I can make a small double turbine wheel by getting a 30 mm dia inconel
blank, putting it in an indexing table on a cnc mill, and cutting the blades
using a program.
Side on the wheel looks like this:
->-> -direction of shaft rotation
|| - - - - - - - shaft
|)) ) ) ) ) ))| - - -> 1st turbine wheel
|)) ) ) ) ) ))| - - -> 2nd turbine wheel
But how do I make guide vanes?
They are like a short tube, about 35 mm outside and 25 mm interior diameter,
with vanes cut on the inside surface.
It works a bit like this, gas flows down:
->-> -direction of shaft rotation
| || |
|\\ \ \ \ \ \\| -nozzles, the gas flows that way \ down
|)) ) ) ) ) ))| ->-> -first turbine wheel, the gas flows / down
|(( ( ( ( ( ((| -stationary guide vanes, the gas goes / down
|)) ) ) ) ) ))| ->-> -second turbine wheel, gas now goes \ down
|// / / / / //|
| || |
I can think of a few ways around it, but I'd like to make them like that -
is it possible?
What is the purpose of the guide vanes? Is it just to "straighten out"
the air flow? Or do they create turbulence giving the second turbine
wheel something to "bite on"?
If you are smoothing the airflow, would an analog be something like
rifling in a gun barrel? Assuming that the guide vanes need to cover
the distance from the trailing edge of the turbine 1 (T1) to the
leading edge of turbine 2 (T2), AND that these vanes follow a helical
path from T1 to T2, then your CNC progran shouild be able to handle
the X, Y, Z, & W to trace the path of the vanes. I'm making another
large leap and assuming that you are soliciting ideas about how to
actually cut the vanes. If you were to use something like a shaper
attached to your indexing rotary table, you could push a tool point
from T1 to T2. Do this enough times, and you would have vanes.
Again assuming these vanes are continuous and follow a helical path
from T1 to T2, then you could think of them as a very long period
multiple groove thread. Imagine a tap cut for 30mm x 10 threads. The
tap is about 30mm in diameter and cuts 10 threads over the course of
30mm of travel. Now imagine a 30 x 0.1 tap. Again the same diameter,
but you would only get one thread in 300mm. Now you have 1 vane, so
add a potload more teeth so that you are starting say 94 grooves at
the same time. You would now have a vane every 1mm around the interior
circumference of the hole.
There is probably some rule that says a specialized tap like this
should be called a broach, but whether it is called a spade or a
shovel, you can still dig yourself into a hole.
Just a thought,
Neither. The nozzles start the gas spinning in an anticlockwise direction
(looking at it from the top). This puts clockwise torque on the nozzle and
vanes. The turbine changes the airflow to a clockwise direction, this puts
anticlockwise torque on the turbine. The guide vanes then turn the air back
anticlockwise, again putting clockwise torque on the vane assembly, and the
second turbine ... and so on
So there is clockwise torque on the nozzles and vanes, and anticlockwise
torque on the turbine. The nozles and vanes are held stationary, so the
The last set of vanes shown would not normally be used in a rocket motor
where the output gas is discharged, and it can usually spin if it likes -
but I am using the gas again, and want it flowing straight.
Turning the airflow rather than smoothing it - and that's not a bad analogy.
Sadly, no. The top and bottom edges of the vanes have to be thin and sharp
to catch the gas flow smoothly, and the middles are thicker. It's a bit more
complex as this i not a pure reaction [*] turbineand the blades are wider on
the bottom, but that part can be done by the cnc milling cutter.
[*] where the gas is expanded before it gets to the turbine wheel itself,
and the gas doesn't change pressure or temperature afterwards and the
turbine just extracts the kinetic energy which the nozzles have impartd to
the gas. In this one the gas expands while in the vanes as well, to increase
the speed of the gas for the second turbine.
BTW there, I am planning 3 versions of this engine - 19, 26 and 39 mm
turbine diameters. The 39 mm dia one should give 1 kN / 220 lb of thrust,
and extracts ~ 7 kW of power from the gas flow. The turbine assembly
(turbine/vanes/nozzle/casing/shaft/one bearing/seals/gas generator) for that
should weigh about 150 grams, but it will probably end up around 250 grams.
Somewhere in the 15-30 HP per lb region. I know the professionals can get
100 HP/lb and more, but that's enough for me for now. Why couldn't I have
set myself an easier target? Ah well, that's rocketry.
I'm pretty sure it will be able to do that.
( I haven't bought a cnc mill yet, still saving up, Got £360 actually put
aside, and will perhaps spend another £500 when I buy. Got a zillion
computers already, and I used to be a bit of an electronics wizz back in the
old days when fixing TV's and VCR's was worthwhile and profitable, so I plan
to build my own interface.
Computer power supplies can be really cheap, and if you know what you're
doing they can be adapted for motor power supply, so it's just the mill now.
Taig... Chinese ... Boxford CNC??.. I'm not going for old iron for this
though, I need small capacity but high accuracy in hard metals like
stainless and inconel. and CNC. Advice / suggestions / experiences
Some nice ideas, but not with the required shape of the vanes. I'll try,
this is enlarged and just shows three vanes edge-on, look at it a bit
. \ \ \ .
. |, |, |, .
. )) )) )) .
. |' |' |' .
. / / / .
The tops and bottoms are thin, the middles are thicker. It's quite easy to
cut the very similar turbine blades when you have access to the outside, eg
put a blank in a rotary or indexing table on a mill table and cnc-XYZ the
individual blades, rotating the table for each blade - but when you only
have access to the inside?
I have, thanks.
The better ones cast their inconel wheels and stainless vanes, or rather
they have them cast and supplied with x-rays and certification - all good
stuff - except I am still in the experimental phase, and the minimum order
is 100 wheels (or 20 if you pay nearly the same amount).
I am thinking of developing my own casting capability to inconel and
stainless, it's only about 1100 C at the moment - this is at least partly
for other reasons - but for now I want to cut one wheel at a time to it's
own unique pattern and experiment with it.
I have though about using their ready-made wheels, but first they are a bit
cagy about selling them, and second they are designed completely differently
- for instance if you look at one of their wheels from the top, the blades
do not overlap - high pressure ratio turbine blades overlap two or three
BTW, the early developers of the mini-turbojet used wooden compressor
wheels, and cut and bent their turbine wheels from sheet. Still got a
thrust-to-weight of two or more. Impressive. Very.
When making the diaphragms for impulse turbines at work we cast the vanes into
an inner diaphragm and then weld the inner diaphragm into an outer diaphragm
with submerged arc welding. It may be possible to make copper shields to form
the inner, outer and upstream faces of the diaphragm, insert the vane tips
through holes in the inner ring and then MiG or submerged arc weld directly
onto the blades. Use a motorised rotary table to get an even deposition rate.
Make the vanes with round pins at inner and outer ends, this allows you to set
the blade angles with a jig and makes the fixturing simpler, this is what we
do for development test turbines.
If it would help, I've got a few kg of 1mm Inconel 625 MiG wire that I
'rescued' from work (brilliant for stainless bike exhausts), although that
might be a bit on the thick side for your needs.
I'd thought about that, and came to the conclusion that it's probably the
only way unless I cast the whole thing.
Fixing the blades to the outer shell/diaphragm is now the problem - two
possibilities I can see, either tig welding (I'm not bad at that, but even
so I don't think it would be at all easy) or diffusion brazing - which seems
a better prospect, there isn't a lot of strength needed (compared with eg
the turbine), and the temperature is not that high.
I don't really fancy that, it sounds a bit too complex for me.
Interesting idea, although I think I'd still prefer to machine the blades
and inner diaphragm in one piece from solid. I'll have to think about that -
how big are the turbines you build?
Yes, although perhaps it might be useful as a tig filler wire. I shall email
you offlist, and thanks.
What would be really useful in the materials field is a reasonable source of
Inconel, one that doesn't either have a huge minumum order or charge the
Think MIG but with powdered flux instead of gas.
The smallest would be boiler feed pump turbines, 18" diameter wheel, but the
live steam (as opposed to lower pressure bled steam) blades are only 1/4"
high! The biggest wheels we see are typically 14' in diameter with 46" blades,
remember that these are doing 3000 rpm, force on the blade roots can be
probably make goof TIG filler, too big for little MIG work, although it worked
in the bike (Yamaha didn't make the headers any more, got some 35mm 303
stainless but couldn't bend it so used two compound mitre joints one each
Most of what I can get it High temperature steel... good for 560 centigrade
with a good (decades) creep life, but not necessarily good for gas turbine
use. It's a right bugger to machine though, it seems to work harden a few
millimetres before the tool gets to it.
I know exactly what you mean .. I'm almost convinced that somehow-sharp
moving air from the hard, sharp tool edge does it. The only alternative I
can think of is that the metal knows it's going to be worked, and cringes in
These turbines don't need that kind of creep-free life. An hour's total life
would be more than enough, they will operate for maybe 30 seconds per flight
maximum, more like 5 seconds usually, and they will get lost long before
they complete 120 flights!
They will operate at at least 700C though, and higher would be better.
I'm thinking of stainless for the less-stressed parts and Inconels for the
turbines, and also looking at an electroformed cobalt-tungsten-iron-nickel
alloy that is supposed to have tensile strength of 57 kpsi at 1,000 C.
Only problem with the last is I can't work out how to electroform the
required shapes (yet). But it looks interesting!
This seems to be a standard gas turbine, using a combined impulse /
reaction turbine blade shape, with an impulse shape at the root and a
reaction shape at the top.
In gas turbines the stator blades and NGVs have two purposes.
They direct the gas onto the next turbine disc and
they form a convergent duct to change the low velocity high pressure
gas coming from the combustion chamber or previous turbine disc into
high velocity low pressure flow onto the next turbine disc. (Bernoulli
The different behaviours are usually dictated by the turbine blade
shape, rather than the stator shape in small gas turbines.
The clearance for the rotor blades is critical, but as the stator
blades are fixed and have no centrifugal load, they can be manufactured
individually and mounted in an inner and outer ring. This ring can be
split so that each half can be mounted round the rotor armature . The
inner ring doesn't touch the armature and the outer ring is clamped by
the engine case.
I'm no expert, but would it be possible to cut the vanes as follows:
Assuming they are circular, with the concave side having a larger
radius than the convex could cylindrical cutters, sort of like a hole
saws but without a pilot arbour be used to cut from the side. The
following diagram explains hopefully.
If the diameter of the saws is large enough, they should be able to
clear the adjacent blade
Please ignore my last two comments... I'll learn to read the whole
thing before shoving my oar in.
I'd be inclined to cut each blade segment individually and then fit
them together inside a ring - if they are static guide vanes then that
shouldn't be too much of a problem. This has the advantage of letting
you easily replace damaged blades.
... and me with an aeronautical degree too!!
I thought about doing that for the turbine wheels, it would save a lot of
CNC work. The technique could probably be used for cutting the vanes too -
they will almost certainly have to be cut from the outside anyway.
However there is a problem, the bottom of the gap between the vane blades is
not flat / parallel to the axis - the gap gets bigger as it goes towards the
I might be able to redesign the blades so they can be made that way though,
I have been thinking about it. The limitation to cylindrical surfaces only
isn't that big a deal. Within reasonable limits the efficiency of the
turbine doesn't matter very much - nothing gets wasted.
Actually, looking at my own diagram, that would mean that the trailing
edge of the vane is too thick. If the right hand side of the trailing
edge were cut right at the edge of the cylinder and then the inner side
(back as you look at the diagram) were then cut cylindrically....
...I might be confusing the issue, but it's been fun trying to work it
Scrub all of that... making the blades in one piece is going to be
difficult. What about making them as per my first idea (i.e. two
cylindrical surfaces) and then cutting the cylinder at the required
angle each side. The resulting blade could then be attached to the
casing by some mechanical means...
... this would mean that the annulus surrounding the blades could be
easily made continuous and you could fairly easily alter the blade