How to fix SUSE after installing a new motherboard (new disk controller)

I am running SUSE 11 on one of my computers and had a motherboard die on me. I got a new and different motherboard, and figured that I could just attach my old disk and boot up the system, hoping that SUSE would somehow recognize the new hardware and seamlessly update. NO!!

The system would boot, but partway through it would stop with a message like:

Trying Manual resume from disk sda1 
Want me to fall back to /dev/sda1 (Y/N) 
Waiting for device /dev/sda1 to appear ……… not found — exiting 
to /bin/sh 

#

Running system repair from the DVD didn’t help.

I considered re-running the installation from DVD, but I was worried that somehow my existing system would be re-formatted, re-partitioned, etc, losing everything for me. Every guide for re-installing starts with “first be sure to back up your current system” — something that wasn’t easy for me to do.

Here is how I got my system working again:

1) Boot from the dvd, select “Rescue System”

2) Login as root

3) Mount the hard disk:

  #mount /dev/sda1 /mnt

4) Fix the /dev directory (this one stumped me for a while):

  #mount –bind /dev /mnt/dev

5) Chroot to the disk system:

  #chroot /mnt /bin/bash

6) Fix the initial ramdisk (the source of the problem):

  #mkinitrd

7) Reboot — now the disk is recognized and the system boots. Yay!  However, the graphics card is not recognized, so X will not start.

8 ) Run the X configurator:

  #sxa2 -r

 

Now I am back up and running with a new motherboard!

Hope this info helps anyone else with a similar problem.

Solar Energy and Concentration

I think the near-term future of our energy supply should be solar energy — which is amusing in the sense that I’ve always thought that the future of our energy supply has been fusion energy, and solar energy *is* fusion energy.  The fusion reactor is already there, we just need to harness it.

I think the fundamental challenge in harnessing solar energy is the fact that it is so diffuse — both in space and time. As any quick estimation can tell you, at ~1000 Watts per square meter times the entire surface of the earth, the total solar energy available is enormous. But on the other hand, a square meter of anything, with the possible exceptions of water and dirt (a.k.a. farming), is expensive compared to the amount of energy available. 

So, the obvious solution is to somehow concentrate the solar energy so that it can be harnessed by a smaller, and therefore cheaper, convertor. This leads to the already established “concentrator photovoltaic” and “concentrated solar thermal” technologies:

 

Solar Concentrator

These use some sort of *optical* concentration of the solar energy, and I’ve had some thoughts on other optical concentrators. Optics are expensive, though.

However, here is an idea:  *Physically* concentrate the sunlight. “How?!?” you might ask…

Well, there are materials that could, in principle, temporarily store the energy in light. If we spread one of these materials out in the sun (maybe with a dump truck and a grader) and let it soak up the solar energy, then scoop up the material into a tank or even just a pile, and somehow trigger the rapid release of its stored energy (probably as heat, maybe catalyzed by high temperatures) we might be able to use this heat to make steam and run a turbine or something similar. A nice feature of this idea is that we could make steam night and day — steam turbines don’t like to be throttled up and down, which seems to be one of the major problems with current solar thermal technology.

The big obstacle to this idea is the energy storage material, and its cost. One simple idea for a material that could reversibly store solar energy is azobenzene. It undergoes reversible photoisomerization — i.e. expose it to UV light, and the trans-azobenzene converts to cis-azobenzene. Expose cis-azobenzene to blue light and it converts back to trans-azobenzene. If we trust wikipedia, then the energy difference between cis and trans azobenzene is 50 kJ/mole.  (Compare this to 100-200 kJ/mole for burning hydrocarbons.)

How cheap would azobenzene need to be (or how quickly would we need to recycle it) in order to be economical (ignoring all other difficulties).  50 kJ/mole divided by 182.22 g/mole gives about .275 kJ/gram. A kilowatt is 1000 Joules per second, so we need 3.6 grams/second per kilowatt. If we recycle the material once per day, we would need about 300 kilograms per kilowatt at 100% efficiency or more like 6000 kilograms per kilowatt at 5% efficiency (perhaps a more likely estimate). If we allow an up-front capital cost of $1 per watt ($1000/kw) the azobenzene would need to be 16 cents per kilogram. Contrast that to an estimated 72 cents per kilogram for gasoline (assuming $2/gallon).

To stick with one cycle per day (allowing 24 hour energy production) and assuming we could make azobenzene as cheap as retail gasoline, we would need 22.5% system efficiency.  Big steam turbines can exceed 50% efficiency, but I don’t think the azobenzene could get hot enough to run a modern steam turbine (azobenzene’s flash point is 476C, so we can’t exceed that without an inert environment, which would add to costs.)

This is a materials problem, a *chemistry* problem. The rough numbers suggest that things are not off by more than an order of magnitude, which is encouraging.

Gas Bag Stirling Revisited

I’ve been thinking more about low-temperature stirling engines, and one area that can be exploited to lower the cost and increase performance is to use materials and methods that function well at lower temperatures but not at high temperatures, and therefore have never been considered in traditional stirling design. Rubber, for example, would not work at high temperatures but could be an ideal material to allow displacement without sliding contact. 

 

My current thinking is to use something like this as the “piston”:

 

Air Bag

 

Another option is to use a fan to move gas around, rather than using a “displacer.” The working gas could be alternately circulated inside the “piston” through a pair of small radiators, causing the “piston” to expand or contract.

Putting some numbers into the “Simple Performance Prediction Method for Stirling Engine” page at http://www.bekkoame.ne.jp/~khirata/academic/simple/simplee.htm leads to a cycle time of about 60 rpm, or about 1 cycle per second. For 1 kilowatt at 1 cycle per second, the force would be large (thousands of pounds). I haven’t decided what would be a good way to convert this slow, high-force motion into electricity. Perhaps using hydraulics, like this http://www.greencarcongress.com/2009/02/mit-students-de.html

Gas-bag Stirling Engine

Stirling engines have the ability to extract work out of relatively small temperature differentials. According to Carnot, the efficiency will be low, but that’s OK because otherwise unused low temperature heat sources are abundant — hot springs, solar, etc.

There are people working on these types of engines. An example of a 1 kilowatt engine is shown below:

1 kilowatt low temperature Stirling engine

As you can see, such an engine has to be huge in order to produce any meaningful power. And here is where we run into another problem — huge engines with enormous castings and large precision moving parts are not cheap. But energy is (still relatively) cheap. My ballpark estimate of the maximum cost for a new energy technology is about $1 per watt, or $1000 per kilowatt, if the thing can run 24 hours a day (continuous output). Solar cells require no fuel and cost at least $3-$4 per watt but run less than half time, putting them somewhere in the ballpark of at least $10 per watt continuous. However, they are proven technology with low maintenance and long lifetimes, so I expect a competing technology needs to be at least an order of magnitude better to really be successful.

If you look at the engine in the above picture, weighing in at 2 tons, the raw materials alone cost more than $1000 (maybe more than $10,000), even without fancy casting and machining. So it is never going to work, except as a research project. The cost has to come way down.

So here’s the idea: instead of enormous pistons, use a pair of nylon (or nomex) “gas bags,” similar to what is used in hot-air balloons. A flat plate can “squash” the air back and forth between the bags via a regenerator tube.

Even better, it would seem to me, is to use a pair of heat exchanger tubes with one-way valves — one tube is hot, the other is cold, so traveling from bag “A” to “B” the gas gets heated via the “hot tube” and then travelling the other way, it gets cooled via the “cold tube”. A sketch of what I am talking about is below:

 

Sketch of Gas Bag Stirling Engine

 

With a “gas bag” you could have an enormous displacement at relatively low cost. The other components, such as the “squasher” and heat exchanger will certainly be a bigger cost than the bag, however these do not need to be very precise or sophisticated, so I can imagine a multi-cubic meter displacement engine still coming in at a relatively modest price. And I can imagine some ways to possibly simplify and perhaps cheapen the engine further, such as maybe stacking the two “gas bags” on top of each other in more of a beta-type Stirling configuration, or squashing the bags between hinged plates like a fireplace bellows, or even using a twisting motion to squash the bags rather than using a flat plate.

I’d love to connect one of these things to an evacuated-tube solar collector, such as:

 

Evacuated Tube Solar Thermal Collector

 

Then you could have free hot water and free electricity! (But you might still be spending a long time paying down the loan needed to buy the things)

“Particle” accelerator fusion

Achieving nuclear fusion is easy. Getting out more energy than you put in is difficult.

Here is the idea: What if one tried to generate fusion energy using a “particle” accelerator?

I’ll explain more about the quotes later.

My thinking about this idea stemmed from the fact that to achieve fusion you need a high energy collision in order to overcome coulomb repulsion, but making these collisions by applying heat and/or pressure seems an awfully brute-force way to do it. Rather than heating something up and hoping all that random motion results in enough lucky collisions before too much energy leaks out, why not direct the motion so you get lots of high energy collisions?

This is what particle accelerators do, and they easily achieve fusion. According to wikipedia, deuterium-tritium fusion needs about 10 keV per particle to achieve fusion — compare that to the LHC which is making 7 TeV per particle. 

However, some things must be standing in the way of making a traditional particle accelerator be a reasonable path to fusion power, or people would have done it already — particle accelerators have been around a long time. 

One problem is flux. If you are going to build a power plant, construction costs are an important point. You don’t want to spend 10 million dollars building a “power plant” that produces 10 milliwatts. With particle accelerators, as you try to increase the intensity of your beam, you start running into space charge effects where the individual particles repel each other. So this limits how many particles you can accelerate. As far as I can tell, flux rates in the milliamp range are considered high for a particle accelerator, so even if you are getting millions of eV’s out per particle, the number of particles is so low that you are getting kilowatts of power out at best (even before subtracting any input power). 

A second problem, apparently, is Bremsstrahlung (“Braking radiation”). This is the radiation given off when you rapidly accelerate or decelerate charged particles. In the realm of fusion collisions, the Bremsstrahlung radiation is in the form of x-rays. So, if you shoot charged particles into a target, a lot of those particles that do not undergo fusion or other reactions (which is most of them) will be rapidly slowed down and emit x-rays in the process. These x-rays can be useful for x-ray crystallography, but are not very easy to capture, reflect, or use for energy. So, most particle accelerators end up being x-ray generators.

So here’s my idea: don’t accelerate “particles” in the physics sense, but “particles” in the macroscopic sense. Particles such as gold spheres (or “gold nanoparticles”) encapsulating a mixture of deuterium and tritium. Electrostatically charge up these particles to 50,000 volts or however high you can charge them and use a linear accelerator (or some kind of accelerator) to fire them into a target. On impact, most of the kinetic energy of the particle should be turned into heat and pressure as it penetrates into the target.

So, we are back to extreme heat and pressure, which was what I was trying to avoid initially, but it seems like it could be an efficient way to generate it, maybe more efficient than a giant laser or an enormous capacitor.

How fast would these “particles” need to be going?

If most of the “particle” mass was in the form of gold atoms rather than hydrogen atoms, then my guess would be that if we gave each gold atom 10 keV of energy, that would be enough to heat up the whole “particle” to 10 keV temperature on impact. I have no idea if this is really true, but it seems like it would be in the right ballpark. How fast does a gold atom need to be going to have 10 keV of kinetic energy? Well, as lon as we aren’t going relativistic, then should work. Rearranging:

So, if I did my math right, that’s on the order of earth escape velocity, but not anywhere close to the speed of light.

Flywheel-in-wheel Vehicle

A major obstacle to electric vehicles has always been (in the absence of an extensive “third rail” or similar network) finding a sufficient method for energy storage. Batteries have long been the obvious choice, and with the advent of better lithium batteries, such as the new LiPO4 batteries, will probably be the actual method used if electric cars ever become successful. 

However, an idea that has been around for a long time is to use a flywheel to store energy. Two of the major concerns about flywheels in vehicles are safety in a crash and detrimental effects on vehicle handling due to gyroscopic effects. Proposed solutions I have seen involve composite flywheels, gimballed mounts, counter-rotating flywheels, etc.

My thought: Put the flywheel(s) inside the drive wheel(s). As wheels themselves need to be rather structurally robust devices, the added weight for containment could be lessened. Also, if you think of how a motorcycle handles, an added flywheel could potentially enhance handling rather than damage it. And finally, strangest of all, by coupling the flywheel directly to the drive wheel, one wouldn’t need to exert a torque on the vehicle itself during acceleration or deceleration, potentially allowing a functional 1-wheeled vehicle (or a side-by-side 2-wheeled vehicle). Crazy.

 

Flywheel in wheel

 

 

So, once you have your flywheel spun up (at home, connected to the grid and maybe a frame to hold the vehicle in place), to accelerate the vehicle, you would simply need to apply a brake between the flywheel and the main wheel. This “brake” could actually be an electric generator — so you would be generating electricity as you accelerate the vehicle. However, to slow down, you would want to “spin up” the flywheel, so you would need to “apply power” in order to slow down. Sounds dangerous — but of course there is nothing preventing you from adding some sort of normal braking system to slow down in case of emergencies, assuming you have more than 1 wheel on the vehicle. 

Here is where it might get interesting — suppose you have a motorcycle configuration, with a flywheel inside of the first wheel and a normal wheel for the second. Each wheel is connected to a motor/generator i.e. flywheel->first wheel is connected via motor/generator and second wheel->frame is also connected via motor/generator.  You charge up the flywheel at home.

Now, to accelerate, you generate electricity from the flywheel/wheel assembly and use this to drive the second wheel. To slow down, you generate electricity from the second wheel and use this to “spin up” the flywheel/wheel assembly. The advantage is that motor/generators can be relatively efficient, maybe in the range of 80-90+ % and can handle large currents without losing much efficiency. Charging batteries gets much less efficient at high rates of charge (maybe down to 10% or so). Ultracapacitors could help the batteries handle brief bursts, but are currently rather bulky and expensive.

Unless I am mistaken, you also ought to get twice the acceleration/deceleration for a given amount of power as you would with a motor/battery setup.

Also, a flywheel + motor/generator is relatively cheap, especially if you make the flywheel low tech. A large-diameter steel flywheel does not have to spin very many rpms (therefore you can use ball-bearings) in order to store a large amount of energy. Plus steel is cheap. On the downside, a big steel flywheel is very heavy, but efficient regenerative braking should mitigate the biggest downside of a heavy vehicle.

And last of all, a really cool feature is that your flywheel-powered motorcycle could be self-balancing even when you are stopped. So then you could make an enclosed two-wheel vehicle.

Superconducting Magnetic Energy Storage Rocket

The idea is that the energy stored in a magnetic field is proportional to

 

where V is the volume of the magnetic field and B is the magnetic flux density (e.g. Teslas).

But the mass of a coil producing this magnetic field is proportional to

where l is the length of wire in the coil and is the density of the wire — this scales (dependent on coil geometry) as an area instead of a volume. So, if you make the thing big enough, you could theoretically have any energy-to-mass ratio you wanted. 

So we know that for a given magnetic field, the energy scales as length cubed and the mass scales something like length squared, the bigger this thing is, the better. So the real question is, how big does it need to be in order to do better than current methods (e.g. the space shuttle)?

Some equations:

For a single-layer air coil:

(r & l in millimeters)

For a brooks coil (maximum inductance per unit length of wire):

where a is the average radius of the coil, wound close-packed out of wire with a relatively small diameter. The cross-section of the close packed windings should be a square with an edge measuring 2a/3 on a side.

 

Taken from http://www.fnrf.science.cmu.ac.th/theory/magnets/

 

It is clear from the brooks coil that an optimal shape much more resembles a hoop than it does a single-turn solenoid.

Regardless of the structure chosen, obviously we want to have as strong a magnetic field as possible, corresponding to as much current (or as many turns) as possible, but there is a limit. Superconducting wire is limited by how much magnetic field it can be immersed in (colder=more field) and by how much hoop stress it can withstand. In this way, it is much like a “pressure vessel” with the magnetic field being the pressurised gas.

Since I am not a supercon magnet designer, I will have to extrapolate from existing magnets. These specs are kinda hard to come by, but as a data point www.cryoindustries.com listed a 15 Tesla magnet with 1.0 mm diameter wire carrying 300 amps having 33 kpsi hoop stress.

If we simplistically use this data for a large brooks coil, let’s see what we get. For example, 1 million turns of 1 mm diameter wire would make a square cross section 1 meter on a side. Making a brooks coil out of this would require a coil form with an average diameter of 1.5 meters. The inductance would be:

The energy storage would be:

Assuming supercon wire (mostly niobium) is about 9000 kg/m^3, the mass of this coil would be about

This gives us an energy density of about 1.3×10^6 J/kg. Compare that to the energy density of burning hydrogen + oxygen which gives us about 13×10^6 J/kg — so we are currently down by an order of magnitude in energy density, and even worse our coil is already enormous.

It is difficult to imagine realistically scaling up this coil by orders of magnitude. But let’s try anyway — if the coil were scaled to 10^7 turns, it would have 3.7×10^13 Joules and weigh 2.8×10^6 kg, giving an energy density of about 13×10^6 J/kg, on par with hydrogen + oxygen, and it would require over 6% of the world’s annual production of niobium (estimate taken from www.roskill.com). 

The killer is that I think the assumption of 300 amps/mm^2 is not realistic in these coils. If we look at the magnetic field inside the first of these coils using

83 Tesla is far beyond the highest fields any superconducting material currently known can withstand at any reasonable temperature. I am sure the stresses inside such a magnet would also be very high.

You might be able to get away with a really large diameter loop of superconducting wire, but such a geometry does not seem very compatible with a normal “rocket.”

Perhaps if we are building a ship to travel to mars, superconducting magnetic energy storage could provide both protection from radiation (in the magnetic field) and energy needs.