J record attempt using Loki J1026

[Buckeye]

Did you determine flutter from the data or from inspecting the fins after the flight? I wonder how flutter would show up in data.

...and is this the same rocket as the flutter analysis in this thread?

https://www.rocketryforum.com/threads/counter-intuitive-aerofinsim-lite-result.181255/

 

[tfish]

@Adrian A

years ago..18+ out at Black Rock a couple of guys Inman and Wamamoto altitude/record guys..Played around with the swept back fins. They had a lot of trouble keeping the fins on (the swept back style) or flight returning with the swept tips broken off. I think of the tips as being unsupported (directly).

Since that time, I've always been leary of trying swept back fins.

A few years ago Wildman came out with the 'Falcon" kit..which has swept fins (my guess is David Reese had some input to the fin design. (?).

In the last few builds I've been slowly been giving them a try.

Including this years Balls project...The picture makes them look more swept then they actually are..




Tony

 

[Adrian A]

Did you determine flutter from the data or from inspecting the fins after the flight? I wonder how flutter would show up in data.

Both. See the acceleration plot upthread.

...and is this the same rocket as the flutter analysis in this thread?

https://www.rocketryforum.com/threads/counter-intuitive-aerofinsim-lite-result.181255/

Yes.

Here is the gyro data for launch until apogee. I'm plotting the gyro data for both Blue Ravens. One has a solid line and the other has a dashed line.



The roll rate (yellow line) was pretty moderate until the fins started fluttering at about 1.8 seconds, and then it railed at 2292 deg/sec. Below is zoomed into the first 2 seconds:


The noisy part at the beginning is the rocket bouncing around inside the 12' tower. Tower exit was at 0.12 seconds. The velocity at that time was about 250 feet/second. There may be a hint of the fluttering to come at 0.6 seconds

Below I zoom in even further to show the first 0.4 seconds:


The gyros in the 2 units have good agreement with each other when the dynamics are slow enough for the 500 Hz sample rate to capture. 500 Hz is pretty fast sampling, but it's not fast enough to do a good job of capturing all the motion of the rocket rattling inside the tower at 120 mph, as you can see from lots of triangular points in the first 0.12 seconds when the two gyros aren't capturing the same data. I'm pretty sure this is why the tilt estimates from the two altimeters disagree by about 6 degrees as soon as the rocket left the tower:


With more normal flight profiles the sampling rate is fast enough to do a much better job of keeping track of the attitude.

The next two plots are the accel and gyro measurements zoomed in to look at the onset of fluttering:




There is a hint of it starting at 0.586 seconds. At that point the rocket was going about 1900 feet/second. I think it could be that the fins were oscillating at that time and there could have been some damage starting to build up in the carbon fiber.

Here is a closeup of the column buckling failure mode of the outer plies of the fin stock:

 

[mikec]

A few years ago Wildman came out with the 'Falcon" kit..which has swept fins (my guess is David Reese had some input to the fin design. (?).

The Falcon fins are not swept nearly as much as Adrian's were.

Don't forget, obviously, that thickness is a big factor also.

 

[kjhambrick]

Adrian --

I am sure you'll post what you have when you have it but, there appear to be some valuable nuggets in with all the tailings ...

Any idea when the column buckling damage occurred ?

The Acceleration -vs- time plot you posted yesterday was very interesting between 1.6 sec and 4.5 sec or so when the 'noise' suddenly leveled out.

I wonder ... there was a noisy high-rate of z-axis deceleration in the 1.6 to 4.5 sec time frame -- could that be when the fins failed ?

What was the corresponding speed -vs- time plot in that same period ?

Wow !

-- kjh

 

[tfish]

The Falcon fins are not swept nearly as much as Adrian's were.

Don't forget, obviously, that thickness is a big factor also.
View attachment 596499

I agree!
at some point the fins shape (in this case rear sweep), fin material/thickness and the rockets speed..don't work.
I think the Falcon is a shape that obvisously works...to a certain point.
and Adrian's combination did not..

I've been slowly working to ....not sure where to.... point where my materials choice, build techniques, my rear swept shape and desired flight profiles all work..

I wish I had seen/read and followed this thread. I would have relayed my above story...instead of the Monday morning quarter back reply ..after the fact!

Tony

 

[Adrian A]

Adrian --

I am sure you'll post what you have when you have it but, there appear to be some valuable nuggets in with all the tailings ...

Any idea when the column buckling damage occurred ?

I suspect that some microscopic damage occurred for the whole high speed portion, but the visible column buckling probably started when you see the big increase in roll rate at 1.8 seconds in.

The Acceleration -vs- time plot you posted yesterday was very interesting between 1.6 sec and 4.5 sec or so when the 'noise' suddenly leveled out.

I wonder ... there was a noisy high-rate of z-axis deceleration in the 1.6 to 4.5 sec time frame -- could that be when the fins failed ?

Almost certainly. After 4.5 seconds, the speed was low enough that the fins stopped fluttering despite the damage

 

[ECayemberg]

Cool attempt!

You had mentioned that the fins were mostly .060 uni. Are you saying one single uniform fiber direction? Or multiple layers of uni fabric layed up in several fiber orientations?

I have a bit of experience in this arena... I would recommend using unidirectional if there are various intentional fiber directions; I would not recommend using uni if multiple layers have the same uniform fiber direction.

 

[Rschub]

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:

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.

 

[kjhambrick]

I suspect that some microscopeic damage occurred for the whole high speed portion, but the visible column buckling probably started when you see the big increase in roll rate at 1.8 seconds in.


Almost certainly. After 4.5 seconds, the speed was low enough that the fins stopped fluttering despite the damage

I wonder if the velocity around 4.5 sec has any correlation to the prediction that AeroFinSim made in your other thread: Counter-intuitive AeroFinSim Lite result?

 

[watheyak]

This is a great thread! Lots of great info from great people.

I guess the collective consciousness of altitude junkies have all been not-so independently moving toward swept fins, me included. I have also been thinking quite a bit about flutter, fin thickness and fiber orientation. Seeing the creases on Adrian's fins is a great and educational visual- you can imagine exactly where and how the flutter happened and perhaps be better informed on a fix.

Here's two projects I'm working on with swept fins. They are both works in progress. The first is my Balls project for this year that will fly on a long burn N motor. I'm not super concerned about flutter on this one. They're thick, have a super dense core and many layers of high modulus unidirectional CF recessed in the fin. The outer layer of bi-directional 45 degree bias was originally just for looks, but after seeing Adrian's creases, maybe the bias is a good thing. All the epoxy and cores are high temp as well.



The second one is a 54mm K altitude attempt rocket I've had on the back burner for a few years. I've noticed that the projects that scare me a little bit get back-burnered for a few years as I mull the possibilities. My L record rocket was that way. First off, the fin shape is completely bonkers. They look like they're ready to flutter just sitting there, and at only .067" thick, it's not a good start.

But! They are high modulus Dragon Plate. And! They are going to get a tip to tip layer of some high modulus, albeit very thin, spread tow CF on a 45 degree bias. I'm still pretty skeptical. On top of that, we could probably have a whole discussion on high vs low modulus CF when it comes to flutter resistance. In the end, this one may just end up being desk art.

 

[mikec]

The second one is a 54mm K altitude attempt rocket I've had on the back burner for a few years. I've noticed that the projects that scare me a little bit get back-burnered for a few years as I mull the possibilities. My L record rocket was that way. First off, the fin shape is completely bonkers. They look like they're ready to flutter just sitting there, and at only .67" thick, it's not a good start.

I assume you meant 0.067?

FWIW, I've made many attempts to fly rockets faster than 3000 FPS (see a writeup in SPORT ROCKETRY from a few years back.) The only time I've lost fins was with a 38mm rocket with clipped delta fins made of high-quality commercial CF that were 0.062 thick with tip-to-tip fiberglass. So I'd be concerned about these purely from a thickness perspective, let alone the sweep.

 

[wsume99]

Trying to learn as much as I can from all the information being shared in this thread, lots of valuable info here.

One question, why use swept fins? Is this to move Cp as far aft as possible to minimize the total weight required to maintain stability?

 

[Adrian A]

Cool attempt!

You had mentioned that the fins were mostly .060 uni. Are you saying one single uniform fiber direction? Or multiple layers of uni fabric layed up in several fiber orientations?

I have a bit of experience in this arena... I would recommend using unidirectional if there are various intentional fiber directions; I would not recommend using uni if multiple layers have the same uniform fiber direction.

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.

The plate that I used for the fin stock had 0/90/0 layup. When you bend the plate, it's very obviously stiffer in the 0 direction than the 90 direction, as expected. I aligned the outer ply fiber direction with the trailing edge so that I would maximize strength and bending stiffness in the direction of the worst cantilevered leverage. I was concerned most about breakage from bending. In the photo you can see the the crease is oriented such that as some outer ply fibers started to fail, they overloaded their neighbors and they failed too. The inner ply's fibers weren't helping at all in that direction, so that's one of the main reasons why you see the creases running in that direction.

The best fiber orientation for torsional rigidity for a non-swept fin would be to have the fibers on the outer layers be +/- 45 degrees from the direction of flight. When the fins are highly swept it's not so clear to me whether the +/-45 degree angle also should be adjusted. My trailing edge angle is already pretty close to 45 degrees from the airframe, so it might be that I just needed more thickness in that outer ply direction. But if the ideal fiber direction is +/- 45 degrees from the average chord, then my fiber orientations were less ideal.

 

[Adrian A]

Trying to learn as much as I can from all the information being shared in this thread, lots of valuable info here.

One question, why use swept fins? Is this to move Cp as far aft as possible to minimize the total weight required to maintain stability?

Usually, clipped delta fins work just as well or better than swept fins for efficiently providing stability, because usually, the driving case for stability margin is dealing with a cross-wind coming out of the tower, when the rocket is at its lowest speed. The Cp moves back (gets better) as the velocity goes up through the transonic region.

But with increasing speed, eventually the Cp starts moving forward again. At ridiculously fast speed, according to RASAero, the high speed stability can become the worst case for stability margin. When I play with different fin geometries in RASAero, swept fins do a lot better in the Mach 3-4 range than delta or clipped delta fins. This is especially beneficial on my sustainer because the "better" the fin is at anchoring the Cp of the rocket at low speed as a single stage, the worse it is for the 3-stage configuration where the sustainer's fins are far in front of the CG. So swept fins give the high speed stability needed without moving the Cp as far forward at lower speeds when it's in the 2 or 3 stage stack.

Here's a little study I did a few weeks ago in RASAero on the effect of fin sweep. The key figures:

 

[watheyak]

I assume you meant 0.067?

FWIW, I've made many attempts to fly rockets faster than 3000 FPS (see a writeup in SPORT ROCKETRY from a few years back.) The only time I've lost fins was with a 38mm rocket with clipped delta fins made of high-quality commercial CF that were 0.062 thick with tip-to-tip fiberglass. So I'd be concerned about these purely from a thickness perspective, let alone the sweep.

Ah yeah .067. I corrected it in the original post.

Oh but this one only goes a pedestrian 2300fps But seriously, I'll stop trying to justify the insanity....

I'll look for the write up. I may even have it.

 

[manixFan]

As others have mentioned, this has been a fantastic thread. It also demonstrates how going 3000+fps in itself is a great design challenge, but it’s even more so when the vehicle is the sustainer of a three stage speed demon – it becomes a very technical balancing act between what’s good for the sustainer individually vs. the stacked rocket.

And I love seeing all the data from the Ravens, it really adds another dimension to the flight, especially when it’s so artfully interpreted. Add in all the experience shared by others and this thread is a mini-course on high speed design and survivability.


Tony

 

[0011001100]

Both. See the acceleration plot upthread.

Yes.

Here is the gyro data for launch until apogee. I'm plotting the gyro data for both Blue Ravens. One has a solid line and the other has a dashed line.

View attachment 596480

The roll rate (yellow line) was pretty moderate until the fins started fluttering at about 1.8 seconds, and then it railed at 2292 deg/sec. Below is zoomed into the first 2 seconds:

View attachment 596479
The noisy part at the beginning is the rocket bouncing around inside the 12' tower. Tower exit was at 0.12 seconds. The velocity at that time was about 250 feet/second. There may be a hint of the fluttering to come at 0.6 seconds

Below I zoom in even further to show the first 0.4 seconds:

View attachment 596482
The gyros in the 2 units have good agreement with each other when the dynamics are slow enough for the 500 Hz sample rate to capture. 500 Hz is pretty fast sampling, but it's not fast enough to do a good job of capturing all the motion of the rocket rattling inside the tower at 120 mph, as you can see from lots of triangular points in the first 0.12 seconds when the two gyros aren't capturing the same data. I'm pretty sure this is why the tilt estimates from the two altimeters disagree by about 6 degrees as soon as the rocket left the tower:
View attachment 596483

With more normal flight profiles the sampling rate is fast enough to do a much better job of keeping track of the attitude.

The next two plots are the accel and gyro measurements zoomed in to look at the onset of fluttering:

View attachment 596486
View attachment 596487

There is a hint of it starting at 0.586 seconds. At that point the rocket was going about 1900 feet/second. I think it could be that the fins were oscillating at that time and there could have been some damage starting to build up in the carbon fiber.

Here is a closeup of the column buckling failure mode of the outer plies of the fin stock:
View attachment 596488

Seeing this makes me take back the aeroheating concern. I agree with others about the fiber orientation. For your case 0/45/90/45/0 might be a better layup with the 90 being perpendicular to the body tube. I don't know if this can be achieved and still keep it as thin as you would like, but the extra thickness would help too.

 

[wsume99]

Usually, clipped delta fins work just as well or better than swept fins for efficiently providing stability, because usually, the driving case for stability margin is dealing with a cross-wind coming out of the tower, when the rocket is at its lowest speed. The Cp moves back (gets better) as the velocity goes up through the transonic region.

But with increasing speed, eventually the Cp starts moving forward again. At ridiculously fast speed, according to RASAero, the high speed stability can become the worst case for stability margin. When I play with different fin geometries in RASAero, swept fins do a lot better in the Mach 3-4 range than delta or clipped delta fins. This is especially beneficial on my sustainer because the "better" the fin is at anchoring the Cp of the rocket at low speed as a single stage, the worse it is for the 3-stage configuration where the sustainer's fins are far in front of the CG. So swept fins give the high speed stability needed without moving the Cp as far forward at lower speeds when it's in the 2 or 3 stage stack.

Thanks for taking the time to shed some light on your design considerations and the analysis you've done to end up with the design you have. The reduced drag from swept fins is another bonus from that geometry when going for a record attempt.

 

[Adrian A]

Seeing this makes me take back the aeroheating concern. I agree with others about the fiber orientation. For your case 0/45/90/45/0 might be a better layup with the 90 being perpendicular to the body tube. I don't know if this can be achieved and still keep it as thin as you would like, but the extra thickness would help too.

After thinking about this some more, I have three conclusions

1. The fin layup was strong enough against bending that when impacting at 60 feet/second with a pretty heavy case, the fins were strong enough to rip apart good fillets and even delaminated part of the body tube without any other damage on the fins themselves. No tip breakage, etc. This tells me that although I will need to add thickness to resist flutter, the current fin stock orientation and thickness already covers the needed the bending strength and trailing tip reliability

2. The flutter broke fibers running parallel to the trailing edge but didn’t bother the fibers going the other direction. The damage to the fins are telling me what fiber direction needed more strength/stiffness, and rather than over-think this I should just beef up the fiber direction that tried and failed to resist the fluttering.

3. Both of the fins that broke off in impact had the fracture go through the fillet, and at least one of the fins took a chunk of the body tube with it. This tells me that the bonding prep and initial gluing of the fins was good, the fillet adhesion was good, and if I want to make the fin attachment stronger I should do a unidirectional tip-to-tip layup over the top of the fillet like I have on previous rockets.

 

[Adrian A]

As others have mentioned, this has been a fantastic thread. It also demonstrates how going 3000+fps in itself is a great design challenge, but it’s even more so when the vehicle is the sustainer of a three stage speed demon – it becomes a very technical balancing act between what’s good for the sustainer individually vs. the stacked rocket.

And I love seeing all the data from the Ravens, it really adds another dimension to the flight, especially when it’s so artfully interpreted. Add in all the experience shared by others and this thread is a mini-course on high speed design and survivability.


Tony

Thanks for taking the time to shed some light on your design considerations and the analysis you've done to end up with the design you have. The reduced drag from swept fins is another bonus from that geometry when going for a record attempt.

The valuable experiences shared by others, and the positive feedback, is why I enjoy posting here.

 

[pugachu]

Adrian,

Do you think you'll be able to repair this in time for your Stratospear attempt next month?

 

[Adrian A]

Adrian,

Do you think you'll be able to repair this in time for your Stratospear attempt next month?

Yes. I only started the aft section about a week ago, including making my fin alignment jig, and I also had a lot to do on the nosecone and av-bay in parallel. I need to balance rocket building with Featherweight Electronics development work more evenly than I have over the last week, but there should still be plenty of time for a 2-stage attempt at Airfest, let alone BALLS.

The next step is to sand off the remaining fillets so I can see if I want to re-use the tube. If I do I'll swap ends so that the fins can be attached to the undamaged end. Since the front end of the rocket is attached to the front end of the motor case, the forward end of the tube is just there for aerodynamic continuity. Also, I need to make this tube shorter anyway to expose 2"-3" of motor in the back when I use it as a sustainer.

 

[Adrian A]

Yes. I only started the aft section about a week ago, including making my fin alignment jig, and I also had a lot to do on the nosecone and av-bay in parallel. I need to balance rocket building with Featherweight Electronics development work more evenly than I have over the last week, but there should still be plenty of time for a 2-stage attempt at Airfest, let alone BALLS.

The next step is to sand off the remaining fillets so I can see if I want to re-use the tube. If I do I'll swap ends so that the fins can be attached to the undamaged end. Since the front end of the rocket is attached to the front end of the motor case, the forward end of the tube is just there for aerodynamic continuity. Also, I need to make this tube shorter anyway to expose 2"-3" of motor in the back when I use it as a sustainer.

Well, it turns out I was too late to apply for a high-altitude flight for Airfest.

 

[mrwalsh85]

I may have missed it in the conversation somewhere, but does the location of the swept fin have direct impact on potential flutter etc, with respect to the end of the motor/airframe?

For instance, if any component of the trailing edge of the fin is below the end of the motor, could there be impact there?

Sorry for the thread drift, Adrian.

-mike

 

[Adrian A]

I may have missed it in the conversation somewhere, but does the location of the swept fin have direct impact on potential flutter etc, with respect to the end of the motor/airframe?

For instance, if any component of the trailing edge of the fin is below the end of the motor, could there be impact there?

Sorry for the thread drift, Adrian.

-mike

Not as far as I know, but I would be surprised if it did. Of course, it does affect the landing impact tolerance quite a bit.

 

[kjhambrick]

<<snip>> ... Something I read years ago is that fins where the thickness tapers from root to tip can defeat resonance ...<<snip>>

Adrian --

This is not the paper I recalled reading long ago ( still googling as fast as I can ) but it is short and interesting and there may be some practical take-aways.

Info Central: Fin Flutter by Duncan McDonald

Duncan's paper discusses root-to-tip chord taper and pros and cons of different aspect ratios but the paper I am trying to find discussed variable fin thickness from root-to-top.

I've not compared the equations to the equations referenced here in this thread but there is a working link to an Excel Worksheet as well that I've not had a chance to mess with.

Anyhow ... HTH ...

-- kjh

 

[Adrian A]

Adrian --

This is not the paper I recalled reading long ago ( still googling as fast as I can ) but it is short and interesting and there may be some practical take-aways.

Info Central: Fin Flutter by Duncan McDonald

Duncan's paper discusses root-to-tip chord taper and pros and cons of different aspect ratios but the paper I am trying to find discussed variable fin thickness from root-to-top.

I've not compared the equations to the equations referenced here in this thread but there is a working link to an Excel Worksheet as well that I've not had a chance to mess with.

Anyhow ... HTH ...

-- kjh

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.

 

[kjhambrick]

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?

 

[robopup]

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