3D printed Tesla turbine

Metal, Wood, ... the heavy stuff can be found here
Downunder35m
Posts: 484
Joined: Sun Aug 16, 2020 5:32 am
Location: Australia
Contact:

3D printed Tesla turbine

Post by Downunder35m »

No problem finding a Tesla turbine to prin in the usual places.
But actually printing one is a different story ;)

I did some research and playing with them many years ago, in the times when computers were still a rare thing.
You can imagine the limitations of using soda cans for the thin metal parts and just brass bushings and shaft from some discarded DC motor (rotor removed)....
But I learned some vital lessons about working with thin aluminum sheets.....
Anyway, already back then I was fascinated by the simple design.
If you are only after the story about how I plan to desing it and make it work please skip the rest of this post and wait for the update further down.
► Show Spoiler
Enough old stories and theories....
Lets go to making such a thing printable.
Exploring the works of the old inventors, mixng them up with a modern touch.
To tinker and create means to be alive.
Bringing the long lost back means history comes alive again.
Downunder35m
Posts: 484
Joined: Sun Aug 16, 2020 5:32 am
Location: Australia
Contact:

Re: 3D printed Tesla turbine

Post by Downunder35m »

Taking what is hidden in the above spoiler we know we have a big problem.
Plastic is way less sturdy and forgiving than metal.
Ever seen these videos where people destroy CD ROMS and DVD's by spinning them until they fail?
They do so way below 50.000 RPM even if from the best company out there.
So using something like this is a bad idea.
Printing it means we do worse than the 3 layers of a DVD - we print in lines and have to hope the layer and perimiter bonds will survive.
Nothing you should try is you consider your 3D printing skills to be below ADVANCED levels.

Axles and bearings are relative as we know we can't use a printed bearing as it would just melt away.
Using a metal axle is certainly an ption but it can be quite tricky to get a proper bond.
Plus: We might have to face the fact that metal and plastic expand differentls.
So a compromise between bearing size and axle diameter needs to be made if we want to stick to a printed axle.
608 bearing and 8mm axle sounds like a fair compromise.
And no problem getting the bearing in an open and dry type either.
I could not really find any 8mm or close to plastic rods of a suitable kind, but if you know a way then let me know ;)

The housing needs to be able to at least hold a failing disk halfway inside.
Since it does not require much else using ABS or PETG seems to be the obvious coice, apart from Nylon if you dare.
The bearing mean we have to go quite big for everything already.
My first models used the flawed slot approach, not because I like it but because those tiny connections between the slots provide a convinient safety feature.
Instead of a disk fully crushing you have one that seperates and starts to grind everywhere, bringing the stems to stop quickly without breaking through the housing.

So much for the first and most basic approach that I won't use for this project.
I works great if you are after a trubine getting to around 20 - 30.000 RPM as a max that is supposed to drive a little motor or such at around 10 - 15.000 RPM.
Let me know if you need something like this and I upload my old designs for you.

What I want for this project as a final goal is a similar design to what Tesla used but modified to what a 3D printer is capable of and what plastic can do.
It is not intended to break the speed barrier so please DO NOT TRY! - the plastic won't make it past 60.000 RPM even in PETG.
The goal is bring the amount of air required down while increasing the efficiency.
A proof of concept for my old metal models if you like.

Initial design considerations:
Using a spoke/fork system will not work with plastic unless you go bigger in size than what most printers can do.
But we can print in ways that were not really possible in an easy way to do back about 100 years ago.
An axial turbine uses nicely shaped blades to create required airflow.
Our air comes out of the disks like a vortex already.
Instead of using forks it would make sense to print the mounting cage upright so we can create three or maybe 4 curved walls going from the axle to the disks.
Acting like a wave guide for the outcoming air so that instead of one vortex three individual ones are formed.
With plastic we won't get away with one set of spokes.
We need some support either end of our disk pack.
I makes now even more sense to consider designing the mounting system like a vortex chamber.
Meaning instead of just three curved walls going directly to the axle we could let them make half a turn or more to get there.
This would also allow to adjust the slope so that the air will have a prefered direction to go.

The housing can be closed on one side but needs a suffient outlet size on the other.
Preferably without a bearing blocking the way.
Only feasable option is to have a hole for the axle big eough for the air to eascape and shartly after the more or less open mount for the other bearing.
Considering the printing limitations on the disks it means the housing is a multi part approach.
You can onyl go so small in size until the disk would need to paper thin....
Thinking a bit in the direction of an open, conical cage to hold the beraing over the outle area....

The really tricky bit is the inlet.
We need a positive pressure that in a perfect world would be equal on all disks.
The air also needs to come onto the disks at the highest possible speed.
Printing such fine details in one go is tricky at best, impossible if you need a smooth surface.
So the inlet needs to be two or even three part block to allow for a flat rocket motor like inlet nozzle.
If you consider how quick a cheap compressor runs out of air pressure with just a 5mm I.D. hose.....
I am guesstimating that I would need a gap of less than half a millimeter for a 8 to 10 disk pack to be efficient enough.
Preferably less to avoid running out of air too quickly.
Getting below 5mm² for the inlet area is a goal that should be possible.

Speed wise I am aiming for an operation speed to drive a little motor at a max of 12.000 RPM.
Ideally so that a cheap compressore will be able to keep up.
Let me know if you are interest in more details along the design process, otherwise I just post in stages.
Exploring the works of the old inventors, mixng them up with a modern touch.
To tinker and create means to be alive.
Bringing the long lost back means history comes alive again.
Downunder35m
Posts: 484
Joined: Sun Aug 16, 2020 5:32 am
Location: Australia
Contact:

Re: 3D printed Tesla turbine

Post by Downunder35m »

While I wait for testprints to confirm whether or not my printer is usable again I started on the design on the slotless turbine.
As our average and favourite 608 won't be big enough for the task I designed aroun the standards 6203 bearing - 40mm X 12mm with a 17mm hole.

The first housing will be plain simple with just a single inlet port.
As the dislodging of disks at high RPM can be a real issue I decided to go with a more trianglur shaft for this model.
The forces onto the bearing area shouldn't be too great to create a real issue with the strength I hope.
As the shaft will be hollow I have a few designs in mind to test for printability.

General idea behind:
Eliminating the need for slots should allow for a much better airflow, especially since we no longer have this huge gap towards the axle to keep the disk in one piece.
The downside is that there must be a proper match between the air coming in and coming out.
And well, I was too lazy to do all the required calculations for flow rates diameters and such.
Means I used my Nickel plated eye gauge and calibrated thumb to guesstimate the values required based on experience.
Most obvious downside of the better airflow is the severely compromised strength of the shaft holding it all together.

Design ideas to overcome the limitations a filament printer has:
The part of the shaft holding the disks is basically just 3 thin bars connected to a hollow tube either end.
Only way to print this with suffient strenght is flat, so it can't really be round for accuracy reasons anymore, not to mention the support problem.
Right now I hope that well calibrated bridging and the thing gaps allow for a good print in one piece.
If that fails my last resort is to design a round axle in two halfs that require joining.

Known fears:
The bearings won't like the speed so for now I leave them closed and grease filled.
Downside of that is that they overheat quickly and either ooze out hot oil that once once stick grease or the surrounding plastic goes soft.
Vibrations usually destroy either the glue connection of the disks or break the axle.
Balancing a printed rotor system isn't always easy so I tried to limit what amount of air can come out in the hope it will create a natural limit for the RPM's.


Well, if you get distracted and forget to post.....
Ok, just continue then ;)
I managed to print some really consistend and solid 0.5mm thick disks.
The axle is still a pain and I will have to print it solid and upright to get the details.
While testing the parts from the previous model again, the one that disintegrated on the inside, I realised that the cause what not an inbalance in the disks but a slightly loose axle in one of the bearings.
The disks started to vibrate enough to bridge the 2mm gap to the housing walls....
I try now a triangular axle with three push inserts to wedge it in the center and to keep it free from any movement.
Not sure how well this works out in terms of keeping all 100% alignet so maybe the bearing ends of the axle need some cutting and replacing to make it work better.
Only thing I know for sure is that it is not really that straight forward to design a high speed thingy in plastic to be printed LOL
Exploring the works of the old inventors, mixng them up with a modern touch.
To tinker and create means to be alive.
Bringing the long lost back means history comes alive again.
Downunder35m
Posts: 484
Joined: Sun Aug 16, 2020 5:32 am
Location: Australia
Contact:

Re: 3D printed Tesla turbine

Post by Downunder35m »

The printer is now basically running 24 hours a day to pump out the parts I need for my drawing machine but I drop in some parts for the turbine here and there.

I managed to fix the issue of propely aligning the disk on the shaft.
If in doubt use spacers while the glue sets ;)
Still working on the right airflow so the thing actually spins.
Tried to somehow make the overall outlet size for the spce between the disk match the inlet area.
Never works 100% the first time so once all is glued a testrun will show how it worked out.
The bearings are a bit of a bummer in regards to their friction with all the grease but hopefully the turbine will be strong enough to overcome this.

Once I had testrun, working or not, I will upload some STL files in case someone wants to give it shot on a nice resin printer os so.
Exploring the works of the old inventors, mixng them up with a modern touch.
To tinker and create means to be alive.
Bringing the long lost back means history comes alive again.
Downunder35m
Posts: 484
Joined: Sun Aug 16, 2020 5:32 am
Location: Australia
Contact:

Re: 3D printed Tesla turbine

Post by Downunder35m »

I did a first test run today after cleaning out the bearings a second time to reduce the friction.
And I have to say the thing did not behave as I expected, maybe due to being printed....
Compared to a standard tesla turbine the thing is almost silent, the whining noise from the bearings gets louder then the air working in it!
Did not aim for extreme speeds so it won't be destroyed too quickly but how it gets to speeds is rather fasciating, have to see if I can get a long hose on the compressor to get the noise out of a video.
Unlike a normal Tesla turbine it speeds up really slowly, almost as if it is fighting it.
And you don't really get the impression of speed as the screaming noise is missing.
All you hear is that it just keeps getting faster, little by little.
Does not seem to use up more air at higher speeds either.
I can upload the current files if anyone wants to try it out with some 6203 bearings...
Don't like to glue the housing together though and might have to design a better one now that I know the basics work.
And yes my admin, I know would really like to have a nice resin printer for those fine parts ROFL
Exploring the works of the old inventors, mixng them up with a modern touch.
To tinker and create means to be alive.
Bringing the long lost back means history comes alive again.
User avatar
Orngrimm
Site Admin
Posts: 308
Joined: Wed Jul 29, 2020 8:02 am
Location: Switzerland
Contact:

Re: 3D printed Tesla turbine

Post by Orngrimm »

Normally, people use a tesla-turbine with water. The nature of air having a bad "stickyness" may explain the slow start... Water sticks to the disks MUCH better than air...

Stupid Q, but your casing IS volute, right?
Builder of stuff, creator of things, inventor of many and master of none.
Tinkerer by heart, archer by choice and electronics engineer by trade.
Downunder35m
Posts: 484
Joined: Sun Aug 16, 2020 5:32 am
Location: Australia
Contact:

Re: 3D printed Tesla turbine

Post by Downunder35m »

Love the guy - he is like the Borat of science !
Found him by accident months ago and think I watched every single one of his videos just because of his werid and funny appraoch to things.
He would be really great running his own show on TV for kids and school science things.

Anyway:
I guess I figured out why it is so slow to react.
Those 6203 bearings, even fully cleaned out still have a lot of friction compared to dry ceramic bearings.
And there is a 15mm thick rotor filled with 0.5mm thick plastic disks - the thing is surprisingly heavy.
Important thing I found out during the testing is a flaw with the rotor design.
While the triangular axle is easy enough to print and mount the disk is has severe limitations for the venting slots and hole through the axle.
You can only go that fine with gaps when using filament :(
I have a design in mind that should work well on a resin printer but impossible using FFF.
I considered cutting slots in some aluminium tubing and using the good old 608 bearings but that won't really help with the problem of fixing the disks properly to it.
Will have to check again if I can't get a sturdy enough axle down in two halfs instead.....

As said before, the noise level is low, very low.
And right now there is a huge pressure building up in the housing while it is speeding up - I take it as a good sign as with the higher speed the pressure drops significantly.
Means the effeciency goes up with the proper airflow and well, both airflow and speed level out with what my compressor can supply.
I was aiming for around 30.000 RPM but the turbine levels out around 22.000 RPM, still not too bad for a guesstimation attempt and considering the bad friction from the bearings :)
But now actually knowing how much the system can handle without blowing apart - no screws!!, means I can reduce the number of disks and use a rotor with only 6 or 8 disks.
I would like to increase the diameter a bit more but fear the forces on the disks will be too great, so only other option is to get a better airflow on the outlet side and more speed on the inlet side.
Had a few extra holes in the testmodel to check how much air is venting out or being sucked it.
Did not reach the speed for sucking but there is a clear need to allow for some sort of channel around the disks.
One to allow to incoming air to go around the disks with less restriction to keep the speed up.
Another one, filled with some fluffy stuff on the inner walls of the housing sides.
Maybe some curved channels on the surface are enough but the airflow in this void needs to be interrupted.
I noticed in earlier models that once the thing makes the speed jump in this void a negative pressure thing happens, pulling the outer two disk against the housing.
Used a bigger gap on this test model but that is not a good solution.

Since you said Tesla turbines are for water....
It is true they can be used fine with liquids but then they have a much wider spacing if used as a turbine.
Most however are used as pump and here the disk space is determined by the viscosity of the medium and particle size of possible contaminants.
If you still have a bunch of DVD's or CD's try a little trick :
Grab a suitable axle or print one so the disks can spin freely on it without having a real gap around the axle.
You want them to be able to just run a bit with the axle if you spin it.
Now stack them closely and blow on them while you spin very slowly - they all come apart.
But if you now spin the axle and get fast enough they all come back together ;)
Same happens in a Tesla turbine, hence the need to spin them up slowly and let them come down slowly or at least with a limited airflow.
Not a problem for those tiny toy ones but try it with something having a disk diameter of more than 30cm and things get hairy LOL
Exploring the works of the old inventors, mixng them up with a modern touch.
To tinker and create means to be alive.
Bringing the long lost back means history comes alive again.
Downunder35m
Posts: 484
Joined: Sun Aug 16, 2020 5:32 am
Location: Australia
Contact:

Re: 3D printed Tesla turbine

Post by Downunder35m »

First little testrun today.
Am pleased to say it reached a bit over my target of 20.000 RPM.
As you can see in the video, the bearings still have too much friction but at least the concept works fine :)
Torque is suprisignly strong as well in the working range of around 10.000RPM.
Will have to make a stand with motor mount to check how much power it can deliver at this speed.

Anyway, here is the video....

And in case anyone wonders why I prefer a self limiting turbine when printed:
Exploding Tesla turbine - not me and not mine!
Exploring the works of the old inventors, mixng them up with a modern touch.
To tinker and create means to be alive.
Bringing the long lost back means history comes alive again.
Downunder35m
Posts: 484
Joined: Sun Aug 16, 2020 5:32 am
Location: Australia
Contact:

Re: 3D printed Tesla turbine

Post by Downunder35m »

Bit off topic in terms of the design criteria but I did not want to start a new topic as it might be included in my design one day anyway:

I looked at the flow patterns of my current design and realised there should be room for more....
As good as the Tesla design is, there is very little useful info anywhere on what exactly makes or affects these surface effects that create the binding between airflow and disks.
With the standard disk design the air resitance is minimised as there is nothing obstructing or interrupting it like in a normal blade turbine design.
Relying only on whatever the disk surface surface provides means a flat and shiny surface can't really be ideal.
I am stating this because I noticed a distinct difference between disks printed on buildtak compared to those printed on Kapton tape.
The later had a much lower efficiency but in return reached a higher top speed.
Had my baffeled for a while as you would think a "rough" surface would actually somehome affect the vortex flow in a negative way.
Then I remembered my old experiements with the spinning golf ball.
If you never tried it then use a golf ball and some airpump or vacuum cleaner where you can connect to hose to the outlet and place the golf ball on it.
If too much air use some makeshift adapter with some vent holes.
On a perfectly flat pipe outlet use some sandpaper to make a tiny indentation in one area - about as wide as your finger and no deeper than half a millimeter.
The golf ball will start spining at rather impossible speeds if balanced properly and the airflow is not too much.

I did some hand drawings to get an idea on how a circular airflow could be affected properly by these dimples you find on a golf ball.
If you read my stories about Tesla and his way of doing math or some of the things I wrote about vortex energy in general you might wonder why I bothered with a drawing.
Well, I did the same once I got near something I considered to look almost ideal.
What I got was looking like a sunflower ROFL - Should have known that from the start, shouldn't I ?
Thing is that the air around the disks must follow the same principles like any other vortex in nature.
Meaning the airspeed has to increase towards the centre while the the diameter keeps shrinking, resulting in both a higher pressure and higher speed.
Printing dimples in resin is easy, doing it in filament and layers....
The already thin layers need to be calibrated properly otherwise the "holes" are just crap.
Preferably those dimples would have to be on both sides of the disk - impossible on some FFF machine :(
Going even thinner and gluing two disk toagether to form a whole seems to be the only feasable option, but questionable if you consider the forces at over 20.000RPM.

Why dimples and not just some tiny walls or linera guides on the surface?
Geometry, symmetry and keeping a free flow area between the disks.....
A golf ball is no stable in the air because those dimples create like an airpocket around it, reducing friction while promoting the initial spin given when hit.
A shark skin like approch would work fine as well but is even harder to accomplish on a normal printer.
Ideally I should use some metal disks and etch the patterns into the surface but right now I neither have the money, nor the right tools to properly deal with such thin sheets.

One thing I might have to try before making things complex is a golden spiral printed as single surface layer.
Using an even and alternating disk arrangement I would have one slot with and one slot without the spirals on the surface.
Should at least provide some clues in terms better or worse effinciency/torque.
Exploring the works of the old inventors, mixng them up with a modern touch.
To tinker and create means to be alive.
Bringing the long lost back means history comes alive again.
Downunder35m
Posts: 484
Joined: Sun Aug 16, 2020 5:32 am
Location: Australia
Contact:

Re: 3D printed Tesla turbine

Post by Downunder35m »

Well, as it turned out there is an alternative to dimples, holes ;)
The printer is still running to make more disks and then a matching axle but I tried on in comparison to the old design.
I know, not really that scientific with just a single disk in a basic case and a straw but still...
The new design offers way more friction for the airflow, spinning up faster and if not mistaken with much more starting torque already.
I also had another look at my old Repulsine designs and thing some of it should be implemented.
We can't utilise the heating and cooling effect with printed plastic but we can use design elements of the Repulsine to improve on the airflow in the Tesla turbine!

One of the big problem preventing disk or bladeles turbines to get into the market is the lack of torque during the start phase.
-Unless you can drastically reduce the load a clutch system is required to drive a bigger generator or pump.
This is mostly due to the missing surface effects, which only come into play at a certain RPM range - depending on the disk diameter.
In theory a Tesla turbine with, let's disks of 1 meter in diameter, would have no problem starting with a generator if said generator is kept without a load during the start.
The problem comes with angular velocity and centrifugal forces - the bigger a disk the greater the forces and resulting radial speed at the circumference.
You can calculate this but I think it is enough if I say that is a very good reason why a grinding disk is limited to 10.000 RPM....
Being unable to reach supersonic speeds means the air available needs to be able to travel in the best possible way to help with the driving forces of the disks.

A housing designed so instead of a single inlet we have multiple inlets that are directed tangential towards the disks seems to be the most logical way out.
That is until you start to check how the airflow is affected by the rotation of the disks.
At operating speeds the airflow around the disks is already faster than what is possible to realise with printed inlet nozzles.
Instead of forcing a high flow speed through tiny inlet nozzles it would make sense to provide a guided, slow speed inlet - all around the disk stack.
After all, the aim to utilise the surface effects to gain the speed and efficiency, not the inlet speed of the airflow ;)
I did some intital calculations using my weird math approach and came to the conclusion that a perfect airflow can not be forced, it has to happen naturally and with as little guidance as possible.

A big problem with the axial vented desing is the axle itself.
With solid metal you can machine some stainless steel to have just tine spokes connecting either end.
Even a resin printer would struggle to create enough mechanical stability in the same dimensions.
During my test runs I realised that the design of the axle is critical to the overall funtion of the turbine.
For example, straight slots spin rather nicels even without any disks - no power but they spin and indicated a vortex airflow is formed in the housing.
However: If the slots are oriented with the airflow the axle does not move at all.
While if you turn the axle around, so the slots go against the airflow it spins even better.
Now, if the airflow over the disk is clockwise then the rotor spins clockwise as well.
As it is a closed system it makes no real difference in which direction the air exits.
I did a test with a "reveresed" axle.
A much smaller hole but the clockwise airflow was directed in a curve to become counter-clockwise when entering the hole through the axle.
To my utter surprise the system performed much better with the reversed airflow.
Have not figured out yet why that is but I intent to use it for the next design.

The bearings.....
As much I would love to say I got them working as planned, these big bearing just have a lot of friction.
I am starting to suspect that the pressure in the housing has some affect on the rubber seals, adding to the friction problem.
To avoid using the big bearing I tried printing an axle to fit into a normal 608 bearing - resulting in just a 3.5mm hole through the axle.
Thing is that printing with support structure is a real pain if you need an accurate fit.
And sanding complex parts or filing them is no real fun either.
However I noticed that even with the losses an open 608 bearing has, a covering axle or at least spacer, work quite well once the turbine is up at speed.
You just need the right balance of pressure and outgoing air.
Plus: the escaping air around the bearing somewhat fixes the issue of turbulences between housing and outer disks.
Still, I don't like the idea of losing air through the bearings as it reduced efficiency, unless I can compensate on the design so that there is actually air going in through the bearings instead of going out.
Have some ideas for that some but not quite ready yet ;)

Stay tuned as I am now quite sure I will be able to solve the problem of speed, torque and efficiency in disk turbines ;)
Exploring the works of the old inventors, mixng them up with a modern touch.
To tinker and create means to be alive.
Bringing the long lost back means history comes alive again.
Post Reply

Who is online

Users browsing this forum: No registered users and 0 guests