Fire in the hole!

[MB]

Thanks to all who contributed in the thread titled, "How many things could be wrong?" That thread has wondered off to additional subjects, but I wanted to share our success.

The alterations we made were that we bent the valve retainers closed a little, so the valves couldn't open quite so wide; and we fabricated an inlet tube.

Our text fixture consists of a structure of 2x4 lumber with 2" angle-irons for a thrust platform.

The straps over the fuel nozzles are to disperse the fuel from a single squirt line into a splash pattern.

We were a little nervous about the thing ripping out of the fixture and flying away from us...

[MB]

We played with it for a while with precisely the same results as before: we turned on the ignition sparker, gave it some air with a leaf blower; and all we got was some flame and a few puffs and booms.

It occurred to me that when holding the blower into the inlet, there was quite a blast of air coming right back out of it. So I put my hand over the inlet to force more air through the engine, and Dang if it didn't just roar to life!

We were elated. We hooped and high-five'd and just watched it rumble.

We both had ear plugs (nrr33) and muffs (nrr 22). I didn't hear much, but one of our team remove them just to see how loud it really was, and quickly put them back on. "It was VERY loud!", he said.

As it ran we observed the sound wasn't as steady as we anticipated, and although there wasn't any flame coming out the back, (standing back a ways) there was yellow flame visible inside the combustion chamber. I swooshed my hand through the exhaust stream, and indeed there was some genuine thrust going on.

[MB]

After about 30 seconds the engine just stopped. We figured, "Ok. We have some tuning work to do." But upon closer inspection we saw dramatic evidence of catastrophic failure. We tore it down, and discovered parts of the aluminum valve grid had melted -- in fact burned -- right through. Note that the aluminum actually bonded to the spring-steel. That thing got HOT!

[MB]

Lessons learned, and tasks planned:

The fuel dispersal vanes were in fact splashing fuel directly on the valves, so the flame was occurring on the valve grid.

• Replace the nozzles with orifices that spray a cone pattern, and which will direct the flame away from the valves. We expect this will keep the temperature lower on the valves -- thus not only protecting the grid, but the valves themselves.

There is no quantitative feedback.

• Add some electronic instrumentation so we can measure the behavior of key components and document the engine thrust and lifespan -- thus not only observing to create better characteristics, but to reproduce them in future generations of the engine.
• Modify the test fixture to incorporate a load cell to measure the thrust. (Not to mention adding more robust anchoring.)
• Add a flow meter in the inlet to measure the air of the input air velocity (to derive the mass-flow).
• Add a pressure sensor and a flow meter to the fuel line to measure exactly how much fuel is flowing (to derive a correlation of the pressure to flow rate). Then we can calculate the air/fuel mixture, and compare it (or modify it) to published ideal mixtures. We can also use this feedback to test different fuels to see how they perform and compare.
• Add a few thermocouples and infrared thermometers to monitor temperatures on the grid, nearby, and elsewhere around the engine.
• Add a pressure sensor to the combustion chamber so we can monitor the effectivity of the tuning. We expect to derive a simple correlation between variable input parameters and output performance; thus minimizing (redundant or more expensive) instrumentation in the future.

The only variable is the fuel pressure.

• Re-design other critical parameters to be variable(ie, adjustable) – such as valve closure pressure and inlet and exhaust nozzle shape – so we can interactively modify the geometry or other physical characteristics and measure the effect.

[MB]

Oh, and I am still working on posting the video. It turns out the video capture card I have is one I used on MS/Win 95, and the vendor never updated the drivers to run on later versions of MS/Win -- bummer! So I gotta go out and get a new one afterall -- busted!.

[Eric]

OH NO! THE VALVE GRID IS ALUMINUM!

I really wish I saw that.

With liquid propane you can set it up so that the fuel tubes expand right against the injector plate and a good deal of cooling will happen from the expansion.

With propane vapor the engine can get extremely hot.

Whenever I run valved engines on propane vapor for any time I hook up a water injection system. It greatly cools the valves and also produces some extra thrust for no extra fuel.

Your valves will still get extremely hot even if the fuel doesnt spray on them with vapor. I would go with running it on liquid propane, using a smaller tube to deliver fuel to the engine, and having larger expansion tubes so the propane will expand and cool, sucking up lots of heat in the process of getting injected.

Also if you make flat plates between the valve openings, so that they are touching the valve plates, they will absorb heat from the plate and provide lots of surface area to transfer the heat to the incoming air.

Eric

[evildrome]

Congratulations!

I am really looking forward to the video.

I am also concerned about your grid melting as I also wish to use aluminium. I think the flame can be kept away from the grid by extending the valve keeper plates to form a venturi. So your valve keepers look like this:


/ \


the venturi would look like this:


\ /
/ \


The air speed through the venturi would be above the flame speed until well after the valve grid. The V1 Argus uses this. If the German engineers thought it was necessary I'm damned well sure I'll be using it on my engine.

Cheers,

Wilson.

[Eric]

If you make the aluminum plate like this then it has a lot more cooling area.

Eric

[MB]

Oh, these close-ups of the valves are precious.

[Eric]

Oh yea, I almost forgot. You should at very least put aluminum foil on those 2x4's to reflect heat, otherwise they WILL burst into flames.

What thickness spring steel are you using? Hopefully they wont deform as much when cold.

Eric

[MB]

We've tried 6, 8, and 10 thousanths. These are 10's. The wooden test fixture is only temporary. We have designed a more generalized steel one that bolts onto a standard test stand we have access to, but it'll have to wait a few more weeks for funding (dang CR steel is expensive!). The rear-end of the previous version of this thing actually did completely burn off. It was really a funny sight.

[Jonny69]

You could try stainless steel instead of spring steel. I found it doesn't deform like spring steel does but performs roughly the same.

[evildrome]

MB,

Go to your local scrap yard or dump. There will be all sorts of metal objects you can cull square & round section tube from to make a test bed.

Or you can go round to a local fabricator or tube dealer and ask for 'drops' (off cuts that no one wants).

Cheers,

Wilson.

[Eric]

You could also try to layers of 6 and maybe even a layer of 8 and 6.

With a good set of heat sink fins the valve plate will be very much cooler and the valves will also be cooled by contact with it a lot more.

Eric

[MB]

Here's a short video clip of our first successful firing last weekend. This was our 5th attempt, so we were pretty excited that we finally got everything right. The engine died just as one of our team members was going to drop a rag into the jet stream to show its velocity. 28 seconds: not exactly a record...