Post flight analysis, anything to be learned?

Not really sure where to put this post, but it's a scratch built rocket, so I'll post here.

This morning on my way to work I stopped and launched a rocket that I've launched many times before, usually on 24/40 reloads, and sometimes D12s for small fields. It is a scratch built, three-fins-and-a-nose cone, dual deploy rocket in BT-55. Very simple and straightforward.

Today's motor was a D15 (blue thunder) reload, a motor that I've used before multiple times and I launched from a 3/16" rod. Almost right off the pad, seemingly while the rocket was still under thrust, it went side ways and both ejection charges fired. It floated down from about 100 feet.

I'm not expecting to know exactly "what happened", but I have data from the flight, and I'm always interested in seeing what I can learn from it. The data/simulating/testing aspect of rocketry is what gets me the most excited.

Here are plots of the altitude derived from barometric pressure, and the filtered velocity calculated from it. The red lines are from a very normal launch previously, and the blue is from today's launch.



The burn time of the D15 is 1.2 seconds, according to thrustcurve. So it seems it went sideways while under thrust.

I only have altimeter vent holes on one side of the body tube. I think that what happened is that the vent holes were on the low pressure side of the rocket as it got sideways, and that resulted in a false altitude over 300 feet. It certainly didn't go that high.

What I don't understand, is why the rocket went sideways to begin with. The margin of stability is more than 2 calibers, but maybe it suffered a bad case of "long rocket syndrome".



According to the altimeter, both ejection charges fired at 2.15 seconds, after it got sideways and stalled out. So it wasn't a case of premature deployment. The fact that the rocket got sideways was the initial cause.

I probably should add a minimum of one vent hole on the other side of the rocket, but I'm not sure why the rocket got so sideways to begin with.

The only thing I did differently this morning was hold the igniter in place with a short piece of wire insulation, because I didn't have the red cap that one normally uses. It didn't hold the igniter very firmly, so I wonder if it got ejected from the motor before it had built up full pressure. But usually Blue Thunder lights so fast...

So what do you think?

Hard to tell without video or pictures. Were the launch lugs clean and undamaged post flight?

Nytrunner said:

Hard to tell without video or pictures. Were the launch lugs clean and undamaged post flight?

Click to expand...

I really wish I had a video of that one. I filmed the first few launches, but the rocket was so predictable I stopped with the video camera. The launch lugs are cleaned and undamaged. I just checked, and they still slide smoothly over the launch rod. Maybe there was some bad rod whip that kicked the rocket around as it left. It just a flimsy Estes low power pad, probably too small for this rocket.

Pic of the nozzle, is it ovoid / lopsided / did a chunk come out?

dhbarr said:

Pic of the nozzle, is it ovoid / lopsided / did a chunk come out?

Click to expand...

Oi, cleaned out the case at work. I glanced at it and tossed it, I didn't notice anything. It's probably still sitting in the trash next to my desk, I'll take closer look if it is still there.

A couple of questions - does the airframe look like it 'crimped' in flight, or did it look like it was wiggling around from the ground? It could be that the rocket sorta bent in half and that's what caused the pitch-over. Again, hard to tell without video of the flight or pictures of the reload nozzle and rocket post-flight. Looking at the data seems to support your conclusion that the single vent hole is what caused the altitude spike on the altimeter - I would also hazard a guess that the oscillations that you see are either the rocket rolling as it pitched over, or oscillations during that pitch-over. Would agree with your assessment to have multiple holes evenly spaced.

The airframe looks to be intact. The joints seem to be solid, there isn't much play.

I do remember a flight some time ago on an E18 where it looked like there was a rapid, small amplitude oscillation on the way up, during thrust. I didn't think much about it at the time, but maybe it is related.

Also, in Rocksim, if I plot "angle of attack" for simulations, I can get large oscillations for some wind conditions. I'll post some plots if I can reproduce them. Maybe this rocket has a high moment of inertia and low damping (despite decent stability margin).

Well, nothing dramatic, the flight angle seems to be similar to other rockets for similar wind conditions. It was a quite calm day, actually.

mbeels said:

The airframe looks to be intact. The joints seem to be solid, there isn't much play.

I do remember a flight some time ago on an E18 where it looked like there was a rapid, small amplitude oscillation on the way up, during thrust. I didn't think much about it at the time, but maybe it is related.

Also, in Rocksim, if I plot "angle of attack" for simulations, I can get large oscillations for some wind conditions. I'll post some plots if I can reproduce them. Maybe this rocket has a high moment of inertia and low damping (despite decent stability margin).

Click to expand...

That would be a bit unusual, unless the fins are really small. Never seen it before personally...

Thanks for the replies, probably falls under the category of things that happen when you put a controlled explosion on the end of a cardboard tube.

mbeels said:

Thanks for the replies, probably falls under the category of things that happen when you put a controlled explosion on the end of a cardboard tube.

Click to expand...

Except that it’s not an explosion, controlled or otherwise. That’s a common misconception that cost us hundreds of thousands to prove to the ATF in a court case that lasted nine years. [emoji851]

Steve Shannon said:

Except that it’s not an explosion, controlled or otherwise. That’s a common misconception that cost us hundreds of thousands to prove to the ATF in a court case that lasted nine years.

Click to expand...

Ah, well, thanks for the correction. I'm glad you stuck to it and won that one for the sake of rocketry.

mbeels said:

It was a quite calm day, actually.

Click to expand...

Kills my theory; at 2.75 calibers I was going to suggest good old fashioned weather cocking. The design doesn't look tall enough, to me, to suffer from excessive Cp migration. The fins look a little small to me, but I must be wrong since you've got a large static margin.

Did you have an accelerometer on board? I guess not, or you'd have posted that data already. Too bad, as it would let us get the speed off the rod much better than the barometric does.

In the simulation your speed off the rod looks to be about 140 ft/s, which obviously would be plenty. But that's simulation.

The oscillations shown on the simulation look like a lot to me. Come to think of it, I guess I don't really know what's typical beyond that really fast damping out (critical damping or more, or close to it) is neither common nor desirable. Still, I had expected something like 4 to 6 cycles of significant amplitude, not 10. Am I wrong here?

In any case, that's not the culprit, at least not by itself. Lots of oscillation, all else being equal, would have had it correcting, over correcting, going back and forth, which is not what you describe happening.

The double ejection event makes perfect sense. The apogee charge (I assume one was set to fire at apogee) fires because you were at apogee, and the main fires because you were at or (more likely) below its programmed deployment altitude and descending. The fact that both of these conditions were met at once means that both charges fired exactly as the should have.

dhbarr said:

Pic of the nozzle, is it ovoid / lopsided / did a chunk come out?

Click to expand...

Ah, I forgot to follow up, I did find the nozzle, and it looks round and intact. Seems ok as far as I can tell.

jqavins said:

Kills my theory; at 2.75 calibers I was going to suggest good old fashioned weather cocking. The design doesn't look tall enough, to me, to suffer from excessive Cp migration. The fins look a little small to me, but I must be wrong since you've got a large static margin.

Click to expand...

Perhaps there were gusts aloft that I didn't feel on the ground.

Because of the weight of the altimeter, I was able to make the fins smaller than I would have otherwise, according to the software. I wonder if this pushes some of the assumptions about the contributions to the CP from tubes and fins closer to a limit where the assumptions are less valid. But it seemed like a pretty vanilla rocket design to me.

jqavins said:

Did you have an accelerometer on board? I guess not, or you'd have posted that data already. Too bad, as it would let us get the speed off the rod much better than the barometric does.

Click to expand...

I wish I had! My next flight computer will be something like the Eggtimer Proton. Having accelerometer on board would be interesting for other things as well, like calculating Cd.

jqavins said:

The oscillations shown on the simulation look like a lot to me. Come to think of it, I guess I don't really know what's typical beyond that really fast damping out (critical damping or more, or close to it) is neither common nor desirable. Still, I had expected something like 4 to 6 cycles of significant amplitude, not 10. Am I wrong here?

Click to expand...

I don't really know either! I do recall observing the oscillations on one flight with an E18, but it damped out quickly. It certainly doesn't seem desirable to me. Perhaps the first oscillation exceeded some critical limit where the fins stalled out and became less effective. I will definitely video record the next launch.

jqavins said:

The double ejection event makes perfect sense.

Click to expand...

I had wondered what the Eggtimer classic would do in a situation like this, turns out it did a reasonable thing. I did witness someone else's dual deploy flight that never deployed the mains, because it never reached their "main" altitude.

Bat-mite said:

Chaos theory?

Click to expand...

Actually that flight sort of resembled the Lorenz attractor! Up and down in a nice quick arc.

Thanks for all the input, I'll get a video of the next launch.

mbeels said:

Perhaps the first oscillation exceeded some critical limit where the fins stalled out and became less effective. I will definitely video record the next launch.

Click to expand...

I doubt it. Oscillation (in any system) is the interpay of some sort of restoring force and inertia (or an inertia substiture such as hystersis). With the fins stalled you wouldn't have the restoring force and it'd just start tumbling. Unless I'm overlooking something, which is always a possibility.

jqavins said:

I doubt it. Oscillation (in any system) is the interpay of some sort of restoring force and inertia (or an inertia substiture such as hystersis). With the fins stalled you wouldn't have the restoring force and it'd just start tumbling. Unless I'm overlooking something, which is always a possibility.

Click to expand...

I shouldn't have used the word "oscillation" for this particular flight, during this launch it just went sideways, stalled out, and came down. During a previous launch on an E18 I did observe an oscillation and it eventually damped out on its way to 1600 feet.

I did some reading on dynamic stability, and ran more simulations in Rocksim. I suspect that although this rocket has an adequate static margin, it seems to have poor dynamic stability. That is something I gave no consideration to while designing this rocket. So if nothing else, I've learned something new about dynamic stability of rockets.

Based on the simulations and my observations, I think it is susceptible to oscillations. The rocket has a relatively high longitudinal moment of inertia and a relatively low damping ratio, so a perturbation to the flight path results in a high angle of attack with a restoring force, but the low damping ratio allows the oscillations to persist for several periods.

I can't do much to change the mass distribution (moment of inertia), but I can improve the damping ratio. I've lengthened the fins by about 3/16" (adding 10-15" fin area), I'll video the next launch and see if it has improved.

Yup, that makes sense. Didn't I say earlier that the fins look a little small, and that I must be wrong because of the large static margin? If the fins are too small, and the margin comes from lots of forward mass, then this would be the result. Of course, I didn't think it through that far either.

If all this is the answer, are you sure you can't reduce the moment of inertia? Lighten the nose cone or parachute? What's that mass object just above the CG, and what's the coupler near it for? The nose cone and parachute, being further from the CG, would be better targets for reducing the moment of inertia if they weigh much to begin with. OF course, that shifts the CG aft, which eats into your static margin, which...

Usually, if the rocket looks "normal" then the static margin is all you need to worry about. The dynamic stuff is trickier, and knowing when to worry about it is trickier yet. And that's just one piece of the reason that "rocket science" is the cliché phrase for something complicated and intellectually difficult.

jqavins said:

Yup, that makes sense. Didn't I say earlier that the fins look a little small, and that I must be wrong because of the large static margin? If the fins are too small, and the margin comes from lots of forward mass, then this would be the result. Of course, I didn't think it through that far either.

Click to expand...

I think that is exactly it, in a nutshell.

jqavins said:

If all this is the answer, are you sure you can't reduce the moment of inertia? Lighten the nose cone or parachute? What's that mass object just above the CG, and what's the coupler near it for?

Click to expand...

It would be difficult. The rocket is packed FULL. The shock cord packs inside the nose cone, because there is no room elsewhere. I also switched from a 18" parachute to 15" because the 18" was very difficult to pack.

The mass object is the flight computer and batteries. I suppose the other option would be to remove it, replace it with a coupler, and try a flight with just motor eject. But as you said, that would move the CG back.

It's probably unusual to pack a DD flight computer into a relatively light BT-55 rocket.

jqavins said:

And that's just one piece of the reason that "rocket science" is the cliché phrase for something complicated and intellectually difficult.

Click to expand...

True, and in my opinion, a big part of the fun.

mbeels said:

True, and in my opinion, a big part of the fun.

Click to expand...

No argument there!

mbeels said:

(adding 10-15" fin area)

Click to expand...

I mean 10-15% more area....

What does RockSim have to say?

jqavins said:

What does RockSim have to say?

Click to expand...

It does help the damping ratio, but only slightly. The simulations still have oscillations (just -less-). We'll see, I'm planning on launching it this Saturday.

I'm considering a BT-60 redesign, and building it around an altimeter with an accelerometer.

Well, I have two suggestions for a redesign as lessons learned from this.

1. Make it look "normal" from the outside. Length:diameter:fin size ratios that look normal do so because we're used to them, because lots of successful rockets have been build that way. It's a reasonably good starting point, and if you end up very far from there then look again.
2. Keep the DD system mass as near to the desired and final CG as you can. Sure, if the design needs nose weight then it'd be great to take advantage of weight you need anyway. Other than that, the closer the extra weight is to the desired and final CG, the less it affects the moment of inertia.

jqavins said:

Well, I have two suggestions for a redesign as lessons learned from this.

Click to expand...

Yes, I think that those are both good suggestions. As much as a like "not-normal" rockets, a bread-n'butter design that looks normal is a good platform. Probably there is some convergence of engineering design choices that results in the "normal" look for good reasons.

This week I'm in Michigan visiting family, so I went out to the Muskegon rocketry club and launched it on an E18. It was a very windy day, the gusts were probably over 20 mph. It was only one launch, but to me, the oscillations look significantly improved (smaller amplitude and damped out faster). There is a little wiggle that is apparent in the slow motion, but it doesn't seem too bad to me.

The rocket went completely out of sight and I assumed it was gone, especially with the strong winds. However, I decided to take a random figure-8 walk and by a small miracle, saw the fin can sticking out of some weeds on the edge of a drainage canal. The parachute was kissing the water, had it gone a few more inches, I wouldn't have seen it. I've had some SUPER lucky recoveries with this rocket. It is just too hard to track a 33mm rocket at 1500+ feet.

It was also the first launch of the Mach 1 rocketry Alien Interceptor on a Cessaroni G58 (white). It had a pretty wild deviation off the rail, it looks like it weather-cocked pretty hard into the wind and then over corrected. A more extreme variation of the oscillations I saw previously with Bare Metal. Perhaps it wasn't going fast enough off the rail with the strength of horizontal winds that were blowing.



All in all, I took both rockets back home, so it was a success. Next time I'll try a higher thrust motor in the Alien Interceptor.

mbeels said:

I mean 10-15% more area....

View attachment 388819

Click to expand...

Are you trailing edges square?

If so that’s not going to help your fins produce lift. Square edges make turbulent air. The leading edge looks round which is better, but more pointy is probably better.

The leading edges are rounded, and the trailing edges are tapered down to about 1/16". Previously, they were rounded off. I squared off the edge to add the extra length of balsa, and then sanded the balsa to a wedge. The fin has a roughly airfoil shape, with most of the cross section flat.