J record attempt using Loki J1026

[Rschub]

I agree Adrian.

I feel like fin design is an area where @Rschub's TLAR Method (*) is about the best we can do.

Thanks for sharing all this !

-- kjh

(*) - TLAR - That Looks About Right - Minnie-Magg drag mods for Level 2?

Well, it didn't look right to me which led to my comment.

High strength is not the critical factor in flutter, it is stiffness. There is also a "flagpole" effect on the section of the fin that is not supported anywhere after the end of the flllet on the body tube.

I did look and found some .005 uni carbon from a vendor, someone could make a custom laminate with 12 layers at 30 degrees say, which would have high stiffness at any orientation. The multiple layers would also increase stiffness, somewhat like a phone book is difficult to tear through when flat, but easier when it is torn at an angle.

This is just my SWAG, however.

Semi Wild Assed Guess

 

[wclaybaugh2]

I played with those equations a few days ago and got the same result that I did with FinSim, namely that if you hold the tip chord and fin height constant, and decrease the root chord, it says that the fin becomes more resistant to fluttering. According to the equation, this stays true all the way down to having no root chord at all. To me, this is just a sanity check failure that says that these equations are not applicable for the geometry we typically use, so I'm just going to ignore this analytical approach from here on out.

Adrian:

I too have had questions about the analytic results on flutter, albeit that concern was a decade ago when I was developing the fin designs for my six and nine inch motors.

In the event I chose (for the sustainer) an all aluminum fin that tapers from the root to the tip (0.500" to 0.050") and has only very a small section behind the rocket (see the attached image). Analysis of that fin shape and material showed no flutter through Mach 6 but I remain unconvinced as to the validity of that result. Flight experience to date indicates there appears to be no flutter though Mach 2.4.

I also concluded that the several small advantages of four fins was cumulatively worth capturing. I appreciate the "sunk cost" argument with regard to already having hardware and infrastructure built around three fins but that "cost" might want to be compared to the cost of flight testing that reveals design issues that might not occur with more conservative design....

A question: since more material at the fin tips adds nothing to strength and tends to lower the resonate frequency; is "tip-to-tip" required? Would not reinforcement at the crease location be sufficient, particularly if combined with tapering of the fin thickness?

Bill

 

[Adrian A]

Excellent flight, thanks for the data and pushing the limits! Do you have details on the specific prepreg product used?

I don't. Unfortunately, I don't even remember where I got it, though it was likely Rock West Composites.

 

[kjhambrick]

Bill --

I remember reading a paper years ago about the effects of tapering the fin thickness from root -to- tip.

It seems like it was an old NACA paper but I've been googling to no avail -- I've found lots of references to taper ratio ( C(t) / C(r) ) but no luck with variable thickness.

Do you have a citation for the effect of variable fin thickness on flutter ?

Thanks.

-- kjh

 

[robopup]

Hello Adrian

I am by no means an expert, but I was concerned before the flight. I find it revealing that the damage occurred at exactly 90 degrees to the fiber orientation.
Please consult the chart here, the green outlined area is stiffness at an angle to the fibers:
View attachment 596513
as you can see the stiffness for unidirectional fibers in the first figure is very low, due to the very small cross sectional area the bending forces act upon.

The third figure shows isotropic conditions that would take a minimum of 6-10 layers, which would likely be too heavy for this application.

I believe your best bet would be to use bidirectional material with the 90/0 axis at a 45 degree angle to the "bend line" in the fin, which would increase stiffness dramatically in that area.

I agree that 45's are important for these very swept fins. And you can still keep total thickness low depending on the product you use. On my fast 38's, I've been using uni T700 with a [-45/0/45/90]s layup where the -45's are on the outside and in a similar direction to the leading edge of the fin. The total thickness of this laminate is 0.045 in.

It's relatively straightforward to calculate the stiffness matrix for our laminates and also homogenized engineering constants as a function of orientation. It is important you're working with good test data, but you can still get in the ballpark if you take datasheets with a grain of salt. The part I've always struggled with is how to determine appropriate loads (or some other criteria like flutter) to design to.

 

[Adrian A]

Adrian:

A question: since more material at the fin tips adds nothing to strength and tends to lower the resonate frequency; is "tip-to-tip" required? Would not reinforcement at the crease location be sufficient, particularly if combined with tapering of the fin thickness?

Bill

My current plan is to align and glue on new 0.062" fins on that are identical to the last set, add the fillets like the last time, and then add the following layup over the fillets:

1. A layer of unidirectional fiber parallel to the trailing edge (38 deg from being perpendicular to the tube) that covers the fillet and about 0.5" of the body tube, out to the first ~70% in height away from the body tube
2. One or two tip to tip layers of relatively light woven cloth that is 0/90 aligned with the rocket. This makes it 0/90 on the body tube but approx. 60/30 relative to the mean chord line of the fins. If I do 2 layers I'll try to end the inner layer about 40% of the way out from the tube

I've got to look to see what uni CF options I have in my shop. I think I have a very thin option and a very thick option. I'm not totally sure I want to attempt the trimming I'm describing in #1, but if I wet it out between 2 pieces of plastic, I can cut it to shape with the plastic on both sides to help keep it under control. I'd like for the finished fins near the root to be in the 0.09-0.1" range, but if they're over 0.08 I expect they would be fine.

 

[watheyak]

Would two dissimilar materials in the stack help? Maybe a core of Kevlar jacketed with carbon fiber, in the orientations that's been discussed previously?

My non-engineer observations are that good carbon fiber parts ring like a bell with nice sustain. A mixed-material part might just have a dead thud.

Would the different resonating frequencies of the materials dampen or cancelb? Or would they just resonate and flutter at a different frequency?

Again, this is a great thread. Thanks Adrian for sharing all this and allowing(?) the thread drift for everyone's education.

 

[Adrian A]

I think it's possible that extra damping would help, but the NACA flutter study doesn't indicate/account for that. It doesn't account for the density either, just the stiffness. Also, the speed is given in terms of Mach number. This makes me think the phenomenon is more about the frequency generated by the aerodynamics and the resonance of the fin at its natural frequency isn't a major factor.

If you combine multiple materials into one elastic structure, the structure will have just one resonant frequency for each mode shape; the different materials will combine for stiffness and mass to affect the natural frequencies for the structure as a whole but I don't think they would cause independent mode shapes or resonant frequencies.

 

[wclaybaugh2]

Bill --

I remember reading a paper years ago about the effects of tapering the fin thickness from root -to- tip.

It seems like it was an old NACA paper but I've been googling to no avail -- I've found lots of references to taper ratio ( C(t) / C(r) ) but no luck with variable thickness.

Do you have a citation for the effect of variable fin thickness on flutter ?

Thanks.

-- kjh

I have no formal reference on this.

I found, messing first with with FinSim and then with the underlying equations, results that seemed to imply that lower mass at the tips resulted in a higher frequency of flutter onset, which seemed—to me—to in turn imply a desire for a tapered fin if flutter seemed an issue.

Bill

 

[Adrian A]

I cut out and shaped new fins and attached them with JB Weld. One fin wasn't quite in the right location and seemed a little warped, so I removed it and cut out another pair. I sanded the edges down to match a spare fin from the original set, and when I went about 1 mm too far on the tip chord, I clamped the new fins to the ones that were glued on and sanded them down together so they match.


Below, if you look closely you can see a scar on the top end of the tube where previous fins broke off and took a layer of axial unidirectional fiber with it. This is in a part of the airframe without any real structural loads (the motor sticks out past the front edge), so I'll just fill and smooth it later.



Now I'm just waiting for the Cotronics epoxy I ordered on Aug 1 to ship. (fingers drumming) Well, back to assembling my 29mm Blue Raven production test fixture.

 

[Adrian A]

I decided to do some more in-depth analysis and compute the Cd as a function of velocity for this flight. (Warning, it gets geeky)

Drag force is 1/2 * density * velocity^2 * Cd * Area, and it also equals mass * acceleration

therefore Cd = 2 * mass * acceleration/ (density * velocity^2 * area)

Mass was measured, because I took the time to weigh the unloaded rocket and the mass of the motor after recovery (560 g), bringing the post-boost mass to 907 grams.
Area is easy to compute from the nosecone maximum diameter (1.61")

Air density is trickier. It's a function of pressure, temperature, and humidity. The University of Wyoming has a website that provides balloon sounding data recorded twice a day by the NWS at dozens of locations around the country. I used this data to get the humidity and temperature as a function of pressure for the closest measurement site, to help calculate the density, and I also used it to more accurately convert measured pressure to altitude, which I used later as a check. There are equations for density that you can look up in Wikipedia, but I didn't want to get too deep into the formula, so I used an online calculator to plug in a few points of temperature, pressure, and humidity to get density from the balloon soundings, and determined that starting with the ground-level density and then multiplying by the pressure ratio and the inverse of the absolute temperature ratio got density answers that were very close to the online calculator.

Velocity is provided by the Blue Raven, but it was corrupted by the circumstances of this particular case. First, the raw acceleration values during the 100G burn were pretty different between the two altimeters, indicating that one or both had bad calibrations, while the measurements made by the +/-32G altimeter had good agreement. Second, the vertical velocity calculated by the Blue Raven depends on gyro values that in this case immediately exceeded the gyro measurement range because of the flutter. Knowing that this was a very vertical flight, I decided to just assume a perfectly vertical flight, which lets me just integrate the accelerometer axis that's aligned with the rocket and subtract off gravity.

I adjusted the measured acceleration data during the high-G portion for both units by a fudge factor until the integrated velocity crossed zero at the actual observed apogee time. A second check is to compare the integrated accel-based altitude to the actual altitude measured from the pressure sensor and the balloon soundings. The accel-based altitude was 18143 vs 17784 for the sounding-adjusted barometric altitude, which is close enough (2%) IMO, considering that the rocket wasn't purely vertical during its flight.



With the the high-G accel values adjusted, here is the resulting velocity plot from the two units:

You can actually see the period of major fin flutter where the velocity takes a steeper turn at about 2-4 seconds. The top speed was just over 3200 feet/second, or Mach 2.85.

Here's the result for the Cd vs Mach number for this flight:


Time is going from right to left, starting with the end of the burn (which looks like negative drag), and going to apogee at the left end, where it gets noisy from having very little measurable drag acceleration. The big noisy triangle in the middle is when the fins were cracked and fluttering. There was a brief time in between motor burnout and when the fins failed that I got a meaningful Cd measurement:



Here's what RASAero has for that area:

At Mach 2.60, RasAero has 0.369 after burnout, compared to my measured 0.347. In other words, the rocket was on the way to exceeding the RASAero simulated altitude (and the Tripoli J single-stage record) before the fins failed.

 

[kjhambrick]

@Adrian A --

Outstanding analysis !

The shape of your CD -vs- Mach curve fits the classic profile.

I love it. Can't wait to see the next one with the improved fin can.

Thank you

-- kjh

 

[Buckeye]

Drag force is 1/2 * density * velocity^2 * Cd * Area, and it also equals mass * acceleration

So, no need to include -mg when using acceleration from the Blue Raven?

 

[Adrian A]

So, no need to include -mg when using acceleration from the Blue Raven?

Yes, not for drag measurements.

Gravity acts on all parts of an accelerometer equally, so is not perceptible while it is in flight. (On the pad, the ground is pushing back up to keep the rocket from falling to the center of the earth, and this resistance to gravity is what is measured). Gravity does need to be added to the measured accelerations when computing the velocity.

 

[Adrian A]

@Adrian A --

Outstanding analysis !

The shape of your CD -vs- Mach curve fits the classic profile.

I love it. Can't wait to see the next one with the improved fin can.

Thank you

-- kjh

The shape of the Cd profile could be partly coincidence. If the fins weren't broken at that point, the Cd would peak around 0.7, rather than 1.8.



The blip around Mach 0.6 is makes me curious. Both Blue Ravens picked it up at around 10.5 seconds into the flight. Maybe a little burst of thrust from the delay burning?

 

[kjhambrick]

The shape of the Cd profile could be partly coincidence. If the fins weren't broken at that point, the Cd would peak around 0.7, rather than 1.8.

View attachment 599100

Are you thinking that the higher than normal CD might be due to active 'buzzing' of the fin material or something else or ???

The blip around Mach 0.6 is makes me curious. Both Blue Ravens picked it up at around 10.5 seconds into the flight. Maybe a little burst of thrust from the delay burning?

View attachment 599101

It is kinda odd ... I went back and looked at the a -vs- t in your post #20 in this thread and you can just barely see that blip at about 10.5 sec in the orange z-axis acceleration in the full flight graph.

It might be the delay burning but I wonder what else might cause off-axis accelerations during the coast phase ???

There does seem to be an inflection in the blue x-axis acceleration -vs- time about then too.

And speaking of the blue line, the x-axis acceleration looks 'busy' between 5 and 15 seconds in the same full scale plot.

I can't think of anything normally happening up there during that time, unless the fins stopped buzzing about 10.5 seconds ???

Fun !

-- kjh

 

[Adrian A]

Are you thinking that the higher than normal CD might be due to active 'buzzing' of the fin material or something else or ???

Yes, the fins had failed structurally and they were flapping in the wind.

It is kinda odd ... I went back and looked at the a -vs- t in your post #20 in this thread and you can just barely see that blip at about 10.5 sec in the orange z-axis acceleration in the full flight graph.

It might be the delay burning but I wonder what else might cause off-axis accelerations during the coast phase ???

Once the fins failed, the rocket started rolling like crazy. The gyro in the rocket axis was immediately railed at over 2200 deg/sec and didn't get back into range until almost apogee. So the lateral axes were reflecting the centripetal acceleration.

 

[Adrian A]

I have rebuilt the sustainer, mostly in the last 2 days, since I was backpacking Sun-Tuesday. I'm still in sprint mode getting ready for Airfest, so I'll withhold most commentary for now, but I thought people would be interested in the pictures:


On the left is my first plan until I realized that if I sanded down to the profile I wanted, the tip would be weak. So I changed to the plan on the right.





After a lot of sanding with 100 grit:




I let the adhesive under the foil leading edge caps get too thick, and I got some bulges that I didn't get on the dry run I did with the 2nd stage. It's a bummer that the first real mistake was near the end but it looks worse than it is. I only really care that the leading edge is intact and the fin is smooth, which it is now after sufficient sanding.

The thickest part of the fin is now 0.100"-0.104", and the thickness profile is a reasonable approximation of the bi-convex shape. The fins are super rigid and strong. I'm pretty confident that I won't get fin fluttering on Saturday, but I'm not sure how much all of the hand shaping is going to throw off the straightness. There will probably be a lot more roll than on the last flight at the top of this thread before the fin fluttering happened.

Tomorrow morning (Friday) I do my ground testing, then clean, prep the rocket again, pack, and drive 8 hours to Kansas for a Saturday morning test flight on a J510. I'll be at Airfest Saturday morning through Sunday afternoon and hopefully I'll be done flying by mid-day on Saturday. I'll be staying next to Sharon and Wayne, who are kind enough to save me a spot.

 

[OzHybrid]

Good luck with the flight. It will be interesting to see the fins afterward. What did you bond the aluminum on the LE with?

 

[bad_idea]

Tomorrow morning (Friday) I do my ground testing, then clean, prep the rocket again, pack, and drive 8 hours to Kansas for a Saturday morning test flight on a J510. I'll be at Airfest Saturday morning through Sunday afternoon and hopefully I'll be done flying by mid-day on Saturday. I'll be staying next to Sharon and Wayne, who are kind enough to save me a spot.

Cool! Hope to have a chance to meet you there.

Edit to add: I'm not flying this year, so if you need a spare hand for anything, let me know. Will PM you my mobile number.

 

[Adrian A]

Good luck with the flight. It will be interesting to see the fins afterward. What did you bond the aluminum on the LE with?

I used more of the Cotronics 4461 epoxy, thickened with West Systems colloidal silica.

 

[kjhambrick]

Very Nice Adrian.

Good Luck and have fun at Airfest !

-- kjh

p.s. I am intrigued by @Brainstormz123's forged composite fins: I500 Submin (iHop)

The Easy Composites, LTD video that he linked was worth watching.

I am thinking about buying a kit to play with ... but first I would need 3D printing capability ...

I wonder if an aluminum LE could be placed in the mold while the carbon tow is compressed when forging a set of fins ?

Anyhow ... no distractions for now, have fun !

 

[pugachu]

Tomorrow morning (Friday) I do my ground testing, then clean, prep the rocket again, pack, and drive 8 hours to Kansas for a Saturday morning test flight on a J510. I'll be at Airfest Saturday morning through Sunday afternoon and hopefully I'll be done flying by mid-day on Saturday. I'll be staying next to Sharon and Wayne, who are kind enough to save me a spot.

So you were able to get you a high altitude waiver for Airfest? How high will she go on the J510?

 

[Adrian A]

So you were able to get you a high altitude waiver for Airfest? How high will she go on the J510?

I was too late to get a high altitude approval, which is why this is a single stage rather than a 2-stage attempt. Gotta remember next year to submit the application earlier. The prediction is about 21k. On the J510 top speed should be around 2900 feet/second (Mach 2.5)

 

[Adrian A]

The flight had a failed recovery that broke 2 fins:



The failure was caused by the wires of the apogee charge shorting out when the primary and the backup altimeters tried to fire the charge. This is clearly shown in the recorded data, which captured the battery currents and voltages:


The purple lines show the current, which spiked over 18 Amps and 16 Amps before the current-limiting feature of the Blue Raven kicked in. The backup Blue Raven was set up to fire the apogee charge if the barometric vertical velocity exceeded 200 feet per second downward.

I'm pretty sure the culprit was the aluminum nosecone ejector piston, which has too much room for the inner piston, so the inner piston can shear the wires where they come out when the nosecone is shoved down as far as it can go:


This can be easily fixed by using a longer screw that attaches the outer cylinder to its base.

I think it's time for me to head back to the rocket pasture to provide more customer support, but there's a lot more flight data and other things to show from this flight that I'll post later.

 

[pugachu]

Too bad about the landing damage. I'm curious as to what altitude and max speed you reached. Hope there is enough time to repair it for the 3-stage flight.

 

[Adrian A]

Too bad about the landing damage. I'm curious as to what altitude and max speed you reached. Hope there is enough time to repair it for the 3-stage flight.

It went to 20,150 and about 2500 feet/second. There is a little bit of disagreement between the speed measured by the accelerometers of the two Blue Ravens that I'm in the process of resolving using the GPS altitude.

This was a very straight flight. Quite a bit of roll rate (from imperfect fin shaping) but very little coning. It almost did a tail slide at 20,000 feet.


With the exception of a May 2-stage attempt that was over-stable and weathercocked into a strong crosswind, all of the flights I have made out of my new 12-foot tower have been nice and straight.

The gyro measurement range was briefly exceeded in roll right around burnout, but you can see the extra pitch rates at apogee when it stalled and started coming back down:



The gyro rates have very good agreement between the two Blue Raven units (pri = solid, dashed = backup), as you can see when zoomed into the rates at apogee:



The accelerometers don't agree as nicely, based on the velocity estimates:



I'll try to resolve the differences tomorrow based on other measurements.




In the meantime here is the first cut at the Cd measured by the 2 units (red and blue for primary and backup Blue Ravens):

This compares pretty nicely with the RASAero prediction. I forgot to weigh the motor after burnout before I cleaned it, so it's based on an estimated mass. Here again this was measured during the coast, so time in this plot starts at the lower right when the motor stops burning, and then goes up and to the left. The Mach 2.5 part down to maybe Mach 1.8 is affected by some residual thrust, but this compares nicely with the RASAero prediction:



I think the low-speed part of my Cd curve will flatten out after I fix up the velocity estimates by estimating better calibration coefficients for the accelerometers.

The drag was almost certainly affected by the fact that most of the Al leading edge wrap was torn off during the flight. I should have known better than to try to bond the thin Al to the CF without careful surface preparation of the Al. But I'm no longer convinced that I can make a thin Al wrap that is more resistant to Mach 4 flight than the CF/epoxy fin without it, even if I bond it perfectly, so I think I'm going to just go for the bare fin going forward.

 

[OzHybrid]

Looking at your finished fins in close up, it looks like you have sanded through the Al leading edge into the CF. This would give the high speed air a place to grab. The idea was to protect the LE. If you've got a hole in it, it's not protected.
Great attempt and a lot less fin damage than before. If records were easy, everyone would have one...

 

[Adrian A]

I decided to

Looking at your finished fins in close up, it looks like you have sanded through the Al leading edge into the CF. This would give the high speed air a place to grab. The idea was to protect the LE. If you've got a hole in it, it's not protected.

I was very careful not to sand through on the leading edge. The place I sanded through was on the trailing edge of the leading edge wrap. I don't think the angles were such that the airflow would be directed under the foil. I think that the foil delamination was more likely a combination of CTE mismatch, poor Aluminum surface prep, lack of post-cure, and high, non-intuitive local shear/peel forces.

The epoxy under where the foil was looks perfect. I'm rebuilding now for a re-flight on the J-1026 hopefully this weekend at NCR if the fire risk is deemed low enough to fly. (low winds, good weather is predicted for Saturday). And maybe also fly my 2nd stage if the J-1016 flight doesn't cause damage and if I can get the 2nd stage finished in time.

Great attempt and a lot less fin damage than before. If records were easy, everyone would have one...

Thanks. I'm happy with the fin stiffness and I think the last layup solved the fin flutter problem, though at the expense of some precision in the fin alignment.

 

[Adrian A]

I had 2 extra fins that I had cut out and trim sanded with the previous attempts, and I had 2 broken fins, so I decided to take a chance and just replace the two broken fins.

I removed those fins and then used my belt sander with 60 grit to clean up the tube again and get rid of most of the woven tip-to-tip layup. Then I used my fin jig with a little compensation for the extra tube thickness to glue on one fin and then the other, with a 150F oven cure to speed things along. The new fins are on now and ready for fillets tomorrow, maybe along with the reinforcement layup.

They seem to be on straight.