What would I need to make a gas turbine from scratch?

May 5, 2007 5:06 PM

What would I need to make a gas turbine from scratch?

Basically, I'd like to know what tools, raw materials, etc I'd need for building a working gas turbine, roughly the size of one for a small helicopter, by hand, from scratch.

Electricity in the construction is obviously allowed, but no computer-controlled tools. Also, buying things like screws and bearings are allowed, but not any pre-shaped metal, or anything like that. Bonus points for the fewest amount of hugely expensive machines (although quality > price).

And yes, I do know it would be expensive.

posted by devilsbrigade to technology (26 comments total)

A small helicopter can be powered by a piston engine. (The med-evac helicopters they used in the Korean War didn't use turbine engines.)

The problem with a turbine engine is the turbine blades. You can't use aluminum because aluminum gets soft when it gets hot. You can't use normal steel because it burns. You have to use more esoteric materials -- or else you have to not run the temperature very hot, in which case the power yield and efficiency will suck.

Mechanically, a turbine is a lot more straightforward than a piston engine. (Only one moving part, not counting the fuel pumps.) The reason it took so long for them to get used in aircraft was that it was a non-trivial materials problem. The rotor spins very fast and the rotor blades are under a great deal of tension. There's no way to cool them, and if they deform very much then you get some sort of catastrophic failure. (There are several potential failure modes, but they all add up to "explosion, smoke, and fire".)

Are you asking this question hypothetically, or are you actually interested in doing it for real? If you want to home-brew a small helicopter, you're much, much better off going with a piston engine.
posted by Steven C. Den Beste at 5:34 PM on May 5, 2007 [1 favorite]

I'm afraid you'll never be able to machine a turbine accurately enough at home without using CNC. Here's a few resources for smaller ones if you're still interested. The problem with scaling up is the alloys involved. Most successful homebuilts use the turbine from automotive turbo's of various sizes/types.
posted by IronLizard at 5:44 PM on May 5, 2007

And don't forget: absolutely perfect balance is essential or it will shake itself into a million pieces in seconds. Try not to be in the immediate vicinity.
posted by IronLizard at 5:46 PM on May 5, 2007

Point taken. How abouts we switch the question to a 4-stroke piston engine?

The question is mostly hypothetical; I certainly don't intend to build a helicopter. I'm more interested in the tools & materials needed to make someting of that precision/complexity if you were to do it entirely yourself.
posted by devilsbrigade at 6:06 PM on May 5, 2007

As far as tools, for the blades, you could use what the pros use. Many gas turbine engine blades are sand cast. The required materials are lots of sand, some clay or other binder, something to pack the sand (so it holds shape) and a burning hot furnace to melt the metal you will pour into the mold.

But keep this in mind, as far as precision. A couple years ago I visited a foundry in Minnesota which is contracted by something government to cast their turbine blades. Their yield is 10% on a good day. Each cast product is scanned (CT) for internal and external defects. If it isn't good enough, it gets remelted and poured again.
posted by whatzit at 6:19 PM on May 5, 2007

By comparison, a four stroke engine is a lot easier. It's much more forgiving; but it's also a lot more intricate. The pistons, cylinders, crank shaft, cams, and valves can all be made on a lathe. (But it has to be a really good lathe.) The block and head could be machined with a mill, but it would take a long time; casting is probably better, but a bit more high tech.

But there's still no way you're going to do it without buying some premade metal stock. In particular you're going to need tubing, which isn't something you can make for yourself. And the Venturi cluster would be a challenge.

For that matter, the entire electrical system would be a challenge in its own right. Spark plugs are not something you make for yourself. And I can't imagine what a pain it would be to try to build an alternator from scratch, even if you bought the wire.

It can be done, of course; that's how the first ones were made. But even Otto and Daimler didn't try to make everything themselves.
posted by Steven C. Den Beste at 6:29 PM on May 5, 2007

Two stroke is even easier.... no valves or cams required.

Cam machining would be complex, and not only do you need the cam, you need a cam drive, and valves. The valve seals are precise, too, and it's important to have good ones on the intake side as there is an explosive fuel/air mixture on the other side of the cylinder interior.

Also, you'll need an oil pump and conduits for same.

Two-cycle is noisy as 40 hells, but much lower precision on a lot of the parts.
posted by FauxScot at 6:57 PM on May 5, 2007

Big lathe. A small foundry and assorted casting equipment. A metal shaper. A large mill. You essentially want a machine shop + foundry. Then you'll need the plans for it.

Start here and build your (hypothetical) shop from scratch first. Most of the foundry portion is geared to aluminum, but most carries over to steel with some modifications. The valve springs will be exceedingly difficult. You'll also need to pay special attention to the alloys used to make the valves themselves. Try learning to rebuild one first, to give you an idea of what's involved. (I have a few pics of my last rebuild in my flickr account, though there's quite a bit more involved than I took pictures of and I didn't do my own machine work).
posted by IronLizard at 7:02 PM on May 5, 2007

While I'm on the subject you'll need to learn to precision-mold rubber for the seals.
posted by IronLizard at 7:09 PM on May 5, 2007

You might be better off trying a simple steam engine. Simple linkage, no elec. system. You'd still have to machine the piston. For my money I'd build this.
posted by BostonJake at 7:15 PM on May 5, 2007

Bruce Simpson offers plans to build gas turbines and pulsejets with a fairly modest tool set.

Personally, I'm kind of drool-y over his DIY Cruise Missle
posted by Orb2069 at 7:37 PM on May 5, 2007

If a hypothetical hobbiest (say, in a work of fiction) really wanted to build his own helicopter, he wouldn't try to build his own engine. He'd visit junk yards looking for wrecked high-power motorcycles hoping to salvage a working 4-cylinder engine, and then design his helicoptor around that.
posted by Steven C. Den Beste at 7:44 PM on May 5, 2007

Couldn't you use an old turbocharger and some clever gear reduction?
posted by TheNewWazoo at 8:55 PM on May 5, 2007

In a work of fiction, I would build a helicopter out of nothing but spare plumbing bits. And it would have Hellfire missiles. YMMV.
posted by IronLizard at 8:57 PM on May 5, 2007

Spark plugs are not something you make for yourself.

What's so hard about that, by the way? I understand that a spark plug is basically a ceramic cone, two precisely machined electrodes, and a seal. Same goes for the alternator - you need a shaft, a casing, lots of insulated wire, and a bunch of small parts. Any shop that can make an engine block should have no problem making those.

Tubes would be a problem, yeah...
posted by azazello at 9:55 PM on May 5, 2007

Spark plugs, hehehe. First make your hollow metal casing with a 'prong' on the lathe. Then make your center electrode. Stuff the middle of the casing with clay and shove the center electrode down the middle. Bake and glaze. Set gap. Watch as your homemade ceramic cracks and falls apart within an hour of use on your rough running, shaking homemade motor. Or not. Getting it right would be an entire huge project in itself.
posted by IronLizard at 10:10 PM on May 5, 2007

The problem with a spark plug is the force of the explosion inside the cylinder. 5-50 times per second it's being hit with sledge-hammer force. It has to survive hours of that.

One made the way IronLizard describes wouldn't last 5 minutes.
posted by Steven C. Den Beste at 10:23 PM on May 5, 2007

I used to work for a major manufacturer of jet engines (3 guesses which one). Here are the parts that are difficult/impossible to make on your own:
-Any of the turbine/compressor blades. These need to be manufactured to less than a millimeter tolerance for two reasons - you want the ends to ride as close to the housing as possible to minimize air leaking through that gap, and you need to account for heat expansion.
-Bearings. 2-stage compressors (most airplane engines have these) require two turbines to power them, which means to drive shafts, which means one rides inside the other. So, you need a bearing that allows two shafts to run at different speeds and connect it to the casing.
-Combustion chamber. This thing gets HOT. Like, thousands of degrees hot. Materials engineering++ here. Plus, you need to design it right so the flame doesn't burn itself out.

There are lots of exotic materials in your standard jet engine that the layperson would never be able to find, also. Vibration is also a big concern. So, there you have it.

(Yes, I'm an Aeronautical Engineer, but I'm not yours.)
posted by backseatpilot at 10:36 PM on May 5, 2007

"Spark plugs are not something you make for yourself."

So make 'er a diesel. No electric parts needed, the combustion works entirely on pressure.
posted by CrayDrygu at 11:15 PM on May 5, 2007

CrayDrygu, diesel engines require fuel injection. No spark plugs, but you have to do injectors, which are even worse.
posted by Steven C. Den Beste at 12:00 AM on May 6, 2007

You could look into to rotor tip rockets, which powered the 1958 Fairey Rotodyne. Here are some recent notes (through April 17, 2007) from a guy doing his own tip rocket powered autogyro in his garage.
posted by paulsc at 7:27 AM on May 6, 2007

SCDB, what ever gave you that impression? I have a number of friends who spent years working on carburetted 2-stroke diesel piston engines that spun at 240 RPM, iirc. Doesn't get much simpler than that.
posted by TheNewWazoo at 12:19 PM on May 6, 2007

NewWazoo, did that engine produce enough power to lift a helicopter?
posted by Steven C. Den Beste at 4:09 PM on May 6, 2007

Indeed - in fact, it produced enough power to move a submarine.
posted by TheNewWazoo at 5:39 PM on May 6, 2007

Was it a little yellow submarine? Because, if so, I want one.
posted by IronLizard at 5:45 PM on May 6, 2007

I had to laugh when I read the "no CNC" part.

My job involves robotics programming for inspecting jet engine parts for defects. I have an intimate knowledge of the tolerances and manufacturing hurdles involved with turbine engines, and there's absolutely no way you're going to have any success without exotic materials or CNC machining.

My specialty is military fighter and helicopter engines, and they're pretty much all made from superalloys (Nickel based metals) and titanium, neither of which machine well with home tools. You'll find that these materials are chosen for several reasons, most notably, they don't fail with large temperature changes, they're insanely (INSANELY!) tough compared to say aluminium or steel, and they don't expand or contract wildly.

As a correction to what backseatpilot said above, modern tolerances are in mils (not millimeters). Mils are thousandths of an inch. Tolerances in turbine blade manufacturing can be less than 10 mils in 3-space. Imagine what a turbine blade looks like, then try imagining machining that to an accuracy of less than 4 human hairs in every dimension. You can't. In fact, there are non contact machining processes like electrochemical machining that do this, so as not to induce stresses into the manufacturing.

As if that weren't difficult enough, the latest innovations are blade shapes that vary with temperature. Imagine a surface shape so complex that differences of temperature cause the attack angle and dispersion surface of the blade to vary, all increasing performance at that engine temperature. Truly insane. Being somewhat old school myself, I was amazed when I realized that these were among the first parts ever produced that won't have traditional blueprints, because there's no way to convey the complexity! Astounding. We live in a computer age.

All of my ranting aside, you could probably, maybe, make a 1950s vintage engine. But why?
posted by kungfujoe at 4:44 AM on May 7, 2007