I could use just a little guidance

The Rocketry Forum

Help Support The Rocketry Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Access for the flight computer is important. Must have been a real PITA before. For next-level access, if you can, add Bluetooth connectivity so you don't need to use a cable. That makes things so much easier on my system. It also means less chance of breaking off the fragile USB connector on my FC. Was a bit of a nuisance to set up, but well worth it IMHO.
Yes, the system I have been using requires taking off the canards, removing the system from the air frame, and then recentering the canards to reprogram the flight electronics. The new design will avoid those issues. I'll mention bluetooth to the brain trust.

Jim
 
Sir:

I am very interested in a "flies straight up" capability and would be happy to beta test your system.

However, I foresee two issues: my smallest rocket is 6" diameter and my previous analysis has concluded that for my particular requirements a warm gas system is preferred over canards.

If it is of interest, I'd be happy to try and use your electronics to drive a N2-based thruster system designed to fit my six inch vehicle. My next flight is targeted for mid-October and I am finalizing the payload design for that vehicle over the next few weeks.

Bill C.
Maybe tell me what you need and what info you need.

Jim
 
Maybe tell me what you need and what info you need.

Jim
JIm:

In principal, the existing pulse width could be used to drive a fast acting solenoid valve and the problem would then be to size the force from the gas jet to match the force generated by the canard from the same pulse width. The immediately obvious issues would be what range of pulse widths' the system outputs and whether a reasonably priced solenoid valve can be found that handles that range of pulse widths--practically speaking, I think that because a question of what minimum pulse width the valve has to handle. Sizing the gas system to generate a similar force to the canard at that pulse width implies a constant pressure supply (blowdown will not work); my previous peroxide rocket used a 6000 psia burst pressure commercial tank running at 4500 psia, I suspect a similar but shorter tank would work fine with a regulator to provide a steady, say, 500 psia for the about three minutes of control that will be eventually required.

I'd need to understand the minimum pulse width that the valve(s) need to handle based on the existing system and would need a map of how pulse width translates to force (I realize that changes with altitude). I'd also need to know the voltage at which the PWM operates.

I'm working on a two stage project (R to P in HPR world) that is simulating to around three minutes to peak; I'd like the guidance system to maintain orientation throughout ascent. (The six inch upper stage flying alone is about 45 seconds to peak but I would size for the longer mission.)

Bill
 
JIm:

In principal, the existing pulse width could be used to drive a fast acting solenoid valve and the problem would then be to size the force from the gas jet to match the force generated by the canard from the same pulse width. The immediately obvious issues would be what range of pulse widths' the system outputs and whether a reasonably priced solenoid valve can be found that handles that range of pulse widths--practically speaking, I think that because a question of what minimum pulse width the valve has to handle. Sizing the gas system to generate a similar force to the canard at that pulse width implies a constant pressure supply (blowdown will not work); my previous peroxide rocket used a 6000 psia burst pressure commercial tank running at 4500 psia, I suspect a similar but shorter tank would work fine with a regulator to provide a steady, say, 500 psia for the about three minutes of control that will be eventually required.

I'd need to understand the minimum pulse width that the valve(s) need to handle based on the existing system and would need a map of how pulse width translates to force (I realize that changes with altitude). I'd also need to know the voltage at which the PWM operates.

I'm working on a two stage project (R to P in HPR world) that is simulating to around three minutes to peak; I'd like the guidance system to maintain orientation throughout ascent. (The six inch upper stage flying alone is about 45 seconds to peak but I would size for the longer mission.)

Bill

Bill, when I'm talking about a pulse width, it is in the context of a standard RC servo, with pulse widths ranging from 1000 to 2000 usec centered at 1500 usec. I don't mean the length of a pulse of a gas jet for example. I'm not familiar with how gas jets are intended to operate, so I just want to verify we're on the same page there. I also can't answer the question about the voltage that PWM operates at.

The current system, if limited to yaw/pitch only and not roll, generates a PWM signal from 1500 +/- 250 usec (half the output range) for x and y tilts from 0 to 7.5 degrees. These gains and values can be changed to any desired values.

Calculating the force generated by the canards I have used is relatively easy. However, I'm generally trying to move rockets around (i.e., intentional and rapid changes in tilt and bearing). For maintaining vertical orientation, the forces required can be much lower.

I think a three-minute control period might be a tall order. The control board that we use is inexpensive and has inexpensive gyros. We are upgrading, but the system would still be defined as inexpensive. I have always used flight profiles that limited the time after launch that the system was operating. For this reason, I typically "guide" a sustainer from below during a coast period and then have it drop off upon sustainer ignition - a flight period of around 30 seconds maximum. It's possible that reasonable drift would be maintained for that period, and that can be tested, but I'm not sure how to verify that for a long flight.

Jim
 
Jim:

I should have recognized that you would be using the 1 to 2 ms pulse structure of an RC servo; in theory, I could build a logic board that translated that pulse to an "On" duration for a cold gas RCS but that moves toward adding too much to my to do list.

If you have not seen it, there is good top level discussion of an RCS system for a six inch rocket here: https://github.com/psas/reaction-control/blob/master/pubs/AIAA RCS Manuscript_FINAL2.pdf

I now see why you are not guiding for the second stage burn; while I fully understand your trade, it is not something I am willing to risk: the dispersion's during second stage burn (around 45K feet to 75K feet) are potentially significant to the splash pattern (factor of 2 or more) and I am looking to keep the impact zone as small as possible in order to try and avoid a national range.

Bill
 
So, work continues on the construction of the control system and spin can. I've made some progress on the spin can part of this. It's been slow going just due to the amount of labor involved in cutting & squaring the tubes - and there are a lot of tubes involved. The pic shows the tubes required for 4 spin cans. It also doesn't help that the "red" fiberglass tubing is not particularly dimentionally consistent. The video shows that there is still a little "sticking" on one of the tubes. Unfortunately, I had some shoulder work done a few weeks ago and it will be a few more weeks before I can consider trying to sand anything. So, I won't be able to finish these up until then.

Jim



SQBR2851.JPG
 

Attachments

  • Spin Can 2.mov
    2 MB
So, work continues on the construction of the control system and spin can. I've made some progress on the spin can part of this. It's been slow going just due to the amount of labor involved in cutting & squaring the tubes - and there are a lot of tubes involved. The pic shows the tubes required for 4 spin cans. It also doesn't help that the "red" fiberglass tubing is not particularly dimentionally consistent. The video shows that there is still a little "sticking" on one of the tubes. Unfortunately, I had some shoulder work done a few weeks ago and it will be a few more weeks before I can consider trying to sand anything. So, I won't be able to finish these up until then.

Jim



View attachment 507610
Jim,
I can stop by and do the sanding if you'd like. Or, if you bring them to the launch tomorrow, I can sand them there. I have a 2'x2' square piece of stiff plywood that I can use for flat sanding, as well as an aluminum sanding T etc. We can talk more about it offline if ur interested.
 
Jim,
I can stop by and do the sanding if you'd like. Or, if you bring them to the launch tomorrow, I can sand them there. I have a 2'x2' square piece of stiff plywood that I can use for flat sanding, as well as an aluminum sanding T etc. We can talk more about it offline if ur interested.
Thanks Josh. It isn't all that urgent to get these sanded down. 3 weeks from now will be fine, but if I still can't sand at that point, I may take you up on your offer. I really just wanted to make sure at this point that the spin cans would work. They are using smaller bearings than the previous versions, and I just wanted to make sure that wouldn't be problem.

Jim
 
There has been just a bit of confusion over how the beta testing and the fin can will work. I'm hoping to have at least 6 beta testers, although this may need to be expanded a little based on interest to date. Each will receive a control section (with the electronics and canards) and a spin can if they want one. I think the spin can will be very helpful, but beta testing without a spin can is valuable too. Anyway, I have four spin cans produced at this time on the assumption that not everyone will want one. More will come, but they take time to make.

For those that want to use the spin can, they will need to couple the spin can with some length of air frame to complete the fin can section of the rocket. Such a tube is shown in blue in the picture. There is a 4" section of coupler tubing at the top of the spin can for this purpose. Then, the user will need to mount the fins of their choice, plus any reinforcement, onto the carbon fin tube (and then reassemble the spin can). I should also say that the spin can will be capable of holding a 75mm motor (but not 98mm). The user will also need to construct a motor mount of their choice. My recommendation would be a thrust plate and a couple of centering rings - but not a motor tube - held in by forward retention. But, motor retainers could be used as well.

Jim

Beta spin can.png
 
People have also asked how the spin can works. I have posted the design and done videos on this in the past. However, a diagram of the current version is attached. I've tried to show to which tube (the smaller coupler tube or the oversize carbon tube) the middle tubes are attached to. The location of the bearings is also shown. It's a simple design, but certain tolerances are important.

Jim

Spin Design.png
 
So, now the fun starts.

As I mentioned at the start of this "beta" topic, one importnt objective is to develop a vertical orientation sytem design that can be 3D printed. For the past 6 weeks or so, Bryan Mashall (Bryrocket) and I have been working to transform my PowerPoint-designed, plywood concept into a Fusion 360, carbon-infused, functional orientation system. We have had so much fun, and words really can't convey how pleased I am with how this is turning out. It turns something that takes months to make into something that can be produced on the order of hours. That opens up access to the technology for others, and I want to thank Bryan for making that possible (and yes, he's a beta tester!). We have just a few final tweeks left to come up with flyable system.

On a parallel path, Bill Premerlani and Frank Hermes have collaborated to produce a control board that offers improvements in performance relative to the board I have used up to this point. And, I think Frank has some ideas for an even better board down the road (once we can actually get some parts for it).

Even though this stuff is coming together, it will still be a month or two before I can get equipment out to the beta testers. My shoulder surgery is really slowing me down (can't sand, can't solder, can't work a screwdriver) and I still have quite a bit of information to assemble before this is ready. But, we're going to get there! Attached are a few pics of a recent prototype.

Jim

66804116908__DD549D1F-E258-4647-9274-26C59380A23B.jpeg66804117780__B6F179A9-5D42-418F-B907-8B1BE41F27A9.jpeg
 
So, now the fun starts.

As I mentioned at the start of this "beta" topic, one importnt objective is to develop a vertical orientation sytem design that can be 3D printed. For the past 6 weeks or so, Bryan Mashall (Bryrocket) and I have been working to transform my PowerPoint-designed, plywood concept into a Fusion 360, carbon-infused, functional orientation system. We have had so much fun, and words really can't convey how pleased I am with how this is turning out. It turns something that takes months to make into something that can be produced on the order of hours. That opens up access to the technology for others, and I want to thank Bryan for making that possible (and yes, he's a beta tester!). We have just a few final tweeks left to come up with flyable system.

On a parallel path, Bill Premerlani and Frank Hermes have collaborated to produce a control board that offers improvements in performance relative to the board I have used up to this point. And, I think Frank has some ideas for an even better board down the road (once we can actually get some parts for it).

Even though this stuff is coming together, it will still be a month or two before I can get equipment out to the beta testers. My shoulder surgery is really slowing me down (can't sand, can't solder, can't work a screwdriver) and I still have quite a bit of information to assemble before this is ready. But, we're going to get there! Attached are a few pics of a recent prototype.

Jim

View attachment 508774View attachment 508775
The print looks slick. People could even use a service like sendcutsend etc if they needed.
 
I wired up the current prototype and did a full-function test including configuration of the servos. What is nice about the setup is that each servo can be configured by simply selecting the servo with a switch. The process is much simpler than my current system. Here's a short video of the first powered test.

Jim

 
Sir:

I am very interested in a "flies straight up" capability and would be happy to beta test your system.

However, I foresee two issues: my smallest rocket is 6" diameter and my previous analysis has concluded that for my particular requirements a warm gas system is preferred over canards.

If it is of interest, I'd be happy to try and use your electronics to drive a N2-based thruster system designed to fit my six inch vehicle. My next flight is targeted for mid-October and I am finalizing the payload design for that vehicle over the next few weeks.

Bill C.
If you haven't read it, there's a good book that has a lot of the math worked out. The author goes over a canard design as well as a cold gas thruster design, I know you mention "warm" gas but there may still be some useful information in the book for your case. The technology used is pretty dated but the theory and math the author goes over is super useful.
https://www.amazon.co.uk/Vertical-Trajectory-Systems-Rocketry-Educational-ebook/dp/B07KKN814D
 
Going off on a brief tangent for a moment: How effective would this "spin can" be (on a rocket without any active guidance) at reducing roll, say for use with a camera? Since most of the roll comes from imperfect fins, and most of the moment of inertia comes from the rest of the rocket, just mounting a spinning fin can might reduce roll to a tolerable level?
 
How effective would this "spin can" be (on a rocket without any active guidance) at reducing roll, say for use with a camera?

Should work well as long as the other irregularities in the airstream [cameras, buttons, etc.] don't induce a spin.
This was the genesis of the invention - avoid spinning a camera platform.
 
Last edited:
So, progress has been made over the past few weeks. We have 6 eager beta testers, who are busy designing and/or modifying rockets. I'm having trouble keeping up actually. Yesterday, we completed the first flight-ready version of the control system. Woohoo! I suspect I'll fly this one for a test launch sometime in April, and assuming no issues, provide equipment to the testers shortly thereafter. I continue to be very happy with the way this came out. I'm sure there will be future improvements, but the approach we've taken reduces the production time from a matter of months to a matter of hours (except for the spin can - that takes a while longer!). I'm expecting this to be a fun summer as everyone does their flights and posts their results.

Jim
 
I was able to perform a test of the revised Vertical Orientation System, Version 2 (VOS2) at Hearne over the weekend. The purpose of the test was just to verify that the revised design works as advertised. This design has a modified support approach for the servos, and the servos are smaller than what I've used in the past. There are many other changes, so the idea was to conduct a test flight before releasing systems to the beta test group.

The system can operate in three distinct flight modes. Mode 1 is roll control only, which is intended to provide a stable platform (I hope!) for video. We haven't tested this mode yet, but I'm confident it will work well. Mode 2 is vertical flight and Mode 3 is near-vertical flight, where the rocket flies at a 12° tilt. Modes 2 and 3 can be used in combination with an upper-level wind forecast to provide a more predictable landing location (and hopefully, away from hazards and/or closer to the pad).

The demonstration flight was performed using Mode 3, with a 12° tilt. One objective of testing this mode was to show that the system will maintain a tilt near the 12° set point until just prior to apogee. This gives a "soft" deployment with a modest horizontal velocity because the rocket doesn't arc over as it approaches apogee. Overall, the demonstration flight was very successful. Here's a video documenting the flight.

Jim

 
Nice flight Jim! Working quite well, except for the gain being a little too high early on and giving oscillations. They settled as the airspeed dropped and control authority of the canards reduced, reducing the overall control gain function. I have seem similar on my system and have reduced the gain by using smaller fins. Yet to retry the system on reduced gains.
 
Nice flight Jim! Working quite well, except for the gain being a little too high early on and giving oscillations. They settled as the airspeed dropped and control authority of the canards reduced, reducing the overall control gain function. I have seem similar on my system and have reduced the gain by using smaller fins. Yet to retry the system on reduced gains.
There do appear to be some oscillations, but on the other hand, the rocket is flying at a 12-16° angle of attack in a strong crosswind. Overall, the gains were on the high side and the canards were my largest. I wanted a relatively severe test for the mechanical system, which is one of the things that is different in this design. I'm pretty happy with how it worked.

Jim
 
Well, I haven't posted anything since the test flight, but I've been busy completing the units for the six beta testers. The six control units and spin cans are complete, except for the last unit that we'll assemble as a group. After that, it's up to the group to complete their rocket designs and then report back on design suggestions, documentation, etc., and of course, their flight reports. Should be a lot of fun!

Jim

IMG_2912.JPG
 
So, I was looking a little closer at the data from the VOS2 test at Hearne. The flight plan was to fly at a 12° tilt to the west as shown by the yellow arrow in the figure below. The predicted apogee is to the north due to drift on ascent. However, the actual apogee is even further to the north. When I looked at the data, I found that the actual rocket bearing was not zero (oriented towards the west), but rather, about 25° on the average clockwise. So, that would explain why the actual apogee was further north than expected.

As I was considering this, it occurred to me that a rocket trying to fly to the west in a 25 mph cross wind from the south is going to see a signifcant angle of attack through most of the flight. Roughly, if the average velocity is 400 ft/s, and the cross wind is 37 ft/s, the angle of attack would be 5.3° (being much higher at launch and near apogee, and about half of that value at maxiumum speed). Most of my flights have been into or with the wind, and this is the first flight with a significan cross wind. What I'm wondering is if the canards "in the shadow" are less effective such that the rocket is aerodynamically unable to hit the required bearing? I also wonder if this unusual flight condition is what led to some of the oscillations noted in prior posts.

Just more interesting data to ponder.

Jim

Hearne path pic.png
 
Shadowing of the canards in a crosswind would result in less system gain and reduce the chance of oscillations. I don't think you can correlate those two as a cause of oscillations. I am happy to be corrected by another theory if it is applicable.
 
So I have more data. The graph shows the "Vert" values for the flight. The VOS2 system is designed to keep the canards in an "X" shape relative to the ground rather than a "+" shape. So, the vert x and vert y values follow the "X" shape. If the rocket is oscillating in tilt, vert x and vert y will move together (going positive and negative together). This is what the rocket did for the first 1.5 seconds. From about 1.5 to 2.5 seconds, however, vert x and vert y are moving in opposite directions, and there is relatively little change in tilt. Here, the rocket is oscillating from side to side, but at the same tilt. This is the oscillation that you see in the video (the side to side movement is easy to see from the vantage point of the camera). From 2.5 to 5 seconds, vert x is changing but vert y is not. Here, the rocket is oscillating at a 45° angle (along one of the legs of the "X"). I have never seen these kinds of movement in previous flights. Not proof for sure, but I think the chances are good that this is the effect of the cross wind.

Jim

Edit - For clarification, the "vert x" and "vert y" values are the tilts from vertical in the x and y directions. The net tilt is the composite tilt from vertical.

Vert values.png
 
Last edited:
A rocket or plane moving in a body of air doesn't know or care what the wind is doing on the ground. Crosswind is a ground concept. You may have motion in the air, and the rocket or plane would move relative to that body of air. Not relative to the ground. Entering layers with different velocity vectors will cause brief periods of non-axial flight, but given positive stability, planes or rockets will reorient to align with the relative freestream direction.

If a vehicle resist this realignment via active controls, it will potentially maintain pointing vector, and experience lateral acceleration until the vehicle is again aligned with the local freestream.

The vehicle angle of attack for a rocket moving at flight speeds (not right off the tower and not near apogee) is unlikely to be significant short of coning. Statistically that is... such situations would not be common. Our rockets are flown in calm clear weather and not in the wake of airlines. Even the boundary of a thermal will only see wind shear of a few m/s. I fail to see how what is being called shadowing of the canards can occur more than briefly and mostly at low speeds and high wind shear, outside of coning.

Note, shadowing as it is being called can occur on air-to-air or ground-to-air missiles which are deliberately making extreme maneuvers to hit a target. The canards are located rather close to the nose of the vehicle to improve the ability to maneuver, reduce the total drag for executing a given maneuver, and to reduce shadowing.

Gerald
 
Last September, I tried the "Python" two-stage flight at Airfest. It didn't go well, with the motor finally coming up to pressure just after all of the chutes blew. It was not a good flight, but I repaired everything and finally got a chance to repeat the flight over the weekend. It went pretty well, except that part of the flight was not visible due to a cloud, and I also didn't get all of the planned video. Still trying to figure out what happened there. Anyway, copied below are two YouTube video links. One is of the entire flight and the other is a zoomed-in version where it's a bit easier to see what the rocket was doing.

I posted some data in the first video showing how the rocket generally followed the flight plan with respect to tilt and bearing. Not perfect, but not too bad.

One thing you can see in the videos is that the rocket motion wasn't all that smooth, particularly when the rocket was turning upward against gravity. I think part of the reason for that is that I allowed more time for those turns due to uncertainty in the turning rate when turning upward. I think I allowed too much time. Since each turn is a discrete step, having too long of a step makes the turn look less "continuous". The other thing that I noticed in the data was that the canards never reached the limit of their travel range. I have them set up to turn proportionally up to 7.5° for angle errors up to 7.5°. The steps in the flight program were at 4° increments, and the canard angles never exceeded 4°. That means that the rocket never "fell behind" the flight program. I was concerned that it might fall behind in the upward turns, but that didn't happen (and in fact, the opposite happened).

I managed to break a fin on the spin can again - crap. I suspect I will fix it at some point, but it may be a while before that happens. In the meantime, there should be some beta group flights coming up over the next few months!

Jim



 
Last September, I tried the "Python" two-stage flight at Airfest. It didn't go well, with the motor finally coming up to pressure just after all of the chutes blew. It was not a good flight, but I repaired everything and finally got a chance to repeat the flight over the weekend. It went pretty well, except that part of the flight was not visible due to a cloud, and I also didn't get all of the planned video. Still trying to figure out what happened there. Anyway, copied below are two YouTube video links. One is of the entire flight and the other is a zoomed-in version where it's a bit easier to see what the rocket was doing.

I posted some data in the first video showing how the rocket generally followed the flight plan with respect to tilt and bearing. Not perfect, but not too bad.

One thing you can see in the videos is that the rocket motion wasn't all that smooth, particularly when the rocket was turning upward against gravity. I think part of the reason for that is that I allowed more time for those turns due to uncertainty in the turning rate when turning upward. I think I allowed too much time. Since each turn is a discrete step, having too long of a step makes the turn look less "continuous". The other thing that I noticed in the data was that the canards never reached the limit of their travel range. I have them set up to turn proportionally up to 7.5° for angle errors up to 7.5°. The steps in the flight program were at 4° increments, and the canard angles never exceeded 4°. That means that the rocket never "fell behind" the flight program. I was concerned that it might fall behind in the upward turns, but that didn't happen (and in fact, the opposite happened).

I managed to break a fin on the spin can again - crap. I suspect I will fix it at some point, but it may be a while before that happens. In the meantime, there should be some beta group flights coming up over the next few months!

Jim




Still an amazing shot of the very obvious "S" curve made by the smoke trail. Fantastic.
 
Still an amazing shot of the very obvious "S" curve made by the smoke trail. Fantastic.
Looking at this, I think the rocket actually stayed in front of the cloud. You can see a smoke trail in front of the cloud if you look closely. Also, the up-facing camera shows the rocket never actually went through a cloud.

The timing of visibility is also interesting. The rocket took 34 seconds to reach apogee. It went "into" the cloud at 15 seconds. However, burnout should have occurred at 14.5 seconds, and the smoke charge would have burned out at 18 seconds. So, even if the cloud wasn't there, the rocket would not have been visible much longer. The main thing that was missed was the rocket going veritically to apogee.

Jim
 
Back
Top