Model Turbines

I disassembled Radial Turbine 1 and took some photos and measured some key dimensions. The following photo shows the major parts and the following drawing shows the key dimensions. The turbine ball bearings are 5x10x4 (105-2Z) with shields on both sides. The speed reducer ball bearings are 5x8x2.5 (85-2Z) with a 9x.6 flange and shields on both sides.

When I reassembled Radial Turbine 1, it was supplied with three 0.1mm shims and two 0.5mm shims. It required the three 0.1mm shims between the rotor face with the blades and the outer ball bearing to barely clear contact with the outer cover. The two 0.5mm shims between the back side of the rotor and the inner ball bearing removed almost all the play and provided just enough clearance for thermal expansion or contraction. With the shims in this position Radial Turbine 1 repeated the same performance it had before it was taken apart. Although the performance didn’t match the best of the turbines I have tested, it is a very nice model. The machining is very good giving it a pleasant appearance. For someone that just wants a model turbine to run on an airbrush compressor it will run up to speeds that give a nice turbine whine with my smallest airbrush compressor and any of the propellers I used. Because of the high windage losses with the radial blades, it can’t spin fast enough without a load to do any damage even with my largest airbrush compressor. I ran Radial Turbine 1 without a propeller and with my largest airbrush compressor at its maximum output and it reached a top speed of approximately 40,000 rpm. This makes it very safe for running without a load. You must be very careful running with the propellers since they have sharp edges and run at high speeds. I plan on leaving the gear speed reducer off since I only wanted it to test the efficiency of this type of gear speed reducer. This single stage gear speed reducer had an efficiency of approximately 80% as mentioned in the 23/01/2023 post.

I'm not entirely clear about the bearing shims, but when I worked on high speed gearboxes many years ago we used bearing preload washers to lightly axially load the bearings. This prevented skidding between the balls and race, but allowed for differential expansion as things warmed up.prof Chadfock used a more sophisticated but similar setup in the Quorn

Edited By duncan webster on 27/01/2023 23:00:17

The shim washers I have been using are flat with an ID that is a close fit to the shaft and an OD that is small enough that it has clearance with the outer race of the ball bearing. These were what was used in Radial Turbine 1. My last turbines are using the ball bearings used in the Star high speed dental handpiece. The rotor in this handpiece uses wave washers to accomplish what you described. The wave washers used in the Star handpiece are available at a very low cost and would work in all my turbines using the same bearings. I've noticed some noise that I think is caused by the rotor moving back and forth in the small (<0.005" gaps. The wave washers would probably eliminate this noise and extend the life of the bearings.

Edited By Turbine Guy on 28/01/2023 15:51:07

It should be noted that although Radial Turbine 1 does not have a very high efficiency, the small steam engines typically used don’t do much better. The most common type of steam engines in small sizes use the oscillating cylinder. I found in my Testing Models thread that the Miniature Steam Models (MSM) Tyne shown in the following photo had a maximum efficiency of 6.0%. The Stuart Turner ST even with several modifications to improve the performance only got a maximum efficiency of 6.9%. These oscillating cylinder steam engines had their maximum efficiency at speeds of approximately 1,000 rpm and maximum powers of approximately 2 watts. That is a lower speed than needed for model boats. The maximum overall efficiency of Radial Turbine 1 with the 5.5 speed ratio gears was 5.7% with an output speed of approximately 2,900 rpm and a power of 1.7 watts. This output speed would probably be low enough for many of the model boat propellers. Michael Fueg used a overall speed ratio of 40:1 in his model train described in the posts starting with 11/01/2023 post shown on the previous page. These posts show how the same turbine as Radial Turbine 1 bought with a two-stage gear speed reducer was mounted in his train and links to videos that show the train running on steam. For Radial Turbine 1 with the 13.5:1 speed ratio gearbox that came with Michael’s turbine, the output power would be approximately 1.4 watts at a speed of 1,700 rpm using my largest airbrush compressor. The overall efficiency would be approximately 4.6% based on 2 stages of speed reduction and 80% efficiency for each stage. Michael’s reversing gearbox had an additional speed reduction plus idler gears that made the overall speed ratio 40:1 and would further reduce the efficiency to about 3.7% and the power to about 1.1 watts. This is based on running on air, the energy available with steam is much larger and if the steam is dry or superheated the efficiency would be about the same but the power would increase.

Since Werner Jeggli has tested on steam, I will try to confirm my last statement in the previous post that a good boiler could supply much more energy. In a test using the boiler of his PRR-S2 locomotive he obtained an average power of 11.3 watts with two 0.8mm (0.031) nozzles that had a combined throat area of 0.00151 in^2. The nozzle throat area of Radial Turbine 1 with the 0.037” nozzle is 0.00107 in^2, so his boiler could easily supply the steam for the larger nozzle used in the test with air. The steam temperature at the entry to the turbine was approximately 186C (367F) with a pressure of 3.5 bar (66 psia). The saturated temperature for 66 psia steam is 299F, so the steam was superheated 68F. The mass flow in his test was 25 g/min (3.3 lb/hr) and enthalpy drop was 115 btu/lb, so the energy available to his turbine was approximately 113 watts. This is over 3 times the amount of energy available with the tests using my airbrush compressor. The overall efficiency of his turbine in this test was 10.0%. My tests of Axial Turbine 2 that uses the same rotor reached an overall efficiency of 13.4% running on air. There are many differences between our turbines, so this only shows approximately what to expect.

Edited By Turbine Guy on 03/02/2023 15:35:12

I ran a test of Radial Turbine 1 and Tangential Turbine 5B with my Stuart 504 boiler. The following photo shows the turbines, and the following test sheet shows the results. I used the GWS EP 2510 propeller for these tests because I thought that it would keep the turbines from over speeding with the energy available. It kept Radial Turbine 1 from over speeding but I had to periodically remove and slide back in one of the wick burners when I ran Tangential Turbine 5B. Since I had to remove some of the energy when Tangential Turbine 5B started to overspeed, the energy and speed were not constant so what is shown in the test sheet for this turbine is the approximate averages. I picked Tangential Turbine 5B for comparison with Radial Turbine 1 because it has the easiest to machine rotor and has almost matched the maximum efficiency of the best turbine I have tested. This was the first test using steam for the bearings that came in Radial Turbine 1 and the dental bearings used in Tangential Turbine 5B. They both seemed unaffected by the steam. Tangential Turbine 5B spun the GWS EP 2510 propeller to its maximum operating speed of 28,000 rpm with air at 20 psig immediately after the test with steam. Comparing the performance of these turbines running on air or steam, Radial Turbine 1 seemed to get the best benefit. It had a maximum power with air of 1.6 watts that increased to 3.1 watts with steam. The maximum power of Tangential Turbine 5B was about the same running on air or steam. Radial Turbine 1 is the first of the turbines I have tested to get a substantial increase in power with steam from the Stuart 504 boiler when compared to the performance with air from the Master TC-96T airbrush compressor. The Reynold Number and wet steam losses have offset the extra energy available for the other turbines. Apparently the wide spacing and small number of blades of Radial Turbine 1 are not affected by these losses nearly as much as the other turbines.

I saved for last, the discussion of what I think is one of Radial Turbine 1’s best features. This is the way they made the nozzle a separate piece. With this type of nozzle both ends are exposed, so the drilling, reaming, and forming requirements of each end can be met. The nozzle can be converging only like in this turbine or a truly supersonic nozzle with a diverging exit. Even used as a converging only nozzle, it has the big advantage of allowing space for the gas to expand before striking the rotor. My testing and Werner Jeggli’s testing have shown the need for this space if a converging only nozzle is used for gases expanding to supersonic velocities. The following copy of a portion of a page from ‘Steam Turbines by Edwin F. Church, third edition’ explains how the gas exits the nozzle. The enlarged portion of the drawing in the 24/01/2023 post added below shows the space provided with Radial Turbine 1’s nozzle. Another advantage of this nozzle is that the hole through the wall of the housing is relatively large compared with the nozzle bore. I have broken several small drills or had them wander when trying to go through a thick wall. Drills as small as 0.6mm used by these tiny turbines are extremely delicate. If I had seen this type of nozzle before I did my last turbines, I would have done something similar.

I wanted to make some tests of my model turbines with my Stuart Twin Drum boiler. This small simple boiler is what I hoped could be used with model turbines and be competitive with the small model steam engines. This boiler did not come with a throttle valve and was intended to be used with the Stuart ST oscillating cylinder steam engine like shown in the following photo. When I made the first nozzles for some of my model turbines they had a throat diameter of less than 0.028” so that they could be run with the Stuart Twin Drum boiler without a throttle valve. The best results of any of the model turbines I ran with the Stuart Twin Drum boiler were with Axial Turbine 2 that had a nozzle size of 0.024”. When I got my larger airbrush compressor and started testing with the Stuart 504 boiler, the optimum nozzles sizes became larger, so I had increased the nozzle sizes on all my turbines. I decided to add an orifice in the outlet tube of the Twin Drum boiler to act as a throttle and run Tangential Turbine 5B and see how the throttled performance compared with the best I have got without the throttle. The following test sheet shows this comparison. The output power turned out to be about the same even though the throttling lost over half of the energy available to Tangential Turbine 5B. This illustrated the increase in turbine efficiency running at sonic velocity compared with running at supersonic velocities. The axial turbines ran better on steam than my tangential turbines because a space between the nozzle and the rotor could be provided that would allow the steam exiting the nozzle to expand as needed to go supersonic. Moving the axial rotor away from the nozzle worked but was a compromise. Even with the compromises of the turbines shown in the first two tests of the following test sheet, they appeared to exceed the performance of the Stuart ST oscillating cylinder steam engine the Stuart Twin Drum boiler was intended to be used with. The test I made with a Stuart ST steam engine with everything the way it came and run with the Stuart Twin Drum boiler; it turned an APC 8x6 propeller at approximately 1,500 rpm. This propeller requires 0.4 watts to run at this speed according to the manufacture’s propeller performance sheets. The turbines would require a 12:1 speed reducer to run at the Stuart ST speed. This would probably require two stages and assuming the 80% efficiency per stage I found with the test of Radial Turbine 1, the power of the turbines would be reduced to 0.7 watts. This is still quite a bit better than the 0.4 watts of the Stuart ST steam engine.

The third test was running the steam without being throttled. Even with the large nozzle size, the flow out of the boiler was steady but at a very low pressure. Although the overall energy output of the boiler was less than half than with throttling or a small nozzle, the power and efficiency increased significantly. These tests indicate that a better way to achieve supersonic velocities is needed to run at the higher pressures.

The type of nozzle shown in the last post would provide this space without increasing the leakage caused by the gaps and should help any of the turbines that run on supersonic air or steam.

Edited By Turbine Guy on 25/02/2023 18:50:11

I am going to try to add an insert like shown in Detail C in the following drawing to Tangential Turbine 5B and call this configuration Tangential Turbine 5C. This will give a minimum distance of approximately 0.050” for the gas to expand to supersonic velocities before contacting the rotor. I’ll start with the throat diameter of 0.020” shown in the drawing and increase the size in steps until I get the pressure below 40 psig running on the Stuart Twin Drum boiler. I believe this will increase the performance shown in the last post for Tangential Turbine 5B.

I tried several times to make the insert shown in Detail C of the drawing in the last post. Each time the drill broke in the bore of the insert. Apparently, the wear of my 37 years old lathe has reached a point that I can’t keep the drill and insert in the necessary alignment. I was able to turn the OD to the correct diameter, make the external and internal tapers, and machine the insert to the correct length. The internal taper was made with a center drill with a 60-degree taper and a 0.020” drill bit size. The maximum length the drill had to pass through was less than the 0.110” shown in the drawing of the last post. The following photo shows the furthest I was able to get. The drill broke the instant it passed through the insert. There was a very tiny amount of the end of the broken off peace extending beyond the end of the insert. I bumped the extended part with a hammer hoping to break it loose but even though the end moved back to even with the face of the insert, it wouldn’t break free. The insert is so small that the photo required a large amount of zoom, but I think you can see that the drill had moved off center. The maximum drilling depth given for the 0.020” drill was 0.150” so I didn’t exceed that recommendation. I ran the drill at 4,000 rpm which is the highest speed of my lathe and moved the drill into the insert very carefully. After four tries, I must concede that this is beyond my skill and the accuracy of my lathe.

I was able to purchase a length of surgical tubing with a nominal OD of 0.094” and ID of 0.024” made of 316 stainless steel. The actual ID was 0.027” so I thought that it might be small enough to work for Insert 5C that was described in the last post. Even though the tube was welded, the inside and outside were very smooth and concentric. Having the hole all the way through eliminated the problem of breaking the drills I described in the last post and the insert was very easy to make. The following drawing of Tangential Turbine 5C shows the updated dimensions and the following test sheet shows the performance running with the Stuart Twin Drum boiler. Allowing the space for the steam to expand without increasing the clearance of the sides or top of the rotor appeared make a substantial improvement.

Edited By Turbine Guy on 08/03/2023 21:03:22

I noticed that the nozzle size shown in the drawing and test sheet of the last post were different, so I measured it again. The correct size is 0.029” as shown on the test sheet. The following drawing of Tangential Turbine 5C has been up dated to show the correct size.

When I ran the test of Tangential Turbine 5C shown in the last post, I noticed the speed was periodically dropping and then returning to the maximum speed. This occurred several times until late in the run and then the speed stayed at the maximum until the water ran out in the boiler. This indicated that water was being carried over during the first part of the run. I checked the average moisture content for the entire run and it was approximately 20%. During the time when the speed was increasing and decreasing the maximum speed was about 25,000 rpm. The maximum speed when it stayed steady, was the 27,000 rpm, shown in the test sheet. This change in speed was due to the steam being drier at the last part of the run.

The performance of Tangential Turbine 5C appeared to exceed the performance of the Saito T-1 steam engine described in the 12/10/2020 Post of the Testing Models thread. This test of the Saito T-1 steam engine with everything the way it came and run with the Stuart Twin Drum boiler; it turned an APC 8x6 propeller at approximately 2,400 rpm. This propeller requires 1.7 watts to run at this speed according to the manufacture’s propeller performance sheets. Tangential Turbine 5C would require a 11.3:1 speed reducer to run at the Saito T-1 speed. This would probably require two stages and assuming the 80% efficiency per stage I found with the test of Radial Turbine 1, the power of Tangential Turbine 5C would be reduced to 2.4 watts. This is still quite a bit better than the 1.7 watts of the Saito T-1 steam engine that was the best performing of any of the small steam engines I tested.

On 25/02,Turbine Guy refers to gear efficiency as 80%. This seems low to me, when I worked on gas turbine gearbox design we assumed 97% for each stage. Of course they were very well made gears,and running in ball bearings.

Duncan, I agree the 80% efficiency is way below what well made and precisely aligned gears running on quality ball bearings would get with higher power outputs. The tiny turbines I have been testing have so little torque that grease lubricated ball bearings like used in the gear reducer that came with Radial Turbine 1 have a much larger effect on them. I ran a test of Radial Turbine 1 with and without the gearbox described in the 23/01/2023 post on page 22 to confirm this.

My original intent of adding Insert 5C to Tangential Turbine 5C was to get the pressure up to 40 psig with the Stuart Twin Drum boiler like I was able to do with the insert added to Axial Turbine 2 shown in the test sheet of the 08/03/2023 post. As I mentioned in the previous posts, I was not able to get the nozzle size down to the 0.024” diameter Axial Turbine 2 had when it ran at 40 psig with the Stuart Twin Drum boiler. I traced the problem of not being able to drill the nozzle size down to 0.024” to misalignment between the headstock and tailstock of my lathe. I found a used tailstock for a Unimat 3 lathe that was described as being in excellent condition on Ebay. I purchased this tailstock and hope that it will get the alignment good enough to drill the very tiny holes. If the new tailstock allows me to make an insert with a nozzle size of 0.020” as I originally planned, I will start with that size and make tests with very small increases in size until I find what works best. I had no problem removing the existing Insert 5C and Turbine 5B performed the same as before I tried the insert, so I have nothing to loose trying the smaller nozzle sizes.

I would like to find out why Tangential Turbine 5B with the overlapping pockets almost matched the performance of Axial Turbine 4A with the traditional impulse blades. I based the design of Tangential Turbine 5B on the guidelines given by Dr. Balje in his report ‘A STUDY OF HIGH ENERGY LEVEL, LOW POWER OUTPUT TURBINES’ made in 1958. The rotor used in Axial Turbine 4A was made by Mike Tilby and based on later guidelines that match or exceed the guidelines given by Dr. Balje. I made the following drawings that show the details needed to compare these turbines. Dr. Balje compared the performance of Axial Impulse turbines with Terry turbines that are tangential flow but with blades. His report showed that the Terry turbine could match the performance of Axial Turbines with single nozzles and a small number of blades running at low speeds. I used the guidelines given for the Terry turbines for the design of my tangential turbines and have found that my overlapping pockets performed better than the non-overlapping pockets used by Stumpf turbines. I have not found anything showing the use of overlapping pockets although I doubt I’m the first to try this. You can see on the following drawings that the size of these turbines are almost the same and the energy available to the turbines shown in the following portion of the test sheet on the 23/01/2023 Post is also fairly close.

The first item I looked at for comparing the performance of Axial Turbine 4A with Tangential Turbine 5B is the nozzle efficiency. Both of these turbines have a 60 degree included angle at the nozzle entrance. The flow coefficient shown in the following chart that was copied from ‘Nozzle geometry variations on the discharge coefficient’ by M.M.A. Alam, T. Setoguchi, S. Matsuo, and H.D. Kim made in November 2015 shows a discharge coefficient of approximately 0.93 for a conical convergent nozzle with 30 degree per side as shown in the following drawing. That flow coefficient would apply for a nozzle with a very short throat length. Axial Turbine 4A has a relatively long nozzle throat length of 0.222” as shown in the drawing of the last post resulting in an estimated 1.7 psi pressure drop reducing the nozzle discharge coefficient to approximately 0.80. This shows the importance of keeping this throat length as short as possible. Tangential Turbine 5B has a nozzle throat length of 0.113” as shown on the drawing of the last post resulting in an estimated 0.6 psi pressure drop reducing the nozzle discharge coefficient to approximately 0.88. Both of these turbines nozzles have a spouting velocity close to sonic velocity so the losses due to expanding to a supersonic velocity are negligible.

The next item I reviewed for comparing the performance of Axial Turbine 4A with Tangential Turbine 5B is the rotor velocity coefficient. I used the method described in the 21/12/2019 Post to determine the stall torque and the equations shown in that post to determine the rotor velocity coefficient. I made some improvements to the test setup. The first was to use a lighter weight at a longer distance as shown in the following photos. This reduces the error in measurement of the distance from the center of the turbine to the point the weight is attached. The second improvement was to use thin string to hold the weight. This allows the whole length of the load to be close to inline. The weights were the only thing keeping the turbines from spinning in both these pictures even though they were at full test pressure. I use propellers for the lever arms since they are balanced in weight each side. I made work sheets to tabulate the results that I will shown in the next post.

The air pressure used in the test of Axial Turbine 4A described in the last post was the same as shown in the following table. The weight of the load used was 0.094 oz. The radius to the load was 3.32” so the maximum static torque was 0.312 in-oz. I did the same test with Tangential Turbine 5B with the same weight and using the pressure shown in the following table. The radius to the load was 2.84” so the maximum static torque was 0.267 in-oz. This large difference in maximum static torque surprised me since these turbines spun the GWS EP 2508 propeller to the same speed of 28,000 rpm. I ran both turbines with the APC 8x6 propeller I used for the static test and they both had a maximum speed of approximately 1200 rpm. Both turbines ran both the large and small propellers at the same maximum speeds even though they had a relatively large difference in maximum static torque. I will try to explain in the next posts how this could happen.