Moment of Inertia Breaks the Sound Barrier

I went out and flew Moment of Inertia last Saturday. It took off really fast and flew fine.  The shock cord did break so it came down in two pieces, but I recovered both pieces. I did not hear a sonic boom separate from the noise of the motor, but the motor was really loud. I’ve heard that you might not hear a sonic boom since you are standing behind the direction of travel of the rocket. The max altitude beeped out by the altimeter in the field was 1059 meters.

When I downloaded the altimeter data it said the max speed was 195 m/s. The point of this rocket was to be a mach breaker, and 195 is well below the speed of sound. My original OpenRocket simulation said it should reach 399 m/s. I had added a little weight when I fixed the broken centering ring from the first flight, and I hadn’t re-weighed it to update the simulation. I was busy and I figured the altimeter would tell me how fast it went. But 195 seems really slow for this design so it’s time to dig into the data and find out what happened.

There are several possible things that could have happened:

1) OpenRocket doesn’t do transonic or supersonic drag right. There could have been a lot more drag causing the max speed to be slower, but 195 m/s is well below where even transonic effects should show up.
2) The thrust curve of this particular motor may have had slight variations from the average for this motor type. Just a longer burn time would have resulted in a slower max speed.
3) The altimeter is just a barometric sensor. It could have been fooled somehow. Maybe low-pass effects from air trapped in the body tube made the altitude change seem slower. I did make the vent holes a little oversize to try to avoid this, but everything happens really fast on this rocket so maybe it wasn’t enough.
4) The velocity reported by the altimeter has to be derived from the pressure/altitude data. Maybe there is some smoothing or aliasing occurring in the derivative algorithm.

And, of course, it could be some combination of multiple of these effects.

Here’s the overall graph from the altimeter. The height in black shows the ascent, a pressure spike when the ejection charge fired (labeled drogue), then linear falling at terminal velocity. It looks like a rocket flight.

Something unusual happens on ejection. The rocket seems to instantly gain ~75 meters of altitude. Was this an effect of trapped pressure in the body tube that got released on ejection? It’s possible that this resulted in a low-pass filter that made the max speed seem slower. Also, there’s some weird pressure readings on takeoff.

Zooming in on liftoff, this just doesn’t look like good data. Maybe shockwaves fooling the sensor? Starting at around 0.3 seconds at 100 meters in the air looks like good data. The altimeter is saying that the max speed occurred at 0.7 seconds on the graph, but this motor only has a 0.4 second burn time after which it should start decelerating rapidly from drag. That would be an unusually long burn time to be a random variation from the average for this motor type. So that points to low-pass effects slowing the pressure rise.

Another question is when did liftoff actually occur? The 0.0 time on the graph is when the altimeter detected liftoff because there was sufficient pressure change to arm the ejection charge. But the graph seems pretty flat before -0.35 seconds and then things start happening. That may have been the actual ignition and liftoff. If that’s the case, then the period of weird data corresponds approximately to when the rocket should have been supersonic.

So I re-weighed the rocket and went back to the simulation. The repairs only added two grams so it didn’t affect the simulation much. There’s one number from the altimeter that I’m pretty sure I can trust, and that’s the apogee altitude after ejection. Then the pressure should be a correct reading of ambient.

The altimeter beeped out 1059 meters, which appears to be the highest single data point excluding the ejection pressure spike. However, that data point looks like noise. 1025 meters looks like a better stable value.

I tried to adjust the surface finish settings in my simulation to get less drag so that the simulated apogee was 1025 meters. Even with the smoothest finish it wouldn’t go that high. It was a hot summer day so the density altitude was probably higher than the field elevation of 1556 meters, but I had to go to 4100 meters to get the apogee that high. That seems unlikely. So maybe the motor did have a slightly longer burn time than what’s in the simulation file. This flight profile has massive drag losses so a longer burn with a lower max speed would result in a higher altitude.

Overlaying the simulation on the altimeter graph gives good general agreement about the shape of the curve with the altimeter reading either being lower or later.

One final piece of evidence is that the supersonic phase in the simulation does correspond roughly to the period of bad data from the altimeter. The bad readings could have been caused by shockwaves.


Here’s the verdict on each of the hypothesized causes.

1) Extra supersonic drag lowered max speed

Verdict: Unlikely

Evidence: Extra drag would have lowered the apogee altitude. Instead, the altimeter apogee exceeded the simulation apogee until I changed the simulation to lower the drag.

2) Longer burn time on motor

Verdict: Likely a contributing factor

Evidence: Trying to get the simulation apogee to match the altimeter apogee by reducing drag required an unrealistically high density altitude. A slightly longer burn time on this particular motor is a more realistic explanation for a raised apogee.

3) Barometric sensor was fooled

Verdict: Likely a contributing factor

Evidence: The sudden jump in altitude on ejection points to trapped air in the body tube. The bad data early in the flight corresponds to the time period when the simulation says the rocket should be supersonic. One lingering issue on this is why the discrepancy between the altimeter and simulation kept growing throughout the flight. If it’s a low-pass effect, one would expect that the discrepancy would grow when the rocket is moving fast and then shrink again when the rocket slows down near apogee. Possibly the altimeter line on the graph is not just low, but also late because of a longer burn time on the motor.

4) Calculation of velocity from altitude data by altimeter

Verdict: Possibly a contributing factor

Evidence: The altimeter software graphed a smooth curve for the velocity even through the time preiod when the altitude data was bad. So they must have some algorithm that interpolates over bad data.

And the overall conclusion is that I think I can confidently say that the rocket did break the sound barrier. Except for the period of bad data, the altimeter data matches pretty closely a simulation that says the rocket should have exceeded mach 1, and the period of bad data corresponds to the time when the simulation says the rocket should be supersonic.


Flying Again This Weekend

It’s been over a year since I’ve blogged.  My life just got busy and I stopped for a while and never got back to it.  Since then, I went to the Argonia cup with the University of Wyoming 17220 Rocketeers, taught two rocketry classes for kids, and have been to a few NCR launches, but not much progress on my projects for a while.

That’s been changing lately.  I finished up Moment of Inertia, and in June launched it on an F motor as a shakeout flight.  It flew fine until ejection ripped out the centering ring that holds the shock cord so it came down in two pieces.  The fiberglass body survived the lithobraking and the rest came down on the streamer and was fine.

I repaired and reinforced the shock cord mount and this weekend I’m going to launch it with the H410 VMAX motor.  It should break the sound barrier at T+0.3 seconds and 50 meters AGL with a top speed of Mach 1.3.  Hopefully, we can hear a sonic boom.  I’ve heard that sometimes you don’t hear it because most of the sonic energy is projected upward in the direction the rocket is going.  Also, will it be loud enough to hear over the motor?  Still, breaking the sound barrier at that low altitude I’m hoping will help us hear it.

I’m also going to launch Stretch Mustang again.  No particular milestone for this one.  Just another launch.

The Force has been on the back burner, but I hope to get working on it again soon.

Stretch Mustang First Flight

I got out to the NCR launch yesterday and did the first flight of the stretch mustang rocket on a G250 VMAX motor.  The altimeter reported an apogee of 1428 feet.  My original simulation predicted 1529, which is not too far off.  I can get an apogee of 1457 feet out of the simulator if I switch from smooth paint to regular paint, but there are probably other factors too.

My girlfriend also got her level 1 certification with her Wicked themed rocket “Defying Gravity”.

New Rocket Weight

The 6″ rocket is close to being all put together so I decided to get an updated weight for it.  It’s now 42.8 lbs.  The target is 45 lbs so I’m in good shape.  Both of those numbers are not including the motor, which according to the manufacturer’s web site is 5.8 lbs.  Here is the list of things that still need to be added:

  • Two CO2 cartridges for the parachute ejection system.
  • Electronics, including batteries.
  • Nuts and bolts to hold the coupler section to the body tubes.  I haven’t drilled the holes for those yet.


Recovery System Progress

I’ve been making progress on the recovery section of my 6″ diameter rocket.  The basic design is a CO2 cartridge, a solenoid, and some tubing blow off the nose cone.  Then a drogue chute pulls out the main chute.  Eventually, I’ll have something hold the shroud lines of the main chute so the rocket can fall on the drogue until low altitude.  But the first flights won’t be to that high altitude so I will let the drogue pull out the main parachute right away.

I got some brackets and mounted the CO2 cartridge/solenoid assemblies in the rocket.  There are two redundant systems.  It’s possible I might eventually use this rocket for my level 3 certification, and that requires redundant recovery system deployment.  The space is tight enough that I figured I should mount two from the start so I won’t get into the situation where I have to rework something to make it fit later.  Also, redundant deployment is good in its own right.

One nice feature of a totally pyro-free CO2 deployment system is that I don’t need to use kevlar webbing to resist the heat of a pyro charge.  Nylon will work just fine.

Why, yes, that is a ten centimeter square.

One of the nice things about a 6″ diameter rocket is that the diagonal of a ten centimeter square is 5.57″.  That means a cubesat can fit inside the rocket.  I’m building a payload bay into my nose cone that can hold a 3U cubesat.  I made a removable endcap for the nosecone that screws into a ring that’s epoxied into the nose cone.


My plan is to offer a free ride whenever I go out to fly to any educational or non-profit organization that has a cubesat.  The rocket will only go up a few thousand feet (up to a few tens of thousands if I complete all of the upgrades on my wish list.)  Still it will be a cool experience for students to see their cubesat fly on any rocket, and with the popularity of cubesats there may be a shortage of free rides.

In other news, life intervened and I couldn’t make it out to the launch this month either.  That’s still on the to-do list.