sgsguru: (Default)
 Machines lie.  Everybody knows this; the Automated House of the Future is going to do everything for us, but can't make consistent toast.  One thing I may have figured out, however, is how to put a thermostat on a shower, without any electronics (Batteries!  Electric shocks!).

The basis of a mechanical thermostat is a bimetallic strip.  It bends one way if it's hot and the opposite way if it's cold.  So if the water is too cold, it opens the hot water a bit and closes the cold water a bit.  Too hot, less hot, more cold.  We set the temperature by physically turning the strip.  If it can't generate the proper temperature, it goes to the end of its travel, where it hits a lever that pops up a little flag on the shower head.  I'd say blue for "too cold" and red for "too hot".

Problems:
  • Scale buildups.  Ideally, the strip bending from the temperature changes would cause scale to flake off.  In practice, who knows?
  • Force.  A bimetallic strip doesn't produce much force.  Is it enough to work the hot and cold valves?
  • Corrosion.  This is a definite problem as we have dissimilar metals by definition.  I'd say just coat the whole thing with plastic.  A sacrificial anode would probably be overkill.
Other Ideas:
  • We could direct the water straight down the drain if it's too cold or too hot.
  • If we wanted to get "just a little bit electric", we could put a little generator in the water flow.  We could get enough juice to run a couple of LEDs so we could see if the shower was ready from outside the shower.  Or hook it to the Internet (only half kidding!)
  • To get really fancy, we could hook the whole setup to a demand (tankless) water heater.  (It's a waste of energy to keep a big tank full of hot water in the basement all the time, in case we might need it.)
sgsguru: (Default)
 The New Scientist has an article about creating gasoline from air and electricity.  There are two reason this is important :
  1. Most alternative energy sources are not constant.  There has to be some way of storing energy for later use.  The flip side of this is that alternative energy sources can produce too much energy; sometimes the producers have to pay somebody else to take it off their hands.
  2. Liquid hydrocarbon fuel is far and away the most efficient way of fueling transport.  Should be; we've been doing it for over 100 years now.
I've been kicking around similar ideas for a couple of years now.  Unfortunately, I lack the chemical engineering knowledge to even know if it is possible.  Unlike the system in the New Scientist article, I assume that liquid carbon dioxide and water would be supplied, rather than extracted from the air.  Water is water and CO2 is a byproduct of "carbon sequestration", which will be needed to cut down emissions from current fossil-fuel power plants.  In other words, power plant operators will pay us to take it away.

So here's the plan.  We have three chambers A, B, and C.  There is a proton-exchange membrane between A and B.  There is a layer of zeolite catalyst between B and C.  A contains water and a bit of sulfuric acid.  B contains liquid CO2.  C contains the generated hydrocarbon.  Ideally, it'll be pure 2-2-4 trimethylpentane (AKA octane, AKA 100-octane gasoline); how pure depends on how good the catalyst guys are.

An electric current through A, with the negative pole at the membrane, electrolyzes the water to hydrogen ions and oxygen.  The oxygen is "waste" for this process.  The H2 ions pass through the membrane into B, where, with a proper catalyst, they pull an oxygen atom off of a CO2 molecule, converting it to CO.  Now, a mixture of CO and H2 is called "process gas", which is an input to all sorts of useful chemical syntheses.  In WWII, the Germans made synthetic gasoline from process gas via the Fischer–Tropsch process.  This uses a fairly simple nickel catalyst and produces complex hydrocarbon glop that can be dumped into a conventional oil refinery.  I'm sure that our catalyst guys can come up with something a lot better.  Last time I looked, zeolites were the top catalysts for petroleum chemistry.

Separating the CO from the CO2, transporting the CO to the reaction zone to convert it to gasoline (or at least hydrocarbon glop), separating out the water resulting from converting CO2 to CO, and transporting the water back to chamber A are things that I have no idea how to do.

Anyway, it's a half-baked idea that I wanted to get down in electrons.  Perhaps somebody else can do something with it, or let me know why I'm full of little red ants.

sgsguru: (Default)
This is a topic that I don't really know all that much about.  Corrections welcome.  This is very much a work in progress.

Speech recognition sucks.  There's been a lot of research; why isn't it any better?

Speech consists of an assortment of hisses and buzzes that are interpreted by the brain.  The current software approach is to try to go directly from the basic noises to words, using software brute force:
sound → words
Seems to me, if we break this process down, we can get a lot more accuracy.  The tool that linguists use is the "international phonetic alphabet" (IPA), which expresses individual phonemes:
sound → IPA → words

This gives us, right at the start, speaker independence.

For that first step, we can use something I call "predictive filtering" (I'm sure there's a "real" name for it).  We classify the sounds coming in to one of a finite set of "base sounds"; some small combination of buzzes and hisses that make up a specific sound.  To combine these base sounds into a phoneme, we look at the set of all possible sequences of base sounds that start with that base.
  1. Get initial base sound
  2. Generate set of all possible sequences starting with that base
  3. Cache the "most probable" sequence(s)
  4. Get the next base sound
  5. Prune off the sequences that don't have that sound as their second element
  6. Cache "most probable"
  7. Continue until we have a single sequence matching a single phoneme.
  8. Output phoneme
  9. Go to step 1
Generating the set of sequences and the pruning process will parallelize like crazy.  It's a classic map-reduce function.  We're doing our pattern matching one small step at a time.  Also, I expect the set of sequences to converge very rapidly; most steps will simply verify the input against the "expected value" in the cache.

This filtering technique can also be used for going from IPA to words.


sgsguru: (Default)
So Google is coming out with "virtual reality" goggles  that project computer data onto whatever you're looking at; sort of a personal "heads up display".  These have actually been around for at least fifteen years, but they've been very awkward and expensive.

I've wondered for a long time why the computer display people don't do more with optical systems.  For example, when Osborne came out with the first "portable" computer, the biggest complaint was that the screen was tiny.  However, it was very clear and bright.  You could get the effect of a much larger screen by simply putting a Fresnel lens over the display.  But nobody did it.  Nobody even tried it and reported that it didn't work.

It'd seem to me that, with an integrated design (ie, one not tied to existing display hardware) that you could get the effect of a very large screen with a tiny display and an optical system generating a virtual image.  It would have a very narrow angle of view, but for a laptop or palmtop, this is an advantage -- think of privacy.  You want wide angle viewing?  Get a projector.  The only problem I see is that the angle of view for practical optics would be too small -- it obviously has to be more than the angle subtended by your eyes.  (Unless you wanted to go 3D.  Hmmm ...)

Another optical system that I've wondered about is with DSLRs.  These have both the traditional "moving mirror and penta roof prism" of the film SLR and a digital sensor.  You get a digital image on a screen on the back of the camera, plus a "through the lens" optical view through the eyepiece.  Seems to me you could eliminate the moving mirror (mechanically complex and puts severe constraints on lens design) and the penta roof prism (honkin' big piece of precision glass) and replace it with a simple electronic display and a couple of lenses.  You have to have an electronic display there anyway to show status info and such.

Topic for research -- how small can you make an electronic display with a decent (720P) resolution?  The main area of research seems to be in making them bigger.

ETA:  Turns out that there are SLR-like cameras that work as I described -- they're called "mirrorless" or "brdge" cameras.  Most of them look sort of SLR-like with a honkin' big zoom lens.  Fuji even makes one that looks like a Leica M4 (IMHO the best 35mm camera ever made), but it costs a small fortune.

Profile

sgsguru: (Default)
sgsguru

October 2016

S M T W T F S
      1
2345678
9101112131415
1617181920 2122
23242526272829
3031     

Syndicate

RSS Atom

Most Popular Tags

Style Credit

Expand Cut Tags

No cut tags