressurized cargo and 1700kg of unpressurized cargo to the ISS. This leaves ~6200kg of free uplift capacity for the booster. With the cargo out of the equation, we're left with the spacecraft and the propellant/oxidizer. If we come up with a reaasonably close estimate for the P/O -- then the remainder would be... Dragon.<br /><br />Looking at the Soyuz TMA -- from Astronautix, it masses ~7,220 kg with 900 kg of propellants for a total of ~8,120kg. If we assume that the Dragon masses at the full 9,300kg capacity of the F9, and a similar iSP, and that the propellant requirement scales up reasonably on a 1:1 basis to mass (a shakier assumption, but the mass numbers aren't far off, so we shouldn't miss by <b>too</b> much), then the Dragon's propellant would mass around 1000kg for a dv capability comparable to Soyuz. Subtracting this out of our figure above, we have a mass of ~5,200 kg for the Dragon capsule (plus the 'trunk' structure).<br /><br />Essentially every one of these numbers has wiggle-room. I expect SpaceX to be deliberately understating pressurized and unpressurized cargo mass numbers. Like

 

[mrmorris]

I have somewhat more sympathy for NASA's woes with the Mars Climate Orbiter (i.e. the Great Metric Mishap). My dv numbers a few posts back were way high because the numbers in the NASA doc I pulled them from were in ft/s and I used those numbers unchanged with m/s as the scale. Anyway -- I've decided to use a different method for determining the dv requirements for Dragon in any event. The MODAP document was just a concept study. It would be better to use dv numbers from the Shuttle and Soyuz -- seeing as they've both actually *made* the trip that Dragon needs to. I'm going to use ft/s in this post, but my application is all metric so far, so I'll convert the results to metric when using them there.<br /><br />The shuttle orbiter OMS pods each can provide 500 ft/sec for a total of 1000 ft/sec of dv. They can shift propellant/oxidixer from the rear RCS thruster storage for somewhat more, but we'll use 1000 ft/s as our 'low' dv required for Dragon. The Soyuz has 1,279 ft/s, which makes for an arguable 'high' value since it's actually got somewhat more dv than is required for ISS operations. From STS-63 and STS-74 (the only two I've found detailed dv budgets for), the orbit-matching burns averaged about 240 ft/s and the de-orbit burns from Mir averaged about 440 ft/s. The 1,000 ft/s dv of the orbiters allow for a reasonable (if not extravagant) amount of maneuvering, whereas the Soyuz has conserably more 'wiggle room'. I couldn't find any dv usages specifically for ISS operations.<br /><br />Since I'm setting up my application to calculate this crud -- I'm going to set 'low' and 'high' values to be calculated. Initially I'll assume the dv will be between 1,000 ft/s and 1,300 ft/s -- with propellant volumes, masses, etc. calculated for the two extremes using a mass of Falcon 9's mass-to-orbit limit. <br /><br />One of the problems that surfaces as a result of my conversion blooper is that the dv required for the LES is now essentially at the upper end of *all* the

 

[mdodson]

I stand corrected! 20% http://en.wikipedia.org/wiki/Delta-v_budget<br /><br />It's a significant amount, though.<br />

 

[mrmorris]

<font color="yellow">"I stand corrected! 20%"</font><br /><br />The only thing I can find at that link regarding atmospheric drag is:<br /><br /><i>"Launch to LEO — this not only requires an increase of velocity from 0 to 7.8 km/s, but also typically 1.5–2 km/s for atmospheric drag and gravity drag"</i><br /><br />1.5 / 7.8 = 19.2 %<br />2 / 7.8 = 25.6%<br /><br />However, the note is combining atmospheric and gravity drag. Gravity drag will far exceed atmospheric drag. Is there something else at that URL I'm missing indicating 20% for atmospheric drag? That still seems very high to me.