1. The SD card is not working in its current form. I’m debating if I’m going to continue trying to hack that into this system. When you count up all of the CPU Cycles used by the sensors, the serial outputs and the GPS decoding (Major CPU time spent cracking GPS strings) the sd card becomes starved for resources.  We’ll see.
2. I’ve been able to source some bmp085 barometric pressure sensors for $4 and rewrite the code to eliminate the DS18s20P temperature sensor. This does 2 things for us, the first is that it lowers the part count which lowers the cost and code complexity. The second thing is that it makes the sensor package that has to be exposed to the elements very small, just two surface mount sensors so roughly the size of an eraser head.
3. The code now handles loss of GPS conditions. If a satellite can’t be locked onto, you will get an “!” in the lower righthand corner of the LCD display and the serial port data will go to default GPS info that can be easily removed at a later date.
4. The anemometer continues to be the most expensive part of this build, I think I can make one cheaper. More to come on that one.
5. I’ve removed all of my hacked together bmp085 code and replaced it with a library from the excellent folks at adafruit. It didn’t save any memory space, but it’s much cleaner

A few major bug fixes

• Removed an overrun condition where once windspeeds exceeded 70mph for over 10 seconds, the rpmcount variable would overflow and you would get very strange speed readings roughly every 5 seconds
• removed a lock condition that would keep the serial data from updating on a regular interval. Serial data is now written every 1 second
• Made RPM updates, LCD updates, Serial updates and GPS updates all independent and placed on their own timer. ie rpm updates are real time, LCD and serial updates are every 1 second, GPS updates are every 10 seconds by default

Let me know if you run into any trouble with this stuff. Enjoy!!

The SPC Day 1 image shows what areas are expected to be impacted by what types of severe weather. They break the image down into categorical, tornado, wind and hail. Categorical covers all categories. it essentially says “severe weather of some sort will happen here” while tornado, hail and wind are self descriptive.   The probability percentage is fairly straight forward, we’re going to use the tornado probability as our example. In this image, the St Louis area has a 5% chance (the brown area) of a tornado occurring.  That sounds like a fairly low chance, but what the SPC is trying to convey is that you are 5 times more likely today to be impacted by a tornado within 25 miles of you then you are on any other storm day.  Tornadoes are pretty uncommon and your chance of being impacted by one is normally very small, so a sudden jump from virtually 0 to 5% is fairly significant.  As you look towards OK and TX, you’ll notice that parts of the area are filled in with a blue hash mark. The hashing is used to indicate that, not only are people in this area at risk of a tornado, but they are further predicted to experience a tornado of EF2 or greater.  So right now, people in the the red hashed areas in OK are at a 15 times greater risk of experiencing an EF2 or greater tornado then on a typical storm day.   The SPC has a great page with descriptions on how this information is broken up that you can find here http://www.spc.noaa.gov/misc/SPC_probotlk_info.html  .   Have fun, stay safe!

Next came the real work.  All of December and January was spent trying to design a way to protect the glass while making sure that we never impede a passengers ability to get out of the car, even if we’ve rolled. What I came up with and started to build in February was a hinge style system that would allow us to drop mesh cages over the windows and suspend caps over the front and rear glass.  First I unbolted the roof rack system and reinforced all of the mounts with steel plates.  I then mounted 1 inch carbon steel tubes running the length of the car to the roof rack system, forming the fixed portion of a giant hinge.

The pivots are made from industrial light mounts used for concert equipment. They are nylon, quick connect and fairly inexpensive. I was also able to machine them to fit some 1/2 inch steel tubes that I threaded to fit into the built in bolts and some 1 inch square aluminum tube that makes up the bottom of the cage frames.  Each frame has 1 inch x .5 inch galvanized steel mesh press fit around it. The press fit is a fairly important design piece because it prevents hail from pushing into the windows, but is pressed in such a way that if we are trying to get out of the car, the mesh will pop right out.  Welds or bolts would have potentially left us trapped inside the car.

With all of the cages built and the remaining steel cut to fit, I did a dry run and discovered that the mesh flexed too much and would allow large enough sized hail to actually push the mesh into the glass.  I took everything apart and took it to a local powder coater while I tried to figure out a way to get the mesh off the glass without leaving too large of a gap.

What I came up with was building hinged shock mounts using hydraulic shocks off an RC truck and some very strong magnets.  Steel plate anchors the shocks to the cage frames and allow the shocks to hinge 90 degrees. The bottom side of the shock was modified to fit into some steel plate that was tapped to fit the shock threads, making the shocks part of the plate. 65lb industrial magnets are mounted to the backside of the shock plate and allows the whole assembly to stay attached to the car during high winds. The zip ties are just to cut down on some rattling at high speed.

The shocks absorb 15lbs of impact at the bottom of the cage and 40lbs at the top, relieving some stress from the rest of the assembly and making the whole thing look very cool.  When we are traveling to a chase, the cages fold up into the roof and the shock assemblies fold flat into the frames.  This picture doesn’t show it, but their is a cage for the sunroof that also keeps the side cages off the glass. An air-dam on the front of the windshield keeps the air off of the mesh, keeping things quiet and leaving my gas mileage intact.  The front and rear windshields couldn’t have a drop down cage like the side windows because of visibility laws in some of the more litigious counties in the midwest.

Instead, they received a cap.  The caps are part of the structure and provide a lot of the rigidity that the frame needs in addition to covering the windshields from direct hail impacts. The front cap extends 36 inches over the hood and the rear 24 inches.  The rear cap is also covered in foam insulation and bright caution tape to keep people from running into it head first.  If you see me on the road, ask me to show you my scar.   You might also notice the black diamond plate in various places. I use it to fill in the gaps between the body of the car and the cages over the window and as air deflectors.

With everything down, fuel economy drops to about 18mpg, but noise is very limited unless you open the windows. As soon as you open a car door, I have the frames built so that the door frame actually pushes the cage up and back into the roof so no hassles getting out in a hurry.  The entire system can come off the car with 4 bolts and then hang on my garage wall in the winter.   On the inside, things are a bit more dull.  On the dashboard we have

• An MS Lifecam Cinema
• At least 1 GoPro style camera
• An android tablet running RadarScope
• A laptop with GR3, GPS and a full offline install of MS Streets and Trips
• A Baufeng UV-5R+ HAM Radio
• A midland weather radio
• A DC-DC converter and several USB chargers

We also run around with a well stocked stormcase full of more emergency oriented gear.

• Correct sized tools for every nut and bolt on the car
• Jumper cables
• spare upper and lower radiator hose
• spare serpentine belt
• black, duct and tin tape
• CERT style first aid kit
• flashlights, dear lord do I have flashlights.
• extra power inverter
• lots of other small necessities.
• fire extinguisher
• lots of rolls of clear plastic tape in case we do lose that window
• big metal pry bar

So far we haven’t had to use any of the emergency gear and could probably get by with just the tablet, but options are always good when you find yourself in an urgent situation.

We’ve tested the vehicle in golf ball size hail and haven’t needed to even slow down.  I suspect we could make it all the way to baseball before we would need to stop. We’ve also successfully survived a 4×4 block of wood, 2 large rocks and swings from a baseball bat without hitting the glass. My neighbors think I’ve lost my mind.   Our goal was never, and is never, to intercept a tornado. That just seems like a bad day all around. The cages on the car allow us to get closer and keep moving when the hail comes, and should we need to get away a busted window won’t slow us down.

The green line is still our dewpoint, the Red is the temperature and that purple line is the mixing line that got us to our LCL. Once we have the LCL plotted, we can draw a line starting at the LCL and following a moist adiabatic line. Remember that we are using a moist adiabat because our box full of dry air was pushed into the sky until it reached its condensation point and now has to be treated like a box full of wet air.   Usually when we draw this line, we end up with a little triangular space that we call CIN or Convective Inhibition. You’ll usually hear weather weenies call this little box The Cap. That would be Capital T  Capital C. The Cap is a bit of warm air that our box of air has to be able to push through in order to get to its Level of Free Convection or LFC.  Imagine a balloon full of 80C air that gets placed into a 0C walk in freezer. That balloon will sail towards the ceiling. Now imagine that the same walk in freezer has a heater midway between the floor and the ceiling that is pumping out air that is 100C. The balloon will rise from the floor where it is 80C above ambient temperature to the level of the furnace air where now it’s 20C below ambient, making it want to sink.  In order to get to the ceiling of the freezer, our balloon will need to come up with 21C worth of energy or an external force will need to push it through the heater air until it can get back to the 0C ambient temperature and start rising on its own again (yes I know the hot air would convect to the ceiling, leave it alone, it’s a thought experiment).  If our box of air can’t push through The Cap, we’re not going to get much of a storm. But assuming that The Cap breaks and we make it through, our box of air will start to accelerate faster and faster towards the upper atmosphere.  The yellow region starting from the LFS and going up to the Equilibrium level (EL, not illustrated) is what we call CAPE or Convective Available Potential Energy.  We like CAPE.  CAPE is responsible for a lot of the energy that goes into a storm.

Skew-Ts are a fairly amazing bit of graphing in my opinion. Armed with just a few bits of information you can figure out how strong a storm is likely to be, if it has the right cloud base for tornadoes, how much inhibition has to be overcome in order for the storm to initiate and a few other items that we haven’t talked about here.   You can find an excellent and more indepth training, for free, from ucar here that is nice and easy to digest and will fill in a lot of blanks that I’ve left in this series.

The LCL, or Lifted Condensation Level, is where the base of the clouds form and is simply the intersection of the dew points mixing ratio line and the temperature.  The purple line follows the mixing ratio line near 20C until it runs into the red temperature plotting at about 28C. At this point our box of air has reached it’s saturation level and has to condense in the form of a cloud.  When you hear chasers talk about high based or low based storms, the LCL is usually what they are talking about.  When the cloud base is around 2000m or lower, that gets most chasers into the cars and on the road.  Why? we’ll cover that in the final part of this Skew-T discussion.

Our box of air starts it’s life on the ground and follows the dry adiabatic line towards the sky. Depending on the amount of moisture in our box of air (humidity) our air will eventually get cold enough that it can no longer hold onto the water in it and it will have to condense it out (dew point). Now our box of air is still climbing, but since it has condensed it has to follow the wet adiabat.

So far, we have defined what each axis of the Skew-T represents,  lets see how these things work together.

When you take a box full of air (you’ll usually hear it called a parcel, but box is more awesome) and you lift it high above the ground, a few things happen to the air inside of it.  Remembering back to the first axis of the Skew-T that we talked about (isobars), as you rise higher into the atmosphere the air pressure gets lower.  You probably learned in middle school, or will learn if you aren’t there yet,  that molecules pushed close together will generate more heat then molecules that are spread far apart.

So thinking back to our box, when we filled it full of air it had around 14.7 pounds per square inch of pressure on it from all sides.  By the time you’ve lifted that box to 900mb that box only has 13 pounds per square inch pressing on it. At 800mb it’s 11.7 psi and at 600mb it’s down to nearly half at 8 psi!  With less weight on our box of air, that air will start to expand and as it expands, it will cool.

These dry adiabat lines are showing us the rate of that cooling with height and plotting that decrease based on  our starting temperature.  Looking at our Skew-T then, if we started our bucket of air off at 1000mb on a day when it’s 30 C outside, and then lifted that box of air up to 600mb, that bucket of air will have cooled down to -10C or at roughly 1 C for every 100 meters. But usually we have humidity to deal with, which takes us to the moist adiabat.

But first we have to get air into the atmosphere and that starts with the dry adiabatic.