Proposal for a "starting point" overall OHKLA design | Project planning and misc | Forum

As I discussed recently in another thread, I've started to think we would do well to try to quickly get together a "first attempt" or "starting point" overall design for OHKLA, rather than taking our time trying to make each decision one at a time, and making the optimal choice first time, each time.  With a first design in place we can use iterative refinement and modelling to improve the design and fix any problems.  This will be relatively easier than doing each decision individually "in a vacuum" because all the core features will be specified.  If we want to try changing one feature, we'll have the knowledge we need to do before-and-after analyses of that change to see if it's an overall improvement.

If other people think this is a good idea, I'd like to propose the following design as a good starting point for this process.  It is based on everything we've learned in the past few weeks from our more-unified-than-usual research direction.

Here's a quick run-down
:

• The nose cone has the Von Karman shape, which is a special case of the Haack series shape.  This shape has the lowest drag for transonic flight, which was the main reason I chose it.  However, it isn't over-optimised for transonic flight, the VK nose shape also performs very well at subsonic and supersonic speeds.  I think that over the course of the entire flight this shape will result in the lowest nose-related drag overall, which is important.
• The nose cone is made from fibreglass which is applied over a lathed wooden plug.  Fibreglass was chosen because it's lightweight and has excellent thermal properties (doesn't melt, highly insulating). The guys at PSAS have some experience doing this (see here, for example) and could possibly help with the actual fabrication.  I have spoken to Natronics briefly about this via email and he has said they've done this several times before – it's a lot of work but a relatively cheap process (about US$50 per nose cone).  I'm going to ask him more about this and hopefully get some discussion happening on the forums.
• The core avionics (i.e. the system which logs all the sensor data to a storage medium and broadcasts GPS coordinates after landing) are fitted into a tube which is inserted into a hollow area in the nose cone.  The rationale for this is that the avionics unit needs good GPS reception and good coordinate broadcasting abilities, so sticking it in a part of the rocket which is made of RF-transparent materials makes sense.  If it went inside the aluminium aerostructure, reception might suffer.  This approach also reduces the amount of dead mass present in the wooden plug of the nose cone.
• The payload compartment, which holds arbitrary payload within the size/mass limit, is located immediately adjacent to the core avionics module so that, at least in future versions, electronics inside the payload can share a power supply with the core avionics and also receive data from the core avionics (so that payloads that need, say, real time GPS data, don't need to have a second redundant GPS receiver in them).  Possibly in a more advanced version we could even have payloads be able to broadcast data during the flight using the core avionics' system for broadcasting GPS coordinates (which isn't needed during the flight).  Project Daedalus has a system similar to this.
• The recovery system is a dual deployment system, as I recently described in this thread.
• The aerostructure is just thin walled aluminium tubing.  Chosen because it should be strong enough to support all the stuff mounted to it but is also fairly light.
• The oxidiser tank is attached to the aerostructure via skirts at either end, so that it can have a diameter equal to the inner diameter of the aerostructure, wasting no space.
• The combustion chamber is attached to the aerostructure via two centering rings welded to it, so that we have more control over its diameter.
• The fins have the standard trapezoidal shape, because these are fairly easy to fabricate and Sampo has suggested there is little compelling reason to consider other shapes.
• The fins are made of aluminium.  Chosen because it should be strong enough to survive shear forces and avoid fin flutter but is also fairly light.
• The fins are attached to the aerostructure using bolts, not to the combustion chamber, because penetration of pressure vessels should be as limited as possible for safety and reliability.
• The fins are canted slightly to induce a rolling motion for stability (not pictured).
• The exhaust nozzle is inside the aerostructure so that a rough landing is less likely to damage it than if the nozzle were exposed.

This design is certainly not optimal, but based on all we have learned so far and on my reading up on lots of university and/or amateur sounding rockets, it's certainly not ridiculous either.  No part of its design is radically different from things people in the know have done before.  I think this provides an appropriate starting point for a program of more detailed analysis and incremental improvement.  However, if anybody wants to comment on anything about the overall form, please feel very, very free.

Some things that I would like to see more discussion of are:

• I've drawn the exhaust nozzle as having a countered outer-surface, like many nozzles do, rather than having a cylindrical outer surface like many other nozzles do and like I discussed briefly in this thread.  This isn't because I think this is the better choice, it's because I haven't come to a strong opinion on this yet.
• I've said nothing about the system to be used to separate the sections of the rocket to facilitate parachute deployment because I haven't done much reading on this aspect of things yet (see this thread for a starting point).  At this stage it's probably okay to treat this process and its hardware as black boxes, but we should get on it.
• The recovery system (stuff inbetween the parachutes) is responsible for activating whatever the separation systems are.  This means it needs to "know" when the rocket is at its apogee (to deploy the drogue) and once it descends below a certain altitude (to deploy the main).  The core avionics unit knows this stuff, but it's relatively far away the recovery system.  After the first separation event any connection between the two will be broken.  This more or less means the recovery system electronics need to be largely independent, i.e. have their own accelerometers and barometers.  This is kind of a waste and stops me from feeling super happy about the whole thing.  The obvious solution is to move the core avionics down into the recovery section between the parachutes so that we can use the same sensors for data logging and parachute deployment.  However, the problem with this is, as discussed earlier, our radio gear is now in the middle of a long aluminium tube which will act as a partial Faraday cage.  It seems like there are drawbacks to both approaches.  I would love to find an elegant solution to this problem.

Once we have decided we are happy with this overall form, our first step should be to complete the mass analysis we are already working on, using the overall form of our initial design, because once that is done we will be able to put approximate physical dimensions on this thing.  Once that is done we can begin doing things like:

• Model the rocket in OpenRocket and make sure the rocket is aerodynamically stable and the expected apogee altitude is over 100 km.
• Optimise the fin design based on the above model.
• Model the rocket in ProEngineer or the like and make sure the selected materials can stand up to the required stresses.
• Do a fin flap analysis to make sure this won't be a problem.
• Optimise materials based on the above two analyses (can we replace aluminium in the aerostructure or fins with carbon fibre?)

Finally, once we are happy that everything is looking pretty good, we can design scaled down versions and actually get to work on them.

What do people think of this approach (and starting design)?

joe.haydu said:

I beleive we can solve the problem with a few wires run down the side of the payload compartment and the drogue chute cable. That allows us to reuse the avionics hardware, without having to move it. We would probably want a failsafe device in the recovery system module which will activate the main chute automaticially if the wire is broken, but that should be a relatively simple item to set up, just a voltage monitor and a timer circuit or some such.

That's one possibility.  I am a little reluctant about running wires along the chute leads, it sounds fiddly and likely to fail, but I could be wrong, perhaps it's routine.

Another thing that I thought of today is that maybe we could use aluminium tubing for the aerostructure of the propulsion section (which needs to be able to hold up the weight of the engine and also to hold the engine in place while burning) and then use fibreglass or something for the uppper sections (which don't have to hold as much weight or endure as much stress).  That way the recovery section would be RF-transparent and we could just put the avionics in there.  This does mean that further down the road we couldn't interface payload with the avionics, but that's really more of a luxury than a necessity.

These images look really, really cool. :)

I definitely agree on using socket cap screws or something similar to minimise drag-causing protrusions from the top of the areoshell.  How does one go about creating the counter-sunk holes?  Are there drill-bits that can achieve this?

Also, with regards to attaching things to the ends of the combustion chamber (like the injector manifold and the nozzle), I suspect it will be necessary to use O-rings or something like them due to the relatively high pressures of hot gasses inside them.  So when designing the manner in which all these things fit together (like in your final graphic above) we should probably take this into account.

With regards to moving forward, I think now is perhaps actually a good time to move toward a natural stopping point – we should make what decisions we want to/can from the basis of our modelling so far (like the requirements listed in this post) and write some reports and officiate some of the decisions we have made (e.g. use of 6061 Al T6).  This lets people following the blog etc. know what we've been doing and also puts all of this forum-based word somewhere more accessible and permanent.  Then we can move forward, making sure we keep everything consistent with the decisions/requirements in these reports.

That said, if we can agree on requirements quickly, I am happy to start work on the reports and let you go ahead with modelling stuff like heat transfer (maybe also, if you can, thermal expansion of the chamber?  We haven't spoken about this before, but I guess it's kind of important)?  I'm really starting to get out of my depth engineering-wise anyway, and want to start moving on to more organisational stuff, so doing the reports is a step in that direction.