Model Turbines

I received the printed aluminum 3 Blade Rotor 2 described in the 04/05/2023 post from Shapeways. The following drawing shows the actual dimensions compared with the design dimensions and the following photos show a closeup of the blades. Shapeways tolerance for printed aluminum is +/-0.2mm (+/-0.008" ) for dimensions less than 10mm and +/-1.5% for dimensions over 10mm. The width of the channel varied from a maximum of 0.035" to a minimum of 0.028" so the variation was within tolerance but the minimum width was over 0.008" less than the design width of 0.040". You can see in the second photo the difference in width of the channel each side of the blade. The third photo shows the biggest problem where parts of one of the blades don't line up. Since this print does not meet Shapeways tolerance and has the error in one of the blades, I could have them make another print. I decided not to try another print of 3 Blade Rotor 2 since the variation in width of the channel would be too large even if it is within tolerance. The precision required for the channel of this type of rotor would probably require CNC machining.

Edited By Turbine Guy on 22/05/2023 12:55:41

I looked through Shapeways guidelines for printed aluminum and found that the minimum thickness should be 1mm. The thinnest part of the blades shown in the photos of the last post had a design thickness of 0.5mm. This is probably what caused the problems. I got some quotes for CNC machining that were over $200 and had tolerances of +/- 0.1mm to +/-0.13mm. I have used printed nylon parts for housings and covers that have worked well if they were made thick enough to withstand the loads and limit the deflections. I decided to try printing a nylon version of 3 Blade Rotor 2 since the propellers I am using are made of this material and designed to spin at speeds up to 28,000 rpm. I realize that the tight clearance on the OD of the rotor required for 3 Blade Turbine 2 will need to allow for the radial growth from running at high speed. Shapeways give the following guidelines for prints of their Versatile Plastic (Nylon 12). The minimum thickness is 0.7mm. The accuracy is +/- 0.15mm +0.15% of the longest axis. The longest axis for the Nylon version of 3 Blade Rotor 2 will be the 1.252” (31.8mm) OD. The tolerance is +/- 0.15 + .0015 x 31.8 = +/- 0.2mm (0.008” ). Since the channel needs to be at least as wide as the nozzle diameter of 0.035” it will need 0.008” added to account for possibly being at the minimum of the tolerance. I decided to make the channel design width 0.050” to allow for being at the minimum of the tolerance and movement required for thermal expansion or contraction. I also provided a large diameter to grip while turning the OD of the rotor. The following drawings show the design dimensions.

Edited By Turbine Guy on 26/05/2023 16:56:38

I decided to add recent steam tests to the test sheet shown in the 29/05/2021 post and the latest air tests to the test sheet shown in the 22/03/2023 post. The following updated test sheets compare the performance of the various turbines when running on air or steam. There is an album in my photos for each of the turbines shown in these tables with photos, drawings, and additional information. The test dates indicate when the tests were performed and posts made near these dates have discussions on the intent and results of the test. 

Edited By Turbine Guy on 01/06/2023 15:59:02

I received the Nylon prints for the revised 3 Blade Rotor 2 described in the 26/05/2023 post from Shapeways. I ordered one print with the standard finish and another print with the optional smooth finish. The following drawings show the actual dimensions compared with the design dimensions and the following photo shows a closeup of one of the blades of each print. All three blades of each print appeared to be done correctly. The print on the left side of the photo had the smooth finish. The regular finish was slightly rougher but still much smoother than printed aluminum. A 0.040” diameter pin would slide down the entire length of each channel of both prints with a low pull force. The channel width was approximately at the minimum of the tolerance and the smallest hole through the center of the rotor was quite a bit under the minimum tolerance. In most cases a smaller than tolerance dimension is better than a larger than tolerance dimension since material can be removed.

I decided to give the drawings shown in the 26/05/2023 post new names since changing the rotor from printed aluminum to printed nylon was more of a change in rotor than a revision. The following drawings show the new names and the dimensions shown are the actual dimensions and position of the rotor in the housing after all the machining. I had to make several compromises in making this turbine that will be explained in the next posts.

I had a few problems making 3 Blade Rotor 3 and 3 Blade Turbine 3 shown on the drawings in the last post. When I machined the OD of the rotor, the cutting tool that was very sharp did not scrape all the nylon off but caused some of the material to wipe over the edges of the channels. When I tried to clean up the edges with a razor some of the wiped over material still would slide under the razor. The first photo given below shows an example of the best I could get the edges. The second photo shown below is a picture of a 5 blade rotor used in one of the reports I read discussing this type of turbine. You can see in the second photo how sharp and smooth the surfaces of the channel need to be to get the maximum performance. I will describe another problem found in the next post.

Another problem I found was the bore of Housing 3 SD Gap that 3 Blade Rotor 3 turns in is not concentric with hole for the ball bearings. When I enlarged the bore of Housing 3 SD to make the gap, all I needed was clearance on the OD of the rotor to allow for expansion to supersonic velocities. My setup for increasing the bore was simple and quick since I was not concerned about keeping it concentric with the ball bearings. The press fit of 3 Blade Rotor 3 on the shaft is strong enough that I could grip the shaft with my collet chuck to turn the rotor OD down to its final size, so it is very concentric with the ball bearings. The drawing of 3 Blade Turbine 3 given in the last post shows the housing bore is 1.254” and the rotor OD had to be turned down to 1.238” to slide into the housing. This is a maximum eccentricity of 0.016”. The following photo shows where I could insert a 0.010” thick shim in the area the nozzle is spraying into. This is a clearance between the rotor OD and the housing bore of approximately 0.010” which will allow a lot of leakage. The problems I discussed in this post and the last post describe two major reasons why 3 Blade Turbine 3 can’t give me the performance this type of turbine is capable of. In the next post I will give the test results.

I decided to revise the spreadsheet that I use to compare the performance of the turbines before adding the test results for 3 Blade Turbine 3. I made the spread sheet in two pages. The first page is what I have been using with the Imperial units. The second page converts the data on the first page to Metric units. This makes it relatively easy to show the test results in both units. I hope this makes it easier for those that prefer the Metric units to read the results. The following are the test sheets adding the performance of 3 Blade Turbine 3 in Metric and Imperial units. I think 3 Blade Turbine 3 did surprisingly well considering the problems I discussed in the last two posts.

I mentioned in the 10/06/2023 post that the bore of Housing 3 SD Gap that 3 Blade Rotor 3 turns in was not concentric with the bore for the ball bearings. I decided to try to make the bores more concentric and then make another 3 Blade Rotor 3 with the unused printed Nylon rotor. I made the bore in the housing more concentric by pressing an oversize dowel pin in the ball bearings bore and gripping the end of the dowel pin that extended out of the housing with my collet chuck to turn the rotor bore. The housing rotor bore increased from the 1.254” (31.85mm) diameter shown in the drawing of the 10/06/2023 post to 1.268” (32.21mm) diameter. The printed diameter of the unused rotor was 1.273” (32.33mm) as shown on the drawing in the 07/06/2023 post. I pressed this rotor on the shaft used for the first 3 Blade Rotor 3 and gripped the shaft with the collet chuck to turn down the rotor OD. I made cuts of 0.001” (0.03mm) per side until the rotor would enter all the way into the housing with the shaft passing through the ball bearings. I had to turn the rotor OD down to 1.263” (32.08mm) for it to spin in the housing. The maximum speed for the 3 Blade Turbine 3 with this rotor OD was 24,000 rpm at 29 psig (2.0 Bar). This was the same performance I got with the first rotor even though the total clearance between the rotor OD and the housing was much smaller (.005” [0.13mm] vs 0.016” [0.41mm]) There was still some resistance to spinning when the propeller was spun by hand with no air pressure, so I thought that more clearance was needed. After taking a few more cuts on the rotor OD the resistance never changed so I replaced the ball bearings. The rotor spun freely with the new ball bearings but the gap between the rotor OD and the housing had apparently increased enough that the extra leakage dropped the performance with the new ball bearings. I don’t know what caused the ball bearings to fail or when they went bad, but I failed to get a good test of 3 Blade Turbine 3 for the second time.

I am convinced that the concept of the 3 blade rotor given in the link that was shown in the 18/04/2023 post is good but the simplicity is very deceptive. To work correctly, the surfaces must be smooth, the clearances very small, and must have minimum deflection. I have tried to come up with a low cost method to accomplish these goals but as I explained in the last few posts I had problems for each thing I tried. The finish of printed aluminum was way too rough. The wiping over edges rather than cutting off cleanly was a problem machining the printed nylon. The cost of CNC machining was way too expensive. The rotor would have to be supported on both sides to keep the deflection low enough. My goal for any of the turbines is to have a design that can be made at a reasonable cost and only require relatively simple machining so this is getting too complicated. Since 3 Blade Turbine 3 was able to get as good performance as some of the turbines I have tested even with the problems my rotors had, it shows the potential of this concept.

Edited By Turbine Guy on 08/07/2023 16:08:35

I ran Tangential Turbine 5C on steam from my Stuart Twin Drum boiler a few times and each time the maximum pressure got lower. I assumed there was leakage around the nozzle insert, so I removed it. The nozzle bore in the insert had increased above the 0.024” (0.61mm) size it had in the first tests. This was probably due to my passing the drill through the insert bore before each test to make sure it was clean. Apparently I opened the insert up a little each time I cleaned the nozzle. With the insert removed, the turbine was changed back to Tangential Turbine 5B as shown in the following drawing. I ran Tangential Turbine 5B on air with the GWS EP 2508 propeller to see how its performance compared with the test shown in the 10/06/2023 post. It only required 12 psig (0.8 Bar) to turn the propeller 28,000 rpm instead of the 14 psig (1.0 Bar) required in the test shown in the 10/06/2023 post. This was closer to the best performance I was able to obtain with Axial Turbine 4A. The advantages this test of Tangential Turbine 5B had over the previous tests was the ball bearings had been run longer, the air temperature was 10 F higher, and the nozzle bore size had increased to 0.043” (1.09mm).

Like checking in on ths thread and reading the latest.

Thanks for following this thread. I have really enjoyed your videos. I hope we can get a few more people to share their thoughts about model engines and model turbines.

When I started the test of Tangential Turbine 5B described in the 20/07/2023 post it slipped out of the holder when the speed was over 20,000 rpm. Remarkably, the propeller was not broken. Since this was my only GWS EP 2508 propeller, I went ahead with the test but decided to check if the air pressure required to turn the propeller at a speed of 28,000 rpm was correct. The first thing I tried was to see if the APC 8x6 propeller would spin at the approximately 1,240 rpm that Axial Turbine 4A and Tangential Turbine 5B were able to do with the same pressure needed to turn the GWS EP 2508 propeller at 28,000 rpm shown in the 23/03/2023 post. Tangential Turbine 5B required approximately 14 psig (1.0 Bar) to turn the APC 8x6 propeller 1,240 rpm. This was the same pressure required before and appears to be the correct pressure. I also ran a test at 14 psig with the GWS EP 2510 propeller and it reached a speed of approximately 26,500 rpm. The GWS EP 2510 propeller requires approximately 4.2 watts of power to turn at this speed. This is the same power required to turn the GWS EP 2508 propeller 28,000 rpm so it also confirms the 14 psig (1.0 Bar) is correct. The increase in nozzle size running at the same pressure increases the energy, so the 0.041 (1.04mm) nozzle size is the most efficient for Tangential Turbine 5B running on air from the Master TC-96T airbrush compressor.

I received the new GWS EP 2508 propellers and ran a test with Tangential Turbine 5B like the one shown in the last post. It only required 11.5 psig (0.8 Bar) to turn this propeller to a speed of 28,000 rpm. This was even a lower pressure than required with the propeller used in the last post. I used the new propeller on Axial Turbine 4A and it still required 12 psig (0.8 Bar) like the test shown in the 10/06/2023 post. Apparently the 12 psig (0.8 Bar) pressure is correct for Tangential Turbine 5B, even though the speed of the propellers requiring more torque were not as high as obtained with Axial Turbine 4A with this pressure. I updated the test sheets shown in the 10/06/2023 post to show the latest test of Tangential Turbine 5B.  Tangential Turbine 5B gets the same power as Axial Turbine 4A at a pressure of 12 psig (0.8 Bar), but uses more mass flow so is not as efficient. 

Edited By Turbine Guy on 01/08/2023 19:15:03

Hi

I sent you a couple of messages to your e.mail address a few weeks ago. I'm just wondering if you received them since I've not recieved any reply.

Mike

Hi Mike,

I haven't seen anything from you for several weeks. I will check if your Emails got to me and I missed them. If I find them, I will respond, if not I will let you know I couldn't find them.

Hope you are doing well,

Byron

Hi Byron

I just received your message and have just re-sent my earlier message. I double-checked the e.mail address I used was correct so if this one does not make it through to you it I don't know what to do.

Best wishes

Mike

I purchased from Amazon the synthetic oil made for dental handpieces shown in the following image. This was designed to be applied about once a day and run air turbines in dental handpieces several times on one application. I thought that since this was designed to run tiny turbines on air that it might work well for my turbines that require lubrication. Tangential Turbine 1 and Tangential Turbine 2 were never changed to use the maintenance free dental bearings so I ran tests using this oil to compare with the tests shown in the test sheet in the 01/08/2023 post. The tests of Tangential Turbine 1 and Tangential Turbine 2 shown in the test sheets used Krytox GPL 102 oil. With the Syntek Oil the speed increased from 19,000 rpm to 22,500 rpm for Tangential Turbine 1 and from 20,000 rpm to 23,500 rpm for Tangential Turbine 2. This raised the output power from 1.3 watts to 2.2 watts for Tangential Turbine 1 and from 1.5 watts to 2.5 watts for Tangential Turbine 2. This shows the importance of the oil in very small turbines.

Hello Byron,

Your proposed oil obviously is designed for ambient temperature, for a short sterilisation at 100°C at most, A model steam turbine is supposed to operate at steam of say 4 Bar. The wet steam temperature at this pressure will be 151°C. Your tests therefore should be run with live steam!

I cannot find any information in fhe internet about the temperature behaviour of this proposed oil. I therefore will stick to my well proven Krytox 105.

regards Werner