heat sinks, radiators, and happy handwaves of the heat problem

[BwapTED]

I like the idea that Traveller starship and power plant fuel usage, capacity, and refueling requirements includes heat sinks. Just fold heat sinks into the category of ''fuel'' the way reaction mass is folded in (assuming reaction M drives)

Related to that, I suggest shipboard computers are big because their size includes large heat sinks.

Does this make sense?


Does TNE ship design include radiators?


Anybody just blow off the heat problem?
Or use a more exotic handwave?

 

[mike wightman]

Heat sinks tied in with the gravitics and acceleration compensators.

And yes, TNE does indeed cover radiators.

 

[Condottiere]

Computing will progressively grow more efficient in hardware, and probably software.

Trying to predict it seems the likely reason volume allocation was eliminated in Mongoose, as the iFone would be at least the equivalent of a factor one.

 

[BwapTED]

Computing will progressively grow more efficient in hardware, and probably software.

Trying to predict it seems the likely reason volume allocation was eliminated in Mongoose, as the iFone would be at least the equivalent of a factor one.

True.

Computers could also be big because they are ''retrotech."

Big old clunky computers like we see in a lot of early and mid- 20th Century sci fi.

Traveller is (usually) set in the future. But not necessarily our future.

 

[mike wightman]

I've never seen the problem with the size of ship computers - they are the equivalent of supercomputers not desktops :CoW:

Can an I-phone solve an n-body problem? In real time? Now add a hyperdimensional element; that is what a model 1 computer must be able to do.

While also controlling the fusion reactor, gravitics, acceleration compensation, environmental systems etc...

Back to the topic of heat sinks - Traveller power plants put out an awful lot of energy if you go by the HG2/Striker conversion rate of 1EP is 250MW. There is going to be an awful lot of waste heat to get rid of.

 

[Condottiere]

Indicate the ram and processing power required.

A technological level seven factor one can do that, and they only cost thirty thousand CrImps and take up a room.

And probably still use magnetic tape.

 

[nobby-w]

Computing will progressively grow more efficient in hardware, and probably software.

Trying to predict it seems the likely reason volume allocation was eliminated in Mongoose, as the iFone would be at least the equivalent of a factor one.

If you look at LBB3, A Model/1 is available from TL5, and an iPhone has substantially more computing power than a TL7 supercomputer such as a Cray XMP. Avionics need relatively little computing power to run per se. The machines that ran the Apollo command module or space shuttle had about as much computing power as an 1980s vintage home computer.

However, If you think of a ship's computer, it's not just a computer. It is a suite of computers, avionics and fire control systems, ECM systems, sensor arrays and antennae, shielding, conduits and wiring throughout the hull, cooling systems and redundant backup power supplies. Being rated for space it's all rad-hardened and fault tolerant, and much of it is at least quad-redundant so you can lose an element and still use a byzantine generals calculation for error correction.¹

For a jump capable starship, you could also assume that controlling a jump drive in jump space requires a quantum computer, which adds insulation and a bulky cryogenic cooling system into the mix.

¹ The byzantine generals algorithm works by comparing the outputs of three or more calculations and eliminating one that does not agree with the other two. Quad redundant systems are needed so that the byzantine generals calculation still works if one node in the system goes down.

 

[Carlobrand]

Since waste heat correlates with power output, I figured HG '08 folded waste heat management into the power plant allocation and into the allocation for those devices that needed dedicated waste heat management systems.

(Communicating by cell today. I'll try to avoid obvious errors, but I have big fat man-size fingers and these keys are so tiny. I don't know how the kids do this without getting cramps in their hands.)

 

[nobby-w]

Can an I-phone solve an n-body problem? In real time? Now add a hyperdimensional element; that is what a model 1 computer must be able to do. [ . . . ]

In practice, for the purpose of navigating a starship, the mass of the starship will be low enough that it will not significantly perturb the orbits of planet-sized masses. This makes the notion of solving N-body calculations in realtime somewhat irrelevant, although it can certainly be approximated quickly enough for me to enjoy a nice evening playing Kerbal Space Program.

Once you have solved the orbital mechanics of the star system, you can use approximations such as sphere of influence, or run the calculations in advance of doing the actual burns.

KSP uses spheres of influence as an approximation for N-Body physics, although there is at least one plugin that puts in a better N-body simulation than the one in stock KSP.

[ . . . ] While also controlling the fusion reactor, gravitics, acceleration compensation, environmental systems etc...

These will have their own embedded controllers and network wiring, which, of course, contribute to the total mass of the ship's 'computer.'

Back to the topic of heat sinks - Traveller power plants put out an awful lot of energy if you go by the HG2/Striker conversion rate of 1EP is 250MW. There is going to be an awful lot of waste heat to get rid of.

Yes. Not one of the high points of the OTU canon, I'm afraid.

 

[Condottiere]

Current practice is to try to minimize heat through minimizing power the computers need, which we know helps them go faster, and why piling on speed tends to give diminishing returns.

I believe the Cray One (mid Seventies) is listed at eighty megahertz and the iPhone 7 / 7 Plus: 2.34 GHz 64-bit quad-core Apple A10 (2x Hurricane + 2x Zephyr), plus three gigabytes of RAM.

Technically, we should be somewhere within level eight, not having discovered yet how to manipulate gravity or spark off a contained fusion reaction.

If you need to spread around computing power, use a dozen iPhones, or better yet, a Samsung, or even a Chinese rip-off, which would be a third to half cheaper.

As I recall, the Space Shuttle used four eight sixes, which probably had the same speed as the Cray One.

 

[mike wightman]

The space shuttle wasn't nuclear powered - look to the size of the computers in a nuclear sub or CNV to get an idea of how big the computer of a starship should be.

The server in my school fills a room...

 

[aramis]

Heat sinks tied in with the gravitics and acceleration compensators.

And yes, TNE does indeed cover radiators.

Which make converting CT capital ships interesting, to say the least. Some of them are unable to make CT armament levels as there's no non-radiator surface area left....

 

[Condottiere]

Control equipment for fusion reactors is likely included, with shielding, in the volume overhead.

 

[Ex_Pendable]

Any shipboard system will usually be more massive that its civilian planetside counterpart as shipboard systems are very ruggedized to protect and shield the internal components from radiation, shocks, and impacts. The case will look more like a small floor safe rather than a flimsy metal box.

Using synthetic diamond rather than silicon wafers will greatly extend any computer's ability to handle heat, but regulating the temperature is important - it can't be too hot or too cold. Gas and liquid cooling may be used, the later hooked up to a plate heat exchanger to transfer the system heat to one of the ship's cooling loops.

 

[nobby-w]

[ . . . ]
As I recall, the Space Shuttle used four eight sixes, which probably had the same speed as the Cray One.

The original avionics systems on the Space shuttle had rad-hardened 6809s running a RTOS. A 6809 was an 8-bit chip with about the same computing power as an Apple-II. The system was quad redundant.

Integer performance of a 486 would have been 100x or more the CPU speed of the shuttle's avionics computers. The Apollo missions got to the moon on even less powerful hardware. The fly-by-wire system in a F-111 had 4k of memory, and the avionics in a F22 have a CPU that's roughly as powerful as a Playstation 1 (although the DSPs in the Radar would be considerably faster).

Cray-1's had either two or four vector units with a peak throughput of around 170 MFLOPS. Their integer unit was fairly quick but the machine was optimised for floating-point vector bashing. A 486 would have had comparable integer performance to a Cray-1, but much lower floating point performance. A modern GPU like a GTX1080 has a peak floating point throughput of about 10,000-50,000 times that of a Cray-1.

As an aside, the gate count of the CPU on a Cray-1 was about 200,000 - which was a lot for a machine built in the mid-1970s, but was not much more than half of the gate count of an intel 386 and about 6-7% of the gate count of a 1st-gen Pentium. IIRC it had 8MB of RAM where a typical mainframe of the mid-1970s would have had something like 256K-1MB and a typical minicomputer would have had between 16K and 64K.

 

[Condottiere]

No one doubts redundancy would be built in, which made combat damage degradation such a questionable issue.

Getting stuck in space because your iPhone components are sourced from the lowest bidders is somewhat dumb, but at technological level nine, you could probably have an application installed on it, and you could dock it on the bridge and run the spaceship with it, while listening to your music mix.

It may be just more attuned to your personal preferences on the spaceship control interfaces, while there are a dozen back up iPhones at default setting spread around the vessel.

It would be one of the reasons that there are minimum factors required to run larger spaceships, so a factor two would be a Samsung Galaxy.

 

[BwapTED]

No one doubts redundancy would be built in, which made combat damage degradation such a questionable issue.

Getting stuck in space because your iPhone components are sourced from the lowest bidders is somewhat dumb, but at technological level nine, you could probably have an application installed on it, and you could dock it on the bridge and run the spaceship with it, while listening to your music mix.

It may be just more attuned to your personal preferences on the spaceship control interfaces, while there are a dozen back up iPhones at default setting spread around the vessel.

It would be one of the reasons that there are minimum factors required to run larger spaceships, so a factor two would be a Samsung Galaxy.

Can starships be hacked like phones?

Maybe this is another reason for crews. You want manual override on everything, and certain functions always under human control.

 

[Condottiere]

If you have a cyber flavour adventure, definitely.

 

[kilemall]

Can starships be hacked like phones?

Maybe this is another reason for crews. You want manual override on everything, and certain functions always under human control.

Which is why IMTU the robot brains are not allowed to operate ships unless they are operating from human UI controls physically, ship computers are kept offline and isolated (so business/entertainment/comms are on separate computers), and there is no such thing as remote control for port landing or anything else.

Even a hacked small craft can do epic amounts of damage, accidentally or intentionally.

The overriding principle is the ship's pilot and commanding officer is responsible for any course or actions, even if robots are doing the driving.

 

[Carlobrand]

Which is why IMTU the robot brains are not allowed to operate ships unless they are operating from human UI controls physically, ship computers are kept offline and isolated (so business/entertainment/comms are on separate computers), and there is no such thing as remote control for port landing or anything else.

Evan a hacked small craft can do epic amounts of damage, accidentally or intentionally.

The overriding principle is the ship's pilot and commanding officer is responsible for any course or actions, even if robots are doing the driving.

Oh, no one would ever hack HAL.