Same thing, every year...

I thoroughly enjoy being a TARC mentor; working with the teams is great, and I am thrilled to be part of the whole experience that is Team America Rocketry Challenge. However, mixed in with the enormous amount of good, there is always a little bad - small annoying things that crop up every year, things you try to eradicate through workshops, coaching, and cajoling. But no matter how hard you try, they appear year after year, like great cosmological constants of the universe. Last minute qualification attempts are one of these, but the one I want to focus on in this post is especially vexing to me - the tendency of most TARC teams (and mentors) to not understand the limitations of the Open Rocket or Rocksim simulations.

Every year, I join the teams at Pegasus during their practices, and every year someone always complains that their rocket is not reaching the altitude predicted by the simulations, falling short by as much as a couple hundred feet. Such a huge mismatch must be due to some cause, and according to the team, it is usually a bad batch of motors or lots of drag on the rail, which everyone knows is not accounted for in the simulations. And nearly 100% of the time, the team is wrong - it ain't the motor, and it ain't the rail.

It's the sims - they put too much faith in the altitude predictions.

Open Rocket and Rocksim are fabulous tools, and they are absolutely the things to use in designing hobby rockets in this day and age. However, both programs make some simplifications and assumptions, with the net result that they almost always overestimate the altitude the rocket is capable of achieving on a specific motor. I can prove this with a specific example - my Geezer TARC entry for this year, the Eggsploder. Here is the altimeter profile of its flight from September of last year:

Eggsploder PNUT altimeter data (Click to enlarge).

You will note that the peak altitude is around 840 feet, 40 feet off the 800 foot goal (not bad for the first flight, eh?). This is the actual performance; now let's pull up the Open Rocket simulations:

Open Rocket simulations for Eggsploder (Click to enlarge).

Take a gander at simulation 15, which is an exact match to the weight of the rocket as flown under zero wind conditions. It predicts a peak altitude of  976 feet, 136 feet higher than that of the actual flight. Simulations 16 and 17 are for wind conditions of 5 and 10 miles per hour; even they are over 75 feet higher than actual.

So Open Rocket over predicts... What about the other guy on the block, Rocksim?

Rocksim simulations for Eggsploder (Click to enlarge).

Note the last simulation - 1000 feet! This is what Rocksim predicts if I let the program calculate the drag coefficient, which is the default when you create a rocket in the software. This is not too different from the Open Rocket results (~30 feet difference) and still way higher than the model actually flew. However, Rocksim allows you to manually adjust the drag coefficient (You can somewhat do this in Open Rocket by changing the roughness of the surfaces, but this is very crude). This enables you to match the sim to the altimeter readings, which is what I tried to do in the first two sims. Sim #1 is for a drag coefficient of 0.63, and simulation #2 is for a drag coefficient of 0.65; both these numbers are higher than the Rocksim determined drag coefficient.

Bottom line - If I choose a motor combination for Eggsploder that gives simulated altitude of 800 feet, I would be very unhappy with the performance, as the rocket would struggle to make it to 700 feet, well short of the goal. This is why, year after year, I always tell the teams to sim high - you can always add weight to bring the altitude lower, but it is very hard to raise the altitude without changing to a more powerful motor, which can mess with stability (bigger motor = more weight in back = less stable).

So I write this post, hoping that next year's teams will read it and pay attention. Maybe next year will be the year.

But I'm not going to get my hopes up...