Richard’s chamber separator disk geometry, using delrin

An analysis of a CSD based on geometry specified by Richard Nakka. In this run delrin was used as the material, with the following properties:

Elastic modulus in X: 448803 psi
Poisson’s Ration in XY: 0.35
Tensile Strength in X: 10805 psi
Compressive Strength in X: 16824 psi
Yield Strength: 9355 psi

Loaded from the bottom with 1000psi, displacement plot:

Loaded from the bottom with 1000psi, FOS plot:

Loaded from the top with 1000psi, displacement plot:

Loaded from the top with 1000psi, FOS plot:

In short, it is insufficiently strong.

Minor revision to electronics bay

The three antennas are moved a bit farther from the inner face of the nose cone, and the battery bay is gapped 2″ from the bottom of the larger ring antenna.

3D PDF: avionics 2012.10.06 3D (12.7MB)

Rendering: (2.3MB)

A new approach to the delay plug, first pass

The current concept for Sugar Shot to Space uses a two burn solid rocket, a novel approach to getting the most altitude out of a single stage. The rocket is two chambers of solid propellant, separated by a delay; the first fires for a few seconds, the rocket coasts for as much as 15 seconds, and then the second firing occurs.

This reduces drag losses compared to a single longer burn, both because the maximum velocity is lower and because more of the burn happens at higher altitude. It also slightly increases the specific impulse (similar to the fuel efficiency of a rocket) and thrust of the second burn because of lower atmospheric pressure.

This is, however, very odd in solid rockets. In most tests the rocket fails during the second burn. It presents a number of challenges, one of which is having something that separates the two charges of propellant but that doesn’t impede the second burn when the time comes for it to do its thing.

The prior approach was a “delay plug” made of slow burning solid propellant. It would burn away sufficiently in the interval between firings that the second could run unimpeded.

Richard Nakka wanted to look at a new approach to it, using a material that is burned away by the second firing. It would have a lower mass, as well as not putting heat into the first burn’s motor casing during the coast phase.

This is the concept specification Richard sent me: DSS Chamber Separator Disc concept. In short, a plastic disk would separate the two motor casings, and would be melted or burned away by the firing of the second burn

It’s a very interesting idea, but I have concerns about the concept. While it would melt and burn away quickly, until it does the hole in the disk would act as a throat. Supersonic exhaust would shoot through the first motor casing until the casing pressurized enough for the real rocket nozzle to act as the sonic choke. It’d be interesting to see it tested, though.

I’ve started on it, and wanted to share the first very rough drafts before I get to a final concept that the modeling says should work.

The material in these analyses is PMMA, a kind of acrylic. It vaporizes well, and has a relatively low melting point, so it would burn away quickly. It’s of average strength for a plastic.

The outer part is a steel sleeve that couples the two motor casings together. It is mild steel. It is sized for the “DSS Boilerplate” engine, it has an outer diameter of 6.35 inches. I modeled a pressure difference of 1000 psi between the lower casing and the top, which creates a fairly huge force on the disk.

First, this is a scenario where a 0.5 inch disk of plastic is bonded into the coupler, both on the edge of the face and the sides of the cylinder. It isn’t strong enough, I’d like to see at least a 1.5x factor of safety (FOS), and ideally 2x.

I wanted to see how it would change if the disk was only bonded on the edge of the face, and not around the sides of the disk. It makes it weaker, but loads the coupler more realistically. Sorry for the color scheme change, but this style is better for a number of reasons. I also removed the screw holes, as they aren’t relevant to the issue being examined.

This is the same study as the last FOS plot, but this is showing how the material will move. It shows it bending in by nearly a half inch at the center. In reality, it would have broken.

This is a different approach, with the disk turned into a dome. Domes hold pressure better, which is why the ends on your air compressor tanks aren’t just flat plates. The situation is significantly improved, with nothing under a FOS of 1, so in theory this would actually hold the pressure.

From here I’ll be doing further work on the dome concept, as well as one with a flat plate reinforced with steel gussets as described in the specification document. I’ll also be looking at other materials, as Richard asked in a subsequent email. I’ll include dimensioned drawings in the next update on this, as well as more FEA images.

Electronics bay modeling

Modeling electronics isn’t one of the core competencies of Solidworks, however it is quite handy for laying out the positions of components in an assembly to make sure that everything is going to fit in the space available.

That has been the objective of modeling the electronics bay of the next big Sugar Shot rocket. While the propellant guys are working to tame the rocket propellant, the electronics team is working to get all of the desired flight data recorders and event triggers working together.

Here are a couple renders of the “e-bay” as it stands now. Click for 2000×10000 pixel images.

Isometric view

Other side

Loadcell adapter plates and the challenges of load constraints

Richard Nakka asked that I run FEA on some plates he’d designed to convert a button-type loadcell into something that could be used for 5000lbf tension testing. It consisted of three 0.375 aluminum plates. Two identical ones on the outside that I named “bread”, and one larger one on the inside that I named “meat”.

I wanted to show the difference in FEA between loading the easy way, into faces of the single part, vs. a more accurate way, into a simulated bolt and loadcell button.

First, here is the meat plate loaded with 5000lb stretching it, with the entire inner surface of the drilled hole applying 5000lb to the right and the rectangular surface of the rounded square hole fixed:

Factor of safety:

Tensile stress:

The plots show that the factor of safety is generally greater than 1, so it won’t deform the part. But while this is a very easy simulation to set up, such that you can do it in the Solidworks “SimulationXpress”, it isn’t a realistic simulation. Essentially, what you’re saying here is that when you put a bolt through the hole it will not only push on one side of the hole, but also be bonded to and pull on the other side of the hole. Bolts don’t do that.

To set up a more realistic simulation, I created parts to mock up a loadcell button and a bolt through the aforementioned hole. Here are the same plots, but instead with the force applied to the back of the button and the bolt fixed on either end:

Factor of safety:

Tensile stress:

The scales are the same in both sets of images. The latter set shows that applying the force through a button and 0.24″ rod significantly changes the stresses in the part. The aluminum plate will definitely yield behind the bolt, though it probably won’t fail. This is a result you can only really find with the extended Solidworks Simulation, as Simulation Xpress doesn’t allow multi-part assemblies.

Finally, and less interesting, here are the same plots for the ‘bread’ plates:

Factor of safety:


And finally a view of how the part will stretch as a spring, with the colors representing how far the various areas of the part will move from their original location:

 Modifying Young’s Modulus or Tensile Strength of KNSB to Allow Casebonding

In this post I consider the possibility, by modifying the physical properties of the propellant, of reducing the odds of having a KNSB solid rocket engine crack its propellant grain during operation. Cracks in solid propellant result in an increase in burning area, which creates an increase in chamber pressure, which can (and recently did) result in the bursting of the rocket motor’s casing.

Here is the burst that was the impetus for these FEA experiments:

What is KNSB propellant? In short, it is a mix of potassium nitrate, which is an oxidizer, and sorbitol, a sugar-like fuel. Read more about it on Richard Nakka’s website.

What are the physical properties that I look at modifying? The first is Young’s modulus, which is a measure of the stiffness of a material. Materials with high Young’s modulus do not flex easily. Normal KNSB has a Young’s modulus higher than that of MDF, the fake wood used to make furniture: difficult to bend, like a desk top.

The other property I consider modifying is the tensile strength of the KNSB. As described on Richard Nakka’s website, KNSB appears to have a tensile strength of around 1050psi, roughly a quarter or half that of epoxy.

In the following plots the factor of safety, or FOS, is plotted on a star-core rocket grain inside an aluminum casing. 1000 psi of gas pressure has been applied to the inside of the star, to simulate the pressure of the motor starting to burn. The plot shown is a slice through the middle of the grain, where stresses are at their lowest; this is the best case in the entire grain. If the FOS here is below 1, the grain will crack. The FOS is plotted from 0 (red) to 3 (blue); 1.5 is a minimum value that we’d like to see, 3 is much better.

This first plot is basically the same situation in the earlier case cast post; it is a reference image at the given KNSB material properties of 850ksi Young’s and 1050psi tensile. Too much yellow, red, and orange, where it would (and probably did) crack.
What happens if we drop the Young’s modulus to 70ksi, i.e. less than a tenth, making it as flexible as a plastic bottle? A significant improvement, but the outer points of the star are still weak.


What if we had a KNSB that had that modulus of 70ksi and we managed to double the strength of the material to 2000psi? A dramatic improvement! No orange left in the plot, and almost the entire grain has a FOS of greater than 3.



What’s it look like with just the increased tensile strength, and no modification to Young’s modulus? Here it is with 2000psi tensile KNSB. Not as good, but still a dramatic improvement over 1050psi KNSB.


Finally, how about that original 850ksi Young’s but a tripling of tensile strength to 3000psi? Improvement on roughly the same level as the previous best. Most of the grain has good margin.



Reducing the Young’s modulus to a tenth of its original value may require a lot of plasticizer. But tripling the tensile strength might be achievable by incorporating a relatively small proportion of fibers into the propellant.

Both plasticizers and reinforcing fibers are likely to be fuels, so the proportion of oxidizer would have to be adjusted to keep maximum performance. Depending on what was used as plasticizer, it could leave the propellant reasonably castable, whereas the fibers would likely make the molten propellant pastier and harder to cast.

On a more dogmatic level, one of the major objectives of Sugarshot to Space is to reach space with an inexpensive and low tech propellant, for which a two part plasticizer (as used in APCP motors) probably doesn’t qualify.

Either course would be an interesting science experiment, but to characterize the change well would require both physical material testing and instrumented test firing to see how the performance of the blend changes.

In the end, it may be more effective to first attempt changes to the grain geometry to reduce stresses on it during operation. Though such changes reduce the ultimate mass efficiency of the engine, they are more of an engineering problem than the “rocket science” of modifying the formulation.

Case bonded with silicone, lower Young’s Modulus

Another run of the star core KNSB grain bonded into an aluminum casing, but this time with a theoretical KNSB formulation that has a Young’s Modulus of 70ksi, rather than the prior example with unadulterated KNSB of 850ksi Young’s Modulus. The stresses are much less, and thus the factor of safety higher, because the grain can flex more with the casing.

Case bonded with silicone

Stresses in a case bonded KNSB rocket motor, with propellant bonded to the aluminum casing with silicone, shown as factor of safety in the parts. The first plot is a section through the middle of the grain, the second showing an end of the grain.

Compare to a grain cast directly into the casing.

Fully bonded segmented grain stresses

Stresses in a segmented KNSB rocket motor, with propellant strips fully bonded to the aluminum casing using silicone, shown as factor of safety in the parts. The first plot is a section through the middle of the grain, the second showing an end of the grain.

Web bonded segmented grain stresses

Stresses in a segmented KNSB rocket motor, with propellant strips bonded to the aluminum casing by 1″ wide webs on the back, shown as factor of safety in the parts. The first plot is a section through the middle of the grain, the second showing an end of the grain.