brainwagon

A recent issue of Make magazine had an article about launching water rockets via hydrogen/oxygen combustion: basically an electric current is used to break water into its constituent hydrogen and oxygen, which then bubbles up in the rocket, forcing out some additional water. To launch, this hydrogen is ignited, and recombines quickly into water vapor, but also generates a huge burst of pressure and launches the rocket.

And here was my question: how big of a burst of pressure does it really generate? Is it safe? 2 liter bottles have a burst pressure of around 120psi or so. That’s actually reachable with ordinary mechanical means, and it is pretty easy to understand and monitor: the pressure increases slowly, and can be monitored if necessary by an ordinary pressure gauge.

The chemical reaction isn’t so easy to wrap one’s head around. You actually need to know some physics. Sadly, I’m mostly self taught, so I have to work through these things slowly. A discussion with Michael and Tom over lunch the other day reminded me of the ideal gas law. One mole of hydrogen and one half mole of oxygen combine to create one mole of water vapor. But I must admit: I know very little about the realities of combustion physics.

So, I did what I always did: I googled. And I found this interesting page. It estimates the pressure to peak at around 160psi, well beyond the burst pressure of the common 2L bottle. To keep the rocket from bursting, it is vital then to dilute the reaction by the introduction of ordinary air. The nitrogen won’t combust, and should limit the overpressure.

I’ll have to muddle over the details more sometime, but it’s good reading.

Powering Water rockets with hydrogen combustion

There has been a lot of publications lately about water rockets. These are rockets which are usually constructed of empty plastic soda bottles, pressurized by a bicycle pump and launched into the air. I haven’t done any of this, but it sounds like great fun. I even picked up a copy of Soda Pop Rockets by Paul Jarvis, which is a colorful if rather arts-n-craftsy book on the subject. While reading it, I couldn’t help but muse about the physics involved. How do things like diameter of the rocket, diameter of the exhaust nozzle, and the amount of pressure affect the height that the rocket might achieve?

I mentioned this to Loren over lunch the other day, since I know that he has done his fair share of water bottle rocket launches, usually using liquid nitrogen as the propellant. He had a few interesting insights, namely that the nozzle design which is common in ordinary rockets isn’t really useful in water rockets, as water is essentially incompressible. This means that you can use Bernoulli’s equation to compute the force generated. Of course, the mass of the rocket is continuously dropping as water is expelled, but that’s not too hard to deal with in the simulation.

What wasn’t clear to me was that after the water is exhausted, there is still residual air pressure in the rocket, and this pressure is significant and must be dealt with differently since the air is compressible. A bit of research led me to this rather nice website:

Rocket Science.

He has an interesting example of a 2 liter bottle which is pressurized to 100 psi and filled 20% with water. It’s fascinating: the “water burnout” (when all water has left the rocket) occurs only 0.042 seconds after launch, when the rocket has an altitude of only 3.2 feet (!). It continues to accelerate though as the air pressure equalizes. In this example, one third of the velocity of the rocket is obtained after water burnout.

It might be fun to make an implementation of this.

Addendum: Here’s another link to a water rocket simulator.
Addendum²: Another simulator.

KJ6AKQ tackles a project I’ve been thinking about for quite a while: the construction of a little handheld satellite antenna:

Building an IOio satellite antenna – KJ6AKQ.

Poking around on archive.org, I found that Louis Bell’s classic work The Telescope was available for download. It is a pretty nice book, well worth reading if one has an interest in astronomy and telescopes. It is part history, part engineering, and part user’s guide. It also includes some great illustrations such as the one below of Newton’s first reflecting telescope. Enjoy!

Internet Archive: Free Download: The telescope.

I don’t really have the hardware to effectively transmit to satellites in linear transponder mode. To really make it straightforward, you probably would like to have a computer to handle the Doppler tracking and antenna pointing, and a full duplex SSB transmitter (or a pair of ordinary ones). That’s more than I am willing to invest at the moment, but with my little FT-817, you can actually receive these birds pretty easily, using my ever present Arrow antenna and just tuning by hand. So, that’s what I did this morning: here is my recording of HO-68. It begins with a bit of the CW beacon, then shows me tuning around to try to find SSB signals. Sometimes, the tracking is pretty good, but later in the recording, my next door neighbor began mowing his lawn and running a leaf blower, and it was hard to hear (I need to use my over the ear headphones instead of these earbuds I’ve used for FM passes). Anyway, here’s the recording:

2009-12-30-HO68-SSB, recording by K6HX in CM87 using an FT-817 and hand tuning

I have a bizarre fascination with algae. There is a lot of science going on with algae for use in food stuffs, in detoxifying industrial waste, and even as food. But my geek-itude is vastly exceeded by Jared Bouck, whose created an entire site about the cultivation of algae. I’m bookmarking this for later fun.

AlgaeGeek.com -- Site News

Addendum: While reading the algae page on Wikipedia, I found a link to this book on identifying freshwater algae in the United States.

An annotated key to the identification of commonly occurring and dominant genera of algae observed in the phytoplankton of the United States

It was reported that an Iridium satellite and an “non-functional Russian satellite” collided yesterday. I was curious, so I did a bit of digging, and found out that NASA had reported that it was Iridium 33 and COSMOS-2251. A bit more work uncovered orbital elements for both objects, so I was able to plug in their numbers and determine the location of the collision. A bit more of scripting, and I had GMT generate the following map (click to zoom in some more):

According to my calculations, they passed within 100 meters of one another (but my code gives an uncertainty much greater than that.) Each satellite is travelling about 26,900 km/second hour (sorry for the typo, but the math holds). I don’t have the mass numbers for the satellite, but even if you think they are travelling at perfect right angles, each kilogram of the mass generates about 28M joules of energy. According to this page on bird strikes, a major league fastball is about 112 joules, a rifle bullet is about 5,000 joules, and a hand grenade is about 600,000 joules. This collision generated 28M joules per kilogram of mass. Ouch!

Addendum: It’s been a long time since I took basic physics. If you care, you shouldn’t trust my math, you should do it yourself and send me corrections.

Courtesy of hack-a-day, check out the following video illustrating an analog computer that implements the dynamics of a bouncing ball, not using a microprocessor, but just a circuit involving analog operational amplifiers.

Bouncing ball analog computer -- Hack a Day.

Sven Grahn’s Space Tracking Notes talks about his efforts in using radio to track satellites. There is all sorts of really good stuff in here, a lot of it having to do with tracking the secret launches of Russian satellites during the Cold War. Very, very neat stuff.

I’ve been reading up a bit on antenna design, particularly the design of Yagi style antennas, and decided to give it a whirl. In particular, I decided to try to design a Yagi that actually is supposed to work on 137.62 megahertz, to see what is possible. I actually am not displeased with the paper specs for the resulting four element 137Mhz Yagi. I designed it using the suite of programs written by David Kirkby, G8WRB. Here’s the resulting pattern at the design frequency.

Here is the specification for the resulting design…

NOTES Optimised with the genetic algorithm
FREQUENCY 137.620000
MIN_FREQUENCY 137.000000
MAX_FREQUENCY 138.000000
STEP_FREQUENCY 0.020000
ELEMENTS 4
DRIVEN 1
PARASITIC 3
ANGULAR_STEP   2.000000
#DATA_DRIVEN        x         y       length     diameter voltage(r) voltage(i)
DATA_DRIVEN     0.38625    0.00000    1.02524    0.00475    1.00000    0.00000
#DATA_PARASITIC     x         y       length     diameter
DATA_PARASITIC
                0.00000    0.00000    1.08740    0.00475 reflector
                0.95210    0.00000    0.94818    0.00475 D1
                1.78574    0.00000    0.92937    0.00475 D2

And here is the analysis over the entire 137-138 Mhz band.

# Driven=1 parasitic=3 total-elements=4 design=137.620MHz
# Checked from 137.000MHz to 138.000MHz.
  f(MHz) E(deg) H(deg)  R     jX    VSWR   Gain(dBi)     FB(dB)    SideLobes(dB)
  137.000 58.1  79.6  28.25  -4.35  1.790      9.202     26.440      0.000
  137.020 58.1  79.6  28.24  -4.22  1.789      9.205     26.547      0.000
  137.040 58.1  79.6  28.23  -4.08  1.788      9.209     26.650      0.000
  137.060 58.1  79.5  28.23  -3.95  1.788      9.212     26.750      0.000
  137.080 58.1  79.5  28.22  -3.81  1.787      9.216     26.845      0.000
  137.100 58.1  79.5  28.21  -3.68  1.786      9.219     26.936      0.000
  137.120 58.0  79.4  28.20  -3.55  1.786      9.222     27.023      0.000
  137.140 58.0  79.4  28.20  -3.41  1.785      9.226     27.104      0.000
  137.160 58.0  79.4  28.19  -3.28  1.785      9.229     27.179      0.000
  137.180 58.0  79.3  28.18  -3.14  1.784      9.233     27.250      0.000
  137.200 58.0  79.3  28.17  -3.01  1.784      9.236     27.314      0.000
  137.220 58.0  79.3  28.16  -2.87  1.784      9.240     27.371      0.000
  137.240 58.0  79.2  28.16  -2.73  1.784      9.243     27.423      0.000
  137.260 57.9  79.2  28.15  -2.60  1.783      9.247     27.467      0.000
  137.280 57.9  79.2  28.14  -2.46  1.783      9.250     27.505      0.000
  137.300 57.9  79.1  28.13  -2.33  1.783      9.254     27.535      0.000
  137.320 57.9  79.1  28.12  -2.19  1.783      9.257     27.558      0.000
  137.340 57.9  79.0  28.11  -2.05  1.783      9.261     27.574      0.000
  137.360 57.9  79.0  28.10  -1.92  1.783      9.264     27.582      0.000
  137.380 57.9  79.0  28.09  -1.78  1.783      9.268     27.582      0.000
  137.400 57.8  78.9  28.08  -1.64  1.784      9.271     27.575      0.000
  137.420 57.8  78.9  28.07  -1.51  1.784      9.275     27.561      0.000
  137.440 57.8  78.9  28.06  -1.37  1.784      9.278     27.539      0.000
  137.460 57.8  78.8  28.04  -1.23  1.784      9.282     27.509      0.000
  137.480 57.8  78.8  28.03  -1.10  1.785      9.286     27.473      0.000
  137.500 57.8  78.8  28.02  -0.96  1.785      9.289     27.429      0.000
  137.520 57.8  78.7  28.01  -0.82  1.786      9.293     27.379      0.000
  137.540 57.7  78.7  28.00  -0.68  1.786      9.296     27.322      0.000
  137.560 57.7  78.6  27.98  -0.55  1.787      9.300     27.258      0.000
  137.580 57.7  78.6  27.97  -0.41  1.788      9.304     27.189      0.000
  137.600 57.7  78.6  27.96  -0.27  1.788      9.307     27.114      0.000
  137.620 57.7  78.5  27.95  -0.13  1.789      9.311     27.034      0.000
  137.640 57.7  78.5  27.93   0.01  1.790      9.315     26.948      0.000
  137.660 57.6  78.4  27.92   0.15  1.791      9.318     26.858      0.000
  137.680 57.6  78.4  27.91   0.29  1.792      9.322     26.763      0.000
  137.700 57.6  78.4  27.89   0.43  1.793      9.326     26.664      0.000
  137.720 57.6  78.3  27.88   0.56  1.794      9.330     26.561      0.000
  137.740 57.6  78.3  27.86   0.70  1.795      9.333     26.455      0.000
  137.760 57.6  78.3  27.85   0.84  1.796      9.337     26.345      0.000
  137.780 57.6  78.2  27.84   0.98  1.797      9.341     26.233      0.000
  137.800 57.5  78.2  27.82   1.12  1.799      9.345     26.118      0.000
  137.820 57.5  78.1  27.81   1.26  1.800      9.348     26.000      0.000
  137.840 57.5  78.1  27.79   1.40  1.801      9.352     25.881      0.000
  137.860 57.5  78.1  27.78   1.55  1.803      9.356     25.760      0.000
  137.880 57.5  78.0  27.76   1.69  1.804      9.360     25.637      0.000
  137.900 57.5  78.0  27.74   1.83  1.806      9.364     25.512      0.000
  137.920 57.4  77.9  27.73   1.97  1.807      9.367     25.387      0.000
  137.940 57.4  77.9  27.71   2.11  1.809      9.371     25.260      0.000
  137.960 57.4  77.9  27.70   2.25  1.811      9.375     25.133      0.000
  137.980 57.4  77.8  27.68   2.39  1.812      9.379     25.005      0.000
  138.000 57.4  77.8  27.66   2.54  1.814      9.383     24.876      0.000

This was mostly just an exercise, to understand how the software works. It’s kind of silly to use a linearly polarized antenna for the circularly polarized polar orbiting satellites (3db mismatch), but I’d still be interested in hearing any comments about this design from knowledgeable antenna design people.

It’s just that the weather has been too crappy for me to stand outside with my laptop and do the recordings. But today, the weather dawned nice, and so I did manage to record a nice western pass of NOAA17.

I keep wondering if the whole KISS principle (a personal favorite of mine) might be sensible to apply more thoroughly. Diane, VA3DB pointed me at a satellite design I hadn’t seen before: a picosat that would carry aerogel ultracaps as well as traditional nicads. It was dubbed Volksat. I think there are lots of sensible and good ideas contained here.

Well, I’ve made some headway on a project that I thought would be cool to write: porting G3RUH’s Plan 13 Satellite Prediction algorithm to a more palateable language than BASIC. I chose python, and it appears to be mostly working. It reads in the TLE orbital elements (same ones I use in “predict” or “gpredict”) and then allows you to create a bunch of satellite objects, and query their positions over time.

Here’s a screendump of a simple test program that I was running this morning:

Satellites which are above the horizon are marked in bold. They are sorted by elevation. The datafields displayed are elevation, azimuth, latitude and longitude of the subsatellite point, velocity, the Doppler velocity, and the frequency of a signal Doppler shifted from 435.845Mhz (just a value I did to check, since I was using PolySat CP3 at the time, which has APRS telemetry downlinked on that frequency). The code requires some additional cleanup, and once I have it all ready to go and documented, I’ll make it available. I think it will have a lot of uses.

I’ve been pondering potential upgrades to my satellite capabilities. Right now, I’m using a very popular combination: an arrow antenna and a Kenwood TH-D7A. Often, at the beginning of passes, where satellites are still relatively distant, I get very weak signals, and can’t often hear the satellite well until they are maybe 15 or 20 degrees over the horizon. In talking (both on the bird, and via email) with Mark Spencer, WA8SME, I’m beginning to see why his recommendation of adding a preamplifier to the receive side of my setup might result in better overall performance.

Mark wrote a very nice article for the Amsat Journal entitled Why can’t I hear AO-51? which is simply great and clearly explains how you can figure out the path loss between the satellite and you. It turns out that the path loss from you to the AO-51 is probably somewhere around -126db on 440mhz, but the uplink loss from you to AO-51 is only about -112db, or nearly 14 db stronger.

What’s the difference? Well, first of all AO-51 is running only 1w of total power and my Kenwood cranks out above 5. That’s almost 7db stronger right there. What’s the rest of the difference? Mostly the difference in free space loss between 2m and 70cm. The approximation for path loss that is most often used for freespace
path loss is:

path loss = 32.44 + 20 log(d) + 20 log(f)

given in db for a distance of d kilometers and frequency f in megahertz.

Assuming the distance is constant, for a given frequency, the difference is the is about 9.5 db stronger on 2m than on 70cm. (20 * log(145.92)-20*log(435.3) is just about 9.5). It turns out that AO-51 is a little deafer than your HT, but has about 4db more gain in its antenna compared to your rubber duck, so overall, you have a much easier time getting signals to it than it does getting to you.

What this suggests to me is that a receive preamp may very well be a good idea. I’m looking into getting one soon.

Well, during a road trip with my wife to San Diego and back, I managed to begin to type up my notes for an upcoming tutorial article (cross fingers) on reception and decoding of weather satellite imagery. I basically reimplemented what I had, stripping it down to its barest essence, and trying to make it easy to understand, yet still capable of creating good imagery. Oddly enough though, in the process, I inadvertently seemed to introduce some aliasing artifacts which I must admit, are puzzling me mightily. Oh well.

This morning I decided to try to record a low 25 degree maximum pass of NOAA17 that occurred to the east. It starts out fairly noisy because I have a pretty high horizon to the northeast. The first is my “advanced” decoder, which is about six times longer than the simpler decoder I wrote.

The second is the same data file, processed with my “simple”, easy-to-follow decoder. The only serious feature it is lacking is the sync detection and rectification, which as soon as I can figure the simplest way to add, I’ll try to get in. The simple version is two pages of code, not including the data for the filter tables.