Unexpectedly high RASAero estimate for a two-stage rocket.

Princeton's rocketry lab plans to make an attempt at a flight past the 100 km von Karman line for space in May, 2018 using a two-stage rocket. Interestingly, they intend to use commercial motors for both stages. If they succeed, this would be the first time commercial motors available to amateurs were used for a flight to suborbital space.

In that vein I modeled such a rocket using an O8000 booster and N1000 sustainer with both RASAero and OpenRocket. I was quite surprised when RASAero returned a result well in excess of the 100 km line for suborbital space. The OpenRocket result was a more conventional 8 to 9 km though.

I'm just starting to work with these programs however, and I'm requesting some feedback on the validity of these results.

The RASAero sim is available here:

O8000 to N1000.CDX1
https://drive.google.com/open?id=1Frz2nqyrv6l85HoI1VF2N-MX3zwZhqX6

The OpenRocket sim is here:

O8000 to N1000.ork
https://drive.google.com/open?id=1yllGlIGGnOVKJh5tLtA4tUY_3JoVnwS7


Image from the RASAeros sim using a CTI O8000-P for the booster and an Aerotech N1000W for the sustainer:



Bob Clark

How does this thread differ from the other one where you received advice that your weights and finishes likely needed some more tuning?

First mantra learned concerning computer programs......

........... garbage in, garbage out.

What I don't understand is your profile lists you as a mathematics professor at an East coast college.
Yet you continually are making comments that math proves, make no sense.

dhbarr said:

How does this thread differ from the other one where you received advice that your weights and finishes likely needed some more tuning?

Click to expand...

There was a response that I should start a thread separate from the other one on the Princeton rocket and this would result in more responses.
You can’t please everyone I guess.

In any case this post has a link to the RASAero sim. To see the estimated altitude, open it in RASAero, then click on the “Flight Simulation” tab at the top. Then click the “View Data” tab in the next window that opens.

I intend to try the various combinations of the suggestions made by Chuck Rogers to get more realistic answers. However, I expect others who have spent years using RASAero have encountered scenarios where tweaks in the rocket resulted in radical changes in the resulting altitude.

Bob Clark

RGClark said:

There was a response that I should start a thread separate from the other one on the Princeton rocket and this would result in more responses.

[opening a file tutorial]

I intend to try the various combinations of the suggestions made by Chuck Rogers to get more realistic answers. However, I expect others who have spent years using RASAero have encountered scenarios where tweaks in the rocket resulted in radical changes in the resulting altitude.

Click to expand...

Continuing a conversation elsewhere and restarting a conversation elsewhere are not the same thing.

Do the sims posted here take into account the feedback already received in that other thread? People who have spent years writing RASAero have given a specific, measurable, actionable, relevant, and timely list of suggestions.

RGClark said:

The OpenRocket result was a more conventional 8 to 9 km though.

Click to expand...

You had the ignition settings wrong in your OR sim and so the second stage never ignited. You need to set the ignition to "burnout of previous stage" plus the delay you want since these motors don't have ejection charges. You want to separate the stages at first-stage burnout. The sustainer seems marginally stable so I added some nose weight. When I make those changes and a couple of others, it sims to about 70K feet, which is fairly believable. I haven't tried to optimize the design at all.

I also just noticed that your fin thickness was set to 0 inches, which is not very realistic.

Also, you should set the fin profiles to rounded or airfoil, as OR handles square profiles poorly.

Your file's first-stage freeform fins were mangled in the version of OR I'm using (14.06). I would avoid using freeform fins as they may not simulate well. I'm also skeptical of OR's handling of your squat conical nose cone so I changed it, but OR's handling of nose cone shape's effect on drag may not be as good as Rasaero's.

Others have already addressed your unrealistic airframe masses.

In my experience, if OR and Rasaero produce wildly different results, there is something wrong with either or both. I have no convenient way to run Rasaero so I can't speak to those results.

BTW, what you consistently fail to acknowledge is that if flying to 100 km was something that could be done with commercial motors, amateurs would have done it a long time ago. Even 100K feet is surprisingly difficult and has only been done by a handful of people.

RGClark said:

The OpenRocket result was a more conventional 8 to 9 km though.

Click to expand...

For a start the two designs aren't even close to being the same. Nor do they look like the Princeton rocket.

I find the Openrocket design especially amusing. The Estes Mosquito style fins on the booster are a hoot. But I wonder why Openrocket included them in the CP calculation since they are specified as having zero thinckness.

UhClem said:

But I wonder why Openrocket included them in the CP calculation since they are specified as having zero thinckness.

Click to expand...

Openrocket apparently doesn't look at the fin thickness when calculating the CP, which makes sense since I don't think Barrowman uses the fin thickness as a parameter.

Do the sims posted here take into account the feedback already received in that other thread? People who have spent years writing RASAero have given a specific, measurable, actionable, relevant, and timely list of suggestions.

Click to expand...

The subject matter of the thread got moved to the new thread here, so my response has been moved, which is below.


Bob:

For rockets of this type, I recommend using the settings All Turbulent Flow (Pages 37-19 in the RASAero II Users Manual), and Rough Camouflage Paint (Pages 34-35 in the Users Manual). Mach 3+ rockets typically have roughened surfaces from the scarring from aerodynamic heating, which trips the flow to turbulent flow, and increases the surface roughness. From altitude comparisons (predicted altitude compared to actual altitude from flight data) for Mach 3 class rockets, these settings provide the most accurate altitude predictions.

Does the rocket have a rail guide or launch shoe? An appropriately sized one has to be added.

You can see the trend here. As you add more and more realism, the altitude keeps falling.

These types of rockets are very sensitive to weight and Drag Coefficient (CD). You'll see a considerable decrease in altitude from using All Turbulent Flow and Rough Camouflage Paint, compared to the default (All Turbulent Flow Box Not Checked) Laminar-Transition to Turbulent-Turbulent Flow and the Smooth (Zero Roughness) settings. Again, for Mach 3 class rockets these settings have provided accurate RASAero II altitude predictions up to 120,000 ft.

Your second stage empty weight is really inaccurate. An N1000W has a loaded weight of 27.3 lbs. Your second stage loaded weight is only 30 lbs. It's unrealistic that the rest of the second stage, minus the motor loaded weight, is only 3 lbs. I'm guessing your first stage empty weight is also inaccurate. These unrealistically low empty weights will have a BIG effect on the apogee altitude.

Take a look at the empty weight of the Loki Dart first stage, to get some idea of the empty weight of what the first stage or second stage would look like. Of course that would include the Loki Dart motor case, but that will represent the metal airframe over the motor you would like use. If you mount Fin Cans to the motors, then the weight would be conservative. Which reminds me, you'll likely use Fin Cans on each stage. They will cause additional drag, so you will have to add Fin Cans to your configuration.

Could an amateur build a rocket with as low of an empty weight as a Loki Dart first stage?

Add the above to the RASAero II sim, and put in more realistic empty weights, and you'll improve the realism of your altitude prediction.


Chuck Rogers
Rogers Aeroscience

RGClark said:

However, I expect others who have spent years using RASAero have encountered scenarios where tweaks in the rocket resulted in radical changes in the resulting altitude.

Click to expand...

Change the empty weight of the second stage from 3 lbs to 6 lbs, and then to 10 lbs, make equivalent changes to the first stage empty weight, and believe me, you will see a radical change in the altitude of the rocket. Rockets of this class fly like sounding rockets, the higher the propellant fraction, the greater the burnout velocity, and the higher the coast. These rocket are also very sensitive to changes in CD, more drag in the lower atmosphere, lower velocity when you get to the thinner upper atmosphere, the lower the coast distance.

You really need to nail down accurate empty weight estimates for the first and second stages. As I posted, take a look at the Loki Dart booster and other sounding rockets. But remember, almost all high power/amateur rocketeers have not demonstrated those levels of empty weights/propellant fractions.


Chuck Rogers

Another thing to note: You set the nozzle exit diameter for the upper stage to 4" and 6" for the lower stage. Not sure exactly how RASAero uses this information, (probably something to do with expansion ratio?) but changing both of them to real values for the motors in question cuts the altitude down to a third of your original sim. That's without changing anything else, such as the unrealistic stage masses.

eggplant said:

Another thing to note: You set the nozzle exit diameter for the upper stage to 4" and 6" for the lower stage. Not sure exactly how RASAero uses this information, (probably something to do with expansion ratio?) but changing both of them to real values for the motors in question cuts the altitude down to a third of your original sim. That's without changing anything else, such as the unrealistic stage masses.

Click to expand...

In RASAero II, the rocket motor nozzle exit diameter is not only used to determine the power-on drag coefficient (CD) for each stage, it is also used to vary the thrust with altitude for each stage. RASAero II varies the thrust with altitude for the rocket motor in each stage based on the nozzle exit area calculated from the entered nozzle exit diameter, and based on the assumption that the thrust curve data in the rasp.eng file is the thrust curve at Sea Level. RASAero II then compares the atmospheric pressure at altitude from the Flight Simulation atmosphere model to Sea Level atmospheric pressure, and based on the nozzle exit area determines the increase in thrust at altitude.

So the nozzle exit diameter has to be correct not only to get the correct power-on drag coefficient (CD), but also to get the correct increase in thrust with altitude.

The rocket performance is sensitive to CD, changes in the power-on CD will have an impact on the burnout velocity, which will determine the coast distance. Depending on the ignition and burnout altitudes of the second stage, the wrong nozzle exit area can have a noticeable effect on the second stage thrust curve, and hence the second stage specific impulse. Again, like a sounding rocket, small changes in specific impulse using the same propellant fraction will vary the burnout velocity, again affecting the coast distance.

Because the nozzle exit area is used to vary thrust with altitude (and hence specific impulse), and to determine the power-on CD, both of which the rocket apogee altitude is sensitive to, the correct nozzle exit diameter has to be entered for each motor.


Chuck Rogers
Rogers Aeroscience

eggplant said:

Another thing to note: You set the nozzle exit diameter for the upper stage to 4" and 6" for the lower stage. Not sure exactly how RASAero uses this information, (probably something to do with expansion ratio?) but changing both of them to real values for the motors in question cuts the altitude down to a third of your original sim. That's without changing anything else, such as the unrealistic stage masses.

Click to expand...

I also just noticed that the second stage diameter is 4.0 in, and the first stage diameter is 6.5 in. So you've filled-in the base area with the nozzle exit area on the second stage, and almost completely filled it in with the nozzle exit area on the first stage. That means approximately you will have no base drag for power-on during Stage-1 and Stage-2. (Approximately, because the RASAero II power-on base drag model is more complex than just subtracting the nozzle exit area from the base area of the rocket.)

No base drag will really help lower the power-on CD, thus increase the burnout velocity, and thus the coast distance and the apogee altitude.

If you add a higher nozzle expansion ratio to fill in the base area of the rocket with the nozzle exit area, you of course will have to enter a new sea level thrust curve (predicted data or actual static firing data).


Chuck Rogers

mikec said:

You had the ignition settings wrong in your OR sim and so the second stage never ignited. You need to set the ignition to "burnout of previous stage" plus the delay you want since these motors don't have ejection charges. You want to separate the stages at first-stage burnout. The sustainer seems marginally stable so I added some nose weight. When I make those changes and a couple of others, it sims to about 70K feet, which is fairly believable. I haven't tried to optimize the design at all.
I also just noticed that your fin thickness was set to 0 inches, which is not very realistic.
Also, you should set the fin profiles to rounded or airfoil, as OR handles square profiles poorly.
Your file's first-stage freeform fins were mangled in the version of OR I'm using (14.06). I would avoid using freeform fins as they may not simulate well. I'm also skeptical of OR's handling of your squat conical nose cone so I changed it, but OR's handling of nose cone shape's effect on drag may not be as good as Rasaero's.
Others have already addressed your unrealistic airframe masses.
In my experience, if OR and Rasaero produce wildly different results, there is something wrong with either or both. I have no convenient way to run Rasaero so I can't speak to those results.
BTW, what you consistently fail to acknowledge is that if flying to 100 km was something that could be done with commercial motors, amateurs would have done it a long time ago. Even 100K feet is surprisingly difficult and has only been done by a handful of people.

Click to expand...

Thanks for the info. I read of some amateurs that tried to reach the Karman line with staging using smaller motors, commercial or homemade, and it seems that the problem was with the staging. So if a team can iron out those problems with staging perhaps an all commercial motor flight to the Karman line is possible. Note also universities have more financial resources than the independent amateur so can essentially make another attempt every year.

I tried doing some tweaks on the design of the OpenRocket sim to see what altitude I could wring out of it. First, I used a slimmer sustainer motor at 75 mm, 3 inches, diameter to reduce drag. I used the M2245 motor.

Next, I gave it more realistic shapes for the booster fins. Note though the reason in the OpenRocket sim I used those exaggerated fins was I was trying to pull back the CP for stability. With the new fins, the CP is disturbingly close to the CG. I didn't get any stability warnings so I let them stay. I also increased the fin thickness to 0.1 cm. For strength of the fins at this slim thickness I chose the fin material as carbon fiber. I also chose carbon fiber for the nose cone and transition.

About the body tube, most rockets meant to get to space, orbital or suborbital, use the rocket tanks themselves as the body of the rocket. So to reduce weight further I chose "Custom" in the selection for the body material and set the density to 0.0.

The nice thing about OpenRocket is that it'll display its estimated altitude as you are making these changes. I was surprised then how much the altitude changed by changing the nose cone shape and length, and how much it was changed by the transition length. The result below:




The link to the new OpenRocket sim is here:

https://drive.google.com/file/d/1ybAB9So4x2ngVJzI72h6-CrPC4oLOc-O/view?usp=sharing

Bob Clark

What about all the weight for recovery gear, electronics, ect. Also 1mm thick fins will flutter and explode, you need much more realistic fins thickness, weights, lengths (where are you going to put all the necessary stuff).

200km and m6 suggests there's plenty of improvement to be had that doesn't involve adiabatic blowtorch confetti.

I suggest cranking space shuttle propellant through Tsiolkovsky to see where a reasonable upper limit best case scenario might be.

That nose cone shape is awful, so there's clearly a problem there with the drag model. Also, those fins would be incredibly difficult to keep from fluttering - that high aspect ratio and substantial sweep means that the torsional and flapwise stiffness will suck. To solve this, you'll need thick fins, which means a bunch of drag. Basically, this isn't even close to a viable design.

UhClem said:

For a start the two designs aren't even close to being the same. Nor do they look like the Princeton rocket.

I find the Openrocket design especially amusing. The Estes Mosquito style fins on the booster are a hoot. But I wonder why Openrocket included them in the CP calculation since they are specified as having zero thinckness.

Click to expand...

Yes. This is not the Princeton design. The statement in the first post about "in the same vein" only meant it would use commercial motors.
The rather elaborate fins on the OpenRocket booster were a first attempt to bring the CP rearward for stability. They would not be in a final design.

By the way, if you want to inquire of the Princeton design you can ask on their web page or Facebook page:

https://rockets.princeton.edu/spaceshot

https://www.facebook.com/princetonspaceshot/

Bob Clark

blackjack2564 said:

First mantra learned concerning computer programs......

........... garbage in, garbage out.

What I don't understand is your profile lists you as a mathematics professor at an East coast college.
Yet you continually are making comments that math proves, make no sense.

Click to expand...

What math is that?

Bob Clark

RGClark said:

Click to expand...

Bob:

Using Google Images, search Images for "Sounding Rocket".

https://www.google.com/search?hl=en.....0...1ac.1.64.img..9.6.1643....0.DzY3Ub1k4YM

Take a look at the nose cones used on these sounding rockets. The designers of these sounding rockets obviously gave a lot of thought on what L/D to use for the nose cone.


Chuck Rogers

Chuck Rogers said:

Bob:

Using Google Images, search Images for "Sounding Rocket".

https://www.google.com/search?hl=en.....0...1ac.1.64.img..9.6.1643....0.DzY3Ub1k4YM

Take a look at the nose cones used on these sounding rockets. The designers of these sounding rockets obviously gave a lot of thought on what L/D to use for the nose cone.


Chuck Rogers

Click to expand...

Yes. I was quite surprised that OpenRocket kept giving me higher and higher altitudes as I made the nose cone shorter and squatter. Certainly not the shape you would expect on a rocket intended for high Mach speeds!

Perhaps you need the pointed nose cone when you have to have a long nose cone for sufficient volume for the payload. It is notable that for orbital rockets the Max Q point, where the total dynamic and static pressure is maximum, occurs at rather low speeed and altitude. Or it be could be OpenRocket just needs more tweaking in its drag calculations for the nose cone.

It was also surprising that the altitude according to OpenRocket kept increasing as I made the transition longer and longer. It could have gotten higher still, over 200 km, if I made the transition longer but the rocket began to look ungainly. I have no explanation why that should be. Or again it may be OpenRocket needs more tweaking in this regard.

I’ll do a RASAero sim on the same rocket to get a better idea if these are real effects.

Robert Clark

Bob, I do believe that Open Rocket bases its nose cone drag on surface area of the cone. The shorter and smaller the cone the less drag it sees.

Rocksim does the same.

There are at least a few places in OR where the estimates diverge pretty radically from reality. You'll get stub noses, impossibly long fins, excessive transitions, tiny tubefins, etc.

Usual symptoms are impossibly high / fast or low / slow results. As with any simulation, the most important check is behind the keyboard.

FWIW, I tweaked your OR sim and optimized upper stage mass and ignition delay, and got an altitude of 392K feet. So, if you believe OR and your impossibly low airframe masses, you could get above 100 km with these motors. Of course I don't for a second believe that one could build something at this weight that would hold up to this motor combination, much less be recoverable, so I'm not sure what this proves, if anything. (BTW, I was using the filename you originally used which suggests you were using an N1000 for the upper stage, but all of your sims are with the N5800 for the upper stage. The N1000 would be a much easier motor to survive but only sims to 262K feet.)



View attachment O8000 to N1000fix.ork

markkoelsch said:

Bob, I do believe that Open Rocket bases its nose cone drag on surface area of the cone. The shorter and smaller the cone the less drag it sees.

Rocksim does the same.

Click to expand...

I just saw an article (Sport Rocketry?) on nose cone shapes that said that up to ~M0.8, the surface friction drag dominates on nose cones, so a short and round nose is preferred. Or at least, the added skin friction and weight of a long nose outweigh any reduction in nose cone Cd. Above that speed, Mach shock wave effects make a long and pointy nose cone better.

boatgeek said:

I just saw an article (Sport Rocketry?) on nose cone shapes that said that up to ~M0.8, the surface friction drag dominates on nose cones, so a short and round nose is preferred. Or at least, the added skin friction and weight of a long nose outweigh any reduction in nose cone Cd. Above that speed, Mach shock wave effects make a long and pointy nose cone better.

Click to expand...

Well I just learned something, that's for sure. I would have Cd to dominate. Very curious to see that article.

RGClark said:

Next, I gave it more realistic shapes for the booster fins. Note though the reason in the OpenRocket sim I used those exaggerated fins was I was trying to pull back the CP for stability. With the new fins, the CP is disturbingly close to the CG. I didn't get any stability warnings so I let them stay.

Click to expand...

With such an unstable rocket you'd be lucky to reach 200ft. Given that 200km seems a bit of a stretch.

RGClark said:

Yes. I was quite surprised that OpenRocket kept giving me higher and higher altitudes as I made the nose cone shorter and squatter. Certainly not the shape you would expect on a rocket intended for high Mach speeds!

Perhaps you need the pointed nose cone when you have to have a long nose cone for sufficient volume for the payload. It is notable that for orbital rockets the Max Q point, where the total dynamic and static pressure is maximum, occurs at rather low speeed and altitude. Or it be could be OpenRocket just needs more tweaking in its drag calculations for the nose cone.

It was also surprising that the altitude according to OpenRocket kept increasing as I made the transition longer and longer. It could have gotten higher still, over 200 km, if I made the transition longer but the rocket began to look ungainly. I have no explanation why that should be. Or again it may be OpenRocket needs more tweaking in this regard.

I’ll do a RASAero sim on the same rocket to get a better idea if these are real effects.

Robert Clark

Click to expand...

I can absolutely, 100% guarantee you that this nose cone behavior is nonphysical. In reality, you'll want a much longer, narrower nose cone to minimize supersonic drag. It's also notable that for your rocket, max q will occur at a quite high speed at a low altitude, and the Q at this max Q will be exceedingly high.

boatgeek said:

I just saw an article (Sport Rocketry?) on nose cone shapes that said that up to ~M0.8, the surface friction drag dominates on nose cones, so a short and round nose is preferred. Or at least, the added skin friction and weight of a long nose outweigh any reduction in nose cone Cd. Above that speed, Mach shock wave effects make a long and pointy nose cone better.

Click to expand...

Thanks for that. Keep in mind when I kept shrinking the nose cone size, this was because I had literally zero payload. So I could make the cone size as small as I wanted. I was just curious what the max altitude could be.
In a more realistic scenario you do have to have the nose cone of a certain size. On the other hand with today’s microminiaturization you could have a tiny nose cone if you only wanted say a micro camera and GPS transmitter.

Bob Clark

mikec said:

FWIW, I tweaked your OR sim and optimized upper stage mass and ignition delay, and got an altitude of 392K feet. So, if you believe OR and your impossibly low airframe masses, you could get above 100 km with these motors. Of course I don't for a second believe that one could build something at this weight that would hold up to this motor combination, much less be recoverable, so I'm not sure what this proves, if anything. (BTW, I was using the filename you originally used which suggests you were using an N1000 for the upper stage, but all of your sims are with the N5800 for the upper stage. The N1000 would be a much easier motor to survive but only sims to 262K feet.)

View attachment 343854

View attachment 343853

Click to expand...

Thanks for that. By the way, why are airframes(body tubes) used in high power rocketry? Is this a holdover from the Estes model rocket days where you put a little Estes motor inside the cardboard tube that served as the body of the rocket?

When I was first reading about orbital rockets I was surprised to learn that when you’re looking at rockets meant to fly to space you are looking at the actual propellant tanks or motor casings as the outside surface of the rocket. There is no airframe around the propellant tanks or motors. The same is true for suborbital rockets. This is important of course because for rockets flying to space saving weight is paramount.

For high power rockets attempting to maximize altitude the airframes don’t appear to have any purpose. You could just as well attach the fins to the motor casing using a fin can. With sufficient care, so as not to degrade the casing strength, you could also directly weld the fins to the casing and dispense with the fin can.

Bob Clark