A couple thoughts on hybrid design...

[D_Hall]

As some of ya'll may be aware, I'm currently in the middle of a redesign of The Pipe Dream. Why? Because it may actually get built. The design I had was better than back of the envelope, but it wasn't a finished design. So I'm rolling things over in my mind and...

Burst Diaphrams vs. Shear Pins

Has anybody ever used shear pins instead of burst diaphrams? It occurs to me that shear pins may be a much easier route to holding chamber pressures. Just use a sheet of rubber to seal the gases in (like you would with any other burst diaphram) but bolt the projectile in using a varying number of nylon bolts (or other softer-than-your-barrel material). Vary the number of bolts and you vary the "burst" pressure.


High Energy Injection

As you can imagine, with a chamber as large as mine proper fuel/air mixture is a concern. In theory I could simply wait an hour or something and to do so would NOT be a show stopper given my application. However, obviously waits like that would be inconvenient. So I'm thinking... Rather than use a traditional metering pipe I could do direct injection of may air into the chamber. Why not? I've got high precision pressure gages. If I can measure it in my metering pipe I can measure it directly in the chamber. What does this buy me? The ability to use unregulated 3000 psi air as my supply. I can use an oriface to control the flow such that my pressure increases at a reasonable rate (say 6 psi per minute). The point being that I now have one hell of an input stream. Turbulence should abound.... and thus I should have reasonable fuel/air mixing.


Thoughts?

[clide]

I've never seen anybody using shear pins, but I don't see any reason why it wouldn't work. Even if they did cause problems for some reason you could just plug up the holes and go to more traditional burst disks.

I do believe people sometimes use direct air injection (although not high pressure). One thing to think about is that high flow out of a high pressure source tends to get rather cold, do you think it would cause any problems with propane liquefying in the area near the air inlet? I'm sure all the steel around would be able to heat it back up pretty quickly, but if it was causing propane to liquefy then it may not work well for mixing the tank. As for it mixing the fuel I thought I remember you being opposed to adding the fuel before the air. Have you changed your mind on that or am I remembering it wrong?

[D_Hall]

As for it mixing the fuel I thought I remember you being opposed to adding the fuel before the air. Have you changed your mind on that or am I remembering it wrong?

Yes and no.

You're not remembering it wrong but I didn't change my mind either.

What DID happen is that I realized that there's no reason I can't do the entire fueling process by remote control. In fact, in some ways it will be easier for me to do it that way. Thus, since nobody will be within a mile of this thing when it's fueled, it doesn't really matter what order it's fueled. My concerns were safety related and they're totally negated by going to remote operations.



edit: As for your other points....

1) Good point regarding "If shear pins don't work just plug the holes and move on." I'd not thought of that angle and with it, there's just no reason NOT to investigate the use of shear pins.

2) I don't think the air input will be a problem from a thermal perspective. Air comes in, it gets cold.... But as the chamber pressurizes it heats that air right back up. Point being that I don't think it'll be an issue.

[D_Hall]

Another thought regarding instrumentation....

I want to be able to say things about the flame propogation and such. Fast response pressure gages are very expensive and very delicate. If I have one "oops" I've just lost Big Bucks.

So what I was thinking....

Have two fast response pressure gages. One just before the burst diaphram and one just after. This will allow me to make statements regarding mass flow through the oriface and the environment "seen" by the projectile.

But to track flame fronts? Photo cells. Yes, just cheapo photo cells. Put 'em in a tube that's been painted flat black internally. Put a window on 'em to protect the photo cell from the flame. Voila. The cell shouldn't "see" any flame until the flame passes it. At that point it should produce a detectable voltage.

Dirt cheap and easy to do.

Put in a photo cell every foot (or whatever seems appropriate) and I should be able to easily track multiple flame fronts with reasonable accuracy.


Again, thoughts?

[jimmy101]

What kind of photocells? I wouldn't think photovoltaics would work, too slow. I would recomend phototransitors (PTs). They respond much faster than voltaics and are much more sensitive. You should be able to just epoxy/glue the phototransistors into holes in the chamber, heck you might even be able to tap them into threaded holes in a steel chamber. A power source, perhaps a resistor or two and a PC with sound card to record the signal. Sounds very interesting, and doable, but...

I'm not sure how bright the flame front will actually be. IIRC laminar flame fronts are very thin (less than 1mm) so there isn't much of a light source. A turbulent front is much thicker but also more diffuse. Furthermore, propane+air burns with a blue flame, not much light is actually created (just look at the flame of a Bernzomatic torch). Most photodetectors are much more sensitve to the red end of the spectrum so the signal might be weak. Sounds like one of those things that just has to be tried to see if it works.

Lets see, a 4mm diameter PT has an area of ~0.02in². At 100 PSIG there will be 2 pounds of force on the PT. At 300 PSIG there would be 6 pounds. Sounds doable.

I believe combustion engineers usually use conductivity sensors (basically a small spark gap) to locate the flame front. Not sure it that'll help you much since it probably takes a high impedence high gain amp to get a useable signal. (A high impedence OP Amp?)


I don't follow your high energy injection. A small inlet orifice, particular if it is at the end of a tube, may well give very good mixing.

But why would you have to wait an hour after pressurization? Granted, the mix will heat up when compressed into the chamber but the heat capacity of air is pretty low, I would think that it'll cool back to ambient pretty quickly (as in less than 1 minute). Plus, the heat capacity*mass of the gun is a heck of a lot bigger than the heat capacity*mass of the gases, the gun shouldn't warm up by more than a few degrees. I figured that an 80ci pvc closed combustion chamber only warms up by ~7F when fired even though the gas temperature peaks out at ~4000F. Furthermore, the combustion gases in a closed chamber are back to with ~10F of ambient within a second or so after ignition.

The pressure detector(s) sound very interesting. You've got something that'll respond within a couple milliseconds?

[D_Hall]

What kind of photocells? I wouldn't think photovoltaics would work, too slow. I would recomend phototransitors (PTs). They respond much faster than voltaics and are much more sensitive.

Thanks. I admit that instrumentation is not my gig so this is exactly the type of input I was looking for.

the signal might be weak.

I can deal with weak. As long as it's there, I'm good.

I believe combustion engineers usually use conductivity sensors (basically a small spark gap) to locate the flame front. Not sure it that'll help you much since it probably takes a high impedence high gain amp to get a useable signal. (A high impedence OP Amp?)

Hmmm.... Is this a commercially available part? Or perhaps a spark plug could be used? As for what it takes to read the signal, I'm not that concerned. I've quite literally got about $8M worth of data acquisition hardware at my disposal (even if I'm clueless as to how to use it). In other words, I'm sure my data acquistion guy can cobble something together on that end.

I don't follow your high energy injection. A small inlet orifice, particular if it is at the end of a tube, may well give very good mixing.

Locally, I'm sure it would. But would that mixing extend 30' down the tube? I'm not sure. That's why I'm looking at high energy injection; to extend the zone of "I'm sure it's mixed."

But why would you have to wait an hour after pressurization?

Because IF the fuel doesn't mix well (see above), I'm going to have to rely on Brownian motion to do my mixing.

The pressure detector(s) sound very interesting. You've got something that'll respond within a couple milliseconds?

Sure. Of course, they cost about $3,000 each so you can see why I don't necessarily want to use a bunch of 'em.

[jimmy101]

D_Hall

I'm so jealous. I have to cobble stuff together using $3 piezo transducers to get a very crude high speed pressure signal.

So, a chamber fan is definetely out? Some wild ass ideas for mixing.

1. I wonder if you could plumb a loop of rubber hose inside the chamber. Pump high velocity air through the hose (in one fitting then back out of the chamber through another). At certain flow rates the hose will flail about and mix the gases? OK, that's a stupid idea.

2. Add a fitting (with valves) on each end of the chamber. Connect the fittings with hose and a small low pressure, high volume, air pump. Mix the fuel by just running the pump for a while so the air is drawn out one end then injected back into the other end. Probably not the most efficient mixing but should be a heck of a lot better than Brownian. This sounds like it would work. If there is a valve at each fitting then you can isolate the chamber from the recirc loop for firing.

3. Metal chamber so a magneticly linked fan / stirring is out... Too bad, a standard chemistry laboratory magnetic stirrer and a 1" long teflon coated magnet can be gotten spinning at a pretty good clip, but only though a plastic or glass container.

Mount a magnet on a plastic plate that is connected to the inside chamber wall with a hinge? Several coils of magnet wire around the chamber and a power supply. Cycle the power supply on and off (or cycle the polarity) so the electromagnet makes the magnet (and the attached plastic vane) bang back an forth. (Not the most elegant solution but it would be a crack up to hear the mixer going wack-wack-wack...)


Do you have access to a research library at your job? Combustion engineers have been using conductivity detectors to follow flame fronts for a long time. I've got a few abstracts but don't have a good research library (and am too cheap to spend $25 for a reprint) so I don't have copies of the articles. Basically (if I understand them correctly) it is just a very small gap, less than 1mm, connected to an ohm meter. The flame front is a plasma and conducts electricity. It doesn't conduct very well but it conducts better than air. The gap needs to be about the thickness of the flame front. An instrument amplifier is probably needed but it sounds like that would be no problem for you.


More wild ass thoughts;
Put a piezo transducer (~$3) or two (or five...) in the chamber, perhaps at different places, and record the sound of the gun firing. The sound (actually the pitch) in the chamber during firing can give information about the temperature in the chamber. On the lnked page is a piezo recording of a closed chamber being smacked with a hammer (gives a very nice "gong" sound, particularly if the chamber is pressurized well above 1 ATM) and a recording of the chamber being fired.

A piezo near the valve might be able to "hear" the gases moving through the valve. Or, rigidly mount a piezo on the outside of the valve. Can you "hear" the gasses moving through the valve. Can the pitch (frequency) of the sound be related to the temperature or velocity of the gases?

Mount a noise maker (like a peizo transducer, apparently I've got piezo's on the brain) near the muzzle. Mount another piezo as a mic. Blast the first piezo with audio noise. The air in the barrel will resonate at a particular frequence. As the projectile moves through the barrel the length of the pipe decreases and the pitch rises. Kind of like those plastic slide whistles we had as kids. Simple math converts the pitch to the distance between the projectile and muzzle. (Might have problems with the gases in front of the ammo being compressed and heated, which will also shift the resonant frequence.)

[jimmy101]

Another way to mix the fuel. Instead of a recirc pump that is run after the fuel and air are injected, use a fan and inject the fuel as the air in the chamber is being circulated. (see attached gif)

1. open ball valves
2. pressurize the chamber with air
3. turn on fan
4. slowly inject propane into the flowing air stream
5. turn off fan and close ball valves
6. fire

The fan should be sized so that it moves the chambers volume in a reasonable length of time. For a chamber volume of Xft³ a fan that moves 0.1X to perhaps X cfm would seem to be about right. The propane injection time is just the calculated length of time it takes the fan to move the chamber's volume of air. (Of course, it might be a good idea to continue running the fan for a while after fuel injection is complete to insure complete mixing.)

Since the ball valves are closed during firing, the fan is not subjected to the combustion process. Any old fan that doesn't generate sparks should be useable. Perhaps several brushless CPU or computer case fans.

Alternate design: Omit the ball valves. If the barrel is attached at the bottom of the U in the drawing then the gas flow during combustion and firing at the location of the fan is pretty close to zero. The fan shouldn't have any problem surviving.

This also suggests that a toroidal (donut shaped) chamber might be worth considering. Circulation of the gases around the loop can be used for mixing. During firing, there will be a place in the toroid where there are pressure fluctuations but very little gas flow. This would be the spot to install fragile components like a mixing fan.

[D_Hall]

So, a chamber fan is definetely out? Some wild ass ideas for mixing.

Well, I don't know that it's out per se, as much as impractical. Remember, I'm talking about a 30" diameter chamber that's 50' long. Admitedly it's over kill and I'm seriously considering chopping 20' off of it to reduce the chances of DDT, but the point is that it's a physically large system. A little 4" fan isn't going to do much in there. Too, going with a system like that would open me to a LOT of scrutiny from the safety department folks (live electrical circuits in an explosive environment attracts those guys like naked women attract teenage boys).

2. Add a fitting (with valves) on each end of the chamber. Connect the fittings with hose and a small low pressure, high volume, air pump.

I've thought about this but it gets back to either a requirement for energized electrical systems in an explosive enviroment or complex seals and such. One *possible* exception would to make this the fuel/air injection point. IE, to set up the tube as an eductor. Of course, the pumping action would stop when the filling stopped, but it WOULD force a significant circulation.

3. Metal chamber so a magneticly linked fan / stirring is out... Too bad, a standard chemistry laboratory magnetic stirrer and a 1" long teflon coated magnet can be gotten spinning at a pretty good clip, but only though a plastic or glass container.

It *could* be done using an aluminum "window" in the chamber, but you have to wonder how effective a small system would be and how expensive a large system would be.

Do you have access to a research library at your job? Combustion engineers have been using conductivity detectors to follow flame fronts for a long time.

Sure. We even have an entire branch dedicated to combustion research.... but to my knowledge they use window bombs and high speed photography. Granted, I've not asked the precise question. I'm just going on what I've seen of their results.

(And for what it's worth, they haven't been much help as this application is vastly removed from the kind of stuff they normally do.)

An instrument amplifier is probably needed but it sounds like that would be no problem for you.

Yeah, I've got a few hundred of those lying around.

More wild ass thoughts;
Put a piezo transducer (~$3) or two (or five...) in the chamber, perhaps at different places, and record the sound of the gun firing. The sound (actually the pitch) in the chamber during firing can give information about the temperature in the chamber. On the lnked page is a piezo recording of a closed chamber being smacked with a hammer (gives a very nice "gong" sound, particularly if the chamber is pressurized well above 1 ATM) and a recording of the chamber being fired.

Fascinating stuff! Thanks for the link.

I don't know how much useful data I'd get out of my system since my chamber isn't cylindrical, but it'd be so cheap there's no reason not to try it out.

Or, rigidly mount a piezo on the outside of the valve. Can you "hear" the gasses moving through the valve. Can the pitch (frequency) of the sound be related to the temperature or velocity of the gases?

Easier to just use a pressure gage and measure pressure drop. Combine that with some temperature data and you're there.

Mount a noise maker (like a peizo transducer, apparently I've got piezo's on the brain) near the muzzle. Mount another piezo as a mic. Blast the first piezo with audio noise. The air in the barrel will resonate at a particular frequence. As the projectile moves through the barrel the length of the pipe decreases and the pitch rises. Kind of like those plastic slide whistles we had as kids. Simple math converts the pitch to the distance between the projectile and muzzle. (Might have problems with the gases in front of the ammo being compressed and heated, which will also shift the resonant frequence.)

Wouldn't work without a LOT of testing. Adiabatic compression issues. Doppler effect issues. All sorts of issues. Easier to just mount a bunch of strain gages to the barrel at various points and detect when the barrel is strained. That'll tell you that the projectile just went by the strain gage.

[rp181]

Ion wind? Its preety much high voltage electrodes placed apart so they dont arc, The ions travel from one electrode to the other, dragging air with it. Search electrohydrodynamic and high voltage lifters. High voltage in it probably isnt smart, but since its not supposed to arc or even corona, it migh work.

[Pete Zaria]

The idea of shear pins intrigues me; I've always thought burst disks are the "weakest link" in the popular hybrid design. _Fnord's work on piston hybrids is groundbreaking, but I'm wondering how far it can be taken.

One thing I've considered is using a "forcing cone" very much like that in a modern revolver.
The bullet is a few thousandths of an inch too large to fit into the barrel - it must be pushed through the forcing cone, which very slightly reduces it's diameter and simultaneously cuts the rifling groves into the bullet.

A similar principal could be applied to a hybrid, given that the ammo was uniform, spherical, and made of a slightly deformable material (hard plastic, lead, etc...).
The ammo itself would block the barrel, being slightly too large to fit, therefore sealing the chamber. On ignition, the pressure would force the projectile down the barrel.
Experimenting to get this right would be a DANGEROUS process. An entirely sch80 chamber and barrel should be utilized, I think. At first, a reasonably soft projectile (perhaps made of wax, or molded from hotglue sticks) should be used as proof-of-concept.
If gotten right, this method could be extremely simple and efficient - no valve or burst disk is required, a gravity or spring fed magazine could be used to feed additional ammo, the pressure difference in the chamber would quickly seal the ammo against the breech...

Best of luck with your research and development Keep us posted.

Peace,
Pete Zaria.

[DYI]

As long as this is done in small diameters at reasonably low mixes (i.e. below 20x or so), using SCH 80 is complete overkill, and unneccesarily expensive. Now, you might be talking about PVC, but using any PVC parts in a hybrid isn't a very good idea at all.

[Pete Zaria]

As long as this is done in small diameters at reasonably low mixes (i.e. below 20x or so), using SCH 80 is complete overkill, and unneccesarily expensive. Now, you might be talking about PVC, but using any PVC parts in a hybrid isn't a very good idea at all.

I was thinking steel. I like a *big* margin of safety.

Besides that, what do you think of the idea?

Peace,
Pete Zaria.

[D_Hall]

Besides that, what do you think of the idea?

No reason why it wouldn't work, but it requires a steady supply of ammunition that is The Same Every Shot.

[jimmy101]

...it's a physically large system. A little 4" fan isn't going to do much in there. Too, going with a system like that would open me to a LOT of scrutiny from the safety department folks (live electrical circuits in an explosive environment attracts those guys like naked women attract teenage boys).

Ya, a single fan isn't going to do squat, but a fan every 10 feet might do the trick.

To get around the explosion hazard you just have to show that there is no ignition source. A generic electric motor is a huge ignition hazard (even though spudders have never been able to ignite propane + air with one). A brushless fan motor running at 12V is not. Live electrical circuits are used regularly in environments that are potentially explosive. Walk into any chemical storage facility; there are light switches, lights and sometimes circulation/exhaust fans. All those devices can be "safed" in terms of sparking and used in an explosive environment.

Another possiblity would be an air driven fan system. A lot more complex that electric fans but zero chance of sparking in the chamber.

Any idea how long it would take for fuel to diffuse the length of the chamber? My WAG is that it'll take a lot more than an hour, indeed it might take weeks.