Interesting info on real space computers....

[combat squirrel]

Just a random bit of info for you guys, the overall memory available for the entire Apollo flight computer, to fly it to the moon & back was equivalent to 72kb in modern terms....that\'s 0.07031 megabytes, or 0.00007 gigabytes, - if at one point we could write software where 3 people\'s lives depended on it, guiding a 3500 ton rocket at speeds approaching 7 miles per second, why cant Microsoft get it right with unimaginably more PC power ? haha

Just thought id post this as its cool how were now playing a full simulation of flying space vehicles in KSP , albiet it is simulating the entire world it works in, how big is the code for just navigation on auto pilot plug ins for example?

Challenge set for the awesome programmers of KSP (haha I only joke, maximum respect to you guys I love this program!!)

(Also is it possible to donate some funds to the project? I bought it already of course but I dont mind coughing up more money if it will help the development of this game)

[sal_vager]

Here\'s how NASA made the software for the Space Shuttle.

There\'s an Apollo computer simulator on the net somewhere, I\'ll post it if I can find it.

Edit:

Here is a javascript version of the Apollo lunar lander computer, it gives you an idea of what they had to work with back then

Plus more info on the computers used here at Wikipedia.

[LukeTim]

Not entirely sure what you mean by equivalent to 72kb in modern terms

72kb is always 72kb.

But, yeah, cool stuff.

The Computer aboard the Apollo spacecraft was actually so underpowered purely because cosmic radiation dictated that ICs and things had to be designed in such a way as to be more resilient to it... which limited the potential computing power. Computers at that time were not generally that limited.

Also, 72kb is plenty. What a dream, to have so much RAM. Try programming a PIC with 768 bytes of it.

[combat squirrel]

well yeah tis all very interesting, love how they coded the software into string...litterally

[LukeTim]

well yeah tis all very interesting, love how they coded the software into string...litterally

A friend told me about using Inductors or something to store data in a magnetic field... I had no idea how it could work. This is pretty fascinating though.

[thorfinn]

The Computer aboard the Apollo spacecraft was actually so underpowered purely because cosmic radiation dictated that ICs and things had to be designed in such a way as to be more resilient to it... which limited the potential computing power.

Oh, no I think they didn\'t even KNOW about SEU effects back then. It was just the best computer they could fit in the very limited size, weight, power and cooling conditions. Non-avionic computers didn\'t use ICs at all in that time period (at least during the design phase): IBM S/360 was discrete components, as were the PDPs of the time. The manufacture of the AGCs and missile computers was the thing that bootstrapped the IC industry.

Also, it was 72KB ROM, 4KB RAM.

[LukeTim]

Oh, no I think they didn\'t even KNOW about SEU effects back then. It was just the best computer they could fit in the very limited size, weight, power and cooling conditions. Non-avionic computers didn\'t use ICs at all in that time period (at least during the design phase): IBM S/360 was discrete components, as were the PDPs of the time. The manufacture of the AGCs and missile computers was the thing that bootstrapped the IC industry.

Also, it was 72KB ROM, 4KB RAM.

I see. Well, I stand thoroughly corrected.

[thorfinn]

I see. Well, I stand thoroughly corrected.

Don\'t take my word at face value too But I have worked on radiation hardening of electronics (for CERN, not aerospace) and I doubt that the primitive ICs of an AGC were susceptible to radiation effects.

A ionizing radiation particle can cause a 'soft error' in a computer because it is, well, ionizing as I said, and so while traversing the silicon it creates electrical charges of both signs that enter the electric circuit and can disrupt ist operation. But as you can imagine, this is felt much more if the circuit elements are very small: with transistors 65nm wide (or even less!), a modern chip is very susceptible and the problem is now concerning for aircraft computers too; but the comparatively huge features of these early SSI integrated circuits just see a little disturbance in the same event. I\'ve read quite a bit about Apollo computers, and I don\'t remember any talk about radiation issues. Also, most of the problems usually concern the memory banks. Core memory is inherently rad-hard, and that\'s why it was used in the Shuttle even if it was outdated by the time.

[Nibb31]

As someone on the Interwebs once said:

Our mobile phones have more computing power than all of NASA in 1969.

NASA launched a man to the moon.

We launch birds into pigs.

[Max Grant]

The manufacture of the AGCs and missile computers was the thing that bootstrapped the IC industry.

+1000. Without Apollo, neither Intel, nor Apple nor Microsoft would have existed -- at least, not as we know them today. I work in IT and I\'ve been aware of the major contributions of the space program for a long time.

Of course, what I love about this game is that it\'s coming along at a time when I see space exploration in a renaissance. I\'ll never be an astronaut, but I\'ll happily take being a Kerbalnaut, and I sincerely hope this game inspires a whole generation of real astronauts.

EDIT: In the early 1980\'s my dad modified a Chevy van with a bunch of cool things, and one of the things he did was get an onboard computer which we would use to calculate mileage and other things. Playing with that DSKY simulator strongly reminds me of how that little thing used to work. Very cool. I have no idea what kind of computer it was, who made it, or even how much it was. It was just another weird toy my Dad somehow got hold of.

[LukeTim]

Don\'t take my word at face value too But I have worked on radiation hardening of electronics (for CERN, not aerospace) and I doubt that the primitive ICs of an AGC were susceptible to radiation effects.

A ionizing radiation particle can cause a 'soft error' in a computer because it is, well, ionizing as I said, and so while traversing the silicon it creates electrical charges of both signs that enter the electric circuit and can disrupt ist operation. But as you can imagine, this is felt much more if the circuit elements are very small: with transistors 65nm wide (or even less!), a modern chip is very susceptible and the problem is now concerning for aircraft computers too; but the comparatively huge features of these early SSI integrated circuits just see a little disturbance in the same event. I\'ve read quite a bit about Apollo computers, and I don\'t remember any talk about radiation issues. Also, most of the problems usually concern the memory banks. Core memory is inherently rad-hard, and that\'s why it was used in the Shuttle even if it was outdated by the time.

I see. I don\'t know a whole lot on the subject, personally. I just found out that radiation is the reason they use CCD instead of CMOS sensors in space, when writing my Degree Dissertation.

It\'s all about the Vidicon, though. Shit\'s tight.

Can we just have a thread about general electronics in space?

[sal_vager]

Something else to bear in mind is that spare parts and repairs were not available to the crew of the Apollo missions and others, so computer systems were and are still chosen for their proven long term reliability, especially for satellites and probes which will have to operate with no visits from tech support.

Magnetic core memory had only just been patented in 1960, with Gemini and Apollo starting in 1961/62, I have to assume that core rope was invented some time earlier as I can\'t find a definite invention date, google, why do you fail me?

[ManaaniWanderer]

Poking at the wiki article it appears core memory was invented around 1949 so that\'s probably a good a date as we\'ll going to find with out getting an expert to tell us or going through hundreds of Google results.

[xclusiv8]

Something to remember, less lines of code equals less things that can go wrong. Belive it or not but the computes aboard modern airliners are running a Microsoft OS. It\'s around 10000 lines of code and has no bugs. Windows has millions of lines of code.

[jolee1868]

nice i always knew that modern phones blew the apollo software out of the water but i never had looked up the exact amount of memory it took. pretty amazing. I dont think most people in todays world appreciate just how complex yet in todays terms \'simple\' many things were. it really is a one of the wonders of the world.

[Ted]

Something to remember, less lines of code equals less things that can go wrong. Believe it or not but the computers aboard modern airliners are running a Microsoft OS. It\'s around 10000 lines of code and has no bugs. Windows has millions of lines of code.

I find it hard to believe that the on-board computers, more specifically the autopilot, contain just 10,000 lines of code.

In fact, after doing some research I have found out that they, specifically a Boeing 747, indeed, don\'t have just 10,000 lines of code. They actually have around 8 million.¹

Fun fact: on average there is one error per 1000 lines of code, statistically that means there are about 8000 errors within a Boeing 747\'s autopilot. Of course this relationship is exponential as the more parts of a program you have, the more likely they are to be incorrect, due to the scale of it.

Also, the above link shows that a Chevrolet Volt has 10 million lines of code. In comparison, the Linux kernel 2.6.35 has 13.5 million lines of code.^{2 [Warning: .pdf file]}

Finally, while maybe some modern aeroplanes do run a Microsoft Operating System, not all do. The Boeing 787 and Airbus A380, at the time of the publication of the linked article, ran \'Integrity\', an operating system that is also used in many military planes.³

[Foamy]

Another interesting fact about the Apollo computer is that it used the first Real Time Operating System (RTOS) since none of the existing system structures were able to run on such limited hardware. From there the RTOS moved to missile controllers and other military applications and then quickly into mainframes/corporate applications. After that it entered the consumer market until now just about everything with a screen runs one.

[samstarman5]

Came across this, figured it was relevant to the conversation:

[LukeTim]

To be fair to AT&T, they have a very limited bandwidth to play with and an awful lot of signals to handle. NASA basically has the whole radio spectrum and only a few signals to handle...

[Foamy]

Plus, it\'s the receiving equipment that\'s most important.

[Rocketscience101]

We got to the moon on a computer less powerful than your cell phone. Please tell me more how 3g isn\'t fast enough.

[sal_vager]

Bandwidth is nice but when you are communicating over such long distances, you mainly need power and sensitive receiving equipment.

On this page you can find info on what the Soviets used with the Venera space probes, they needed far more power than a modern cell phone is designed to handle.

[Ted]

We got to the moon on a computer less powerful than your cell phone. Please tell me more how 3g isn\'t fast enough.

I don\'t think data transmission rates are that dependant on the processing power of your phone.

[Nibb31]

Something to remember, less lines of code equals less things that can go wrong. Belive it or not but the computes aboard modern airliners are running a Microsoft OS. It\'s around 10000 lines of code and has no bugs. Windows has millions of lines of code.

The only 'embedded' Microsoft OS are Windows CE and Windows Mobile, and these are hardly ever used on critical embedded systems in the aerospace, defense or automotive industries for critical applications.

Serious embedded systems usually run a specialized real-time operating system (RTOS) such as VxWorks or QNX.

One of the reasons that aerospace computer hardware and software use old technology is often because it costs a lot of effort to develop, test, and certify new systems. When a system is robust and flight-proven, redesigning it involves more risk than it\'s worth.

[LukeTim]

I don\'t think data transmission rates are that dependant on the processing power of your phone.

They aren\'t. At all.