politik
Venus
on a crusade against ignorance
Posts: 83
|
Post by politik on Oct 22, 2005 11:52:21 GMT -4
He has a new introduction on his site. Not a new flash movie, just the opening paragraph after the movie. He has drawn a comparison of the shuttle program to the Apollo program, and even mentions the computing power available. Amazing really. He's starting to regress I think...
|
|
|
Post by nomuse on Oct 22, 2005 15:02:45 GMT -4
Hey...if Apollo didn't have the computing power to get off the ground, and the shuttle barely has the computing power to get off the ground, and Moore's Law says computing power doubles every two years -- then by 2018 we should be able to get to the moon with a toaster oven and some bread sticks.
|
|
|
Post by frenat on Oct 22, 2005 15:58:54 GMT -4
Mmmmmmm, bread sticks.
|
|
|
Post by gwiz on Oct 24, 2005 6:54:06 GMT -4
Kids today - they think you can't do anything without a computer. When I started work, the organisation had one computer. It had 64 words of memory, plus a 7k magnetic drum. I/O was via five-hole paper tape. There was a maintenance engineer based in the next room who spent most of his time changing thermionic valves.
Most of the design team never used it - the tools of the trade were slide rules, data sheets and mechanical adding machines. We still managed to produce aircraft that actually flew, went into service, kept the customers happy.
I believe the Gemini spacecraft was the first to actually carry a computer into space, but even then on one flight they demonstrated doing a rendezvous without the computer.
|
|
|
Post by sts60 on Oct 24, 2005 9:50:21 GMT -4
Sibrel doesn't have the excuse of being a kid. He's old enough to know better, but chooses to be ignorant and deceptive.
And this is the guy that margamatix has quoted so approvingly?
BTW, remember margamatix? I suppose he'll show up with some cock-and-bull story about driving to Russia or somewhere, when we all know that flimsy truck couldn't possibly have survived passage through the English Channel...
|
|
|
Post by dwight on Oct 24, 2005 14:51:36 GMT -4
Not only that but how can a truck with a steering wheel on the right hand side, drive on a road designed for left hand drive traffic??? ;o)
|
|
|
Post by Fnord Fred on Oct 25, 2005 17:11:11 GMT -4
And how could the truck survive all the polar bears and the cold? For a truck to have survived that trip it would need the sides to be at least a few feet thick with lead and on fire.
|
|
lenbrazil
Saturn
Now there's a man with an open mind - you can feel the breeze from here!
Posts: 1,045
|
Post by lenbrazil on Oct 25, 2005 22:15:31 GMT -4
He has a new introduction on his site. Not a new flash movie, just the opening paragraph after the movie. He has drawn a comparison of the shuttle program to the Apollo program, and even mentions the computing power available. Amazing really. He's starting to regress I think... How does he explain the Russians and Americans getting into Earth orbit with no [or even more primitive] computers? A friend of mine taught English at one of Brazil's two space centers. Relations between the military types who ran the base and "called the shots" and the engineers who ran the program were tense. The engineers all had to work with old XTs but the general's secretary had a top of the line [for the time] 486 notebook
|
|
politik
Venus
on a crusade against ignorance
Posts: 83
|
Post by politik on Oct 27, 2005 22:13:51 GMT -4
How does he explain the Russians and Americans getting into Earth orbit with no [or even more primitive] computers? He doesn't of course. I'm not sure what all the computing power was needed for in space. I'm sure anything that was needed to be computed could have been done on the ground before hand (LM guidance system notwithstanding). Surly NASA was launching commercial satellites by 1965. Didn't need a big fancy computer on those to get them or keep them in orbit.
|
|
|
Post by gwiz on Oct 28, 2005 3:20:10 GMT -4
Quite right, there were no computers on Telstar, Syncom, Early Bird, etc. Apart from Gemini mentioned above, IIRC, the OAO astronomy observatories had a computer-controlled attitude control system, possibly as an experiment rather than operational, and the Ranger/Mariner lunar and planetary probes had some sort of sequence controller, but I'm not sure if that would count as a computer in the modern sense. More general use of on-board computing, manned missions apart, didn't really arrive until the 1970s.
|
|
|
Post by PhantomWolf on Oct 28, 2005 4:00:52 GMT -4
Actually politik you are virtually totally correct there. The main part of the computing power was on the ground. Those computers crunched the numbers and all the computers on the ships needed were the results. If you read through the flight journals you can see them sending the crews the details of the burns and navigation needs. Basically a string of numbers they had to input into the DSKY which told the computer onboard which direction to point the craft. The computer power might have been limited, but it was enough to do the job it had too.
|
|
Bob B.
Bob the Excel Guru?
Posts: 3,072
|
Post by Bob B. on Oct 28, 2005 9:33:30 GMT -4
One somebody says the computers of the day were incapable of performing the required tasks, I always ask, what were the required tasks? Usually the HB hasn’t even considered what the computers actually had to do. I’m sure Jay can elaborate on this far more than I possibly can, but this way I understand it:
Spacecraft status - such as attitude, attitude rates, accelerations, etc. – was sensed by external instrumentation and this data was sent to the guidance computer. The computer would then compare this data against pre-programmed flight data, and if the actual status was outside the acceptable tolerances, the computer would send the appropriate command to fire a thruster, gimbal the engine, throttle the engine, etc. to bring the spacecraft back within tolerance. All the heavy number crunching was done on the ground, thus the job of the on-board computer was relatively simple. Furthermore, the guidance program was stored in ROM so only enough RAM was needed to perform the temporary calculations.
Jay, if my understanding is wrong, please correct me.
|
|
|
Post by colinr on Oct 28, 2005 10:53:43 GMT -4
Some of us old BAAB hands remember , with varying degrees of disgust, the ramblings of Cosmic Dave , a particularly stubborn HB who trotted the old nonsense regarding Apollo comuting requirements. When challenged by Jay and others , he had to admit that his entire experence of computing , beyone setting up a Web site, involved assembling PC's to run windows, nothing else. He had no experence beyond PC's mainframes- nothing - embedded systems - nothing - and so on .. Becasue the LM computer did have enough power to run a graphical user interface - then clearly is was unable to carryout any meanginful work - Some us older computing hands , (I hesitate to call myself a professional , despite 20 years of PC support), have learned that without silly interfaces, all computers are pretty impressive number crunchers - and are more than capable carying out the necessary calcluations to match velocity against a computed model .. hey my old HP Scientific calculator could run a pretty decent lander sim! In this case its most certinly a generational problem .. computers=PC's - and that's it!
|
|
|
Post by Count Zero on Oct 28, 2005 12:58:36 GMT -4
IIRC, Soyuz didn't get a computer until the '90s. The program sequencer was a rotating drum that simply spooled-out each event, one after the other. To execute a particular operation, such as re-entry, you simply switched-on the sequencer at the appropiate moment.
They had a close-call with this system when Soyuz TM-5 tried to land in September, 1988. There was a problem with infra-red sensors that delayed retro-fire by one orbit. On this second try, they got only a partial burn. While the ground assessed the situation, Afgan guest cosmonaut Abdul Ahad Mohmand broke protocol (sit-down & shut-up) and started hollering. He had noticed that the sequencer was still running, and was less than a minute away from jettisoning the engineering module. Without that, they had no power except for their batteries, and no way to maneuver - either to de-orbit or return to Mir. Commander Vladimir Lyakhov was able to shut it off in time. Soyuz TM-5 returned safely to Earth the following day.
|
|
|
Post by JayUtah on Oct 28, 2005 13:42:48 GMT -4
Bob, you're correctly describing the digital autopilot program. That was one of a small handful of "background" programs run on the guidance computer at any given time.
The AGC hardware was identical between the CM and the LM, except for the form factor of the DSKY. Each ran the same operating system -- a fairly sophisticated time-sharing system that maintained a "foreground" task on the DSKY, as many time-sharing tasks as resources permitted, and a few real-time tasks.
The user interface was operated by the PINBALL program that implemented the "noun" and "verb" abstractions for instructions keyed from the DSKY. PINBALL began as the lamp test program!
There was a low-level software library that implemented high-precision numbers and vector datatypes and algorithms beyond the native types and operations.
Even more astonishing, there was a small interpreter for pseudocode that was mixed in liberally with machine code in the programs -- somewhat similar to IBM's old "I/O channel programs". This was a clever way of expanding the instruction set without redesigning the CPU and without consuming a lot of memory.
The application software suite was COLOSSUS for the CM and LUMINARY for the LM. When speaking specifically of one platform or the other, CGC is the CM's AGC running COLUSSUS and LGC is the LM's AGC running LUMINARY. COLOSSUS and LUMINARY had about 20% common code and were designed to be as symmetrical as possible to operate. For example, P52 was the IMU realignment program in both systems. The famous P6x programs in LUMINARY operated the LM during powered descent. Programs could start or terminate other programs; when one program exited it often specified the program that should be run next in that slot -- e.g., P63 transitioned to P64 without operator intervention. Programs could promote themselves to the foreground and "grab" the DSKY (hence the foreground program number was always displayed on the DSKY).
The various global data structures such as the state vector were stored in known addresses in the computer and could be accessed by programs that needed to examine them or update them. PINBALL referred to many of these through the "noun" shorthand. Noun 43, for example, in LUMINARY was the ground track information, stored in three adjacent locations in erasable memory. You could instruct PINBALL to display "noun 43" continuously, and it would use the three display registers of the DSKY to give a running count of latitude, longitude, and altitude. Another program running on the LGC would update them periodically (e.g., once every few seconds) from basic orbital mechanics equations.
So the CGC and LGC did some number-crunching, but not much.
This same computer organization is used in the space shuttle computers. The computers switch from one major mode to another as the flight progresses. Originally the computers had to be loaded with each major mode as needed, but the modern equivalents (IBM's AP-101S general purpose embedded computer) can contain all the flight programs simultaneously.
And yes, modern computers waste an enormous amount of computing power on flashy interfaces and the various levels of abstraction that make modern home and business computing possible. Such frills are not appropriate to an embedded system, nor is such flexibility necessary. It is possible -- even advisable -- to program to the "bare hardware" in those contexts.
There is indeed the generational problem of computers. Modern software is not written especially efficiently; it is meant to be gotten out the door as soon as possible, not to make as frugal use as possible of the hardware resources. A modern PC bears very little resemblance to embedded controllers of any era.
The other notion is that computers must be used for certain tasks just because it's commonly done. Maybe it's a bad example, but my kitchen currently contains an Apollo-era stove/oven. It employs an analog control circuit (the schematic is printed on the back of the appliance) to regulate the heating. Pretty brainless and pretty hard to screw up. But after more than 30 years of loyal service, it's going away because its cooking temperature is starting to fluctuate by as much as 75 F.
I'm shopping for a replacement. Even the least expensive cooking appliances on the market have some kind of digital microcontroller in them.
When I was younger I used to play around with washing machines. They had a sort of analog sequencer that makes a good example for designing mechanical control systems. We used to use cam/follower designs all the time for automatons in haunted houses; in fact, that's what Disney's audio-animatronics were based on: pretty much the same thing as a clothes washer. A shaft containing a set of cams turns at a prescribed rate, and the followers implemented mechanical actions or closed circuits. So, for example, you'd have one cam that operated the water fill valve. It would have a raised portion to indicate in which parts of the cycle the valve was to be opened. Level sensors provided overflow safety.
Nowadays washing machines are all controlled by microcontrollers. A computer program running on a twelve-dollar microprocessor monitors all those inputs, watches the clock, and issues commands to valves and motors. It's very much cheaper to do it that way, and you have more flexibility in the kind of system you can engineer if you have a general-purpose digital controller.
But the point is: how much computing power does it take to roast a turkey? How much to wash a pair of jeans? Nothing about any of those tasks inherently requires a digital logical solution. It's simply cheaper and easier -- now that computers are cheap and easy to program -- to engineer a control system using those tools. Conspiracy theorists don't comprehend the difference between the best ways to do things and the possible or impossible ways of doing things.
|
|