Mach Madness Extreme Flight Failure Analysis

[dixontj93060]

I participated in the Mach Madness contest at Thunderstruck 2012. The contest results are shown here. Although my rocket experienced an anomaly north of Mach 2.65, I used it as a learning experience as I typically am a low-and-slow builder. I thus have written a document that briefly summarizes the design/build (as requested by some TRF members) and, more importantly (for me at least), delves into potential reasons for the failure--where many of the areas are new learning for me. The document, Mach Madness Flight Postmortem, is located here for those that might want to examine (or critique) my conclusions.

Cheers,
Tim

[dixontj93060]

There has been some off-line private conversation from a couple of folks who opine that, given the acceleration spike at the 1s point in the flight, the root cause of the anomaly may be a motor event. This may be true. I did not originally post the Raven2 data in the analysis as it is a bit convoluted given the folding of the airframe (and the reason I sought input from Adrian A. on the results), but, given that I am not experienced in extreme flights, I thought I would go ahead and post it to get additional input.

The fact remains that there was an airframe failure, and I wanted that airframe to survive Mach 2.5-3.0, so the postmortem analysis is still valid, but if it does turn out to be a motor issue (that can be proven), I would be happy to include it in the analysis and give credit to the one providing the input. I will say that I spoke to several observers of the flight and they did not see anything in the flight pattern/profile to indicate a motor problem. But, in any case, you motor experts, have at it!

[Adrian A]

There has been some off-line private conversation from a couple of folks who opine that, given the acceleration spike at the 1s point in the flight, the root cause of the anomaly may be a motor event. This may be true. I did not originally post the Raven2 data in the analysis as it is a bit convoluted given the folding of the airframe (and the reason I sought input from Adrian A. on the results), but, given that I am not experienced in extreme flights, I thought I would go ahead and post it to get additional input.

The fact remains that there was an airframe failure, and I wanted that airframe to survive Mach 2.5-3.0, so the postmortem analysis is still valid, but if it does turn out to be a motor issue (that can be proven), I would be happy to include it in the analysis and give credit to the one providing the input. I will say that I spoke to several observers of the flight and they did not see anything in the flight pattern/profile to indicate a motor problem. But, in any case, you motor experts, have at it!

Did the nozzle look o.k. post-flight?

[dixontj93060]

Did the nozzle look o.k. post-flight?

Adrian, unfortunately, nozzle/motor/fin can has yet to be located.

[Adrian A]

In that case, a CATO is probably the simplest explanation. Otherwise, we would have to assume that the av-bay flipped around 180 degrees before running into the end of a shock cord to produce the 300+ Gs seen. Instead, I bet that the back end of the motor let go.

[Dwatkins]

Having seen the flight. Motor performed as I would have expected. No anomalies in the the flame or smoke trail. Looking at the data I suspect Adrain you are right. Avbay coupling let go. 1st event. Swapping end at 180* starts the the g increase by inertia. Hits booster or end of shock cord for second event.

Dennis

[dixontj93060]

Having seen the flight. Motor performed as I would have expected. No anomalies in the the flame or smoke trail. Looking at the data I suspect Adrain you are right. Avbay coupling let go. 1st event. Swapping end at 180* starts the the g increase by inertia. Hits booster or end of shock cord for second event.

Dennis

Actually no, the nose had to be the first point of failure otherwise there wouldn't have been a 9" zipper down the side of the payload on the side that the av-bay then was sheared. If the av-bay coupler failed first, the chute and nose would have just fallen out without the significant damage to the fore end and the payload bay.

[dixontj93060]

In that case, a CATO is probably the simplest explanation. Otherwise, we would have to assume that the av-bay flipped around 180 degrees before running into the end of a shock cord to produce the 300+ Gs seen. Instead, I bet that the back end of the motor let go.

I certainly could be wrong, but the flip doesn't look to be done until well after the high-g event.

[mannyskid]

I think that Adrian is probably right. One thing that concerns me is the battery voltage throughout the flight. The max battery voltage was only 3.96 volts. Was the rocket sitting on the pad for a long time? Was it an old battery? Could this just be because the Av-bay was thrown violently from the rocket? Also, before the first event, the altitude goes from 0-180', then from 180' to -600' (all while velocity is still increasing). Could this also be because the Av-bay was tossed about violently? If so, could we belive any of this data? Just speculation.

Manny

[GRIFFIN]

I think that Adrian is probably right. One thing that concerns me is the battery voltage throughout the flight. The max battery voltage was only 3.96 volts. Was the rocket sitting on the pad for a long time? Was it an old battery? Could this just be because the Av-bay was thrown violently from the rocket? Also, before the first event, the altitude goes from 0-180', then from 180' to -600' (all while velocity is still increasing). Could this also be because the Av-bay was tossed about violently? If so, could we belive any of this data? Just speculation.

Manny

It's a 4 volt lipo battery

[mannyskid]

So he was using the power perch. Okay, It makes more sense now. But what about the weird altitudes?

[dixontj93060]

So he was using the power perch. Okay, It makes more sense now. But what about the weird altitudes?

Not a power perch, just a LiPo and switch.

Altitudes were a question on my part, Adrian's email answers to me referred to the accel altitudes as the reference point for the first seconds of flight.

[mannyskid]

From taking a second look at the data, I noticed that even after the 311.14 G spike, the rocket keeps accelerating, leading me to belive that it was the nose cone that came off, and not the Av-bay.

Manny

[Dwatkins]

Actually no, the nose had to be the first point of failure otherwise there wouldn't have been a 9" zipper down the side of the payload on the side that the av-bay then was sheared. y.

When this flipped and knocked the nose cone & chute out these parts would still be moving at near Mach 2! Hence zipper down payload from the abrupt opening of a chute at these speeds.

Dennis

[dixontj93060]

When this flipped and knocked the nose cone & chute out these parts would still be moving at near Mach 2! Hence zipper down payload from the abrupt opening of a chute at these speeds.

Dennis

The fore payload opening would be facing backwards in your scenario (before any acceleration disruption BTW?), if turned backwards why wouldn't the nose and chute just deploy backwards without affecting the payload airframe structure as was observed? And if the payload was loose, how was the payload in such a firm position as to hold against a 1000 lb Kevlar cord? I don't see how it could without it still being attached to the ascending booster.

EDIT: Actually Dennis, I forgot to mention one other thing (which I believe I put in the analysis), the Kevlar cord also zippered ~3/4" into the 5-ply av-bay fore bulkhead (to save weight on connect points the Kevlar was threaded through the av-bay and held in place by expansion nuts). Zippering through plywood just could not have occurred unless the av-bay was firmly attached to the booster at the time.

BTW, before the flight I felt the weak link would be the (perceived thin) 1/8" Kevlar cord. Resulting debris from the flight proved me wrong. Materials in order of strength based on post flight analysis were: 1) Kevlar, 2) carbon fiber, 3) epoxy, 4) plywood, 5) aluminum hardware, 6) laminated phenolic.

[dixontj93060]

Just for clarification (and as an aid to those not reading the postmortem in detail), my proposed sequence of events after analysis is:

1. Internal radial failure of airframe at payload nosecone joint (initial 120 g spike).
2. In quick succession: loss of nosecone, pull of Kevlar out/down side of payload causing wide zipper and release of parachute through zipper area, parachute inflating, end of Kevlar reached where nosecone attach was ripped off and parachute canopy was detached from shroud lines (corresponding to the many g spikes including ones @ 300 g)
3. Above sequence lightened the rocket by more than 1/2 lb and thus the Raven still resident in the av-bay measured an increase in speed due to lighter weight.
4. In quick succession after the end of #2, the av-bay was detached by radial failure in top of booster by pull/drag of cord/payload debris.
5. As the av-bay, with the remaining payload airframe still attached, broke off and folded back the Raven inside continued its baro measurements and swapped signs on its velocity.
6. During steps 4 and/or 5 above the Kevlar cord shorts power feeding Raven.

In reference to the postmortem analysis, I categorize steps 1 and 2 above as "Event 1" and steps 4 and 5 as "Event 2".

[patelldp]

BTW, before the flight I felt the weak link would be the (perceived thin) 1/8" Kevlar cord. Resulting debris from the flight proved me wrong. Materials in order of strength based on post flight analysis were: 1) Kevlar, 2) carbon fiber, 3) epoxy, 4) plywood, 5) aluminum hardware, 6) laminated phenolic.

This is good news. I am using 1/8" Kevlar cord on my 54mm rocket to save space. I am able to fit an ample amount (25') in my drogue section.

[ECayemberg]

Here are some thoughts, but I'm no expert!

1. I enjoyed reading your post mortem report! Good content and great learning material.

1b. I think you're on to something with the nose coming off/failing theory. I believe this happened right at motor burnout. The graph shows that event happening at roughly 1.25 seconds, minus the roughly .15 seconds of idle time at the beginning, this equates to Liftoff + 1.1 seconds. The L2300 burns in earnest for 1.0 seconds, with .1 or .2 seconds of steep tail off following (~1.1sec burn). The anomaly happened right at motor shut down. I do not believe this was a motor failure, especially not at ~T+1.1 when motor shutdown was to occur.

1c. I also agree with Dennis in stating that when things come apart at Mach 2+, that violent event will zipper, crush, etc. lots of stuff!

2. Barometric altitude: while it seems that the charge did not fire, that data is wacked!. I'd be leery of a barometric altimeter that tells you your uber fast rocket achieved a net altitude gain of 200' and a net loss of 2700' or so...unless it wasn't vented (which I'm certain it was).

3. Accelermoter Velocity: I find a few things odd with this curve.
a. While I would expect a spike when the anomaly occured, there is no way there will be a steady positive acceleration post anomaly as shown in the graph from ~1.52 seconds on. As soon as the motor stops burning, the rocket will experience negative acceleration. As soon as an anomaly occurs, after the initial spike, without a significant force applied to the rocket (burning motor), there would not be such a steady acceleration curve. I would expect erratic positives and negatives resulting from the tumbling av-bay that I witnessed.
b. From T+.25 second to motor burnout/anomaly (T+~1.1seconds), the curve is dead nuts straight...printed it out and put a straightedge to it. This is fairly believable with the fairly constant axial acceleration, but the line is perfectly straight while the axial acceleration has small slopes present; I'm not sure these correlate as closely as they should.
c. At 2.0 seconds on the graph, or ~T+1.85 seconds, the axial acceleration drops to a flat line, presumably zero. This contradicts a free flying payload section's ascent and descent which would experience some measurable and I'd expect varied acceleration.

4. Accelerometer based Altitude: Adrian pointed you towards the data from this curve; which looks good until the anomaly occurs. Why would this curve increase following the event? I'm assuming it is based upon the increasing accelerometer based velocity, but again I don't believe it possible that a steady positive acceleration could have occured past anomaly due to dramatically increased drag and a lacking force of thrust. Furthermore, I would expect a "jump" around T+1.1 unless the spikes were dampened out for the sake of the curve?

I do not believe that the altimeter was the cause of failure. I do believe however that the altimeter had some wierd vibes going on from liftoff to motor burnout, and then again after anomaly. That said, it certainly sensed the separation. Brown out possibly? Definitely interesting data; thanks for sharing!!!

Again, just my thoughts, for what they're worth. 'Twas a really cool flight!

-Eric-

[SS/EA 6BBL 71 Cuda]

Tim, as said before many good thoughts and observations on this but,.....maybe just maybe Justin managed to get a shot of this on video?
If you could coax him out of his cave for a short time to retrieve possible moment....maybe we could see this in a little better detail.
It was through his work a couple years back that i narrowed down the loss of "Dream Evil" when it was determined that either the coupler turned into powder or a slight drag separation issue caused the previously said confetti fest.
At these speeds even a one frame shot may prove your hypothesis.....

Remember what i told you as i was leaving Sunday? "Tim you've got the recipe...now you've just got too learn how to cook it."

I'd give the Farrand-man a shout.

[dixontj93060]

Tim, as said before many good thoughts and observations on this but,.....maybe just maybe Justin managed to get a shot of this on video?
If you could coax him out of his cave for a short time to retrieve possible moment....maybe we could see this in a little better detail.
It was through his work a couple years back that i narrowed down the loss of "Dream Evil" when it was determined that either the coupler turned into powder or a slight drag separation issue caused the previously said confetti fest.
At these speeds even a one frame shot may prove your hypothesis.....

Remember what i told you as i was leaving Sunday? "Tim you've got the recipe...now you've just got too learn how to cook it."

I'd give the Farrand-man a shout.

Gus, good idea!!!

[patelldp]

Admittedly I have not read the entire report. However, do you have a picture of the rest of the parts recovered? I have a hard time believing that a Performance clam shell cone could go sideways at mach 2.65 and remain intact, while damaging the glassed phenolic. In my opinion, the 54mm tube is more robust that the cone you were using.

I just don't know how much is left of this rocket to assess the damage. There of course is a likelihood that the booster tube failed, seeing as that has yet to be recovered.

Additionally, to minimize weight addition, I would be sure to double wall any tubing that is spanning open air within the rocket. For example: from the top of the motor casing to the altimeter bay is a single thickness of tube. Epoxy in a coupler to double wall that area and that should strengthen it significantly.

[dixontj93060]

Admittedly I have not read the entire report. However, do you have a picture of the rest of the parts recovered? I have a hard time believing that a Performance clam shell cone could go sideways at mach 2.65 and remain intact, while damaging the glassed phenolic. In my opinion, the 54mm tube is more robust that the cone you were using.

I believe the term I used in the report was that the pictures of the av-bay/coupler combo were "unreadable" (or something like that), this is true, but actually doesn't tell the whole story... I had taken pictures of the debris at the launch and read/loaded it on computer desktop the Monday after Thunderstruck when I got into the office. Early that morning I had been cleaning up some other files on the desktop when somehow I picked/selected those three pictures also and they went into the trash without me knowing. And yes, you guessed it, I emptied the trash without looking. When I realized what I had done I was sick. I ran back to my house to go dumpster diving as only that morning I had cleaned out the truck I took to the launch, but to no avail, as my trash carrier just happens to come on Monday and I missed him by about a half hour .

In any case, no reason to mis-represent. As described in the analysis, the top part of the payload was jagged and it was easy estimate from overall length of 14.5" that I had lost about the first 2.5" which coincides with the shoulder length on the nosecone (actual shoulder length is 2-5/8" with bulkhead plus break varied +/-1/4" or so distributed pretty evenly around the circumference of the payload frame). I would agree that you may think the nosecone shoulder would be weaker (and there was only about a 1/4" chip in the bottom of the shoulder), but there are three things to note: 1) as stated in the analysis, to save weight I only had one layer of 6 oz. glass on the phenolic--that isn't much. My grandson put a big hole in my PML Bulldog with two layers of glass and that was just a tip of the rocket while it was standing on the floor, 2) when the nosecone came off likely it was still protected by the bulkhead on the end 3/16", 5 ply and the 1-1/4" aluminum studs which themselves were attached internal to the nosecone shoulder with epoxy/chopped carbon mix which strengthened those two sides considerably, and, 3) the fiberglass layering on this particular nosecone, again an older one from Performance Rocketry, seems much thicker than the more recent stuff I have received from them (non-wound nosecones I mean).

I just don't know how much is left of this rocket to assess the damage. There of course is a likelihood that the booster tube failed, seeing as that has yet to be recovered.

This is actually true. I assume the booster airframe also had an internal radial failure but I have not seen the booster. There was damage on the aft coupler of the av-bay where the booster attached. When I say damage though it was more of a shear/twisting/scraping type mark on about 1/4 of the outside of the coupler on the av-bay and a sheared off charge location (wires broken/cut). Again though, there was no BP residue of any kind on the aft bulkhead indicative of a charge going off.

Additionally, to minimize weight addition, I would be sure to double wall any tubing that is spanning open air within the rocket. For example: from the top of the motor casing to the altimeter bay is a single thickness of tube. Epoxy in a coupler to double wall that area and that should strengthen it significantly.

Please explain this. Again, I am no expert in extreme flight profile rocket design. In my mind double walling the phenolic is going to add much more weight than if I added a couple more layers of FG, and, in fact, would be even heavier than if I went to a G12 Profusion airframe.

[dixontj93060]

1b. I think you're on to something with the nose coming off/failing theory. I believe this happened right at motor burnout. The graph shows that event happening at roughly 1.25 seconds, minus the roughly .15 seconds of idle time at the beginning, this equates to Liftoff + 1.1 seconds. The L2300 burns in earnest for 1.0 seconds, with .1 or .2 seconds of steep tail off following (~1.1sec burn). The anomaly happened right at motor shut down. I do not believe this was a motor failure, especially not at ~T+1.1 when motor shutdown was to occur.

Eric, actually you are right. I have it at 1.14 to 1.15s but that is right in line with motor burnout (although the Raven2 did not report burnout, and I think I mistakenly have in the paper 1.3s when I should have 1.13). But saying it was at burnout, then what? In the analysis (baring a CATO or other anomaly), I assumed the worst case of nosecone inertia continuing, rest of rocket staying a rest (which of course impossible) placing a force on the shear pins of 30 lbs which is less than 1/2 the force it takes to break them. Again, I am not an extreme flyer, so what else might have happened at this juncture? Or is burnout just so disruptive that I could have gotten a 3 degree tilt going at Mach 2.6 and thus enough to kick off the radial failure scenario?

2. Barometric altitude: while it seems that the charge did not fire, that data is wacked!. I'd be leery of a barometric altimeter that tells you your uber fast rocket achieved a net altitude gain of 200' and a net loss of 2700' or so...unless it wasn't vented (which I'm certain it was).

Well yes. I believe that the only value to the baro data is showing that there was a velocity sign reversal.

3. Accelermoter Velocity: I find a few things odd with this curve.
a. While I would expect a spike when the anomaly occured, there is no way there will be a steady positive acceleration post anomaly as shown in the graph from ~1.52 seconds on. As soon as the motor stops burning, the rocket will experience negative acceleration. As soon as an anomaly occurs, after the initial spike, without a significant force applied to the rocket (burning motor), there would not be such a steady acceleration curve. I would expect erratic positives and negatives resulting from the tumbling av-bay that I witnessed.
b. From T+.25 second to motor burnout/anomaly (T+~1.1seconds), the curve is dead nuts straight...printed it out and put a straightedge to it. This is fairly believable with the fairly constant axial acceleration, but the line is perfectly straight while the axial acceleration has small slopes present; I'm not sure these correlate as closely as they should.
c. At 2.0 seconds on the graph, or ~T+1.85 seconds, the axial acceleration drops to a flat line, presumably zero. This contradicts a free flying payload section's ascent and descent which would experience some measurable and I'd expect varied acceleration.

This is all true. I pretty much discount all the data past 1.5s and I believe I state that it is suspect in the analysis. Saying that, I had a theory on what is going on, but failed to get confirmation of this from Adrian A. and thus didn't include it in the report. At the risk of being labeled a nut, I will relate my theory here...

It might be possible that through the whole sequence out to the 3.8s of data available that the av-bay/payload was still attached to the booster but in an inverted direction (and likely not near the booster so as to bang against the booster to get additional accel spikes, but hanging on the drogue shock cord behind the booster) and subsequently dropped off after the 3.8s. Why do I say that? Take a look at a nominal flight simulation (attached). If you invert the velocity curve after burnout, what do you have? You have the measured steep velocity curve on a linear rate, a slight move up around burnout and then, when inverted, a continued upward movement at a lower slope(!). Could it be that when the av-bay was flipped the accelerometer velocity algorithm in the CPU (or maybe in the FIP program) mis-interprets the slower backwards acceleration as an increase in velocity during that 2+ second window past the 1.5s mark instead of a decrease in velocity?

4. Accelerometer based Altitude: Adrian pointed you towards the data from this curve; which looks good until the anomaly occurs. Why would this curve increase following the event? I'm assuming it is based upon the increasing accelerometer based velocity, but again I don't believe it possible that a steady positive acceleration could have occured past anomaly due to dramatically increased drag and a lacking force of thrust. Furthermore, I would expect a "jump" around T+1.1 unless the spikes were dampened out for the sake of the curve?

See above theory.

I do not believe that the altimeter was the cause of failure. I do believe however that the altimeter had some wierd vibes going on from liftoff to motor burnout, and then again after anomaly. That said, it certainly sensed the separation. Brown out possibly? Definitely interesting data; thanks for sharing!!!

Brown out maybe... I do know that when I opened the av-bay the LiPo was shorted -- wires were stripped and pinched by Kevlar going through bay that had been pulled during the flight. Given that condition I was quick to separate the leads and unhook everything. I did not though check on whether the switch had been operated or not.

On the other hand, I do know though that the other altimeter board battery was not shorted, yet the board lost power. The Stratalogger recorded a loss of power at, guess when, 1.35s (surprise!). It is possible, given the tight quarters inside the av-bay, the position of the pushbuttons, and the violent disturbance suffered, that the switches could have been operated. I cannot remember perfectly, but I do believe that the Stratalogger switch was off and it, as did the Raven2, powered right back up when I applied power. But, on the Raven2 in particular shouldn't it have recorded more than 3.8s of data? In the manual it states that the capacitor will hold power for 7-10 seconds which would have more than covered the full flight (and fluttering of the av-back to the ground). I'm clueless on this power issue for the Raven2--seems like it would have picked up more data?

[rfjustin]

maybe just maybe Justin managed to get a shot of this on video?
If you could coax him out of his cave for a short time to retrieve possible moment....maybe we could see this in a little better detail.

Great idea, and I would be happy to share, but I didn't get this one on video which is most unfortunate.

IMO, I wouldn't put a lot of stock into the data gathered by the altimeter. I agree with Eric, no fault of the altimeter, but data is more than likely FUBAR based on the nature of the failure.

Near as I can tell, the motor worked, looked like either airframe on NC failure to me.



Justin

[patelldp]

Please explain this. Again, I am no expert in extreme flight profile rocket design. In my mind double walling the phenolic is going to add much more weight than if I added a couple more layers of FG, and, in fact, would be even heavier than if I went to a G12 Profusion airframe.

My hypothesis is that you experienced a radial failure of a single walled section of your rocket. I will first explain how I arrived at this, and then I will describe my remedy that I posed before.

You explain in the report that you desired a lightweight tube and you achieved this by designing the rocket around phenolic with a single layer of 6oz glass. Radially, as you showed in your report, phenolic with a single layer of glass is not all that strong.

This leads into my assumptions:

1. The booster section was of a standard design and not zipper less. I assume the rocket was designed in the standard HPR scheme where the rocket separates at the altimeter bay and also at the nosecone.

2. There is some damage that makes it appear as though you "scraped" the altimeter bay against the inside of the booster.

3. Your altimeter bay coupler is more robust than a single thickness of phenolic tubing.

4. The altimeter bay is intact, as is most of the upper section of tubing apart from the zipper.

So now for my idea:

Your airframe experienced a radial failure just above your motor case.

This would explain the relative good condition of everything north of that point, especially the altimeter bay. I would assume that your CG of this rocket was somewhere near or just below the top of the fore end of the motor casing. Should there be any radial force, say a crosswind or any other turbulence, this is the approximate location that the rocket would "rotate" around.

Your rockets' initial failure was right at the top of the motor, just ejecting everything above that point. In this EXTREMELY turbulent environment, the insane deceleration, and centrifugal force experienced by the cord and main chute, it easily sheared the pins in the nosecone and ejected. The quick inflation of the chute resulted in the zipper, but this ejection scenario explains the relative good condition of the nosecone shoulder.

So, how do we fix this given the current material selection?

1. Add more glass (you knew this)
2. Epoxy a coupler into every area of the rocket that is only a single layer of phenolic tube.

I think #2 is a lightweight option that will help significantly. Basically, install a coupler of the appropriate length such that the motor butts up against one end of said coupler, and the altimeter bay butts up against the other end. Further up, the fore end of the altimeter bay butts up against the second coupler, and the nosecone shoulder butts up against the fore end of the second coupler. Essentially, it's a "rocket within a rocket."

Or, you can redesign the rocket in a very minimal matter using a design that Tony A. uses, but using better materials.

[dixontj93060]

My hypothesis is that you experienced a radial failure of a single walled section of your rocket. I will first explain how I arrived at this, and then I will describe my remedy that I posed before.

You explain in the report that you desired a lightweight tube and you achieved this by designing the rocket around phenolic with a single layer of 6oz glass. Radially, as you showed in your report, phenolic with a single layer of glass is not all that strong.

This leads into my assumptions:

1. The booster section was of a standard design and not zipper less. I assume the rocket was designed in the standard HPR scheme where the rocket separates at the altimeter bay and also at the nosecone.

2. There is some damage that makes it appear as though you "scraped" the altimeter bay against the inside of the booster.

3. Your altimeter bay coupler is more robust than a single thickness of phenolic tubing.

4. The altimeter bay is intact, as is most of the upper section of tubing apart from the zipper.

So now for my idea:

Your airframe experienced a radial failure just above your motor case.

This would explain the relative good condition of everything north of that point, especially the altimeter bay. I would assume that your CG of this rocket was somewhere near or just below the top of the fore end of the motor casing. Should there be any radial force, say a crosswind or any other turbulence, this is the approximate location that the rocket would "rotate" around.

Your rockets' initial failure was right at the top of the motor, just ejecting everything above that point. In this EXTREMELY turbulent environment, the insane deceleration, and centrifugal force experienced by the cord and main chute, it easily sheared the pins in the nosecone and ejected. The quick inflation of the chute resulted in the zipper, but this ejection scenario explains the relative good condition of the nosecone shoulder.

So, how do we fix this given the current material selection?

1. Add more glass (you knew this)
2. Epoxy a coupler into every area of the rocket that is only a single layer of phenolic tube.

I think #2 is a lightweight option that will help significantly. Basically, install a coupler of the appropriate length such that the motor butts up against one end of said coupler, and the altimeter bay butts up against the other end. Further up, the fore end of the altimeter bay butts up against the second coupler, and the nosecone shoulder butts up against the fore end of the second coupler. Essentially, it's a "rocket within a rocket."

Or, you can redesign the rocket in a very minimal matter using a design that Tony A. uses, but using better materials.

Dan, OK, I'm good with most of the above.

First Assumptions... 1-Yes, 2-Yes, 3-Yes (coupler thicker and inside laminated with fiberglass), 4-No, not quite; zipper yes, but also top (fore) ~2.5" of payload bay broken off.

I still have a problem with how a zipper occurs in your (and Dennis') scenario. If I am moving at Mach 2.6 (and when loosing mass the rocket accelerates--but ignoring this) and I experience rotation which causes radial failure at the mid-point (aft coupler of av-bay) the payload/av-bay quickly folds back parallel to flight. In that scenario, where is the force vector that pulls the shock cord through an equally, but oppositely force applied (or stationary) payload? Please be specific, because I cannot figure it out.

Finally, I agree with your reinforcement recommendations, but Blur II is redesigned and the architecture is altogether different (parts in UPS transit now).

[patelldp]

I still have a problem with how a zipper occurs in your (and Dennis') scenario. If I am moving at Mach 2.6 (and when loosing mass the rocket accelerates--but ignoring this) and I experience rotation which causes radial failure at the mid-point (aft coupler of av-bay) the payload/av-bay quickly folds back parallel to flight. In that scenario, where is the force vector that pulls the shock cord through an equally, but oppositely force applied (or stationary) payload? Please be specific, because I cannot figure it out.

This is difficult to explain, and may be best for you to pick up a rocket payload section and simulate.

The Exercise
1. Grab the payload section by the coupler, with the shock cord and parachute inside your payload tube, with a nosecone on top.

2. Swing the payload tube like a baseball bat.

3. When the chute and cord eject, the parachute and cord will create drag that will oppose the motion of your swinging. The cord will therefore bear against the side of the tube opposite the direction of your motion.

What happened at full speed
This simulation I describe is really just to get a feel for what happened from a functional perspective. In reality, this happened MUCH quicker, and with MUCH more force. Picture swinging the tube 50x harder.

As soon as the nosecone was ejected, that drag force opposing the direction of the tube swing increases dramatically. As soon as some of the cord and chute got out, it was game over, and there was certainly more than enough of an opposition drag force to zipper phenolic, which is in and of itself easy to zipper. This all happened in a infinitesimally short period of time, as far as I can perceive it.

All while this was happening, the booster was still going. The booster would then start to pull the payload section, cause the chute to oppose movement in the other direction, and most likely shredded the chute and broke the drogue shock cord.

Remember that at this acceleration, these items exert a force of x-times that of their steady state weight, x being the gee-loading at the time.

Finally, I agree with your reinforcement recommendations, but Blur II is redesigned and the architecture is altogether different (parts in UPS transit now).

Glad to hear you are getting back on it. Looking forward to your next effort!

[ECayemberg]

Eric, actually you are right. I have it at 1.14 to 1.15s but that is right in line with motor burnout (although the Raven2 did not report burnout, and I think I mistakenly have in the paper 1.3s when I should have 1.13). But saying it was at burnout, then what? In the analysis (baring a CATO or other anomaly), I assumed the worst case of nosecone inertia continuing, rest of rocket staying a rest (which of course impossible) placing a force on the shear pins of 30 lbs which is less than 1/2 the force it takes to break them. Again, I am not an extreme flyer, so what else might have happened at this juncture? Or is burnout just so disruptive that I could have gotten a 3 degree tilt going at Mach 2.6 and thus enough to kick off the radial failure scenario?

A theory is the nose departed the airframe at or near motor shutdown, at or just after Max Q, when the rocket body began to experience
negative acceleration due to the lack of thrust. The lower relative drag of the nose verses the airframe meant the airframe wanted to slow down faster than the nose. It is possible that the heat from the extreme speed partially melted, or at least comprimised the (nylon?) shear pins. In my limited experience with speedy flights, I've seen the base of the nosecone as one of the areas greatly affected by heat. We saw lots of high-temp paint melting on the Mach Madness flights, if it could melt 500 deg. F paint, it could probably compramise nylon with a melting point around 430 deg. F.

It's a poor example, but I've had the nose come off my Gizmo following burnout unexplainably one or two times, and it gets a trio of #2-56 nylon screws which should have plenty of holding power.

If the airframe (not altimeter) vent hole was partially blocked by cord, chute, or otherwise, the combination of higher internal pressure, potentially heat weakened shear pins, and the negative acceleration coupled with lower drag of the nose vs. airframe, may have been enough to force that nose off prematurely.

It might be possible that through the whole sequence out to the 3.8s of data available that the av-bay/payload was still attached to the booster but in an inverted direction (and likely not near the booster so as to bang against the booster to get additional accel spikes, but hanging on the drogue shock cord behind the booster) and subsequently dropped off after the 3.8s. Why do I say that? Take a look at a nominal flight simulation (attached). If you invert the velocity curve after burnout, what do you have? You have the measured steep velocity curve on a linear rate, a slight move up around burnout and then, when inverted, a continued upward movement at a lower slope(!). Could it be that when the av-bay was flipped the accelerometer velocity algorithm in the CPU (or maybe in the FIP program) mis-interprets the slower backwards acceleration as an increase in velocity during that 2+ second window past the 1.5s mark instead of a decrease in velocity?

Interesting...certainly could be. I would still expect a wide range of varied accelerometric values however, as once the payload departed from the rest of the stable lower airframe, it would likely not be stable and would therefore tumble, oscillate, and otherwise not stay in a uniform axis.

This is fun, but my brain hurts!

-Eric-

[dixontj93060]

This is difficult to explain, and may be best for you to pick up a rocket payload section and simulate.

The Exercise
1. Grab the payload section by the coupler, with the shock cord and parachute inside your payload tube, with a nosecone on top.

2. Swing the payload tube like a baseball bat.

3. When the chute and cord eject, the parachute and cord will create drag that will oppose the motion of your swinging. The cord will therefore bear against the side of the tube opposite the direction of your motion.

I certainly understand the centrifugal force causing the shearing of the nosecone pins and the nosecone being ejected as described above, that really wasn't my basic question--it is what happened after that...

What happened at full speed
As soon as the nosecone was ejected, that drag force opposing the direction of the tube swing increases dramatically. As soon as some of the cord and chute got out, it was game over, and there was certainly more than enough of an opposition drag force to zipper phenolic, which is in and of itself easy to zipper. This all happened in a infinitesimally short period of time, as far as I can perceive it.

This is where I don't quite get it. What I would expect is that the payload and nose would be rotating and incurring the drag force opposite of flight nearly in sync. And as the nosecone ejects from centrifugal force, there is no opposing drag force as you describe above and/or it is extremely small compared to them both experiencing the drag at Mach 2.6. In this scenario I see that dominant drag taking them down together. While the booster moves forward, the nosecone and chute then come out facing opposite of flight direction; parachute canopy and nosecone bulkhead get pulled off as they reach limit of cord. No opportunity for a zipper that I see. Also in this scenario there is no explanation for the cracked/missing 2.5" at top of payload because the pins would simply shear with the centrifugal rotation.

Glad to hear you are getting back on it. Looking forward to your next effort!

Well, yes, thanks. We likely will never know what happened unless I find a video somewhere. To be honest, not a big deal, it was all a big experiment anyway

[dixontj93060]

It is possible that the heat from the extreme speed partially melted, or at least comprimised the (nylon?) shear pins. In my limited experience with speedy flights, I've seen the base of the nosecone as one of the areas greatly affected by heat. We saw lots of high-temp paint melting on the Mach Madness flights, if it could melt 500 deg. F paint, it could probably compramise nylon with a melting point around 430 deg. F.

Hmmm, that's interesting. Most of the paint bubbling was at the base of the cone (and it was 1200 deg. paint!). Maybe the strength of the shear pins was compromised. I will have to look at the characteristics of that nylon.