DIY Tesla Impulse Turbine

Home Model Engine Machinist Forum

Help Support Home Model Engine Machinist Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
I suspect (but am guessing!) that the smaller holes are putting proportionally more air in the turbine slots and at a higher velocity, and less air impinging on the outer metal surface of the rotor where it simply goes around the outside and out the exhaust?
Gas velocity is everything in the turbine. - And putting it where you want it, versus NOT putting it where you do NOT want it is the key to efficiency.
What was the pressure gauge steady at with the larger holes, and what is it steady at now?
K2

As the videos show, there isn't a steady pressure time period; when I open the valve to the compressor's air tank, all the air inside the tank rushes out through the nozzles. Tank pressure starts out at 100 psi and steadily decreases in just a few seconds.
 
I suspect (but am guessing!) that the smaller holes are putting proportionally more air in the turbine slots and at a higher velocity, and less air impinging on the outer metal surface of the rotor where it simply goes around the outside and out the exhaust?
K2

My current theory (and it is just a theory) on why smaller nozzle holes have resulted in increased power from the turbine:

As I mentioned previously, larger nozzle holes required a larger air flow than my compressor tank was able to supply, so I blocked off half the holes in each of the six nozzle arrays. Although this configuration allowed the turbine to spin and do work, an unplanned result was that half the discs in the turbine, which had no air nozzle holes shooting high velocity air at them, were now free to function as a Tesla compressor. I suspect exhaust air from open, working nozzles, once it reached the open, center area of the disc stack, was being sucked into the discs with closed off nozzles and being flung against the inside walls of the container shell, where it interfered with the air jets from working nozzles.

The nozzle arrays with the smaller holes have no holes blocked off, resulting in no section of the turbine discs to inadvertently become compressor sections, sucking power from the turbine.
 
The height of jet - even momentarily still represents the water pressure achieved. So it there is a wall you can hit you can see how high you get for max pressure (probably when the air is at/near max pressure!). Knowing the number of vanes, you can possibly hear a musical note of the pressure pulsations as vanes pass the outlet (pressure pulsations in the water at the jet will generate some sound in the air) - from which you can deduce the frequency of noise, and dividing by the no. of vanes can deduce the rpm... Lots to measure already. Get your thinking cap on.
K2
 
The height of jet - even momentarily still represents the water pressure achieved. So it there is a wall you can hit you can see how high you get for max pressure (probably when the air is at/near max pressure!). Knowing the number of vanes, you can possibly hear a musical note of the pressure pulsations as vanes pass the outlet (pressure pulsations in the water at the jet will generate some sound in the air) - from which you can deduce the frequency of noise, and dividing by the no. of vanes can deduce the rpm... Lots to measure already. Get your thinking cap on.
K2

K2, I applaud both your curiosity and your creative problem solving ideas,....however, at this point in time, I believe my project time is better spent finishing my monotube boiler which will allow me to properly test this turbine at the design pressures of 500 psi.
 
Hi Toymaker. Maybe I am not the best to give advise, but my educated guess is as follows. Gas turbines are extremely hungry pieces of equipment. A small jet engine with the diameter of your turbine delivers maybe even tens of kW, so on reverse, unless you provide constant power/ gas quantities of the same magnitude you won't obtain stable operation on a mid-range workload that allows you to properly test and adjust the turbine. My best bet would be to find a very high flow compressed air source -like you'll find in a plant - to do adjustments. Maybe monotube boiler could give you required gas flow, but monotube boilers tend to be very unstable in operation and complexity of controlling all parameters of boiler-turbine compound could drive you crazy. I don-t mean it can't be done, but I won't advise to. And of course, unless you try, you'll never know.
 
Last edited:
Hi Toymaker. Maybe I am not the best to give advise, but my educated guess is as follows. Gas turbines are extremely hungry pieces of equipment. A small jet engine with the diameter of your turbine delivers maybe even tens of kW, so on reverse, unless you provide constant power/ gas quantities of the same magnitude you won't obtain stable operation on a mid-range workload that allows you to properly test and adjust the turbine. My best bet would be to find a very high flow compressed air source -like you'll find in a plant to do adjustments. Maybe monotube boiler could give you required gas flow, but monotube boilers tend to be very unstable in operation and complexity of controlling all parameters of boiler-turbine compound could drive you crazy. I don-t mean it can't be done, but I won't advise to. And of course, unless you try, you'll never know.

Thanks for the good advice NapierDeltic. Tens of kW is in the range I'm hoping for.

I'm keenly aware of the difficulties of obtaining stable operation from a monotube boiler, which is why I'm using a microcontroller to monitor and control parameters such as fuel and air flow, as well as pressure in and output; the ECU is the part of the boiler I'm currently working on. I briefly spoke about the ECU (Engine Control Unit) here: Electronic Control and there's more on the burner-boiler unit here: Monotube Flash Boiler Design.

Back in the 1970s, a small company known as SES (Steam Engine Systems) built several test cars using monotube boilers, which were electronically controlled; SES
Back in that pre-computer time, the ECU was said to take up the entire trunk space of the car. Fortunately for me, today's microcontrollers are much smaller, easy to acquire and not too expensive.
 
Thanks for the good advice NapierDeltic. Tens of kW is in the range I'm hoping for.

I'm keenly aware of the difficulties of obtaining stable operation from a monotube boiler, which is why I'm using a microcontroller to monitor and control parameters such as fuel and air flow, as well as pressure in and output; the ECU is the part of the boiler I'm currently working on. I briefly spoke about the ECU (Engine Control Unit) here: Electronic Control and there's more on the burner-boiler unit here: Monotube Flash Boiler Design.

Back in the 1970s, a small company known as SES (Steam Engine Systems) built several test cars using monotube boilers, which were electronically controlled; SES
Back in that pre-computer time, the ECU was said to take up the entire trunk space of the car. Fortunately for me, today's microcontrollers are much smaller, easy to acquire and not too expensive.
Thank you very much, Toymaker and I am (will be) following you with much interest. If I am allowed to skew a bit your topic, Traian Vuia -who happens to be my fellow countryman -was one of the first to design research and promote a fast generating monotube steam boiler, starting maybe since 1902, which he called "catalytic burning steam generator" :
https://www.researchgate.net/public...bution_to_the_development_of_propulsive_power
1686154795639.png

1686155290905.png

I obviously read long time ago about his work. Some of is outdated, some too optimistic, some is experimental and systematic and still valid. Anyway, hence my interest.
And also from my belief that steam hasn't shown yet its best! As a non-specialist, I am really interested in your experimentation both for monotube boiler and for Tesla turbine, another technical marvel!
I wish you Good Luck and I will rejoice of it!
 
Just reading all this (Again... Am I becoming obsessed with Tesla turbines?).
Napier Deltic. With such a revered name in producing some of the most powerful (size per size of others) and practical engines (When used for main line locomotives they surely must have proven a reasonable level of reliability and durability? - Favourites of mine!) I hope you have the expertise to aid us lesser mortals?
I have just borrowed a Tesla turbine from the local club.
Testing it on a newly re-commissioned boiler, all be it way underpowered, the most significant issue was simply water. pre-warming the turbine, it started to run with hardly any steam, most of which condensed on the plates and body and rapidly filled the turbine so it was running slowly in "paddle wheel mode" as the 3 exhaust holes near the centre paddled through the water! Unfortunately the (Original) burner was way to small to do anything useful. I expect to need a boiler (probably a flash-tube boiler, probably with huge burner?) of maybe 10kW but the original burner was only made as about a 1kW gas burner...
I am now making a 3 kW gas burner just to try and get a proof of concept model to resolve issues... such as condensate...
I need to add a small hole at the bottom of a side plate to permit condensate to drain, and avoid the paddle wheel action. I am sure (having experienced a similar sort of phenomenon with a compound engine, that was hugely inefficient while cold, but as it warmed the LP cylinder the speed suddenly picked-up as that started working, instead of simply pumping condensate...?) that initially, when "cold", the steam condensate will travel outwards from centrifugal force thus the reverse direction to the steam involute. This must interfere with the gas flow and functional efficiency, so the turbine simply cannot run very fast. Drag from paddle wheel effect in the condensate is also a problem just now, until I can develop enough steam to get to a hot turbine, when condensate should be less of an issue.
I am also concerned that as steam is wet stuff, just vaporised, in most little boilers, the energy is in condensing the steam to water, which thus kills any performance in a turbine. I believe I need very high superheat? Then the superheat energy will ensure dry gas flow through the turbine to steam above condensing temperature/pressure? But perhaps this is where Tesla came unstuck with very hot metal, stressed by the high speeds of rotation, thus becoming stretched?
I think Parsons Turbines were also plagued by condensate until he had very high superheat, as used in modern power stations using high pressure and superheat so the steam is dry throughout the turbine?
Or perhaps the better solution is not to worry about condensate and set the turbine with a vertical axle so that the condensate naturally drains from the central exhaust? - It just feels wrong to have any condensate in a Tesla Turbine reliant upon interaction between steam and discs by surface friction. Water film must surely spoil the whole thing by taking the energy away as 1/2MvSq of water?? I.E. The energy we want transferred to disc metal as 1/2Mvsq.
Any further ideas from anyone may help both Toymaker and I?
Thanks,
K2
 
I also found this comment:
"The blades of an impulse turbine are usually bucket-shaped so they catch the fluid and direct it off at an angle or sometimes even back the way it came (because that gives the most efficient transfer of energy from the fluid to the turbine)."
IF the direction of the steam jet is reversed from the impulse blade... (Max. efficiency), then the direction of steam is contrary to that to encourage the "Tesla" part of the turbine to work.

Which is why I didn't chose a bucket-shaped blade.

Actually, I suspect a Tesla turbine without the impulse blades may be more effective than the hybrid one you have made? - OR, as achieving the high rpm for a Tesla turbine to really become good is very difficult, a proper curved blade Impulse turbine is the best option?
How about it?
K2

I chose the flat-plate "bucket" blade shape not only for it's ease of machining, but also to allow the impacted steam to be easily forced inward between two rotors where it would continue to expand, hopefully in a useful direction, acting as Tesla had designed it. Will it work?? I don't yet know,...which is why I built it and why I will test it.
 
Early on you mentioned lacking the cutters to make tiny divergent nozzle holes?

Three cheap options that I use are:

Use a cheap center drill (easiest)

Or

Just spin a slightly oversized dill bit backwards with a drill against a grinding wheel. Sculpt the profile required and knock back the edges with a hand stone to cut brass without the bit grabbing.
You need to use a bit, for this, that is small enough to still have flutes at it's tip when done.


Other then that, starting with a predrill hole you can gently ream the divergent profile. Just take some silver steel, or a broken drill bit and use just the shank. Grind your profile like above (chucked in a drill against a grinder) and when the profile is right, then grind off exactly half the width of the profile finishing with hand stones. If you used silver steel, harden and temper it. You should have a tool that is round on one side and flat to the axis on the other. You really need to end up dead center with these, in my experience.

You can make awesome reamers for just about any profile this way.


All three ways have worked very well for me to make mini venturi's with specific profiles.

and all these ways are cheap.
 
Early on you mentioned lacking the cutters to make tiny divergent nozzle holes?

Three cheap options that I use are:

Use a cheap center drill (easiest)

Or

Just spin a slightly oversized dill bit backwards with a drill against a grinding wheel. Sculpt the profile required and knock back the edges with a hand stone to cut brass without the bit grabbing.
You need to use a bit, for this, that is small enough to still have flutes at it's tip when done.


Other then that, starting with a predrill hole you can gently ream the divergent profile. Just take some silver steel, or a broken drill bit and use just the shank. Grind your profile like above (chucked in a drill against a grinder) and when the profile is right, then grind off exactly half the width of the profile finishing with hand stones. If you used silver steel, harden and temper it. You should have a tool that is round on one side and flat to the axis on the other. You really need to end up dead center with these, in my experience.

You can make awesome reamers for just about any profile this way.


All three ways have worked very well for me to make mini venturi's with specific profiles.

and all these ways are cheap.

Yes, center drills are by far the easiest way to get a 60 deg cone shaped hole, which should be much better than my plain drilled holes. And I just today found a supplier on AliExpress that sells center drills with 0.5mm tips; before today, the smallest I could find was #1 center drill which has a tip diameter of 3/64" (0.046"),...too big for my needs.
 
Something I've always wondered about is popping a wee small axial turbine on the exaust of a tesla turbine and using it to run an oil pump for the tesla bearings. It wouldn't need to scaveng much power... with your CNC abilities, maybe you can do something to take advantage of the wasted energy.

Full ceramic bearings don't require lubrication,...just remove excess heat and they'll last forever. I use full ceramic bearings for most application.

A second tesla turbine, in series, designed to be condensing and put out water, not steam, would also be a feat. Given that there are now some solid papers on spacing the blades, maybe Making the blades have a mated divergent profile would let you plan for a dT through the turbine that's sufficient to make it act like a condenser?

A two stage system like that may be key to unlocking the full potential.


If you're impulse idea does not give you the increase in power that it might, maybe it would give an increase in efficiency as a tesla blower and be suitable as your burner blower?


Edit: posted this before your post came in on moving the discussion, my apologies.

I have no knowledge on placing back pressure on Tesla turbines. I do know that radial turbines don't work well with back pressure.
 
Going the opposite way, a water ring pump would be easy for you to rig up with your tooling and would increase the pressure differential across the tesla system, provide a way to pump water to your feed pump and condense residual steam to water, further increasing the pressure differential.

A water ring pump added after the turbine might just act as an economizer and a reverse turbo booster(pulling instead of pushing), vs adding in a radiator or something to condense remaining steam.

Bonus over a air cooled economizer is the waters heat would get preserved.
 
Just reading all this (Again... Am I becoming obsessed with Tesla turbines?).
Napier Deltic. With such a revered name in producing some of the most powerful (size per size of others) and practical engines (When used for main line locomotives they surely must have proven a reasonable level of reliability and durability? - Favourites of mine!) I hope you have the expertise to aid us lesser mortals?
I have just borrowed a Tesla turbine from the local club.
Testing it on a newly re-commissioned boiler, all be it way underpowered, the most significant issue was simply water. pre-warming the turbine, it started to run with hardly any steam, most of which condensed on the plates and body and rapidly filled the turbine so it was running slowly in "paddle wheel mode" as the 3 exhaust holes near the centre paddled through the water! Unfortunately the (Original) burner was way to small to do anything useful. I expect to need a boiler (probably a flash-tube boiler, probably with huge burner?) of maybe 10kW but the original burner was only made as about a 1kW gas burner...
I am now making a 3 kW gas burner just to try and get a proof of concept model to resolve issues... such as condensate...
I need to add a small hole at the bottom of a side plate to permit condensate to drain, and avoid the paddle wheel action. I am sure (having experienced a similar sort of phenomenon with a compound engine, that was hugely inefficient while cold, but as it warmed the LP cylinder the speed suddenly picked-up as that started working, instead of simply pumping condensate...?) that initially, when "cold", the steam condensate will travel outwards from centrifugal force thus the reverse direction to the steam involute. This must interfere with the gas flow and functional efficiency, so the turbine simply cannot run very fast. Drag from paddle wheel effect in the condensate is also a problem just now, until I can develop enough steam to get to a hot turbine, when condensate should be less of an issue.
I am also concerned that as steam is wet stuff, just vaporised, in most little boilers, the energy is in condensing the steam to water, which thus kills any performance in a turbine. I believe I need very high superheat? Then the superheat energy will ensure dry gas flow through the turbine to steam above condensing temperature/pressure? But perhaps this is where Tesla came unstuck with very hot metal, stressed by the high speeds of rotation, thus becoming stretched?
I think Parsons Turbines were also plagued by condensate until he had very high superheat, as used in modern power stations using high pressure and superheat so the steam is dry throughout the turbine?
Or perhaps the better solution is not to worry about condensate and set the turbine with a vertical axle so that the condensate naturally drains from the central exhaust? - It just feels wrong to have any condensate in a Tesla Turbine reliant upon interaction between steam and discs by surface friction. Water film must surely spoil the whole thing by taking the energy away as 1/2MvSq of water?? I.E. The energy we want transferred to disc metal as 1/2Mvsq.
Any further ideas from anyone may help both Toymaker and I?
Thanks,
K2
Hi Steamchick. I have little experience with Tesla turbines - I mean even from informal point of view.
From what I have read, it seems many had poor experience with it (I mean low efficiency even in ideal conditions - dry steam or alike) while few have hit the bull's eye. I think the efficiency of your purchased turbine can be equally verified with compressed air (you'll need a very strong air source with the output pressure close to your application parameters). Assembly facilities have such high output sources an they could be delivering 9-10 bar. No chance to make a test in the plant you have activated before? If you measure the flow and know the pressure, and if you can measure output power using an electrical generator in the right configuration, you have your answer in digits. Or at least you can see it spinning and you have the eye to evaluate☀️ or 🌧️. But my impression is that the turbine has low efficiency at partial loads (steam/gas flow) so the right test should be made only on full load.
Water ring for me is a no go; Tesla turbine concept is based on micro-adhesion between gas and disks, so the friction between water and disks being much higher, it will never reach the correct speed to work properly.
How I see it, is working as a first stage in a multi -expansion system. (though, The Ignoble Troll mentioned Tesla turbine does not like back pressure, so I don't know...). Anyway, I think, first, turbine should work with dry steam (overheated) and I believe it should , relatively fast (up to 10 minutes :) reach the working temperature of dry steam and eliminate gradually condensate (either spinning or not in this interval). I would anyway think to try - at least as a test - a following expansion stage; maybe including even an intermediary steam reheating. I don't have in mind now a more detailed picture of how such setup would look like to be both cheap, practical, relevant ... Maybe using the turbine stage from an old supercharger. Both stages shouldn't be mechanically linked. I think you could even use the secondary turbine (in certain limits) - like hidraulic torque converter is used in automatic gearbox - to alter the working conditions of Tesla turbine. You could adjust secondary load from free spinning to completely stopped?
Funny thing (though not relevant, I'm afraid) - uniflow steam motors have the same ability, like Tesla turbine, to dislike backpressure.
 
Last edited:
One thing that doesn't fit my mind is why Tesla turbine doesn't like backpressure! As I figure it (and was described by its inventor), the mechanism is based on micro-adhesion between steam molecules and turbine disks so a direct transfer of kinetic energy from steam to rotor; no other aerodynamic forces with p-V change processes involved in power extraction -other than those imposed by geometry of steam path.
 
Last edited:
One thing that doesn't fit my mind is why Tesla turbine doesn't like backpressure! As I figure it (and was described by its inventor), the mechanism is based on micro-adhesion between steam molecules and turbine disks so a direct transfer of kinetic energy from steam to rotor; no other aerodynamic forces with p-V change processes involved in power extraction -other than those imposed by geometry of steam path.
The behaviour of any turbomachine has a lot to do with the pressure ratio across it.

Tesla turbines are impulse turbines, in that the pressure drop occurs entirely in the nozzle and the resulting jet of high speed gas then transfers its momentum to the turbine wheel. The velocity of the jet will be a function of the pressure drop (assuming subsonic flow). So if you add backpressure, at a fixed steam supply pressure, your jets of steam will be slower and thus have less momentum to transfer to the turbine!
 
Another thing that concerns me at Tesla turbine: I know from wind power plants and similar applications that there is a certain proportion between gas speed (wind) and rotor speed at which energy transfer (efficiency) is maximum. There is a certain formula I can't recall. I don't know how this matches with TT as I have impression this proportion is variable inside rotor, being related to radial location. Actually I think steam flow should keep much of its speed while tangential speed of rotor decreases towards center. Of course this is a simplist view.
 
A couple of points. The Tesla turbine I am working to get running as a demo for shows is one made by our late chairman. Not a commercial job. It needs proper bearings, as the existing ones are a bit second-hand.
On air at 100 psi it ran up to 20,000rpm, just, before the reflective tape flew off. When my laser meter suddenly read zero... 😁
The (im) balance and bearing friction limited this .
Tesla turbines do not convert direct impulses (momentum) the way a water impulse turbine does.
Tesla turbines take energy from the 1/2 M v squared kinetic energy of the fluid stream, by skin friction drag between the fluid and discs. E.g. hold a CD on a horizontal shaft - a pencil held horizontally will do - then blow tangentially on the edge of the CD. It spins - exactly the same mechanism as the Tesla turbine..
Using wet steam, as it transfers the press release before the nozzle to velocity after the nozzle, water droplets appear as the pressure and temperature drop. Presure and temperatureand Latent heat of vaporisation are the energy source at the nozzle that becomes kinetic energy of the fluid. I think the water droplets add to the energy transfer of some of the kinetic energy as they settle on the discs, only to be shed due to centripetal forces, so they take their kinetic energy away again, to be lost against the casing .
Thus water "steals" power.
As the steam jet is limited by the speed of sound, the max speed of a turbine cannot exceed this at the perifery. But as a jet of fluid engages with the turbine in a close spiral, as the jet slows it also spirals to a smaller radius until ejected near the shaft, back-pressure reduces the pressure drop and thus power that can be extracted. From James Watt to today's steam turbines in power stations we have condensers to get more power from the steam by eliminating as much back pressure as possible (even 14.7 psi of atmosphere!).
The efficiency of a Tesla turbine comes from the disc rotational speed approaching the speed of the fluid jet, so the spiral has hundreds of turns to get to the core exhaust point. But this means a 3 in dia turbine needs to be capable of at least 100, 000rpm. Which this one isn't!
I have seen a >10kW flash boiler power a Tesla turbine up to over 100,000rpm (speed meter limit) with no-load. A slow speed increase up to over 70,000rpm. Then it suddenly accelerated rapidly as it came on-song and ran away to top speed. Top speed may be limited by the speed of sound, or lower if the friction and extracted power does so. When disc surface is at the same speed as the jet it cannot extract energy from the jet.
Fun engineering!
K2
 
Fun fact: dental turbines use gaseous bearings. If you have low side loads, they are optimal for extra-high rpm-s.
Don't know if they work well with TT, either alone or in certain combination.
For a home-built model, though, ceramic bearings should be preferred. They offer high precision and tight tolerances required ready made and can maybe work easier with a not perfectly balanced rotor.
 
Last edited:
Back
Top