Tuesday, November 03, 2009

16 hrs of sleep (in a single night!) is helpful....

I slept 16 hours last night and I am feeling much better.
Thanks to everyone for the kind words.

Over the next month or two I'll take some time and describe some of the projects on the
"want to do next" list. I'm going to have to pick one or two off the list. The list is large enough that it will have to undergo significant triage. A smattering of ideas:

Do I work on guidance and controls? (I have a magazine that wants me to publish a seires on the helicopter guidance and controls) I'd like to build a full up system to go on existing airframes that duplicates most of the capabilties of the dragon fly

Do I work to build a hardware in the loop simulator tailored to become an open toolkit for others? I have some really good ideas in this realm ,but I really need to find someone that can help with the core math and transforms part of this. I can do all the basic physics simulation in 2D, but when I try to model things like what a 25 lb gimbaled engine deflection does to to the vehicle in 3D IE accouting for the 3D rotational intetias of both the motor and the rest of the vehicle when the momentum transfers between the two peices are constrained to 2 axis by the gimbal joint, add to this that the deflection changes the rotational inertia of the system in the roll axis my head explodes. As the gimbaled mass is more than 15% of the vehicle weight I don't think you can just ignore the effect. My rocket parts bin does not have any spherical frictionless cows so I can't make thoose assumptions. (If that makes no sense to you you probaly had a different college physics class than I did)

Do I finish my electric pumped hi pressure rocket demonstrator.

Do I do work on better catalysts. I got a really good download of information on Beal's catalyst development and would like to pursue this but the basic equipment to do or hire the flame spray metal application is outside the present budget.

Do I go back to permangenate and static mixing tubes to skip the whole catalyst. (I have all the parts on hand for this experiment so the necessary budget is close to zero)

Do I go back to Lox and ignitors? (The helium cost is a killer)

I really like the small pump work done at LLNL I'd like to duplicate that.

The low cost turbo pump ideas described on Charles Micro Launchers site are of interest.

I have a hankering to build a 10K lb tube wall motor that would be usefull as the first stage in a
reusable booster.

Glide back airframes and or parachutes are of interest.

Do I work on full up vehicles?
This is the most expensive option if I do this I think I will use Wellmate WM35WB for peroxide and WM-6LP for fuel as the WM-6 I used for a fuel tank on silver worked really well.
Using a derivative of silvers propulsion system this could go to 50K ft and then soft land, or well over 100Kt and parachute back.

Any customer(s) with $ to spend could push any one of these items to the top of the list....


simon said...

The chopper work sounds the cheapest and you've got a a magazine willing to pay...

heroineworshipper said...

Skip the contests & go for a reusable suborbital ball.
Why haven't Masten, Carmack, or Unreasonable ever just gunned the throttle to see how high their ball could go?

Anonymous said...

I'd enjoy and learn a lot from working with you on the mathy part of any sim (if they let me here, but I think that should be fine). I definitely take you up on that offer. That said, I think that working on pumps, (electric, peroxide fed, anything) would set you apart from anybody else and would be a HUGE contribution to the industry. I'd say pumps.

Sebastian said...

I was following your blog for a long time. What you have acomplished is incredible, and I hope to read more.

As yet another software guy, I'd repeat the recomendations of others (under the prievious post of Yours) to try Mercurial or Git (both are in fact interchangable to large extent).

Eric M. Collins said...

They all sound like worthy projects. I think pumps are likely to be a important part of a fully reusable launch vehicle. Any reasonable investment there is likely to pay dividends later.

The GNC problem you describe also sounds like an interesting challenge. Conservation of mass, energy, and momentum are the bread and butter of any useful sim. If you can precisely account for the change in the gimbal geometry with respect to time, then I can probably get you an expression for the torques and forces imparted. (Yes, I think we may have had similar physics courses.)

Let me know if you decide to pursue the hardware-in-the-loop simulator. I think I would enjoy the challenge.

Carl Tedesco said...

Paul, as offered before, I can help you with pump design. I have a CAD design mostly done that I can scale depending on the requirements. I'll talk to you off-web of some ideas.

Carl Tedesco said...

Also, thank you for the privilege to work with you. Truly inspiring I must say!

Joe Stanton said...

@heroineworshipper Because only Paul has access to FAR, a relatively unique location that has virtually unrestricted altitude with minimal overhead (pun intended). Carmack or the others would need to buy vast tracts of private land in a remote area to get the same benefit.

Carl Tedesco said...

To heroineworshipper:

Armadillo has been publisjing videos on aRocket of their test program showing boosted hops going higher and higher.

Masten will soon be on their heels. Ben Brockert assured me that after taking some well needed time off they will start altitude flights, which has been their business model all along.

I think they haven't just "gunned the throttle" because they are trying to test their systems responsibly by slowly expanding the flight envelope so they have a chance of recovering from any unexpected "gotchas"!

kert said...

For the 3D sim i would look at existing ( lighter-weight ) solid-body physics libraries around and integrate one.

Bullet, ODE, Newton are the popular candidates.

No need to reinvent the wheel.

dave w said...

Just out of curiosity - how much of the hardware from the "silver ball" turned out to be salvageable? (Initial word on the web was "tank punctured, engine intact...") -dw

Paul Breed said...

On the silver...damage:

Main Tank punctured,
One Ball screw actuator bent.
Subsequent power up to open main valve for dumping propellants cooked the motor control for bent actuator as it jammed.

James Bauer said...

"I got a really good download of information on Beal's catalyst development and would like to pursue this but the basic equipment to do or hire the flame spray metal application is outside the present budget."
I will let you use some oxyfuel flame spray equipment I have. I would like it back when you are done though. I have a selection of various powders, do not know what you need though.
The silver ball- If you are willing to ship it to and from texas, I would be willing to repair it for free (if possible, and in my spare time). Not that I have a lot of spare time. But I would like to see it fly again if possible. Get with me if interested. You have my email...
-James Bauer

Paul Breed said...

Thanks for the offer Since the LLC is over I have almost no interest in rebuilding silver. I want to take silvers "guts" and make an aerodynamic airframe for altitude work. I have COTS fiber tanks picked out and will work on that as funds allow.

Paul Breed said...
This comment has been removed by the author.
dave w said...

If the engine and most of the electronics are still usable, I think it would be a shame not to resurrect them, somewhere down the line - in some form, if not in the specific "lander challenge" vehicle configuration. (Of course, a good bit of your effort was going into "precision landing" capability that was mainly relevant to this specific contest, but a lot of the rest - general flight stabilization etc. - would certainly carry over!)

Carl said...

The version I heard was a spherical chicken in a vacuum, rather than a spherical, frictionless cow.

Glad to hear you got your needed rest. :)

Joe Stanton said...

For those who need more info about the cow reference.
Consider a spherical frictionless cow...

Sebastian said...

Heh, while I'm about 10000 away I'm also from spherical cow group. But our cows had friction :)

Anonymous said...


Are you saying only Paul can fly at FAR? I though it was open to anyone who wanted to support the group? Why can't "others" fly there as well?

Joe Stanton said...
Because only Paul has access to FAR, a relatively unique location that has virtually unrestricted altitude with minimal overhead (pun intended). Carmack or the others would need to buy vast tracts of private land in a remote area to get the same benefit.

Robert van de Walle said...

Congratulations! You are winners, and I'm glad you got some well-deserved sleep, too. That's a prioze worth more than money sometimes.

Paul Breed said...

Actually anyone can fly at FAR.

Anonymous said...

So happy you are planning to continue your extremely valuable work. A lesser person may have been crushed.

Joe Stanton said...

My mistake. I thought the way I read some of the other postings was that because Paul was the only person launching from FAR (perhaps on a given day?) then he could go to any altitude. I'll just retract all that and stay out of this sorry for the confusion.

Joe Stanton said...

On second thought, what I meant, was of the groups listed, e.g. Masten, Armadillo, and Unreasonable, only Unreasonable launches from FAR where the altitude restrictions are slim if any. Not that any of them could not launch from there, they just do not, at least not at this time.

Thomas said...

I have really enjoyed reading your weblog. I'm a bit gutted you didn't win anything, as you've contributed more insight and excitment than Masten, and at least equal Armadillo.

Kert suggested using existing computer-game oriented physics simulation packages. Even if they don't meet your needs directly, the people behind them could be useful to know. I'm sure there will be some who'd love the idea of their knowledge contributing to real rockets.

Your strenghts so far have been having the determination and resources to build real things, get feedback from reality, and repeat. Hooking your control systems up to a simulated rocket might be a good way of consolidating your experience without spending very much money (until you take it back to reality), but doing it alone probably wouldn't be the best use of your time.

If you could continue working on peroxide systems, you seem to be in a good position to extend rather than duplicate other's efforts.

This deep in atmosphere, pumps look very attractive in theory, but how well placed are you to develop reliable and practical pumps?

It seems to me that lander challenge performance remains relevant to anyone hoping to rely on soft vertical landing, so it might still be worth assembling a landerish vehicle from what you have left.

(In the version i heard, the physicist's report on how to make money at the races begins 'Consider a spherically symmetric horse...')

Laurence said...

I would look at this this way. The final suggestion of an all up vehicle should be the end goal. The ideas above it are more like some of the things that would need to be investigated to get there.

Break the end goal down into categories like fuel, propulsion, control/guidance, aerodynamics, testing/validation - then look at what you need to do to understand, develop, master those categories category in the context of the all up vehicle. This might make it easier to see the wood for the trees (or make the goal seem impossible lol).

I've had a couple of other rather random thoughts in no particular order:

If you don't already, could you improve the control software by running the gyro and accelerometer readings through a Kalman-like filter before being fed into the PID loops?

If your valves and actuators have sufficiently little hysterisis and backlash, could you take an approach similar to modern auto engines with respect to the control. I'm thinking of the engine map. Something like looking up up the desired vehicle "reaction" in a "map" - i.e. to do a translation in the x-axis acuator 1 goes to y%, actuator 2 goes to z%. These values are then tuned by the inner control loop - eliminating the proportional factor I guess? Ideally this map would allow for things such as current mass too. Obviously you need to be able to fly the vehicle and record all of the parameters to generate the map in the first place. A map for net engine thrust against fuel and oxidizer valve positions could be developed by testing on a stand I suspect though.

If the valves have a fair amount of backlash or hysterisis, how can this be improved? Can these be actuated hydraulically or pneumatically? This sounds a lot more complicated but may allow better controllability. I know the bigger valves have damping anti-slam valves which would be a good-to-have to mitigate fluid hammering risks.

Is there some reliable, accurate and inobtrusive way of metering the flow of fuel and oxidizer? This would help burn optimization, mass estimation, CoG estimation and the control "map" idea. This could be incorporated into a new pump I guess.

This last one is almost certainly a very stupid idea, but... Could you use some form of gyroscopic stabilization to make a vehicle's behavior more benign? I'm not sufficiently awake to do the back of an envelope calculations of how much mass needs to spin and how fast but I suspect this is a non starter.

Good luck for whatever you decide to do next, though!

P.S. Some links c/f the above

A Kalman-like filter for the Parallax processor:


General info about controlling stiction in PID loops:


industrial equipment said...

Valves vary from the extremely basic to the extraordinarily complex, and they are one of the oldest mechanical designs. Thanks for sharing great information, Hope to hear more updates from you ...