Monday, June 30, 2008

What damage 4 seconds can cause!

I've learned to respect the time unit of one second.  It's amazing how much impact this miniscule time increment can impact our efforts and lives.  I'm getting my second wind on this level 3 thing and tonight I finally committed the zipper damage to blog memory:
Ouch.  That hurts.  The good news is that I'm convinced it'll be better than new when the repairs are complete and that the nose cone remains unscathed.  I plan to cut off all but ~4" of the left tube and roughly half of the right tube.  I'll then epoxy couplers into each and attach a new length of 7.5" airframe tubing.  One cured and fiberglassed the whole rocket will be a bit longer but these repaired joints should be very strong.  
     I'll plan to update my documentation package with a new pre-flight data form and addenda on these repairs.  I'm also gonna make changes to the electronics switches so they can be armed externally.  If Kurt and Dok will be at the July 12th launch at Lucerne I might try again because I'd like to attain my L3 cert before my 40th BD in September.

Sunday, June 15, 2008

Oops! Almost forgot the cool recovery shots!

Kristine must have shot these and they're great but I forgot to tell her I'd switched to manual focus so they could be sharper. That's entirely my fault, however, and I'm grateful she caught 'em!
It all looks fine above right?
Doug's on the left in the foreground and I'm repacking the 'chute in the right background.  You can see my sad face if you squint really hard.  Just kiddin'.

Failure analysis: I think I know what went wrong

I tried to download the numbers from the altimeter today but found no accelerometer data.  The unit obviously had power throughout the flight, however, since the pressure sensor reported a peak altitude of ~4,900 at 17.6 seconds.  Hmmm.  Now I begin re-reading the ARTS2 altimeter manual and I figured out the problem: the unit must be armed when the rocket is in a vertical position!  This may seem obvious but we literally had no other safe way to arm the electronics so this change of plan made sense at the time.  Here's my best guess at a sequence of events:
  • I armed the electronics in a horizontal position so the altimeter's accelerometer never calibrated or started measuring.  In retrospect I thought I'd heard it beeping correctly but I must have been wrong.
  • Since the altimeter requires acceleration data to fire the ejection charge nothing happened at peak altitude.
  • I had set the timer for 22 seconds based on a simulated delay of 21.5 seconds but the actual barometric peak occurred at 17.6 seconds.  This means the rocket continued to arc over for 4.4 seconds longer than was required.
  • This extra speed combined with the extremely heavy rocket parts must have caused the zippers.
When I attempt to certify again I plan to:
  • Move the switches to the exterior of the payload section so the rocket can be easily armed in a vertical position.  I'd originally designed around this because I wanted to avoid drilling holes and weakening the critical sustainer/payload area.  I also didn't want to risk lengthening the switch wires adding complication. I feel much better about all this now and the switches will mount in the payload section not the top of the electronics bay as originally feared.
  • I'll replace the 1/2 " tubular nylon with 1/2" Kevlar with stitched end loops to improve strength after ejection.
  • I'll add two Giant Leap "fireball" anti-zipper pads to both shock cords.
I look forward to another "attempt at success" in fall of this year.

Flight day... Yay then boo!

Alright so first off I want to thank Kurg Gugisberg and "Dok" Hanson for acting as my Technical Advisory Panel members.  I also want to thank my buddy Doug for helping me drag this beast around the lakebed in the heat all day.  His son Brandon shot video of the flight and my friend Kristine shot photos and toted my tripod/1Ds mark III/300mm f/2.8 IS lens around in the sun for hours.  I'm grateful to all of you and this certification attempt wouldn't have been possible without you!
       I rented a minivan on Friday and had it loaded up by about 10PM that night.  Doug, Brandon, and I headed out at 5:50AM the next morning, encountered no traffic, and touched down at Lucerne Valley dry lake bed at 8:25AM.  There was NO wind which is a small miracle these days.
       I immediately headed over to Jack Garibaldi's "What's Up Hobbies" trailer to borrow the 75 mm/5120 N•s casing for my cert flight.  Someone else had beat me to it so I dropped back and prepped everything but the motor.  The electronics wiring and payload assembly came together without a hitch.  I then took the partially assembled rocket over to "Dok" to check the build and authorize the flight.  He looked through my documentation and asked "What's the thrust in pounds of the motor you're flying?"  I froze up a bit because I knew that answer but it wouldn't come to me.  I offered "1297 newtons divided by 4.45 N/pound-force... now if I only had a calculator."  Dok was fine with that answer and after an inspection authorized me to fly.  I then needed the same checkout by Kurt but, sadly, he'd just crashed a very beautiful/complex project and was just heading out to pick up the debris.  I felt so bad for him!  It's so painful to work that long on something, one little thing goes wrong, and the whole thing goes down.
     Now I'm waiting for someone to announce a level 3 certification attempt on an M1297 because this will tell me who has the casing I need.  At about 11AM I hear that Jim Hawk is flying his Polecat Aerospace Army Hawk rocket on an M1297 for a level 3 cert attempt.  Bingo. His flight appeared beautiful to me.  I wandered over to find Jim and let him know that there was no rush but I needed to borrow the casing whenever it was cleaned.  At this point I'm starting to feel a breeze and I begin panicking because the wind almost never subsides once it starts in the desert. Unfortunately Jim suffered a failure known as a 'zipper' where the shock cord tears through the body tube due to the extreme ejection forces involved.  One very important element of any certification attempt is that the rocket be recovered ready to fly without repair so Jim had failed to certify.  After about 20 mins Jim handed the casing over to me and I ran back to prep the motor.
     It's amazing that this huge motor is much easier to assemble than than a hobby store motor! It's a very elegant design and it was built and loaded into the rocket in no time.  The final weight was ~47 pounds with the motor.  Doug was concerned that the whole thing was too heavy and I must admit I was also a bit concerned that all my simulation work was inaccurate.  In the end I trusted Rocksim and we carried the rocket over to Kurt for examination.  He said that the 1/2" tubular nylon included with the kit was small for a level 3 attempt and offered his 1/2" kevlar cord to be attached in parallel as a reinforcement.  After this he signed my paperwork and we proceeded from the range safety officer to the launch control officer and out to the pad.
       Doug and I loaded up the rocket on the launch rail in a horizontal position.  After tilting the rocket to vertical I found that my original plan to pull the payload section off and arm the electronics would not work without a stool to increase my height.  We then moved the rocket horizontal again, armed the electronics, and re-raised the rail.  I installed the igniter and we were all set.
       After some additional waiting the launch control officer announced my flight at about 1PM. The wind had oscillated between 0 and 8-10 mph so I'm stressing at this point.  Right before launch the wind calmed down and the flight was beautiful!:
The delay seemed a bit long to me but everything looked fine as the parachute ejected and inflated.  I was elated!  I ran about 1/2 mile out to the rocket now on the ground only to discover that I'd zippered both tubes!! Dammit!   How had this happened?  I'd never zippered a single rocket in 32 year of flying rockets.  Did it HAVE to happen now?!  Apparently so.  It's interesting to note that the altimeter was not beeping out the altitude as it should have been.  Doug and I carried the damaged rocket back and I had to tell Kurt that I'd failed.  
     I'm convinced that I'd done everything I could to fly successfully so I'll need to do some thinking to determine root cause.  I was planning to buy another kit to try again but I had an epiphany while standing in line: I can rebuilt it better than it was. Better. Stronger. Faster.  As such I purchased two 48" lengths of 7.5" diameter phenolic tubing and three 12" long coupler tubes.  I'm thinking the rebuild should be comparatively quick and should also allow me to put the video camera back in.  Yay!  Wish me another round of luck for October or November of this year, por favor.

Finished, painted beast.

Thanks to my buddy Doug Vannier for shooting this:

Electronics board

Here's the final physical electronics board layout:
The two holders on the left and the two-sided foam tape on the right will hold the three 9V batteries.  I'll also bind all three on the board with zip ties hence the four holes on each side of each battery.

Friday, June 13, 2008

And......... done.

Having arisen at 3:45 this morning I'd certainly hoped to do more posting today.  However we had a company picnic and I was working all day to finish this project up so you'll have to trust that it's now finally complete.  I plan to post the final pics of the electronics board, and the final version of the doc package later tomorrow night or Sunday.  Fingers crossed for some killer flight and recovery shots as well.  I also hope to get some quality HDV video so I'll try to throw that somewhere.  Wish me luck tomorrow!

"Modern High Power Rocketry 2" book

I'm putting the finishing touches on the big "torpedo" as one friend calls it and was searching for the latest information on vent hole and static port sizing.  These holes are drilled into the airframe to acclimate internal and external pressure and prevent failures.  In any event a Google search revealed an almost complete preview of the book Modern High Power Rocketry 2 by Mark Canepa.  I marvel at how thorough this 'preview' is and was so impressed with the content I bought it from Amazon on the spot.  Great stuff, Mark!

Tuesday, June 10, 2008

Documentation packet v2

I just completed an extensive revamp of my documentation packet [updated 8/8/08].  Changes/additions in red include:
  • I corrected the CAD rendering on the TOC page to reflect the actual sustainer tube length of 40".
  • I corrected two schematic wiring errors
  • I added a flight simulation section based on Rocksim simulations using the latest Lucerne Valley weather and Google Earth coordinates and elevation data.
  • I authored my detailed flight plan
  • I added a packing list (don't want to forget anything!)
  • I edited my build plan to reflect changes and additions now that I'm very nearly done.
After Dok/Kurt review and approve this document I'll publish a final version before Friday and print this beast out so I can carry a book out with me.

Rail guides

I was all hot to prime the sustainer and finish this beast but I'm glad I lagged because I almost forgot the rail guides!  These standoffs will compensate for the nose cone bulge near its bottom and prevent scraping during liftoff.  I designed this airfoil-shaped standoff to be cut from 1/2" plywood:


The top shape is the top view and the bottom is the cross section showing the curvature to match the 7.67" diameter tube.  I traced the top design on plywood and did a rough cut using my bandsaw.  I then finished the top-down shape using my disc/belt sander.  I then used the curved end of the belt sander to form the curvature of the bottom shape.  
    I center punched the hole in the top and drilled a 1/8" hole through it.  I then marked the sustainer tube both 4" and 29.25" from its bottom (into the centering ring plywood) using a 90˚ strip of extruded aluminum and center punched markers there.  I drilled a hole in each 1.1" deep and sanded the area around these holes.  I applied epoxy to the bottom of the standoff, mashed it down on top of these holes, and screwed 1-3/4" machine screws through the rail guide/standoff and into the 1.1" holes.  Since these assemblies will need to support 40+ pounds of weight they had to be solid and I'm confident I achieved that goal:

Fiberglassing for fun a profit

A wise manager at work once said "if you work on something it gets better."  He was really speaking of working on issues but I think this applies to me since fiberglassing used to be an issue for me.  As I progress toward the end of this project I'm feeling really good about the journey and the improvement of several of my skills.
    I created a rectangular template in Illustrator (21" x 50") and used it to cut a single piece of glass cloth at a 45˚ angle, again for maximum strength.  After masking the shoulder of payload section with masking tape I laid this cloth strip on the tube and tacked it in place with a few brush loads of epoxy.  It's interesting that the 45˚ cloth deforms in a way that makes it difficult to maintain adequate width and coverage on the 20" payload tube.  I had to continuously stretch the cloth to width every 6" of length or so.  I also learned to brush downward at a 45˚ angle rather than across or lengthwise.  This makes sense in retrospect but I had to figure it out on the fly.  When I'd applied the total length of cloth (2.1 wraps worth) I then squeegeed the excess resin out of the layers.  This minimizes weight without detriment to strength:
I think you can tell that this is an extremely uniform layer.  Next I started on the upper portion of the sustainer but this time I chose 0˚ cloth rather than 45˚.  The reason is that the majority of this 40" long tube is filled with either motor rube/rings or the coupler of the payload section.  There's only about a 4" gap in the tube so I don't really need the extra strength of the 45˚ hassle.  I applied the same technique as above and here's the result:
I allowed both to cure and they're ├╝ber tough now.

Kilz does!

Meet my new favorite rattle-can product:
I read an article on "easy fiberglassing" and in the last paragraph they mention this Kilz spray primer.  It's really amazing stuff because it's thick and soupy but exhibits excellent cohesion and doesn't run when you really lay it on.  It quickly fills sizable gaps and dries quickly leaving a thick layer of material.  It also sands easily so, all of sudden, it's easy to produce very smooth finishes quickly.  Git sum.

Tubular nylon linkages

While waiting for some other parts to cure I secured quick links to each end of both 30 foot lengths of tubular nylon:

I leveraged the Public Missles technique of:
  • tying a triple square knot (it's probably called something else, however)
  • applying a couple of drops of CA adhesive to tack the knot in place
  • applying epoxy to the two nylon interfaces at each end
  • wrapping three pieces of gaffer's tape along these two lengths of nylon at each end to bind them together
  • placing a clamp near the knot at each end
  • allow to cure
Since these assemblies will tether together the sustainer/payload and payload/nosecone it's extremely important that they be robust; hence the overkill on knotting and bonding.

Friday, June 6, 2008

Final internal reinforcement

I've seen many a high-power rocket fail at the coupler between the payload and sustainer sections.  Many folks forget to reinforce the inside of the coupler so that's the weak point that buckles under the tremendous forces of flight.  I cut three 2" wide strips of 45˚-oriented glass, soaked them in epoxy, wrapped them one at a time around a 3/8" dowel, then unwound each inside the coupler against the forward bulkhead:
I pressed all three layers down onto each  other and I think you'll agree that looks much stronger than a bare piece of phenolic tubing.  Keep in mind that the bottom of that bulkhead is about 1" above the payload/sustainer interface so I think I have things covered.
     I also repeated the above procedure as the final reinforcement in the nose cone.  I poured one more pool of epoxy in there and now the nose cone is really ready to finish.

Tuesday, June 3, 2008

Funberglassing... I mean fiberglassing begins.

Actually fiberglassing is becoming less tedious the more I do it.  I began by building a template in Illustrator for a pair of fin surfaces and the body tube in between.  I placed this template on the fiberglass cloth at a 45˚ angle (to maximize strength in the final composite), traced around it with a paint pen, and cut around it with about 1/2" margin all around:
I pressed this piece of glass cloth into a fin pair before committing epoxy:
Finally I mixed up a double batch of epoxy (no filler) and used a disposable paint brush to apply it to the cloth.  I then squeegeed the excess resin off of these surfaces and this new composite reinforcement is curing now:
I'll repeat this for the other three fin pairs and then start on the rest of the tube.

Who needs sawhorses?!

I need to support these  huge parts while I fiberglass them.  I didn't want to buy/store a pair of sawhorses in my already full garage so I had an idea last night.  I cut two lengths of 2x4", screwed them together, then screwed that pair to the vertical support as seen on the left here:
The left white 2x4 is 4" higher than the right one so this allows me to rest another board in this gap.  I then took the remainder of the 10' 2x4 board and set one end in that gap, slid the sustainer over it, and rested the other end on my fold-up table.  Voila!  I can rotate the sustainer as needed and the whole setup stores away in a tiny space when I'm done. 

Monday, June 2, 2008

Refresh this page!

Some web browsers in some operating systems aren't refreshing this blog when you load the page.  Be sure to reload the page each time to read the latest epic advancements.  [It's sad to note that if your browser doesn't auto-refresh then you'll never see this!]

Sunday, June 1, 2008

Payload final assembly

This last step was uneventful and unworthy of a photo but I basically epoxied the coupler 6" up into the 20" payload tube.  One cool new thing I tried that worked well was to pour epoxy into the payload tube and wipe it around the inside 6" with a resin squeegee.  This formed a nice uniform film and then I poured more resin near the entrance to the tube.  Finally I twisted the coupler up in to ensure good contact then propped the assembly upright in the sustainer to cure.

I still need to sand that nose cone shoulder to better fit the payload tube but otherwise this beast is heading for the glassing/finishing department at G. Smith Aerospace (or G.S.A. as I like to now call it)!

Sustainer final assembly

All that's left to do on the sustainer assembly is the two aft centering rings, the motor retainer, and application of a glob of epoxy to solidify this high-impact zone.
       I had to dig out/chisel the foam that arose above the bottom of the fin tabs to ensure the next centering ring would seat well.  That took a surprisingly long time as this polyurethane foam is resilient and bonds well!  I also had to plug the bottoms of the fin slots (epoxy plus tons 'o' silica) to prevent resin from running out.  I then poured in epoxy on the fin bottoms, slid the centering ring into place, and tapped it down with a dowel and a hammer to ensure good contact.  I flipped the sustainer over while the epoxy cured.    As I mentioned in my documentation package I needed to pre-assemble the retrofit-ready Allen bolts into the collar of the very cool motor retainer to match my design intent:
Then I loaded up a batch of epoxy with silica to form gel and globbed this into the bottom inside edge of the sustainer.  I pushed the final centering ring into this epoxy and leveled it with the L-angle.  I then pushed this ring in the rest of the way using the threaded nuts on the motor retainer (~0.375"), centered the motor retainer on the motor tube, and tacked it into place with CA adhesive.  
You can also sort of see that I applied electrical tape around the inner interface between the retainer and motor tube so epoxy would leak through and into the motor tube (this could be disastrous!). Finally I mixed up a big batch of epoxy to fill this end volume:
The masking tape forms a dam to prevent epoxy from leaking over.  Ok so why is it so goopy and bubbly you might ask?  Well that first huge batch of epoxy wasn't sufficient to fill up all the way to the edge so I mixed another half that size and poured it in on top.  Even that was still just shy of the top so I mixed/poured a third smaller and final batch in and that did the trick.  As I began stirring/pulling the bubbles out of this pool I noticed some microbubbles creeping up the sides in a few points.  Then I noticed some light  smoke wafting out above the pool....
     Now I'm a chemical engineer but I haven't built a rocket of this diameter for years now.  Just like the foam mentioned below this epoxy resin is highly exothermic (energetically favorable) and liberates a bunch 'o' heat.  The three batches were mixed only minutes apart but when the bottom-most batch began curing it heated up the layers above and basically boiled them as they began to cure.  Duh.  It's no biggie but I'll need to sand fill that end a bit to achieve my customary aft finish.

Productivity!

I had some weird intestinal bug yesterday (Saturday) and had to lie on my couch, queasy, all day long.  I awoke today feeling great and was highly productive having dodged a weeks-long flu or something.  My goal today was to finish the assembly of the nose cone (below), sustainer, and payload section so I can start external fiberglassing next week.  I achieved these goals (as you'll see) AND hit the gym for 575 kcal of hawt treadmill akshun.  Great success!

Nose cone final assembly

After I finished filling the nose cone with foam (highly exothermic!) I loosely wrapped some masking tape inside the shoulder as a support, wiped a bunch 'o' epoxy above the tape, and inserted the bulkhead.  You can see how irregular the inside of the shoulder is so there wasn't much adhesion:
You might also note that the bulk head is a bit tilted.  The foam pushed up unevenly, formed a high spot and, despite my best efforts, would not level out.  I thought about the forces involved on the U-bolt and decided this tilt won't affect much.  I then poured in some thickened epoxy around the perimeter but the gaps were so large it ran right through.  Danger, Will Robinson! This loose fit caused alarm bells in my head because only this interface will prevent the bulkhead from pulling out of the nose cone.  I then elected to use the remainder of the 1" wide fiberglass cloth and wrapped that around the inside of the shoulder and on top of the bulkhead.  Finally I poured a heap of epoxy in on top of all this and ensured it cured at a tilt that would level the epoxy fill.  This part is now done and ready for external finishing.  Yay!

Tube squaring antics

One remaining nagging point about this kit is that none of the tube ends are even/square.  I think the longest these 7.5" tubes are manufactured is 60" so I would have guessed that at least ONE of the four ends would be square and usable as a reference for the other ends -- alas this was not the case.
     This unevenness looks bad and increases the chance than the payload and sustainer won't distribute loads evenly and worsen the chances of buckling during flight.  Since these tubes are comparatively large I had to be clever in squaring them up.  I tried several things but in the end I settled on an L-square ruler and spirit level (both of those have proven so handy during this build!). I verified that my kitchen countertop is perfectly even with the level then laid the 20" payload tube horizontally on that surface.  I then aligned the long edge of the L-square along the top of the tube so the short edge dangled downward.   This allowed me to find/mark high spots on the tube edge and iteratively sand down those spots and recheck.  When one end was even I repeated this technique on the other end.  
     Once that tube was square I used it and the coupler tube to square off the sustainer tube which was WAY uneven.  Duh.  In any even all four tube ends and the motor tube end are now square and looking nice.