Date: Fri, 27 Nov 92 05:04:37 From: Space Digest maintainer Reply-To: Space-request@isu.isunet.edu Subject: Space Digest V15 #464 To: Space Digest Readers Precedence: bulk Space Digest Fri, 27 Nov 92 Volume 15 : Issue 464 Today's Topics: (!) Shuttle Computer Problems Computer synchronisation by GPS (2 msgs) escape systems Hubble's mirror Shuttle Computer Problems Shuttle replacement (9 msgs) Titan, DC, etc. (2 msgs) US/World News in Brief Nov 26 10 am PST VSA: "The Military Aspects of the Soviet Cosmonautics" ( the new Maxim Tarasenko's book ) What comes after DC-1 Welcome to the Space Digest!! Please send your messages to "space@isu.isunet.edu", and (un)subscription requests of the form "Subscribe Space " to one of these addresses: listserv@uga (BITNET), rice::boyle (SPAN/NSInet), utadnx::utspan::rice::boyle (THENET), or space-REQUEST@isu.isunet.edu (Internet). ---------------------------------------------------------------------- Date: 26 Nov 92 20:58:47 GMT From: "Michael F. Santangelo" Subject: (!) Shuttle Computer Problems Newsgroups: sci.space,comp.arch,comp.arch.storage I'm finding this a wee-bit hard to take that someone at KSC is posting to usenet with what seems like a critical design problem of the shuttle's computer systems and asking for suggestions on how to fix. Well, for what it is worth, I will re-post it to a couple more relavent groups to see what people think. This JUST appeared in our nn database (~3:55PM EST) in group sci.space.shuttle. ---start forwarded article--- waterman@titan.ksc.nasa.gov writes: >Monday November 23 10:20 EST >I'm looking for some help out here in the net. Any physists or experts >in electron physics out there? > The problem we're having is with both the General Purpose Computers >(GPCs) onboard the shuttle as well as the Main Engine Controllers (SSMEC). >Both experience single bit upsets while in orbit. The GPCs have one bit >error correction and log these upsets in what we call a soft error counter. >For the SSMEC, it has no error correction. We dump the SSMEC memory after >landing and determine how many upsets have occurred. Well today when the >SSMECs were powered up on Discoveries three Main Engines, the standby >computer on engine 3 had one bit flipped from the last time it was powered >up (about a week ago). The GPCs have also experienced soft errors while at >the pad. The current theory is that high energy particles striking the >memory cell impart energy which changes the state. My question is if >high energy particles can change the memory on the shuttle sitting on the >pad. Why aren't all the other computers in the world inflicted with the >same problem? Has anyone heard any studies being done on this? >Some info, the SSMEC uses 8K static RAM chips to make up the 64K by 16 bit >main memory. When power is removed the memory is held up by a 3.56V >battery. Todate it is believed that all bit flips have occurred at this >lower battery voltage (power up voltage is 5V). This of course can not >be proved (could have happened at full power in memory that was not being >accessed). >Thanks. >--------------------------------------------------------------------------- >Bob Waterman [Aqua] >waterman@titan.ksc.nasa.gov >--------------------------------------------------------------------------- -- -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Michael F. Santangelo + Internet: mike@cbl.umd.edu Computer & Network Systems Director + mike@kavishar.umd.edu UMCEES / CBL (Solomons Island) + BITNET: MIKE@UMUC University of Maryland + Voice: (410) 326-7237 (direct) + (410) 326-4281 x237 ------------------------------ Date: 26 Nov 92 12:12:47 GMT From: Alan H Jones Subject: Computer synchronisation by GPS Newsgroups: sci.space In article , jeuck@unix.sri.com (Philip Jeuck) writes: |> An interesting footnote to this is GPS time is exactly 8 seconds ahead |> (behind?) UTC time. I have never seen an explanation of why. Most GPS |> clocks correct for this and report UTC time but not all navigation units |> correct for it. So if you have a receiver that was meant for positioning |> you might not want to use it for timing without checking it against WWV or |> some other source of UTC. |> Phil Jeuck |> jeuck@unix.sri.com The reason for the integer offset in seconds between GPS and UTC is that UTC has leap-seconds inserted from time-to-time. This is a great annoyance if you write software that uses time differences, and would probably leave bugs latent until the leap-seconds arrived. For this reason, GPS time does not include leap seconds. To help users that want to convert to UTC, the satellites send down the integer number of seconds offset to UTC with each timing message, so that when the leap second occurs, this integer changes by 1. Personally, I use only GPS time in the computational parts of my software, and convert to/from UTC using the latest offset only when displaying or obtaining values from a user that has requested UTC. To quote from the "Technical Characteristics of the Navstar GPS": >"The Control Segment shall maintain GPS time relative to Coordinated Universal >Time (UTC) to within 176ns >(95%) plus accumulated leap-second jumps. The >difference between GPS time and UTC shall be included in the >navigation >message. >SPS timing accuracy with respect to UTC shall be 363ns (95%)" Cheers from Alan Jones. ------------------------------ Date: 26 Nov 92 12:29:04 GMT From: Alan H Jones Subject: Computer synchronisation by GPS Newsgroups: sci.space With refernce to the GPS time to UTC offset: I should also add that for users that want high accuracy UTC, the GPS satellites broadcast parameters such as fractional offset (in 2E-30s units) and rate of change of offset (in 2E-50s/s units). Cheers. ------------------------------ Date: Thu, 26 Nov 1992 21:26:19 GMT From: Henry Spencer Subject: escape systems Newsgroups: sci.space In article roberts@cmr.ncsl.nist.gov (John Roberts) writes: >-[Apollo] used either one, depending on timing. The escape rocket >-was jettisoned after it was no longer needed. > >Didn't Apollo eventually quit using escape rockets? I don't have photos handy to check, but as far as I know, every Apollo ever flown (the last being the Apollo-Soyuz mission) carried the escape tower. In fact, it was kind of hard to avoid, because the Apollo capsule was covered by a protective cover during early ascent, and it was tower jettison that got rid of it -- the cover went with the tower. -- MS-DOS is the OS/360 of the 1980s. | Henry Spencer @ U of Toronto Zoology -Hal W. Hardenbergh (1985)| henry@zoo.toronto.edu utzoo!henry ------------------------------ Date: Thu, 26 Nov 1992 20:44:21 GMT From: "robert.f.casey" Subject: Hubble's mirror Newsgroups: sci.astro,sci.space >henry@zoo.toronto.edu (Henry Spencer) writes: > >>The point of an end-to-end imaging test, in this context, is not that >>it's an easy way to detect the problem, but that it's so hard to >>argue with. Well, they could have said "oh, the reason it's out of focus is the Earth's gravity field is distorting the mirrors. Has to be that, the big mirror tested great before!" :-( ------------------------------ Date: Thu, 26 Nov 92 19:49:44 EST From: John Roberts Subject: Shuttle Computer Problems -From: mike@starburst.umd.edu (Michael F. Santangelo) -Newsgroups: sci.space,comp.arch,comp.arch.storage -Subject: Re: (!) Shuttle Computer Problems -Date: 26 Nov 92 20:58:47 GMT -I'm finding this a wee-bit hard to take that someone at KSC is posting -to usenet with what seems like a critical design problem of the -shuttle's computer systems and asking for suggestions on how to fix. I wouldn't call it a "critical design problem". And it's interesting that somebody considers Usenet a potential source of useful information. (After all, that's what we want, isn't it? If we post what we consider to be good ideas, we'd like the project people at NASA, etc. to read them and consider them.) You left out sci.electronics. John Roberts roberts@cmr.ncsl.nist.gov ------------------------------ Date: Thu, 26 Nov 1992 18:25:11 GMT From: Henry Spencer Subject: Shuttle replacement Newsgroups: sci.space In article hugh@whio.cosc.canterbury.ac.nz (Hugh Emberson) writes: >Will the DC-{X,Y,1} have gimballed engines? I can understand that the >DC-X might need then since it only has a small number (~4) of engines, >but will DC-Y need them? DC-Y will have more engines but I think they're still going to gimbal them. >Surely when you have a large number (~20) of engines around the base >of a rocket you can steer the rocket by throttling some of the engines >down a bit. Most aerospike designs -- the ones that put the engines around the base -- do seem to do thrust vectoring by differential throttling. I've never seen numbers on the effectiveness of this, but it would seem feasible. However, DC-Y is not going to use an aerospike; it will, by current plans, use more or less conventional bell-nozzle engines with telescoping nozzles. This was considered a bit less uncertain. Nobody has ever flown an aerospike. (Nobody has ever flown nozzles that telescope during burn, either, but it looks more straightforward.) -- MS-DOS is the OS/360 of the 1980s. | Henry Spencer @ U of Toronto Zoology -Hal W. Hardenbergh (1985)| henry@zoo.toronto.edu utzoo!henry ------------------------------ Date: Fri, 27 Nov 92 00:29:28 GMT From: amon@elegabalus.cs.qub.ac.uk Subject: Shuttle replacement > course without smartly contacting the ground. Rockets coming *down* > are an entirely different issue. An engine or control failure means > a crash of a vehicle containing *rocket* fuel, perhaps miles from > intended touchdown, like main street in Disney World for example. > C'mon Gary, don't be silly. A landing DC-1 has nothing on a 747. If you insist on landing the DC-1 at sea, then in all honesty you will have to insist we build off shore mile long floating runways for all jumbo jets: maybe even for smaller jets as well since the DC-1 really won't be all that huge. Won't even top the empennage of the 747 when sitting on its little shock absorbers, and certainly nowhere near the volume of the jumbo. As Henry and I think others have been pointing out, Jet-A is far more explosive than a LOX/LH spaceship, and particularly a spaceship with nearly EMPTY LOX/LH tanks. After all, the Shuttle tanks ripped apart with nearly a full load of fuel and didn't explode. They Challenger was ripped apart by aerodynamics, remember? Where's the beef? ------------------------------ Date: Thu, 26 Nov 1992 21:28:43 GMT From: Henry Spencer Subject: Shuttle replacement Newsgroups: sci.space In article <1992Nov26.160614.19313@ke4zv.uucp> gary@ke4zv.UUCP (Gary Coffman) writes: >>A bit less zenophobia please! what the hell do you think an airliner is? > >Something with *wings* on it. Just think of DC-1 as a high-performance helicopter. -- MS-DOS is the OS/360 of the 1980s. | Henry Spencer @ U of Toronto Zoology -Hal W. Hardenbergh (1985)| henry@zoo.toronto.edu utzoo!henry ------------------------------ Date: Thu, 26 Nov 92 14:02:23 PST From: Brian Stuart Thorn Subject: Shuttle replacement Newsgroups: sci.space We weren't talking about costs per pound to orbit, Allen. You made a claim that the Titan IV with the new solids would close most, if not all of the cargo capacity gap with Shuttle. I was simply pointing out that this is not true. I consider a 10,000 lb advantage to be considerable, not negligible. Of course, Shuttle is much more expensive to operate. I don't argue that. But this was *not* the argument. -Brian ------------------------------ Date: Fri, 27 Nov 92 00:46:50 GMT From: amon@elegabalus.cs.qub.ac.uk Subject: Shuttle replacement > The problem of both systems is that crucial parts are not redundant. > This is nothing new, and the problem is *not* restricted to VTOLs. > The F-111 that lost a wing pivot, early in the program, pinwheeled too. > The problem is inadequate redundancy in lift producers, not VTOLness. > Aye, and not to mention the DC-10 at O'Hare that dropped an engine and did a wingover into the ground. ------------------------------ Date: Thu, 26 Nov 1992 21:50:51 GMT From: Henry Spencer Subject: Shuttle replacement Newsgroups: sci.space In article <1992Nov26.161842.19428@ke4zv.uucp> gary@ke4zv.UUCP (Gary Coffman) writes: >>Have you told the FAA about this? They've certified many types of airliners >>for flight with very full tanks and no destruct systems. > >The proposed DC-1 is nothing like an airliner, airliners have wings and >a limited gliding ability. Actually, DC-1 has limited gliding ability too, although it's not very useful at low altitude and speed. (Mind you, get a 747 below its stall speed and it doesn't glide well either...) >Nor is the proposed DC-1 anything like a >helicopter which can autorotate if there is a power failure. That's right, it has far more engine redundancy than any helicopter, so it doesn't need autorotation. DC-1 is *better*; it can make powered landings where helicopters would have to autorotate (a much dicier procedure). >Nor do >either of these systems require high performance rocket fuel. What matters is not whether the fuel is "high performance", but what the total energy content is. The energy content of a 747 with full tanks is a lot higher than that of a landing DC-1. A hundred tons of kerosene makes a much bigger mess than a couple of tons of LH2/LOX. >On a related note, you said the proposed DC-1 would land on nearly empty >tanks. Does that mean it can't abort an approach and try again? "Nearly empty" here is relative to being 9/10ths fuel at takeoff. There's still plenty for a controlled landing with ample margins. But you miss a point: a VTOL doesn't *need* to abort an approach. Only things relying on aerodynamic lift need to abort approaches, because they have to maintain a minimum speed or they fall out of the air. A VTOL can simply stop, shuffle sideways to correct the misalignment or whatever, and continue. >Licensable aircraft *must* be able to do this. Flying on fumes is >an FAA violation. My understanding is that McDD has talked to the FAA, and they see no fundamental problems with the notion of certifying DC-1 as an airliner. They're unofficially satisfied that the concept could deliver adequate safety. Much depends on the details, of course, and those haven't been nailed down yet. -- MS-DOS is the OS/360 of the 1980s. | Henry Spencer @ U of Toronto Zoology -Hal W. Hardenbergh (1985)| henry@zoo.toronto.edu utzoo!henry ------------------------------ Date: Thu, 26 Nov 92 14:32:12 PST From: Brian Stuart Thorn Subject: Shuttle replacement Newsgroups: sci.space Allen... Actually, returning Solar Max to Earth was an option throughout the mission. Your mistake (and a mistake I've seen from others regarding Hubble Space Telescope) is assuming that LDEF had a cradle. It didn't LDEF was secured to attachment fittings on the payload bay wall 'sill'. When LDEF was returned on STS-32, Columbia did not have a special cradle. Solar Max was placed in a pallet at the aft end of the payload bay. Had the repair been unsuccessful, astronauts would have removed the solar panels and Max turned 90 degrees, parallel to the payload bay floor. Regarding the 'abort switch' just inches away from an astronaut. Well, no, I'm afraid. In 'Men From Earth', Buzz Aldrin describes the Apollo abort switch as a handle below and beside the commander's couch. His hands were not inches away from it. However, I still say that this is beside the point. I don't believe any astronaut would have had time to pull any lever or press any switch in the period between Titan SRM failure and total conflagration (milliseconds) or that an escape motor could have pulled a spacecraft out of said conflagration in the available time. But this argument will be hard to prove either way. Let's just hope this situation never happens. Regarding the RL-10, something went haywire in your message, and I can't figure out what you were trying to say. However, my argument was that the RL-10, which will power the DCX, has failed twice in flight in the past 18 months. New or reusable makes no difference, as the Centaur engines are all new. Further, one of the few other cryogenic engines, Saturn's J-2, had a two-engine failure on at least one mission (Apollo 6, April 4 1967) and I believe also on Apollo 13. I'm not saying that the DCX is dangerous because of it's RL-10s, but these facts should at least give it's engineers pause. -Brian ------------------------------ Date: Thu, 26 Nov 92 14:40:53 PST From: Brian Stuart Thorn Subject: Shuttle replacement Newsgroups: sci.space I'll probably get in trouble for saying this, but it seems pretty obvious to me: RL-10 HISTORY, 1986-1992 1986 - RL-10 is working fine. Centaur is working fine. 1987 - RL-10 gets reworked to accomodate more Atlas-Centaur's use of Shuttle-intended RL-10s. 1992 - RL-10 fails twice in eighteen months. If I were Pratt & Whitney, I'd start looking at the Shuttle mods! Oh, and by the way, exactly HOW MUCH MODIFICATION will DCX RL-10s require? ------------------------------ Date: Thu, 26 Nov 1992 23:27:51 GMT From: Henry Spencer Subject: Shuttle replacement Newsgroups: sci.space In article <70359@cup.portal.com> BrianT@cup.portal.com (Brian Stuart Thorn) writes: >1986 - RL-10 is working fine. Centaur is working fine. >1987 - RL-10 gets reworked to accomodate more Atlas-Centaur's use of > Shuttle-intended RL-10s. >1992 - RL-10 fails twice in eighteen months. > >If I were Pratt & Whitney, I'd start looking at the Shuttle mods! They are. There's nothing very obvious, but the reworked RL10s may perhaps have tighter margins in some areas, tight enough to fail when conditions aren't quite right. (Remember that only one engine of the pair has failed to ignite each time, and indeed on the recent one, the other engine lit properly *twice* -- the onboard controller retried the ignition sequence.) >Oh, and by the way, exactly HOW MUCH MODIFICATION will DCX RL-10s require? Shorter nozzle for sea-level operation, and minor modifications (just finer valve control, I think) for throttling. -- MS-DOS is the OS/360 of the 1980s. | Henry Spencer @ U of Toronto Zoology -Hal W. Hardenbergh (1985)| henry@zoo.toronto.edu utzoo!henry ------------------------------ Date: Thu, 26 Nov 1992 18:09:05 GMT From: Henry Spencer Subject: Titan, DC, etc. Newsgroups: sci.space In article <70267@cup.portal.com> BrianT@cup.portal.com (Brian Stuart Thorn) writes: > Oh, PLEASE! Do you HONESTLY believe a crew would have survived the > April, 1986 Titan 34-D launch failure? Like Challenger, there was > no warning... only close scrutiny of the telemetry indicated a > problem (well after the fact). I doubt even an Apollo LES would > have been able to fire... If I'm remembering that Titan failure correctly, I'd say a good LES should have been able to do it, because it was an on-board destruct system that actually blew up the SRBs (after detecting massive abnormalities in one of them). > assuming, of course that a crew would have known to pull the lever. Why assume it has to be manual? The Apollo escape system was partly automatic, since some classes of failures -- major explosions, for example -- were considered likely to happen too quickly for the crew to handle it. >>If your in Shuttle, true enough. On the other hand if your in a Delta >>Clipper made of much simpler components (less likely to fail) and with >>abort modes throughout the entire flight envelope it is a great deal >>of comfort. > >Except of course, the RL-10 engine, which has now failed in flight twice >in the past eighteen months... Note that each time, it's been only one out of two engines failing. The DC designs can survive a single-engine failure at any time. Furthermore, these are ignition failures, not explosions. Depending on exactly how you do the design, an ignition failure in a launcher that has all engines firing at launch can be just a pad abort. There is also a strong suspicion that these recent failures are the result of rebuilding Shuttle/Centaur RL10s for Atlas-Centaur use. -- MS-DOS is the OS/360 of the 1980s. | Henry Spencer @ U of Toronto Zoology -Hal W. Hardenbergh (1985)| henry@zoo.toronto.edu utzoo!henry ------------------------------ Date: Thu, 26 Nov 92 14:36:53 PST From: Brian Stuart Thorn Subject: Titan, DC, etc. Newsgroups: sci.space No, sorry. The April, 1986 Titan 34D launch ended when one of its two Solid Rocket Motors suffered catastrophic failure and exploded, taking out the whole vehicle. Range Safety and the onboard destruct systems were never fired. I believe the incident you are referring to took place in September, 1985. -Brian ------------------------------ Date: Thu, 26 Nov 1992 22:15:54 GMT From: "Mitchell E. Gold" Subject: US/World News in Brief Nov 26 10 am PST Newsgroups: sci.space In article clarinews@clarinet.com (UPI-Radio) writes: A University of Florida astronomer says a two-mile wide asteroid is approaching Earth and will give astronomers a rare close-up view December 8th. Dan Dyurda says the asteroid will not pose a threat when it crosses the Earth's orbit... coming within 2-point-2 Million miles of the planet. -- Can anyone venture a guess as to the possibility of realigning Galileo to get a shot of this? Any more information on how close it'll be? --Mitch Gold, sundance@rpi.edu ------------------------------ Date: 26 Nov 92 10:28:41 GMT From: "Voevodin S.A." Subject: VSA: "The Military Aspects of the Soviet Cosmonautics" ( the new Maxim Tarasenko's book ) Newsgroups: sci.space Copies of new Maxim V. Tarasenko's book "The Military Aspects of the Soviet Cosmonautics: A Side View" (in Russian) are available now. If you want to get the book please write to vsa@msd.orbi.kostroma.su ( Sergey A. Voevodin ) for further information. ------------------------------ Date: Thu, 26 Nov 1992 22:05:49 GMT From: Henry Spencer Subject: What comes after DC-1 Newsgroups: sci.space In article <1992Nov26.163541.19527@ke4zv.uucp> gary@ke4zv.UUCP (Gary Coffman) writes: >The Russians don't refuel cryo systems on orbit. Big difference. In any case, the Russian refuelling system is not really suitable for a main propulsion system with big tanks. However, there is nothing that awful about pushing cryogenic fluids around if you prepare carefully. If worst comes to worst, link your tanker and tankee nose-to-nose and put 1rpm or so of spin on the combination to provide enough fake gravity to settle fluids. >>>Could we strap on SRB's to the DC-1 to increase payload... >>Well, people have proposed it (not SRB's but strap ons). It defeats >>most of the purpose and adds cost but it might be worth while if DC1 >>turns out to be more marginal than expected. > >We *know* from experience that weight always increases and performance >always decreases on the way from the drawing board to the launching >pad. SRBs may be the only way the proposed DC-1 ever gets to orbit. If so, the project will basically have failed. One hopes it won't come to that. It shouldn't; the project has been reviewed by several groups, and their conclusions have consistently been that weight growth may eat into the payload but won't reduce it past zero. DC-1 with strap-ons wouldn't fulfill the real promise of SSTO, but it would still probably be a much cheaper and better launcher than the current ones. Small solid strap-ons are not expensive, complex, or manpower-intensive. >... Everything about the proposed DC-1 requires >operation on the thinnest of margins. Any system pushing that close to >the limits *all* the time is likely to be finicky and dangerous if it >works at all. That's why SSTO has never been seriously considered before. SSTO has been seriously considered repeatedly, actually. The margins are smaller than for a two-stage system, in the same way that a two-stage system's margins are smaller than a three-stage system's. Considering the fact that 1.5 stages made it into orbit in 1958, one would think that the margins for SSTO should now be adequate, especially if it's not done as a national megaproject that will have failed if the payload is 49999lbs rather than 50000lbs. -- MS-DOS is the OS/360 of the 1980s. | Henry Spencer @ U of Toronto Zoology -Hal W. Hardenbergh (1985)| henry@zoo.toronto.edu utzoo!henry ------------------------------ Date: Thu, 26 Nov 92 20:30:20 CST From: Daniel Pancamo SUB Daniel Pancamo ------------------------------ End of Space Digest Volume 15 : Issue 464 ------------------------------