Tuesday, December 2, 2008
Wednesday, November 26, 2008
Monday, November 3, 2008
This is what I did yesterday. In total, it was about 6 hours of work. This is my first project in stainless steel and I was a bit nervous after having read of the problems others have had. I'm releived to say that it wasn't as bad as I expected. I did shatter two carbide inserts threading the external threads. However, the trick seems to be to take very light cuts.
Saturday, November 1, 2008
Thursday, October 30, 2008
This is the latest version of the valves I use. I have one listed for sale on ebay: http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=270295521784&ssPageName=ADME:L:LCA:US:1123
Sunday, October 12, 2008
Sunday, September 21, 2008
Monday, September 8, 2008
Instead of using a steel rotor and a hall effect sensor, this meter has a plastic rotor with an IR emitter/collector pair. It outputs a nice square wave pulse with frequency proportional to flow rate. My DAQ board (from Measurement Computing) has a counter input which can be configured to measure frequency so interfacing to the existing DAQ software is quite simple.
Wednesday, July 23, 2008
Friday, June 20, 2008
This is a picture of hydrotesting I did on a 6" 316 Stainless steel float from McMaster-Carr (2762K49). It failed at 2150 psi. Interestingly, the failure was a tiny pinhole that opened right at the weld termination. My grease gun was not able to pump enough water fast enough to get the pressure higher higher than 2150 once the hole opened up.
I proof tested a second of these floats to 1900 psi without failure. And 3 more were tested to 1500 psi.
Tuesday, June 10, 2008
The packaging comes out quite nicely. They have several advantages to the linkage actuated version - this will be my go forward design.
Along those lines, I have been wondering if there are people out there that might like to buy a pre-made valve like this one. It was actually more of a pain to build than I thought it would be and personally, I prefer to spend my time building rockets rather than various peripheral devices.
Here are some of the relevant details:
- CV of 1.6 and is rated for 1.6 gpm of flow at 1 psi delta and 5.1 gpm at 10 psi
- rated to 3000 psi
- can throttle fairly well and has fast response (as can be seen on this blog)
- o-ring sealed and can be compatible with most non-cryogenic propellants, including hydrocarbons and hydrogen peroxide
- the servo would be a Hitec digital servo with titanium gears (6.0V or 7.2V input)
If you're interested in buying a valve like this, drop me an email.
Monday, June 2, 2008
I was able to get out and test some modifications this weekend. The good news is that the urethane o-rings seem to be working. I did a long term pressure test with N2O to see if they would remain sealed for several days. Also, after the testing this weekend they showed no signs of damage.
The main point of this weekend's testing was to test a new configuration with the propane/spark ignition with the goal of eliminating the initial pressure spikes. The new configuration seems to have reduced the tendency towards hard-starts. Unfortunately, the ignition sequence which was quite reliable before is no longer. I spent quite a bit of time just getting it to light at all and stopped testing without arriving at a reliably repeatable sequence.
Sunday, May 4, 2008
A common solution is to use ball valves. They have several advantages including reliability, availability, and some throttle-ability. However, the packing around the ball generates considerable amounts of friction, especially when the valve hasn't been moved for some time (sticktion). As a result, ball valves require rather significantly sized actuators which are often heavier and larger than the valve itself.
While browsing through the Swagelok catalog some time ago, I discovered a valve they call a "plug valve". It is similar to a ball valve but with a cylindrical rotating plug that is sealed with o-rings. The result is a valve that has very low friction and thus can be actuated with correspondingly smaller actuators. The primary downside is that the o-rings become a consumable item and prohibit their use with cryogenic fluids such as lox.
This is a movie of some early testing I did with a garden hose and a standard Futaba servo.
If you look closely in the video in this post, you can see the valve moving in the lower right-hand corner.
As I alluded to in the previous post, I've had several o-rings tear when being closed after use at part throttle with N2O. The valves come from Swagelok with PTFE coated FKM o-rings. I assumed they were not rated for lower temperatures so I ordered some PTFE and urethane o-rings. Below is a disassembled valve with a urethane o-ring.
Last night I did some simple testing with the urethane o-rings and to see how cold the N2O is getting downstream of the valve. The picture of the setup is below showing the valve, thermocouple, and injector orifice.
As you can see, it gets pretty cold. At minimum throttle, the temperature reached a steady state value of ~ -84F after 20 seconds or so. The urethane o-rings are rated to -65F and the PTFE rings are rated to -100F. I'm going to try the urethane rings first because the PTFE rings are very hard.
Last up is a neat picture showing the formation of solid N2O at the orifice exit. Visible are both ice-sickles and snow.
In summary, with the right o-ring material, these valves should work for all storeable propellants such as ethanol, kerosene, and peroxide. They should also work with saturated liquids stored under their own vapor pressure, like ethane, propane, and N2O. The odds of working with LOX are slim to none.
Wednesday, April 23, 2008
Last night was the first attempt at using the trailer to test and I was very pleased with how it all turned out. It took almost exactly 1 hour from the time I arrived home from work to the first test. During that time, we loaded the van, drove to the test location (about 10 minutes), and setup the electronics on-site. There were no issues with the test stand and everything worked correctly the first time.
If only the rocket motor were so easy ...
The light sequence is very reliable. I actually had two lights that are not shown in the video - for the day, the motor was 5 for 5 on light attempts. Unfortunately, there is still a fairly high pressure spike at ignition that was triggering a redline on chamber pressure. I had lowered the chamber pressure redline to just a tiny margin above the design maximum - the startup spike is around ~25% higher than that. On the first two tests of the day (not shown in the video), the software redlined at engine start and aborted the test. For the the first two tests in the video, I raised the redline limit and the motor started fine and executed a scripted throttle profile. On the last test, I knew that there was probably a damaged seal in the throttle valve but I decided to initiate a test anyway (the venting between the second and third light is a result of the damaged seal). The end result of that test is obvious - another blown gasket.
I'm not sure yet what the exact cause of the blowout was - the data did not show a pressure spike that was larger than any of the previous tests that day. I have not yet reviewed the data in detail nor have I disassembled the motor to inspect it.
So I have three problems to solve with this motor right now:
1) Pressure spikes at start up - I think I have a plumbing solution to this
2) Blown gaskets at start up - I think solving 1) should help solve this one. Maybe a new gasket material is in order.
3) Repeated failures of valve seals - More on this later
PS - Damien gets 10 points. The propane bottle can be seen in the video at the center of the screen. However, I would not call it a "torch".
Monday, April 14, 2008
PS - the tank at the bottom of the picture wrapped in electrical tape is a red herring.
Wednesday, March 26, 2008
Some time ago, I began researching the idea of initiating nitrous oxide decomposition with spark energy alone. My literature research turned only one relevant reference, documenting work performed by Pratt and Whitney and Rocketdyne.
"Investigation of Decomposition Characteristics of Gaseous and Liquid Nitrous Oxide", Air Force Weapons Laboratory, Kirtland Air Force Base, NM. 1974.
Pratt and Whitney was able to initiate sustained decompositions reactions in pure nitrous oxide with very low spark energies. However, the reaction was only sustainable in large diameter pipes and at pressures of several atmospheres. In addition, diluting non-reacting species such as helium or nitrogen significantly reduced igniteability.
I decided I would try it anyway.
This is the injector head that I designed for this experiment and am still using. The chamber side of the injector (minus the spark plug) can be seen in this post. It has three ports - two 1/4" NPT and one 14mm. The outboard NPT port is for measuring chamber pressure. 14mm is standard spark plug thread.
The spark plug is a standard Autolite platinum plug I had laying around. I'm using a standard ignition module from CH Ignitions. After trying several different approaches, I ended up triggering it with an old HP function generator I had. I found out from Paul Breed - after I ordered my unit - that CH will sell a custom box that sparks continuously when ever it has power.
So, after designing a chamber and a nozzle, casting fuel grains, shaking down a throttle valve, setting up new DAQ software, and getting the spark box to trigger ....
I found out that it doesn't work.
I was unable to measure any indication that the N2O was sustaining a decomposition. No chamber pressure rise, no clearing of the cloudiness in the exhaust. I tried several different throttle settings and spark rates and finally gave up and moved to plan B.
Which I will talk about in Part Two ...
Friday, March 21, 2008
I am considering starting project which is too small for NPT fittings. I would like to use Aluminum brazing materials for attaching the propellant tubes directly to the injector heads. I bought the brazing material from Durafix some time ago for another fabrication test that didn't work out. McMaster-Carr also sells aluminum brazing alloys. Before I machine the injector, I thought I would test out the joining method with a hydro-test - if it doesn't work I'll have to modify my design.
This is a picture of the finished joint on the test piece. The test article is just a 3/8" aluminum tube brazed to a round rod which was left closed on the end. Both the rod and the tube are made of 6061T6.
The test procedure is simple. I fill the apparatus shown below with water and attach the free end to a standard grease gun. I fill the the grease gun with water as well and then pressurize the entire system. I got the grease gun at an Ace Hardware and its really nothing special. The box says its good to 6000 psi, but I've only ever had it up to 3500. It is very important when doing this kind of test to fill the entire apparatus with water - don't use air!
The end result was no leaks, so it looks as though brazing could work as a low cost method for joining aluminum tubing to manifolds.
Sunday, March 16, 2008
Other than the failed gasket, there were no problems with the motor. I made a new gasket and reassembled.
Today I tested out some modifications of the start sequencing. The results were good - 4 lights in a row on 5 tries with no hard starts. Unsurprisingly there is some art to getting a sequence which lights reliably without hard-starting. Moving the valve closer to the chamber gave better throttle response as well. I'm working on a post with details of my ignition system (you'll be disappointed). Unfortunately, I'll be out of town this week on business so this movie of today's run will have to do for now.
Wednesday, March 5, 2008
The only interesting thing about this movie (other than being a rocket motor firing) is that it demonstrates "chugging". The popping at the end of the run is a low frequency oscillation in chamber pressure that is most likely caused by elasticity in the propellant feed system. The issue only ever manifested itself at the very end of a run, when the tank held only N2O vapor and the pressure drop across the injector was relatively small.
Sunday, February 24, 2008
As you can see in the picture, there is fire coming out of the flange that holds the injector plate onto the chamber - the gasket failed when the combustion chamber pressure hit ~1000 psi momentarily. The good news is that there is no other damage to the motor and gaskets are easy to make.
The other bit of good news is that my redline software performed exactly as it was intended to do. I created a module in the DAQ software that triggers when any of the 4 analog channels I'm monitoring (thrust, chamber pressure, injector pressure, tank pressure) exceed predetermined values. When the redline module triggers, it sends a logic signal to the engine controller module and the engine shuts down. In both tests today, the redline was triggered and the oxidizer valve closed. The second test, pictured above, had a duration of about 0.5 s before the valve closed.
Wednesday, February 13, 2008
This first picture is of a single injection orifice flowing water from my garden hose - the stream is very laminar with very little dispersion. Engines with liquid injection usually use impinging streams to accomplish the mixing that is required for good combustion stability and efficiency.
Robert Watzlavick has several good pictures of an impinging liquid-liquid injector on his website, in addition to this one.
The last picture is of the same injector in picture number one flowing liquid N2O instead of water. In this case, the saturated liquid flashes to vapor immediately due to the much lower atmospheric pressure. The resulting plume is easily 10 times the diameter of the injection orifice. Gaseous injection would probably not result in such a large plume but would still expand much more than the water above.
Tuesday, February 12, 2008
Wednesday, January 23, 2008
As you might be able to tell from the picture in this post , I'm using a beam load cell to measure thrust. It has the advantages of being relatively easy to mount and requiring only one load cell. It has the disadvantage of turning the engine into a vibrating, cantilevered beam supported by the load cell. Fortunately, my data sample rate was 10kHz, so I was able to spot the low frequency ringing by comparing the thrust and chamber pressure data.
In order to determine which frequencies were caused by the test stand and which were "real", I did a hammer test. My test stand is mounted on a steel I-beam, so I turned on the DAQ system and hit the beam with a hammer. The next two charts show the raw thrust data and a FFT of thrust data. The FFT shows vibrations at 31, 33, and 85 Hz, so I can use a digital filter to remove that frequency content from my test data. I use and recommend Sigview to do plots, FFTs, and filtering.
Monday, January 21, 2008
That motor turned out to be a good way to do many more things than simply collect regression data. Finding vendors, setting up data acquisition, and learning plumbing best practices are all non-trivial tasks. In keeping with the test a little, build a little philosophy, I find it much easier to incrementally add features as you go. Not worrying about all the features a motor (or any system for that matter) needs to have will often generate more progress than a "single step to solution". Or as my brother says, sometimes you need to just go build something even if you're not sure what it is.
In the intermediate time between when I started and now, a group at Stanford has started publishing data for nitrous oxide hybrid motors. You can read the first page of their papers here and here. Fortunately, their data and mine compares reasonably well as you can see in the chart.
Once I had reasonable regression data and a working test stand it was time to start building a flight worthy motor. However, before you design a motor you have to choose a vehicle architecture. A VTVL vehicle has a much different motor than a sounding rocket. I'll talk a little about motor applications in a future post.
Tuesday, January 15, 2008
moon·light unˌlaɪt/ Pronunciation Key - Show Spelled Pronunciation[moon-lahyt] noun, adjective, verb, -light·ed, -light·ing.
5. to work at an additional job after one's regular, full-time employment, as at night.
You have to get paid for it to be moonlighting. I'm not, so it's not moonlighting.
I've been working on building hybrid rocket motors in my garage for about two years. Having made some small measure of progress and lacking any eminent prospects for starting a rocket business, I thought I would share some of my lessons with the community. I hope that what I can share may be of some small help to others.
To start, here is a picture of my first motor.