# Nose cones

Discussion in 'The Watering Hole' started by Theo J. Swank, Jun 24, 2019.

### Help Support The Rocketry Forum by donating:

1. Jun 25, 2019

### Nytrunner

#### Master of Rivets

Joined:
Oct 15, 2016
Messages:
4,334
755
Gender:
Male
Location:
Huntsville AL
Tangent ogive coming off the airframe, but magic taper to a parabola apex (power series 1/2) or ellipse somewhere in the middle

2. Jun 25, 2019

### jqavins

#### Joseph Avins

Joined:
Sep 29, 2011
Messages:
1,578
253
Gender:
Male
Location:
Bath, NY
No, a parabolic (paraboloid of rotation) tip.

At first, you could put the change-over at any height and any diameter, which gives you two degrees of freedom, but it wouldn't generally be smooth. Adding the smoothness constraint takes away one degree of freedom, meaning you can pick either the height or diameter and then compute the other (or zero in on it in RockSim, OR, or the CAD program of your choice.) I've done the math, but don't have the formula handy. I also don't have any of the above mentioned software available here at work, nor internet at home, so I'll try to draw one tonight and post it tomorrow.
No magic about it. Just pick the change-over such that the slope of the ogive side (circular arc) coming up from the base matches the slope of the parabola coming down from the tip. Freshman calculus or trial and error. With the trial and error method it's easier to keep adjusting the parabolic length until the slopes match; if I want a fixed total length I'll have to find the paper with the formula on it.

Last edited: Jun 25, 2019
3. Jun 25, 2019

### kuririn

#### BARGeezerTRF Supporter

Joined:
Oct 3, 2016
Messages:
1,804
372
Gender:
Male
Location:
Hawaii
Speaking of nose cones, interesting read about the Nike Smoke nose cone in Alway's Rockets of the World. When they replaced the original short 36 deg. nose cone with the iconic longer 10 deg. one it doubled the rocket's altitude in computer sims.

4. Jun 26, 2019

### jqavins

#### Joseph Avins

Joined:
Sep 29, 2011
Messages:
1,578
253
Gender:
Male
Location:
Bath, NY
Well, I made the pictures last night, put them on a thumb drive, and left it at home. (Shoulda used my phone to transport the files.) Sorry.

5. Jun 26, 2019

### jlabrasca

#### Well-Known Member

Joined:
Oct 31, 2016
Messages:
996
255
Well. I will just keep guessing at it until you post something...

Actually, figuring out how to get the parabola tangent to the circular arc of the ogive in F360 is a nifty problem. There is a tangent constraint, but it doesn't seem to work for conic curves. Or, at least, not in the way I want it to work.

Rob702Martinez likes this.
6. Jun 26, 2019

### Nytrunner

#### Master of Rivets

Joined:
Oct 15, 2016
Messages:
4,334
755
Gender:
Male
Location:
Huntsville AL
That's how I imagine it too.

Interesting, what problems is fusion giving you?
When I tried it in S-works, I just set the parabola apex as the tip with unfixed focal point, put a circular arc tangent to (phantom) body tube with unfixed radius, and used tangent constraint to link them.

7. Jun 26, 2019

### jqavins

#### Joseph Avins

Joined:
Sep 29, 2011
Messages:
1,578
253
Gender:
Male
Location:
Bath, NY
Than looks like you've got about the right idea. I put the change-over about half way along the length.

If you get that, lay it over an elliptical with the same overall length and you'll see it's a little skinnier through the middle, a little sleeker. It's not a huge difference, but I like it.
How do you determine where along the length the two curves meet? Mathematically you can place that anywhere along the length then use the tangent constraint to set the diameter, or the other way around.

Last edited: Jun 26, 2019
8. Jun 26, 2019

### jlabrasca

#### Well-Known Member

Joined:
Oct 31, 2016
Messages:
996
255
When I try to apply the tangent constraint it just collapses the conic curve. I should say, I really don't know what I am doing. It might not work with conics. It might be the coincident endpoints of the conic and the circular arc. I've also already got all kinds of other constraints on the drawing (symmetry, midpoint, perpendicular), so F360 might just be choking on some conflict. I'll play with it some more later. I, too, like this not-quite-ellpitical nose-cone and might try to print one this week.

Edit. It was the coincident endpoints...which makes perfect sense. The slope of the conic is slightly steeper than the slope of the circular arc where I trimmed it, so there was no place where the conic curve could be tangent to the arc.

Works a treat with a closed circle.

Well that's a relief. Now, though, I need some other thing to distract from doing actual work.

Last edited: Jun 26, 2019
9. Jun 26, 2019

### georgegassaway

Joined:
Jan 18, 2009
Messages:
4,268
189

10. Jun 27, 2019

### Sooner Boomer

#### Well-Known Member

Joined:
Mar 21, 2011
Messages:
2,419
81
Personally, I like a little tail

cone.

11. Jun 27, 2019

### jqavins

#### Joseph Avins

Joined:
Sep 29, 2011
Messages:
1,578
253
Gender:
Male
Location:
Bath, NY

Blue is the ogive-parabola hybrid, and magenta is an ellipse of the same base and length.

It all started when I was trying to use a rounded ogive in RockSim. The one's you buy have, I think, spherical tips, but RockSim doesn't have elliptical without parallel sides at the base, so a parabolic tip was the best I could do. That had the change-over much further forward. Then I realized I could play with the shape by moving the change-over, and I like the result.

Incidentally, the hybrid shape's Cg and Cp are both further aft. Since the Cp difference is greater than the Cg difference this should improve stability of a whole rocket, but only by a smidgen. Thinking this through on the fly now, if the design needs nose weight, it could be reduced using the hybrid relative to the ellipse. Now, I expect that the drag with the hybrid is slightly higher than with the ellipse, since we all know that the ellipse is the best of the simple shapes. I also expect the two shapes' drags are very close, whichever proves to actually be better. So, the hybrid shape likely gives more drag but lower weight. It's anybody's guess which will give better performance, so there could possibly be some tiny performance advantage. But I'm sure it's really teeny-tiny either way.

12. Jun 27, 2019

### Nytrunner

#### Master of Rivets

Joined:
Oct 15, 2016
Messages:
4,334
755
Gender:
Male
Location:
Huntsville AL
I'm not convinced the CP is affected tangibly. Openrocket does funny things with transitions, and half your nose cone is now a transition

13. Jun 27, 2019

### jqavins

#### Joseph Avins

Joined:
Sep 29, 2011
Messages:
1,578
253
Gender:
Male
Location:
Bath, NY
You raise a good point. My results are from RockSim, not OR, but I certainly can't say that RockSim is better on that. (It may or may not be.)

14. Jun 27, 2019

### opcon

#### Member

Joined:
May 30, 2019
Messages:
7
3
Gender:
Male
Location:
SW Michigan
I grew tired of using MATLAB so I converted the Haack equations over to Excel. Protected worksheet but no password on it, just so you only change the values necessary.

To use:
Update Ratio
Update Units (make sure all units are same inches with inches, cm or mm's etc)
Update Diameter
Choose a Constant

Length and radius will be calculated for you.

It will spit out a series of 50 x,y coordinates. I tested the maths several times, should be good. Let me know if any problems with it.

#### Attached Files:

• ###### Haack Nose Cone Profiler.zip
File size:
25.6 KB
Views:
8
Nytrunner likes this.
15. Jun 27, 2019

### Flyfalcons

#### Well-Known Member

Joined:
Apr 14, 2015
Messages:
2,094
210
Pointy nose cones get da ladies

16. Jun 28, 2019

### Nytrunner

#### Master of Rivets

Joined:
Oct 15, 2016
Messages:
4,334
755
Gender:
Male
Location:
Huntsville AL
Since the tangent constraint doesn't lock in the radius and parabola focal point, What breed of shape were you looking for?

In descending order I present:
-The "Why didn't you just say you wanted a parabola?
-The "Isn't this the power series?"
-The "Magic Midpoint Mod"
-The "Pretty much an old artillery shell"
and -The "Basic blow-molded Ogive"

neil_w, mbeels and dhbarr like this.
17. Jun 28, 2019

### vcp

#### Well-Known Member

Joined:
Jan 28, 2009
Messages:
880
76
Location:
Boise, ID
That's how I finished my printed NARAM models:

https://www.thingiverse.com/thing:3056609
1. What I did... sand, sand, sand. ABS sands nicely and you can sand aggressively, starting with 150 or 180 grit. I just flat sanded the ridges off completely. The benefit to this is that I measured a 15% weight loss from the material removed. Keep sanding until there is no 'click' when a fingernail is run across the surface anywhere. Then finish off with some 600 to get rid of the scratches from the heavier grits. Finally buff with some 1000-1200. After that, you could really get by with nothing else, but I did give everything a coat of liquid wax (Pledge Floor Care) to give a bit of a gloss.

18. Jun 28, 2019

### vcp

#### Well-Known Member

Joined:
Jan 28, 2009
Messages:
880
76
Location:
Boise, ID
I've done some stuff like this with iterative programming - it would be an interesting project in OpenSCAD. Just thinking about it, I think it would run something like this.
1. Choose base shape, base diameter, length of a full cone in that shape, and the overall desired finished length. Also the tip shape and the height above the base you want to switch to that tip shape.
2. Run up the base shape until you get to the switchover height. Get the slope and diameter there.
3. On a 'unit cone' (base diameter = 1) of the tip shape, run the shape from the tip down until you find a matching slope. Get the unit cone diameter there.
4. Scale the tip shape cone shape so it's diameter at the slope match point equals the diameter of the base cone at the switchover point.
5. Scale the length of the combined shape so that it matches the overall desired length.
I guess that's pretty much what you said.

Last edited: Jun 28, 2019
19. Jun 28, 2019

### OverTheTop

#### Forum SupporterTRF Supporter

Joined:
Jul 10, 2007
Messages:
2,665
429
Gender:
Male
Location:
Melbourne Australia
20. Jun 28, 2019

### jqavins

#### Joseph Avins

Joined:
Sep 29, 2011
Messages:
1,578
253
Gender:
Male
Location:
Bath, NY
To do it in OpenSCAD I would dig up or recreate the math and write it into the code.
• Pick the base diameter;
• Pick the overall height;
• Pick the change-over height;
• Apply the formula for the change-over diameter.
The iterative approach works is fine for a quick one-off in RockSim, OR, or the like, but if you're going to code it then, IMO, it's preferable to do the slightly hard stuff up front and keep the code simple.

The next step I'm working on is to eliminate the last degree of freedom by adding a constraint that the two curves' radii of curvature must also match at the change-over. That would force a particular change-over height. I know I can do the math, but it gets messy and I've managed to get tangled up in it each time I've tried (so far). I'm also not sure there is a solution; since having the "change-over" either all the way at the top (all ogive) or all the way at the bottom (all parabola) would satisfy it, there may not be a solution in between. But I need to get through the messy part in order to prove that to myself.

(Am I the only person who engages in recreational calculus?)

21. Jun 28, 2019

### jbuscaglia

#### Rusty-eyed Missile Man

Joined:
Jan 25, 2009
Messages:
418
2
Location:
Hudson, MA
Just make sure you don't drink and derive.

neil_w likes this.
22. Jun 28, 2019

### mbeels

#### Well-Known Member

Joined:
Feb 9, 2019
Messages:
358
103
Gender:
Male
I hope not.

This is turning out to be surprisingly interesting thread.

23. Jun 29, 2019

### OverTheTop

#### Forum SupporterTRF Supporter

Joined:
Jul 10, 2007
Messages:
2,665
429
Gender:
Male
Location:
Melbourne Australia
I sometimes wonder the same thing!

Title
Replies Views
Last Message

Replies:
5
Views:
73

Replies:
4
Views:
63

Replies:
8
Views:
270

Replies:
3
Views:
57

Replies:
48
Views:
1,001