Howto Debug and Start a Homemade Turbojet Engine

First of all I am glad to thank all of You for the persistent interest to the Homemade Tin Can Turbine With 3d-printed Compressor. This instructable is devoted to putting this thing alive once You have made and assembled it. People's experience shows that this process may be not as obvious and straigtforward as I thought earlier, that's why I decided to make a new instructable to provide You with the idea how to do it. Lets get started.

We will need:

• a several feet of garden hose
• some wooden planks
• a glass jar with a tight lid
• an ink
• an (optical) tachometer
• a high speed drill or a hand engraver with regulated rpm

To debug the compressor stage we will need to know how much pressure it can provide inside the engine at a certain rpm. Usually this pressure is well below 0.1 bar and most of the commercial pressure gauges are out of use here. So we have to make our own manometer.
Its principle is simple (look at the 1st picture). If we took an U-shaped pipe and fill it with water, and if we blow into one end of the pipe while keeping the other end open, the water will drop at the side we are blowing in, and it will rise at the opposite side of the hose. The difference of the water levels can give us the information about the difference between the pressures at the inlet side of the pipe and at the open side. Look at the picture for the formula. Rho means the density of the liquid there. The density of the water is 1000 kilograms per cubic meter (MKS).

This pressure difference is oftenly called "head" or "pressure head" and in most cases it is the value to be known. Usually one does not need to know the exact absolute value of the inlet pressure, and thus one does not need to measure the ambient pressure. If one uses the formula and substitutes there the units in MKS system, one will get the pressure head expressed in Pascals. However it is much more comfortable to express the pressure directly in centimeters of water column.

From the practical point of view of building and usage, the U-shaped design of the manometer is not the best. There exist more robust design, based on a jar. To get an idea, how to make it, look at the second picture. The details of the design feel free to elaborate Yourselves. One possible practical embodyment is shown at the next pictures. Note that for better visibility You can colorify the water by some ink.

Hint: take a special care to avoid leaks of pressure, especially between the jar throat and the lid. The reason is that if You will use a fuel line of the jet engine as the pressure port (as I usually recommend to do) the tiny fuel spray holes can not provide enough of air if Your manometer is lossy.

Remember, that our goal is to determine the pressure head that the compressor is able to provide at certain rpm. It means we cannot use an external fan to rotate the shaft. Otherwise we would be measureing the ram head of the stream from this fan rather than the one provided by the compressor. So we need an other reliable mean to rotate the shaft. One can use a high speed drill or a hand engraving tool or some electric motor here. Keep in mind, however, that You can not use a rigid clamp between the external motor and the shaft. If You try one You will destroy Your drill or Your turbine (don't even know which one is worse).

Refer to the pictures above to get an idea how to make a soft connection between a hand engraving tool and the shaft just by means of a piece of silicone hose.

Arrange Your equipment as shown at the 1st picture. Note the lid, closing the exhaust of the engine. We are here to test the static pressure head provided by the compressor. The dynamic pressure tests are much more complicated so we will leave them alone for a time. To perform the static head test we need to stop the air flow through the engine. The turbine and NGV alone can not provide the necessary air resistance, because the air is cold in this test.
If Your engine has a special air port for pressure measurements it is better to use it. Otherwise use the fuel line to connect the water manometer as shown above. Another note: at this step it is better if Your engine has a special window for rpm measurements. If it has not, one may try to aim the rpm measurer at the drill's spindle, but in this case the measurements are complicated by the fact that the engine's shaft may slip behind the drill's spindle.

Spin up the jet engine's shaft to 13 000 rpm. Check the pressure head shown by the water manometer. You should see something like 10-12 cm of water rise. The best value, I got from this design (Homemade Tin Can Turbine With 3d-printed Compressor) was 13.5 cm of water column @ 13 krpm. Usually I get something around 12..13 cm. From the practice I can say that engine runs good when this value is above 12 cm. If it is below 11 cm the NGV and turbine vanes become very hot. And I was unable to get the engine to self sustain if this value was 10 cm and below.

If You can not obtain exactly 13 krpm on the shaft, don't worry: use the law of squares:

 P(rpm1)     rpm1^2
--------- = --------
 P(rpm2)     rpm2^2

or

                     rpm2^2
P(rpm2) = P(rpm1) * --------
                     rpm1^2

In our case let's take that rpm1 = 13 000 and P(rpm1) = 12 cm.

If, for example, You were able to spin it up only to 9 krpm, You will assume that rpm2 is 9 krpm and hence the pressure corresponding to 12 cm @ 13 krpm is now P(rpm2) = 12 cm * (9krpm/13krpm)^2 = 5.7 cm of water column. Since 12 cm @ 13 krpm was a good level of pressure for this type of compressor, it means that 5.7 cm @ 9 krpm will be as good. You can also say that 6.2 cm @ 9 krpm will be excellent and 4.7 cm @ 9 krpm will be bad. From this example You can also see that You should not use too low rpm in this test, due to possible errors of rpm and water height measurements.

If the test above shows that compressor in Your engine performs well, You may proceed further. But what to do in the other case? What are the causes of the low pressure?

1st of all: CHECK THE TIGHTNESS OF THE MANOMETER. If it does not leak badly, then

• check the tightness of the engine itself. It does not need to be vacuum tight, but too large holes may ruin its operation. If there are ones - try to seal them. High temperature joints are good to be sealed with kiln sealant. Low temperature ones may be covered just with a duct tape.
• The other reason for the low pressure may be the low quality of 3D-printing. Too "hairy" parts may cause too high aerodynamic resistance. Arm Yourselves with a knife or filer and try to make 3D printed parts smoother. The most sensitive elements are rotor and stator vanes. The quality of shrouds surfaces is of secondary importance (but still influences).
• Yet another reason of the low pressure maybe longtitudinal error of the impeller installation. The impeller (the fan wheel) forms an air channel between its top and bottom plates. The similar channel is also formed between the stator's outer casing and its inner body. Both channels should match. Look at the picture above. (warning: the picture is simplified. Most of the unimportant for the current task details are omitted.) Since the stator channel has flared lips the system is rather tolerant to the longtitudinal misalignment. In fact it can endure a misplacement of +-1 mm (14% of the channel's height!) and still be operational. (Compare to the typical tolerance of +-0.1 mm for the common turbine designs of this size.)

Since further actions require ignition, it's higly not recommended to perform them indoors.

• Find a well ventilated fire safe place.
• Install Your jet engine there securely.
• Equip Your propane tank with a gas regulator and connect it to the fuel port of the engine.
• Connect the lubrication system (oil) if Your design uses one.
• Connect a water manometer if Your design has an auxilliary pressure port.
• Prepare all the other needed things to be in range of hand access. Don't forget the (starter) leaf blower, butane torch, tachometer (and IR thermometer if You have one).

Now ignite the butane torch and, while keeping its flame near the exhaust end of the engine, begin to blow from Your leaf-blower into the engine's intake. Don't be too agressive, otherwise too strong air stream from the blower won't let the flame to pop in to the combustion chamber. At this step You only need to spin up the turbine to maybe 3000 - 5000 rpm, not more.

When done, begin to open the propane regulator. At some setting You will see a blue flame behind the engine, just like a gas fryer has. Try to periodically increase and lower the propane flow to provoce the flame to pop in. It may be also helpful to wave the blower to some extent (refer to the 1st picture).

You should also know that the engine has four modes of burning:

• gas fryer mode (second picture, part A),
• combustion between NGV and turbine (second picture, part B),
• normal operation (second picture, part C),
• partial flameout (second picture, part D).

Gas fryer mode is usual for the beginning of starting process. When the flame pops in normally, the engine goes to the normal operation. The flame goes to the combustion chamber and sits there. One can note a prominent change in engine's sound and sudden and significant increasement of rpm.

One should avoid popping the flame to only the space between the NGV and turbine wheels. If this happens You have 3 to 5 seconds to shut the propane flow. If You don't the turbine wheel will heat up to red yellow glow and the wheel will become corrupt. In normal mode the increasing of spinning speed allows engine to suck more air and it does not happen.

If Your turbine is hot as hell but speeds up like lazy cow, or does not speed up at all, You probably have partial flameout (second picture, part D). The gas continues to burn between NGV and turbine, resulting in high temperature and poor torque. The reason is that the gas has too low temperature at the approach to NGV, and thus it accelerates poorly. On the other side it continues to burn and becomes too hot when approaches the turbine. There's nothing You can do with this with only altering the starting procedure. If You suffer from partial or full flameouts You will need to debug the combustion chamber (its liner at the first order). Refer to the next section to learn how to do this.

The last mode to mention here is (full) flameout. Usually it happens like this: You increase rpm gradually and suddenly the engine quenches. The engine slowes down and You panically catch the leaf blower to coll the engine down. The reason of (full) flameouts is again in improper combustion chamber assemblage. Please refer to the next section for how to fix it.

To avoid flameouts first of all check if You have placed the fuel spray ring too low (downstream). The fuel needs some space to burn out completely, and this space is comparable to the combustion chamber length in engeine of this size. So try to place the fuel ring as high (upstream) as possible.
Also note, that if You direct the fuel spray holes downstream the engine, the length of flame will increase greatly. Their direction should be orthogonal to the engine axis.

Third thing to note is whether You have installed the fuel spray ring in front of liner holes or not. If You did so, the wind of air coming through the holes will tend to blow the flames out (refer to the first picture of this section).

Check whether You have made the combustion chamber according to the drawings (Homemade Tin Can Turbine With 3d-printed Compressor. Step 14). The sizes of holes and their positions are not arbitrary. Make them too narrow and the combustion chamber will have too much of air resistance for the engine to operate successfully. Make them too large and the air flow will ignore most of them. Yes, the tolerance here is not that 1/10 mm, but still try to reproduce at least the sum area of each belt of holes and the position of each belt of holes as accurately as You can.

The ultimate mean against flameouts is to use a flameholder. In the simplest case it may be an (incandescent) wire just below (downstreams) the fuel spray ring. Look at the second picture of this section to get an idea how it can look like. Even a simple wire flameholder makes the combustion so efficient, that even an engine of this size has a few cm rezerve (the combustion chamber can be shortened by a few cm without affect on performance). The main shortage of the flameholder that it tends to burn out. I haven't yet invented a design that would live more than for a few minutes safely.

Now You should be skilled enough to handle the ignited engine properly. At step 3 we made a static test of its compressor operation. We could not test it in full flow mode, because the turbine has too low resistance for the cold gas. If one wanted to make an attempt to measure the pressure inside the cold engine with its exhaust open, the result would be close to zero because the compressor wuld be "short-circuited" to the ambient air.

But now, when we are ready to use hot gas, we can perform the dynamic test and check whether the compressor does match the turbine or not. For this test You need Your engine to be equipped with an additional pressure port however. The fuel line will be occupied with fuel flow here, and cannot be used for the purposes of pressure measurements. You will probably have to make a pressure port for measurements if Your engiene does not have one already. Good news here are that the turbine rarely requires any debugging, and most probably You already have an operational engine after passing through the previous steps.

To perform a dynamic test set up the things as shown at the first picture of this section. Then spin the engine up with an external motor to some 10 000 ... 13 000 rpm. Put the ignition butane torch close to the exhaust and open the propane tank regulator slowly. Try to adjust the fuel flow to the value, when the flame pops inside the combustion chamber. Observe the pressure rise. Continue to spin the shaft with an external motor. Measure the rpm. Since the procedure needs too many hands it may be wise to ask some person to help You at this step.

By gradual heating the exhaust gases more and more (by careful increasement of the fuel flow) try to obtain the inside pressure more than 8 cm of H2O column at 13000 rpm. Again as at step 3 don't worry if You can not set 13 kilo rpm exactly. Use the law of squares instead. E.g. if You achieve 14.5 cm at 18100 rpm it corresponds to 7.5 cm @ 13 krpm. A bit low but the engine does usually work with this.

The best value I've seen at this step was 11.7 cm of water column @ 13 krpm when the NGV vane was dark red. The critical value is somewhere 7..7.5 cm. 8 and above should be OK.

If You are able to get the proper pressures with reasonable temperatures of NGV and turbine wheels - grats with working engine. If You fail to get it to self sustain, try to increase the rpm when starting and try to give it more fuel. At some point it will start.

If You are UNABLE to get the proper pressures at reasonable temperatures the things are a bit more complicated. The quality of NGV or turbine assemblies may be too low.
Check for the quality of NGV assemblage. Its shroud (the steel stripe wrapped around it) should sit on the tips of blades tightly.

Checks for the gaps between turbine and its shroud. The engine is designed to operate properly when the spacing between turbine blade tips and the surface of the shroud is not over 1 mm (when at room temperature).

Checks for other leaks and possible gas losses.

Finally check for the correct angles of NGV and turbine blades. The simplest way to do this is to photograph the NGV and turbine wheels from the edge and to check the angles on the snapshots in some drawing software. Check the alfa_ngv, beta_ngv, beta_t1 and beta_t2 angles as shown at the 1st picture of this step. Compare these angles to the ones in press molds (refer to the Homemade Tin Can Turbine With 3d-printed Compressor. Step 3 for the stl files). If the angles are gone - press the discs again.

Here I will use the chance to upload the 3D models of the most recent version of the turbine. Here are the parts.
The file heap is a bit messy, unsorted and uncommented, but if Your are familiar with the previous version (Homemade Tin Can Turbine With 3d-printed Compressor) You will certainly understand what to attach to what. This version uses 10 mm shaft, ballraces 6800 and its impeller is carbon enforced (note the grooves to place carbon fiber wrapping there). It also uses three (not two!) ballraces along the shaft.

The result is: it can spin up to 70+ krpm (kilo rpm), the internal pressure reaches 270 cm of water column. The estimated thrust (no direct measurements, sorry) is in range of 1 kg.

Don't ask me to make a separate instructable for it. It is just an intermediate version on the way to higher rpm and thrust. Further on I want to open the angles a bit to obtain more thrust at given rpm.

As it was said in the previous instructable (Homemade Tin Can Turbine With 3d-printed Compressor) this is a REAL GAS TURBINE and it REALLY can spin up to the self destructive rpm. The video above is about how it can explode. Do note that it is pure mechanical explosion (neither chemical nor electrical one) when the kinetic energy of spinning wheel suddenly goes to the kinetic energy of flying fragments.
When using the turbine please wear protective googles or shields and some protective wear (gloves at least). It's also a good idea to have a fire extinguisher nearby.

If You have problems with vimeo You can watch the video at https://flic.kr/p/2hLrKpA