This article appeared in the October/November 2002 issue of High Power Rocketry.

The Babylonian Interstellar
Tactical Cruiser for Hyperspace
by Doug Gerrard, TRA #568

Background
escalate \'es-ke-l_t\: to increase in extent, volume, number, amount, intensity, or scope.

This project started out with such a simple idea; build a night launch rocket that used something other than a flashing xenon strobe. To me rockets launched with a strobe are difficult to follow because of the intermittent light and basically they all look the same. I wanted my night launch to be different. There were a few simple requirements, (1) continuous illumination throughout the flight and (2) no xenon strobe. This project far exceeded these expectations (Photo 1).
I thought of using a simple LED sequencer for my night launch rocket. A sequencer has 10 LED's, but only one is lit at a time, sequentially lighting each of them. By using many LED's without having them all lit at once, moving lights could be used for high visibility, but there would still be continuous light coming from the rocket. This combination could make for a very interesting effect. The following project escalated from this simple idea. I will try to convince you why the rocket was quickly named the Babylonian Interstellar Tactical Cruiser for Hyperspace, also known as, the B.I.T.C.H.

The Design
complex \käm-pleks\ a whole made up of complicated or interrelated parts.

The light from a LED can be very directional, sometimes as little as a few degrees for it's brightest intensity. For light to be visible from any angle LED's would have to go around the rocket and would be lit all at once. This would create a row of LED's around the rocket. With a sequencer, each row would illuminate and sequence down the length of the rocket, so two sets of LED's would be lit at the same time at the ends of a tube and sequence towards the middle. This created an array of LED's where at least one LED was at nearly maximum intensity from any angle the rocket was viewed.
I doubled the number of rows of LED's (except for the center row); so two rows were illuminated at each end of the array at the same time. Twelve inches of the rocket were illuminated from any angle with only four rows of LED's powered at one time. The 684 LED's (38 rows X 18 LED's/row) required only 80 mA. There would be too much drag created from exposed LED's and they needed to be protected so the array was encased in an acrylic tube.
The rocket was designed around a 4" body tube and a 75 mm motor mount. The rocket was made entirely from fiberglass except for the clear pieces made from acrylic tubing. The weight was substantial so the motor of choice was the L850.
The fins were designed to have pods that would also contain LED's to spread out the light. The taillights in the bottom of the pods had a cluster of LED's. These taillights were covered to protect them on landing by using 2 1/4" diameter acrylic hemispheres. The taillights had five strings of 17 LED's. Eighty-five LED's per pod added another 340 LED's to the rocket.
The upper part of the pods had a smaller version of the main array on each pod except instead of sequencing up and down like the main array; the light rotated around the pod. Two opposing columns were lit at the same time so there were 20 columns to match the sequencer. Four pods X 20 columns X 9 LEDs/column added another 720 LED's. It started to get ridiculous, and I was still designing.
I wanted a futuristic design; something that looked like it could be a space vehicle. Above the array, the rocket was reduced to 3" diameter tubing that would house the parachute, but then expanded back to the 4" diameter tube for the nose cone.
With all the lights down in the booster, something was needed higher up to balance the rockets illumination. The nose cone also had LED's. A 4" diameter plate held 272 LED's next to each other comprised of sixteen strings of 17 LED's in series. This made up the headlight that was on continuously during flight. Although not as noticeable when vertical, it looked down during descent glowing against the background of the parachute. A clear acrylic dome covered the headlight.
The base of the nose cone also provided opportunity to add more LED's. The centering ring from the 4" nose cone to the 3" parachute tube could also hold a ring of white LED's that would shine down on the rocket lighting the rocket up when it was vertical, and it would also would shine into the parachute on descent, illuminating it.

The Construction
dismay \dis-`m_\ to deprive of initiative through the pressure of anxiety or great perplexity.

During the design phase, parts and materials were purchased as the design progressed until I realized that this project far exceeded anything initially intended. After about 5 months of designing (and redesigning) the rocket and circuits that would be used, I had spent a lot of money and accumulated a lot of parts, and very little construction had been done. There were a few pieces completed and basically a kit with several hundred parts that had to be assembled.
The fins were made from 1/16" fiberglass. The two sides were laminated with strips of 1/16" fiberglass strips in between them. The spaces left were used for running the ribbon wire to the pods (Photo 2). Tabs at each end of the fins were used for the through the wall design. A custom jig held the fins in place while the glue was drying to ensure each was placed straight (Photo 3). The aft fin support remained in place until the fins were completely secured. After tacking the fins with CA, each fin was attached with carbon fiber weave (Photo 4). Epoxy with Kevlar pulp filler was used at both the motor mount joint (Photo 5) and the outer tube (Photo 6). An internal fillet was also used. The pods were also attached in the same manner using carbon fiber and Kevlar soaked epoxy on both internal (Photo 7) and external fillets (Photo 9).
The upper center ring piece for the lower body tube was also used to center the array tube and the clear acrylic tube. This piece was identical to the centering ring at the upper end of the array and consisted of two centering rings and a section of coupler tube to hold the array (Photo 9). The two centering rings were for the 4" body tube and the 4" acrylic tube since they had different inside diameters. The array tube and clear outer tube were not glued in so they provided very little strength to the rocket. The motor mount tube was connected to a 2 1/4" tube that ran through the array for strength (Photo 10).
I started the array by building small circuit boards that would hold the LED's. Much of my 2000 Christmas vacation was spent drilling the holes in the tubing (Photo 11). Many hours of tedious soldering and drilling holes made me realize that this project would not be finished anytime soon (Photo 12). I realized how much work the project was going to be. The first few months of the next year was working on the circuit boards until Springfest 2001.
I met Dave Flynn at Springfest and was impressed with his insight into high power rocketry. We talked about the night launch rocket, and he seemed very interested. You can imagine my delight when he agreed to help out (although I don't think he realized how much work he was getting into).
Dave had a different idea for installing the LED's in the array, eliminating the circuit boards. He inserted the LED's in the holes in the first row, bent the leads over, and soldered the leads together. After soldering wires to the ends, the string was carefully lifted out of the first row and placed in the appropriate row. Each LED was carefully tacked in with CA. Within a few weeks he had soldered all the LED's for the array. I can easily say that without his help this rocket would have never been completed. His assistance motivated me to continue this project after I was dismayed since beginning construction. I continued with the rocket and Dave worked on the electronics.
After getting the beautiful array from Dave (Photo 13) the wires sticking out of the end of the array looked like a birds nest. There was no way to have the 2 1/4" tube, as well as have the 8 ribbon cables from the pods, pass through the center of it. Since the leads for the LED's were just bent over and soldered there was room under the leads to pass the wires underneath (Photo 14). After the wires were under the LED leads, each row was identified, the ribbon cable was soldered on, and heat shrink was applied to each connection (Photo 15).
The inside of the array was reinforced. It was a fiberglass tube but it had 684 holes drilled in it, and the lights had the fragile leads that could be broken. If any one of the solder connections broke, the entire row would not work. The inside of the array was coated with epoxy filler (Photo 16). Applying the filler was not difficult except for in the middle of the array where the wires passed across the inside of the array. These wires restricted the center of the array, especially, with the epoxy filler applied. It had to be sanded. You guessed it; I sanded through one of the wires for the array. I had to use an inspection mirror and re-solder the wire in the middle of the tube (Photo 17).
The tube above the array proved to be just as difficult. This section of tube is where the master controller and array circuit boards are located, and it provided access to the batteries. Texture was added to the rocket with heat sinks (Photo 18). They did have a practical application to cool the batteries if they got too hot, but that was unlikely. After carefully cutting out the body tube and mounting the heat sinks, it was time to install the assembly (Photo 19).
A generous amount of epoxy with Kevlar pulp was applied to the bottom of the tube and installed on top of the array. I proceeded to fill in significant joints inside the tube and weighted it down. I thought the double center rings, one for the clear tube and the other for the upper tube, would prevent the epoxy from leaking into the array. I was wrong, it did. The leaks on the outside weren't bad they could be wiped away. The leaks inside the clear tube were permanent. Fortunately, the epoxy is also clear and it was not very noticeable from a distance, but this is just another example of how this rocket lives up to its name.
The taillights for the pods were constructed by mounting the circuit board with the lights inside one of the hemispheres. Gluing a 1" length of coupler tube onto the circuit board provided the location for the t-nuts (Photo 20).
The upper section of the pods is called the "chase" lights and was constructed similarly to the main array; except, the LED's were mounted onto small circuit boards (Photo 21). Eighty circuit boards with nothing more than pads close together were made to attach and align the LED's for the chase lights. The boards were mounted across from each other, so the light passed by the viewer at twice the rate that the main array was cycling. Instead of drilling another 720 holes, a vertical slot was used. The circuit boards were tacked in with CA and then coated with epoxy filler like the main array (Photo 22). The wire leads also had to be soldered with ribbon cable and protected with heat shrink. The pods were encased in acrylic tubing and another plastic hemisphere that was frosted.
In order to conform to a futuristic design, Dave machined a nozzle looking device that actually doubled as a motor retainer. It was made from aluminum and weighed 3 pounds. With nearly a mirror finish, it matched the reflective surface behind the LED's.
The reducer was custom made and contained the ALTACC for recovery. The reducer alone contained over a dozen pieces including the t-nuts that had to be mounted to secure it to the rocket (Photo 23). Each t-nut was glued to a 3/4" square piece of fiberglass and sanded to the inside diameter of the tube it was going in. Over 40 were constructed for the pods and reducer (Photo 24).
The nose cone was an interesting challenge because it had to contain its own battery (Photo 25) and control circuitry for its lights. The headlight (Photo 26) was mounted on top inside a short section of body tube. The ring light and centering ring were epoxied to a 3" piece of tube (Photo 27). This was epoxied into a 4" coupler after the wires were attached (Photo 28). For decoration, heat sinks were added to the outside of the tube, and the clear acrylic dome fit perfectly. The bottom of the nose cone also used a piece of clear acrylic hemisphere with a hole cut out for the 3" body tube.
The heat sinks, main array, and nozzle were taped off (Photo 29) and the painting began. The entire rocket was painted white (Photo 30). Since the heat sinks were already anodized black, the color scheme provided a nice contrast.
A 4-foot piece of 3" fiberglass body tube was used for the parachute compartment. It made the rocket about 9 1/2 feet tall, but it was necessary for the 15 foot cargo chute that was needed for a slow descent rate. The completed rocket weighed 35 pounds without the motor.
Details were added to enhance the appearance of the rocket. Heat sinks were added to the main body tube below the main array to counter the ones above it. Screws heads were added to the existing mounting holes in the heat sinks and the rail guides were detailed. Ultimate Rail Guides were used because they had to be stood off from the main body tube. Reflective Mylar that was used throughout the rocket was also used on the standoff and screw head for the rail guides.
The rocket consisted of 6 distinct light circuits: (1) the main array in the main body tube, (2) 4 tail lights in the bottom of the pods on the fins, (3) 4 chase lights in the upper pods, (4) 4 strobe lights also in the upper part of the pods, and (5) the headlight and the (6) ring light in the nose cone. Altogether 2096 ultra-bright LED’s were used. Building it required over 200 structural parts that had to be cut, shaped, sanded, drilled and glued, thousands of holes that were drilled, dozens of circuit boards that were made, and hundreds of feet of wire that were soldered to connect the circuits together. All this work for a night rocket?

The Circuitry
allocate \al-e-k_t\ to apportion for a specific purpose.

The circuitry in the main airframe consisted of a master controller that provided sequence lines to remote circuit boards that actually powered the lights. The master controller was a simple design that consisted of a 555 timer, which provided a pulse to a 4017 counter. The counters outputs switched 10 transistors that provided power to the remote circuit boards. A single 9 V battery powered the master controller, but the power for the LED's was from a battery pack rated at 39.6 V made from Nickel-Metal Hydride batteries. Thirty-three AA size batteries were used in the booster with a capacity of 1100 mAH. The nose cone had it's own battery pack; thirty AA batteries made a 36 V pack which was also rated at 1100 mAH.
Each remote circuit board took the input from the counter and sequenced the appropriate string of lights. The taillight and strobe light circuit boards were identical. They were selectable for when they flashed the lights by use of surface mount DIP switches and diodes. There were up to five outputs and each string of LED's had a LM317 current regulator set up to provide 15 mA of current. Dave designed all the circuit boards to have current regulators for each string of LED's, so each string had the same current independent of the number of LED's in each string.
The main array board was set up using removable diodes so any row could have been turned on for any channel output of the counter. By using this arrangement, the sequence was smoother. For example, on one pulse, rows 2, 3, 4 were lit as well as the corresponding rows at the bottom of the array. The next pulse rows 3, 4, 5 were lit providing a better transition of lighted rows.
The chase light circuit board was also constructed in this manner so that two strings are lit at any one time. The following pulse turned off one string and lit the next string. Since both the array lights and the chase lights were on intermittently, the current regulators provided 20 mA to each string.
The headlight control circuit board brought in power from the batteries and had 16 current regulators to provide 15 mA of constant current to each string. The ring light circuit board was very similar to the chase light board except that instead of only two strings being on at one time and sequencing, all the strings were lit except for two strings. The "off" strings rotated around the ring like the chase lights, one went off as another string came on for each pulse giving a smother sequence.
Time was running out in the summer of 2000, and we were determined to have it ready for LDRS XX. Dave built the controlling circuit boards while I completed the rocket, and we met at Lucerne to assemble everything together.

The Flight
ecstasy \`ek-st_-s_\ a state of overwhelming emotion; esp. rapturous delight.

We met on the lakebed and did the final assembly of the rocket for the first time. All this work we put into this rocket, and we didn't even know if it would light! We carefully coordinated the wiring since Dave was making the circuit boards, but I was connecting the LED's. We powered it up and were amazed at the sight. The illumination from the rocket filled the camper with a bright red glow. The flight was planned for Friday evening, but the winds did not cooperate. Many people gathered around that evening while we were taking pictures against the setting sun behind the mountains and even a few "fly it" chants started. Dave and I were completely in agreement. It was just too windy to fly it. We tried at the ROCSTOCK launch in November and again at Springfest 2002 but once again the winds were too great.
Finally things were looking good for a new launch from Tripoli San Diego called Plaster Blaster. I wrote to Andy Warner and he described the area and told me about the winds. In October, the winds are usually light plus there would be two evenings of night launches doubling the opportunity to fly it. There was a new moon and I verified that an Ultimate Rail would be at the launch. After a year and a half of waiting, it looked good for finally flying this rocket we had worked so hard on.
Upon arrival John Thompson, the Special Projects Coordinator, informed me that the launch rail would not be showing up. Andy quickly reacted. Kevin Harness had a full scale Army Hawk with a 22-foot rail but my rail guides were not the correct size. The rail guides were removed off of the Hawk and Paul Snow provided another pair for the Hawk. After obtaining a couple of bolts the rail guides were glued on with J.B. Weld.
Night was falling and the rocket was prepped. Andy gathered over a half a dozen people to help raise the tower. The lights were turned on and the glow was intense. After 8 months of designing and building and another 18 months of waiting, it was actually going to fly. I was concerned rather than nervous. The pods on the fins added a lot of mass and any jolt to the pod vibrated the fin. Any flutter would cause the fins to fall off during motor burn. With the glued on rail guides and heat sinks, I was convinced that it would not survive going up.
I stayed close to take pictures and Andy joined me on a four-wheeler. The countdown started and I was ready to click away. The motor lit and it quickly accelerated into the night sky (Photo 31). It flew straight up with the crowd cheering and I started breathing again. But only for a short time, it was quickly apparent that something was wrong. The rocket arched over and came down fast, very fast. The chute never came out and it came in ballistic.

Epilog
complacency \kem-pl_sen-s_\ feeling of contentment especially when coupled with an unawareness of danger, trouble, or risk.

There was absolutely no wind, and the impact was less than a hundred feet from the rail (Photo 32). It never occurred to me what had happened only that it was completely destroyed. The destruction of the B.I.T.C.H. at Plaster Blaster was by no means unusual, although the carnage was spectacular (Photo 33). All that survived was the motor casing, the machined nozzle, and minor hardware (Photo 34).
Many experienced flyers suffered losses for a variety of reasons. Some of the reasons are complex and may never be ascertained. But some, like in this case, were from dumb, avoidable, mistakes. The pressure of flying large, complex projects seems to cause an increase in the number of failures due to complacency. In this case I failed to arm the ALTACC. Once launched, it never had a chance for a successful recovery.
Dave and I have discussed how much time complex projects required prepping and using a checklist. One person would perform each step and another person would specifically check off each step. I thought the requirement of generating a checklist annotating everything that had to go in it would be too burdensome and not worth it. "Where would you stop?" I argued. Think about every step that it takes to assemble a simple project!
Step 36a - attach the shock cord to the rocket.
Step 36b - install the chute protector.
Step 36c - attach the other end of the shock cord to the parachute.

Step 150a - load the rocket onto the rail.
Step 150b - raise to the vertical position.
Step 150c - secure the rail.
Step 150d - arm the electronics.

Step 186a - install the igniter completely into the forward end of the motor.
Step 186b - attach the clips to the end of the igniter.
Step 186c - make sure the igniter wires are not touching each other.

Step 187 - Launch, if you have completed every step of this checklist!

Imagine the checklist for a two-staged, cluster rocket with camera payloads! I argued, "Where would it end"? I now believe "It doesn't matter". The more complex (or expensive) the project, the more a checklist needs to be used. It should be excruciatingly detailed not because you may forget a step but because eventually you will forget a step. Because it takes hundreds or even thousands of steps to assemble a complex project is why a checklist is necessary.
I have built other challenging rockets, but nothing like this one. Yet even as Dave and I made the long journey home, we already started discussing the next version of the B.I.T.C.H. Making the rocket larger (with more LED's) would actually make it easier to build. Most of the difficulties experienced had to do with the confined space that we had to work in. However, it won't be completed for a few years.
I would like to thank Andy Warner and the rest of the fine folks of Tripoli San Diego for their excellent support at Plaster Blaster. I also wish to express my heartfelt gratitude to Dave Flynn for all his help. From redesigning the control circuits and building them to assembling the array and machining the nozzle, I can easily say that this project would not have been completed without his help.