Boundary Layer (Tesla) Turbine Page

The Tesla Boundary Layer Turbine

Well...it sure does spin fast...real fast!

(Click on thumbnails to view fullsize images.)

NOTE: My most recent pages are back over at www.stanford.edu/~hydrobay again.

Click here for my polyphase motor/generator page.
Click here for my single-phase motor/generator page.
Click here for my solar powered fluid mechanics lab page.
Click here for my 3-axis CNC machine page.
Click here for my $662.00 supercomputer page.

Home

What it is?...

Well...an article about a Tesla blower from the September 1955 issue of Popular Science provided below will help give you a notion of that. And, from the date, also indicate this isn't exactly new technology. Basically it's flat discs with vent holes near their centers stacked up on a shaft with thin spacers between. These discs are spun by directing some fluid (air, water, burning gas, what-have-you) between them so that adhesion of the boundary layer of the fluid on the surface of the discs drags them around as the fluid travels in from the outer edge of the discs and out through the central vent holes. As the Popular Science article implies, it also makes a good pump by spinning the shaft with a motor, where-by some fluid is sucked in through the central vent holes and expelled out through the gaps between the rotating discs. This concept was patented by Croatian immigrant inventor Nikola Tesla in around the year 1909.

Now, sorry to disappoint you zero-point energy, perpetual motion machine, alternative energy source suppression conspiracy theorist, true-believer type folks. I, sincerely, know your hearts really are in the right place. But, this ain't the Holy Grail. Not even close. Ol' Tesla was a right clever guy, but also quite the showman [...those pictures of him at Wardenclyffe sitting reading a newspaper outside the Faraday cage of his giant active Tesla coil...double photographic exposure is the real hidden secret there...]. As well, sorry to disappoint you Second Law of Thermodynamics touting, CRC Handbook of Chemistry and Physics thumping true-disbeliever type folks, too. But, the minor flaws in your understanding of how these things actually work somewhat outstripe your understanding of the physics involved.

The Tesla boundary layer turbine is just that. A turbine. And, you use the same equations to solve for its power output that you do for any other kind of turbine. The tricky bit is calculating the energy transfer from the fluid stream to the rotating discs via the boundary layer. (More on that to come).

In operation, a Tesla turbine is simply a turbomachine which is reasonably efficient at a fairly high rotational speed, (dependant on several factors, including gap between the discs and disc diameter), but not too efficient to get moving. The particularly good thing about Tesla turbines is they are very easy to build. The better the quality of construction with regard to alignment, balance, materials, and such, the better will be your turbine. But, it's almost impossible to make one that won't work to some degree if you actually want it to.

Tesla blower article from Popular Science September 1955 issue:

The publication date happens to be the month and year of my birth.
...amazing what you can find on e-bay...

What's up?...

Over time I will be documenting here my investigation of various means of constructing compressed air driven boundary layer turbines, and their efficiencies. The first turbines will be constructed using mainly bailing wire and spit techniques, from polystyrene sheet plastic, lexan, machine screws, and glue. As the project progresses, and I settle in on a reasonable design, I'll make a switch to aluminum disc runners. I'll also be looking at how to make compact three-phase motor/generators for something to spin with the turbines. (See the links at the top of this page.)

I have a number of ideas on ways to computer model the energy transfer from the boundary layer to the discs, (and access to some large computers). Eventually, those results will show up here, too.

First test rotor (5.25" diameter, 10 discs):

(c.a., February 2001)

Constructed from 0.03" and 0.10" polystyrene sheet, and 0.125" and 0.375" polystyrene tubing, it's easy to spin with just breath by gripping the bearings, and blowing into the gaps between the discs at an angle a bit more than tangent to the circumference of the discs. Blowing straight in towards the center shaft doesn't work well at all. There seems to be a "sweet spot" angle for air flow somewhere between tangent and perpendicular, closer to tangent than to perpendicular.

First test shrouded turbine (6" rotor diameter, 10 discs):

(c.a., April 2001)

The disc and spacer stack assembly, (i.e., the "runner"), and the volute, (i.e., curved housing which directs fluid from the nozzle to the runner, or vice versa), constructed for this turbine were scaled from figure 1 in Nikola Tesla's 1918 US patent #1,061,142. This patent is actually for a turbopump, but it was the only picture I had at the time. It seems to work backwards just fine.

The turbine housing pieces were cut from standard 0.236" thick lexan sheet material using a hand held electric jigsaw with a fine tooth wood cutting blade. To avoid excessive vibration and cracking of the lexan material it was clamped between two pieces of 1/4" plywood while cutting. Paper cutout patterns were taped to the plywood and followed with the jigsaw. The lexan pieces were assembled by drilling and tapping them for, then screwing them together with 2-56 flat head machine screws. All screw heads were set flush in the lexan sheet by countersinking their entry holes. A sheet of thin, stiff, clear plastic was glued in place to enclose the volute space, using acid free silicon sealer.

A close look at the two pictures above will show how the air inlet was assembled from two pieces of 0.08" thick polystyrene sheet plastic and a piece of 0.375" OD polystyrene tube. And, also how the thin clear plastic used to enclose the volute was set into thin cuts in one of the 0.08" pieces of polystyrene sheet material in the inlet assembly. The air inlet is attached by two 2-56 machine screws, one each in a hole tapped into the end of the two lexan pieces cut to form the volute curve. The curved lexan pieces are attached to the outer turbine lexan frame pieces via 2-56 screws which pass through countersunk holes in the curved pieces and screw into tapped holes in the outer frame pieces.

The turbine discs and spacers, (the spacers are usually referred to as "star washers"), were cut from 0.04" polystyrene sheet with tinsnips, then machined to uniform size and shape using a small drill press with a jig that allows it be used as a vertical shape cutter. The runner discs and star washers were glued in an alternating stack on a piece of 3/8" OD polystyrene tubing. Once the stack was assembled, the central vent holes were further defined following a paper pattern with a Dremel(tm) tool using a high speed hole saw for a bit.

Four discs were cut from 0.08" thick polystyrene, and pairs were glued together without spacers to provide approximately 3/16" thick discs for the two outer runner discs. These thicker outer discs help provide rigidity to the runner assembly.

Prior to assembling the runner, tweleve 0.125" holes were drilled to align around the circumference of each disc. After assembling the runner, 0.125" polystyrene tubes were inserted into these holes and glued in place. These tubes take the place of rivets in the original Tesla design.

The plumbing parts sticking out of one of the bearing mounts allow adjusting thrust on the bearings. Thinwall brass tubing was slipped over the polystyrene shaft tube to back up the bearings and space them from the runner. The brass tubing was cut to length to center the runner in the volute space.

The model spun right up with air from a shop compressor. The runner was statically balanced, but the unit vibrated a lot. This, it seems, was mainly just a "flimsyness" issue with the bearing mounts.

Reworked 6", 10-disc turbine:

Improved bearing mounts and larger bearings:

(c.a., June 2001)

Besides beefing up the bearing mounts and installing larger diameter bearings, the polystyrene tube shaft was reinforced by driving a 1/4" steel rod though its center hole. This rod fits tightly enough so that it does not require gluing or pinning to the polystyrene tube to transfer out mechanical power from from the spinning runner without slipping. One end of the steel rod was threaded, and a 3/16" thick 1-5/8" diameter steel disc with a center hole and four surrounding holes was threaded onto the shaft and then brazed in place to use for a mechanical mount point. (Before driving it into the polystyrene center tube, of course.)

The thin plastic sheet used to enclose the volute space is held in place by double-sided foam tape in the reworked version of the turbine, rather than the clear silicon sealer used in the first version. For extra reinforcement, pieces of 0.125" polystyrene tubing were attached between the turbine housing side plates so that they clamp the thin plastic sheet in place. This is just a feels good feature. There really is no pressure buildup in the volute. You could even punch drain holes in the bottom of the volute if moisture condenstation were a problem with no effect on the turbine's operation.

Added 1.063 kg flywheel:

(c.a., December 2001)

The flywheel is approximately 5.34" in diameter cut from 3/8" steel plate. A coupling nut was turned round over about three-fourths of its length, and a center hole was drilled in the flywheel so that the turned portion of the coupling nut fits closely in the hole. Four more holes were drilled around the flywheel's center hole to match with the holes with the mounting plate on the 1/4" rod inserted in the turbine's polystyrene tube shaft. The flywheel is mounted on the turbine by slipping its center hole over the threaded portion of the 1/4" rod that extends from the mounting disc, then threading the turned end of the coupling nut onto the threaded shaft while aliging the flywheel so the coupling nut passes through it's center hole. When the coupling nut is threaded up snuggly against the mounting disc it provides support for, and centers the flywheel on the 1/4" rod shaft. Then four 1/4"-20 screws are inserted through the remaining holes in the flywheel and aligned with and threaded into the 1/4"-20 screw holes in the mounting plate.

Now the turbine spins smoothly at over 2000 rpm, (measured with a model airplane prop tach), when connected to a compressed air supply at about 75 psi, though a 25 foot long, 3/8" diameter line. Though it really is very smooth, I'm nervous about trying to spin it much faster, given the construction!

[...well, I actually did push it for all it was worth once...but didn't get close enough to try and measure its rpm...concern about being around the flywheel rather than the rotor...it's going damn fast before high frequency vibrations start it skittering about on the floor...]

It takes full flow from the shop air supply to get the turbine up to speed in a reasonable length of time (though it will get up to speed eventually at much lower air flows). Once at speed, it takes comparatively little flow to keep it at speed, even when adding a friction load to the flywheel (toe of work boot on flywheel face).

One of these days I'll get another first stage regulator for my SCUBA tank, a variable area flow meter, and few other odds and ends and use them to set up a flow metering bench. That, a tachometer, and an alternator hanging off the turbine shaft with a variable load should be enough to do some basic flow vs. power experiments.

A preliminary simulation:

(c.a., June 2002)

The code:

NOTE: If you picked up this program prior to 10 October 2005, there was a factor of 2 error in the output. Although I described it in the comments, I neglected to account for runner discs having two sides in the actual code. That has been fixed here now.

This program, TurbTorque.c, based on an article from TEBA News [1] , is an application of straight forward fluid mechanics techniques, and not a sophisticated model of real Tesla turbine operation. None-the-less, it does provide some insight into the effects relative to rpm of two parameters on turbine operation, namely disc diameter, and disc spacing.

The code simply computes values using prederived equations. The equations are obtained by noting that only the tangential velocity component of the flow contributes to the shear stress on the surface of a turbine disc, and that the force acting on a small area on the surface of a turbine disc is equal to the shear stress at that area times the area. Torque on the small area is the force on the area times the distance of that area from the center of the disc. And, power is torque times angular velocity.

By double integration over the surface of the disc for the stress as given by the equation for shear stress from steady flow between two parallel plates, and using the tangential velocity obtained by differentiating the flow given in polar coordinates for a spiral vortex with a central sink in two dimensional plane potential flow, a formula for the total torque on the surface of the disc can be obtained which has only the disc spacing, and inner and outer radii of the disc (vent radius and disc radius, respectively) as variables. The equation solves for the total torque on one face of the disc, from which the power can be derived. Two times this value gives the power contribution from one disc, and the product of that value with the total number of discs minus 1 give the total turbine power. One is subtracted from the disc count to remove the outer faces of the outside discs, which are assumed to make a negligible contribution to the total power generated.

(Yes, I'll eventually flesh this out with equations. For now, refer to the TEBA News article.)

A better solution to the problem would be modeling of the real physics and solving the Navier-Stokes equations for the system. In general solving the Navier-Stokes equations is not easy to do, except for special cases, and exceptionally difficult for vorticies. Large computers and new techniques for solving three dimensional physical systems will help.

The code presented below was written using the StormC 3.0 Professional C compiler for the Amiga computer. There is a lot of Amiga specific code related to CLI input and output and the graphics display you can probably ignore. If you are familiar with the Amiga you will notice some unfamiliar library calls. I've been programming for a long time and have developed a number of my own libraries for simplifying rote graphics and I/O coding. The equation solving section is below the MAIN CODE separator comment.

The results:

The results given below are from running the program for the following cases:

6 inch discs, 0.031 inch gap
6 inch discs, 0.020 inch gap
8 inch discs, 0.031 inch gap
8 inch discs, 0.020 inch gap
12 inch discs, 0.031 inch gap
12 inch discs, 0.020 inch gap

The progam takes user input for disc diameter and gap width. It assumes the same vent diameter to disc diameter ratio as for Tesla's 9.75 inch turbine, and performs calculations for 6, 9, 12, and 15 disc runner stacks with air as the working fluid over a range of 2000 rpm to 10000 rpm. The data are presented in both graphical and numerical form.

As can be seen from the data, faster is better, bigger is better, and closer is better... (one of these days I'll do some percent difference calculations on the numbers...)

A Spreadsheet Version:

(c.a., September 2005)

A spreadsheet version of the calculations from the turbtorque.c code, turbtorque.xls, provides a bit more flexibility in user input than does the original C program. If your web browsing system is set up with a spreadsheet program that supports the .xls (Excel) format, (I use OpenOffice.org Calc myself), then clicking the in-line link in the previous sentence should open the turbtorque.xls spreadsheet for your use. If the spreadsheet doesn't open on your system, then you should be able to save the file and try it on a system that does have a spreadsheet program available.

The screen capture image below will be useful in following the short description of how to use the spreadsheet program that follows:

All parameters of the spreadsheet are free for editing. The simplest use is to change the runner gap width and disc diameter values in the yellow block at the upper left corner of the spreadsheet. Enter your desired data, (in inches), and the spreadsheet will calculate and plot the potential power output (in horsepower-electric) for runners with the given input gap and diameter having 6, 12, 9, 15, and 18 discs (including the wide outer discs) over a rotational speed range of 0 to 27000 rpm. The legend to the right of the plot gives the colors for the plot lines relative to the number of discs in each runner.

Besides varying gap and diameter values, you can also change the number of runner discs and the rpm range. Below the plot are two green data blocks. Changing the "Mindiscs" and "DiskStp" values in the upper block recalculates the "Ndiscs" row values, which give the number of discs for each runner being simulated. The Mindiscs parameter gives the lowest number of discs for a simulated runner, and DiskStp gives the number of discs to increment up from the Mindiscs value for each runner in a simulation set. Similarly, changing the "Minrevs" and "Revstp" parameters in the lower green block will recalculate the rpm values in the "Rev" column, which give the rotational speed values for each simulation.

It is also possible to change the number of runner disc sets and the number of rpm values for a simulation. The red data cells which are associated with the green cell blocks allow this. Changing the "Numrun" value in the upper red block will change the number of runner disc sets calculated from the Mindiscs and DiskStp values. Changing the "Numrev" value in the lower red block will change the number of rpm values calculated from the Minrevs and Revstp parameters. However, for changes in the Numrun and Numrev parameters to have any visible effect, the correct number of cells in the Ndiscs row and the Rev column must be added or deleted as required. The format of the cell formulas can be found in the existing cells. You must increment the subtracted value in the last cell by one in each added cell. Also, for the changes to be seen in the plot, the plot range must be modified to include any added data values.

Major modification:

(c.a., April 2003)

The reason:

Broke it...

An unfortunate bicycle incident, (the physics of which would be hard to describe unless you happen to be familiar with the arrangment of my apartment), resulted in severe damage to the main runner housing and a bent rotor shaft. The bent shaft was easy to straighten, but the runner housing required extensive repairs. Rather than simply rebuild what was, I decided to just keep the original 6" turbine runner, bearing mounts, and base plate, and, using the tried and true hacksaw and hammer techinques from before, build a prototype version of the form of turbine I plan to construct in the near future using pieces cut with a CNC machine.

The salient features of the new form turbine are a genuine inlet nozzle, and an easily removable runner. The runner is removable to allow experimentation with different runner disc spacings, star washer types, and vent configurations. The inlet nozzle is also removable to allow for tests with different nozzle arrangements and alignments.

With the new form turbine being easily reconfigurable, and since its housing is transparent, it will be possible to construct a runner with a few of its outer discs also made from transparent material and do real flow visualization via smoke injection using an inlet nozzle modified for that purpose. (If I can figure out how to make it hold water, I'll do visualization with dye injection instead; applying Reynolds number scaling during testing. With all other scale factors fixed, that would mean just scaling rpm by the ratio of the viscosity of air to the viscosity of water, so, a few hundred rpm of the runner driven by water would be equivalent to a few thousand rpm driven by air.)

The efficiency of the new style housing and nozzle turbine setup is clearly higher than that of the old pump style volute turbine. For the new turbine it is possible to bring the 1.063 kg flywheel up to about 40 rpm with 15 good breaths through the inlet hose, then sustain rotation by easier breaths, and, for a while, not pass out. (Be advised I'm a certified research SCUBA diver with a measured vital lung capacity about twice that of the average person.) With the pump style volute it was possible to get the flywheel to move, but not to sutain rotations by breath alone. If that still doesn't sound impressive, lay a can of soup on the floor, (about 0.5 kg), and try and roll it across your kitchen by breath power alone...;^)

Construction of the new prototype housing and inlet nozzle is described in the text subsections and photographs that follow.

The inlet nozzle:

The nozzle was constructed from 0.091 inch thick clear polycarbonate sheet, 0.375 inch OD polystyrene tube, and clear polyester casting resin. It was made to have a flat top and bottom, with sides straight in the vertical and curved in the horizontal to form a typical nozzle shape when seen in a plan view, i.e., in vertical cross section each section of the nozzle is a simple rectangle with fixed height relative to a common centerline, and having a width that varies depending on the position of the section along the length of the nozzle centerline. The form of the nozzle curve comes from the design of a laboratory demonstration nozzle, the dimensions of which came out of an old aerodynamics lab worksheet I found laying around somewhere a long while ago. (I never took the lab, I just found the worksheet.) The scaled points on the nozzle's curved sides relate to the position of pressure sensor tap points in the demonstration nozzle.

The nozzle width was taken to be the distance between the inner sides of the outer runner discs, that being approximately 0.68 inches, giving a nozzle half-width of 0.34 inches. The half-width provides the distance from the centerline to the maximum distance of the nozzle curve of one side away from the center line. Since the nozzle is horizontally symmetric only one side curve need be dimensioned, and the opposite side simply mirrored from the dimensioned side. The nozzle thickness was taken to be the diameter of the polystyrene inlet tube, 0.375 inches.

With Y representing the distance of the nozzle curve from the half-width line towards the centerline, and X representing the distance of tap points from the nozzle inlet, the scaled curve points, in inches, are:

**Nozzle Tap Points**
TAP	X	Y
0	-0.334	0.152
1	0	0
2	0.334	0
3	0.415	0.160
4	0.482	0.197
5	0.548	0.221
6	0.615	0.231
7	0.676	0.227
8	0.756	0.215
9	0.876	0.187
10	1.010	0.154
11	1.144	0.120
12	1.405	0.057
13	1.646	0
14	2.007	0

Tap 0 is an added position which brings the curve back to meet the outside diameter of the 0.375 inch inlet tube, thus forming an expansion chamber behind the full-width nozzle inlet given by the original demonstration nozzle tap positions 1 through 14.

Using a pattern generated from the nozzle tap position data a core of modeling clay was formed to fill the nozzle void space during the casting process. The clay was kneaded and rolled to the required 0.375 inch thickness and then cut to the proper shape by running an X-Acto(tm) knife blade around the pattern. The 0.375 inch OD tube was stuck into the expansion chamber end of the clay form to complete the void mold piece.

Conveniently, I had a section of square metal tubing which had the precise thickness required for clamping the nozzle assembly's 0.091 inch thick polycarbonate side pieces to so that they ended up spaced apart at the same outside width as the outside width of the turbine housing. So, laying a piece of 0.091 inch thick polycarbonate sheet cut to fit between the clamped side pieces on the metal tube as a bottom for the nozzle assembly, centering the clay core piece on this botton sheet, and drilling a 0.375 inch hole in another piece of 0.091 inch polycarbonate so that the hole's outside edge was 0.091 inches from the edge of the sheet, and fitting this hole over the 0.375 inch OD polystyrene tube extending from the clay core piece to cap the inlet end of the nozzle assembly was all that was required to make the nozzle core mold. Another piece of 0.091 inch polystyrene sheet was cut to make a top piece for the nozzle, but this piece was not actually installed until after the casting process was completed.

Sealing up the gaps with masking tape and pouring catalyzed casting resin to slightly overfill the the mold form produces the actual nozzle curve side pieces. Prior to pouring the casting resin was degassed using the simple vacuum chamber as described in the subsection titled "A simple vacuum chamber" that follows. Degassing helps prevent bubbles from forming in the casting resin as it cures.

Since there are bound to be some leaks and spills no matter how carefully you try and seal the mold pieces, a piece of waxed paper was wrapped around the square metal tube before the mold pieces were clamped in place for final assembly. Without that, once the casting resin has cured, it could end up being very difficult to remove the nozzle assembly from the metal tube it was clamped to.

Once the casting resin was fully cured, the excess extensions of the 0.091 inch polycarbonate sheet pieces were removed by various sorts of clamping, filing, scraping, and sawing machinations; involving square metal stock pieces, small c-clamps, a machinest's vice, an X-Acto saw, a cabinet scraper, and a Dremel tool with a cutoff wheel installed.

With the side and bottom pieces trimmed and squared, and the excess casting resin filed flush with the top of the side pieces, the clay core mold piece was removed by first digging it out, and then scraping the edges of the newly cast nozzle side pieces with an X-Acto blade knife. (Note at this point the nozzle's polycarbonate end piece was left extended to allow easier alignment of the yet to be installed nozzle top piece.)

After cleaning out the clay mold core, the nozzle's 0.091 inch polycarbonate top piece was installed by running a bead of gel style cyanoacrylate glue around the top edge of the cast resin piece and clamping the precut polycarbonate sheet into place. After the cyanoacrylate glue cured, the extension of the nozzle end piece was trimmed flush with the nozzle top piece.

The final result was a nozzle with outside width the same as the width of the turbine housing, and an exit port size of 0.735 by 0.377 inches. The final exit port size is well close enough to the design port size of 0.680 by 0.375 inches for our purposes.

The nozzle was installed into the turbine housing by cutting a slot in both housing side pieces with a hacksaw, using a piece of square metal stock clamped in place as a guide along marks drawn to fit the nozzle. Once the slots were cut, the nozzle was set in place and marked for shaping of the nozzle exit port region to have the same curve as the runner holes in the turbine housing side pieces. The nozzle was then shaped to the marked curve using a small sanding drum in a Dremel tool. Once the nozzle was shaped to fit the turbine housing side piece runner holes, it was fixed in place by drilling three holes each in two pieces of 0.25 inch square metal stock about 1 inch long, the two outer holes tapped for 4-40 machine screws, and the middle hole drilled for loose fit of a 4-40 machine screw. These metal pieces were attached to the outside of the turbine housing side pieces by 4-40 screws inserted through holes drilled above an below the nozzle slot, and the nozzle itself was drilled and tapped to accept 4-40 machine screws inserted through the middle holes in the 0.25 inch square stock pieces. All the screw holes were countersunk for the tapered machine screw heads.

A simple vacuum chamber:

Probably more a matter of just being anal, than being of any great necessity for this project, but, degassing casting resin can be helpful in preventing bubbles from forming in the resin as it cures, which makes for a prettier cast. And, in the case of something like the nozzle being constructed here, degassing can help prevent voids caused by bubbles that adhered to the clay mold core from appearing in the inner sidewalls of the final item when the clay is removed. Degassing is performed by exposing the catalyzed resin to vacuum for about half of its cure time, then pouring the resin. Gas (air) in the resin is pulled into the vacuum so it is not available later for bubble formation.

A simple vacuum chamber was constructed out of a large glass jar, originally containing 3-bean salad, and having an opening diameter of about 3-1/2 inches, plus the threaded sleeve used to run wire through from the bulb socket to the base of a standard household lamp. The threaded sleeve was inserted into a hole punched in the metal screwtop lid of the glass jar, and a nut run up the sleeve to the jar lid from both the top and the bottom, clamping the sleeve in place. The sleeve and nuts were then sealed to the jar lid with a layer of red RTV silicon rubber. A piece of 0.25 inch ID vinyl tubing was slipped over the end of the threaded sleeve extending from the top of the jar lid for attaching the jar to a vacuum source. A clear plastic "party cup" with a wire bridle was used for a resin holder.

The source of vacuum was a hand pump of the type used by automotive mechanics for testing smog systems and break systems. In use, with a fair bit of continuous pumping effort, it was possible to pull about 25 inches of vacuum on casting resin in the cup inside the jar.

To use the chamber you must pump continuously for the degassing period (about 7 minutes in the case of the 15 minute cure time casting resin used here). If you just pump to the maximum possible vacuum level, then sit back and relax, the gas in the resin will expand out to fill the vacuum, but no more will leave after that.

Bubbles will form in the resin during the degassing period, and a little flurry of bubbling when the vacuum is removed is common. Once the vacuum is removed and the bubbling competed, the resin should appear clear with perhaps a few bubbles still on its surface. Since you've already used up half your cure time, pour quickly.

The housing:

The new turbine housing uses the same base plate and bearing mounts as the original housing. But, the mounts have been reinforced by fitting a piece of 0.25 inch thick by 1 inch wide aluminum bar stock between the bearing mount side pieces, attaching them to the bearing mounts by drilling, tapping, and countersinking for 4-40 flat head machine screws. The aluminum bar pieces are drilled and tapped to accept 10-14 screws for attaching the bearing mounts to the turbine base plate, and also drilled and tapped to accept 4-40 screws passed through 0.25 inch square steel stock pieces used to stiffen the turbine base plate.

Side pieces for the new 6 inch turbine runner housing were cut from 0.223 inch clear polycarbonate sheet using a jigsaw with a fine tooth carpentry blade following a taped on pattern. Holes 3.125 inch in radius to allow installation of the turbine runner were also cut in the side pieces by following the taped on pattern with the jigsaw. Slots for the inlet nozzle assembly were cut in the side pieces as described in the section above titled "The inlet nozzle." Spacers for the side plates were cut from 0.25 inch thick 0.75 inch wide aluminum bar stock. The side pieces were drilled and countersunk for 4-40 machine screws to attach the aluminum bar spacers. The aluminum bar spacers were drilled and tapped to accept 4-40 machine screws to attach the side plates. The bottom spacer was also drilled and tapped for 10-24 machine screws to attach the runner housing to the turbine base plate, and drilled and tapped for 4-40 screws passed through 0.25 inch square steel stock pieces used to stiffen the turbine base plate.

Circular vent plates were cut from 0.091 inch thick polycarbonate sheet by clamping the material between 0.5 inch square metal bar stock pieces, scribing and snapping off pieces to form a near round shape, then trimming them smooth using a pair of tin snips. Vent holes 1.25 inches in radius were cut in the center of the vent plates using a holesaw chucked in a drill press.

The side plates were drilled and tapped for eight 4-40 machine screws around the circumference of the turbine runner holes, and in a matching pattern, the vent plates were drilled and countersunk for 4-40 machine screws, for attachment of the vent plates to the side panels.

The volute was sealed by attaching a strip of 0.030 inch thick clear plastic sheet between the side pieces with 0.125 inch thick double sided foam tape applied to the circumference of the turbine runner installation holes, and across the bottom of the mouth of the inlet nozzle. The strip was fit first with the inward facing side of the double sided tape covered so that it would not stick in place before its time. The strip was marked for cutting an opening for the inlet nozzle exit port so that the bottom of the nozzle is flush with the cut opening, but the top of the cut is at the top of the runner hole, not at the top of the inlet nozzle exit port. This allows unobstructed flow out of the nozzle. A small strip of the 0.030 inch clear sheet the cut to the width of the space between the turbine side pieces was glued to the main strip so that it extends back from the top of the runner hole and overlaps the top of the inlet nozzle and completes sealing of the volute. The main strip was cut wide and trimmed off with an X-Acto knife after the inner double sided foam tape surface was exposed and the main strip set in its final position against the tape.

The original bearing thrust screw adjustment was discarded in favor of set screw collars slipped onto the turbine runner shaft inside of the bearings. With the bearings pressed into their recepticals, the turbine runner is centered in the turbine housing, and the set screw collars are then slid up against the bearings and their set screws tightened to retain the bearings and fix the postion of the runner in the housing.

Putting it all together:

As can be seen in the following photo sequence, the turbine can now easily be completely disassembled and reassembled, allowing changing runners for different test. Plus, there is even a spot for a bit of duct tape. No home project is complete without it!

Dimensioned drawings:

Here are dimensioned drawings of the major pieces of the 6 inch turbine housing upgrade components, plus runner disc and star washer drawings. Note the washer drawing shows both short spoke and long spoke varieties. Short spoke star washers were used in constructing the transparent runner assembly described below. [cite tesla long spoke improvement patent here someday]

Click on thumbnails for fixed size jpeg images. Click on the link below a thumbnail for an svg image. You can download svg image viewers for most browsers from www.adobe.com.

SidePanel.svg VentPlate.svg SeparatorPlates.svg BearingMountPad.svg

4SpokeRunner.svg 4SpokeWasher.svg NozzleCore.svg

I can see clearly now...

(c.a., July 2003)

With a transparent runner it should be possible to see streamline flow patters between runner discs by injecting a tracer into the driving gas stream. Even if it proves difficult to characterize, if it works it should at least look cool. So, of course, here is how I built first the jigs to build one, and then built one.

Arc cutter jig:

The arc cutter jig is used to cut circles for runner discs and star washers, and for cutting the arcs necessary for the vent cutter jig. It has two parts, a cutting board with a 0.375 inch diameter metal peg extending perpendicular to it's surface, and a Dremel tool mounted in a router jig with that jig screwed to a 0.25 inch thick flat aluminum plate. A 0.375 inch diameter center hole was drilled near one end of the aluminum plate, and holes spaced down the plate appropriately for each desired radius for the Dremel cutter to extend through. Four holes were drilled through the Dremel router base for 3/16th inch machine screws that attach the router plate to the aluminum plate. Spacing for the router bit from the center hole for each arc is set by the placement of 4 base plate mounting screw holes in the aluminum base plate.

Drilling a 0.375 inch hole in the material to be cut, setting the material with its hole over the metal peg on the cutting board, and placing then placing the 0.375 inch center hole in the aluminum router plate over the metal peg allows the material to be cut with a radius determined by the distance from the center of the metal peg in the cutting board to the edge of the Dremel router bit closest to the metal peg. On thin material it is reasonable to plunge start the cut. For thick material, it is better to drill a starting hole for the Dremel router cutter to pass through.

Vent cutter jig:

The vent cutter jig consists of several parts, including a pattern piece that a Dremel router with a cutter having a guide bearing is run inside of to cut the vent holes, and several alignment and support pieces. To construct the vent cutter jig, the arc cutter jig was used to cut two arcs relative to a 0.375 inch center hole of the proper radii for the inner and outer edges of a runner disc vent hole. To make the proper radius arcs, the excess width of the guide bearing over the width of the router bit must be taken into account. Here, a 5/16th inch OD, 1/8th inch ID high speed ball bearing intended for gas engine powered model race cars is used for the guide bearding. The cutter is a 3/16th inch Dremel straight cutter with 1/8th inch shaft. With the bearing slipped over the cutter shaft, the cutter edge will be 1/16th inch inside of whatever edge the bearing rides one, so, the outer arc must be 1/16th inch outside of the desired outer vent arc, and the inner arc must be 1/16th inch inside of the desired inner vent arc; similarly, the straight sides of the vent jig must be displaced 1/16th inch away from the center of the vent. Also, very important this, the half-width of the cutter must also be taken into account. Besides the 1/16th inch offset for the guide bearing, the thickness of the cutter from its own center must be compensated for in setting the jig arcs. That means, for a 3/16th inch cutter the placement of the router on the arc cutter aluminum plate must place the centerline of the cutter 3/32nds of an inch inside the desired arc radius for an outside cut, and 3/32nds of an inch outside the desired arc radius for an inside cut. This cutter diameter compensation was made for all arcs available, both inside and out, in placing the mounting screw holes on the arc cutter jig aluminum plate. So, only the 1/16th inch compensation for the guide bearing was necessary in sizing the vent cutter jig pattern guide hole.

With proper arcs and straight sides cut, a piece of waxed paper was taped to the bottom of the the arc cutter guide plate under the guide hole region, and the excess arc cut areas were blocked with cut pieces of cardboard. The excess arc cut areas where then filled with 5 minute epoxy resin to provide a continuous pattern surface for the guide bearing to ride on.

Holes for two alignment pins cut from 3/16th inch brass welding rod were added to the arc cutter jig cutting board, one that aligns with one of the runner inner assembly rod holes, and one that aligns with one of the runner outer assembly rod holes. With these, it so possible by rotating and flipping a runner disc on the 0.375 guide peg in the cutting board to make proper alignment for all four vent holes in a runner relative to the guide pattern hole in the vent cutter jig.

To complete the vent cutter jig a clamping plate assembly was construted. Along with the hole for the 0.375 inch guide peg already present, two 3/16th inch holes for the welding rod guide pins were added to the vent hole bearing guide pattern plate. On the end of the bearing guide plate away from the 0.375 inch peg hole, holes were drilled and counter sunk to allow attaching a strip of 0.223 inch material about an inch wide to act as a spacer when cutting verts in 0.223 inch thich discs and as a place for clamping the jig to the cutting board once all the parts and materials are properly set on the guide peg and pins.

When cutting vents in 0.223 inch material, only the cutting board with alignment peg and pins, and the bearing guide plate are required to make a cut. But, when cutting thinner material you need to compensate for the difference in thicknesses. This is accomplished with a strip of material the same thickness as the material you are cutting placed under the 0.223 inch strip on the end of the bearing guide plate, and also another 0.223 inch plate with a hole centered under and larger than the bearing guide hole, plus holes for the cutting board guide peg and pins. This piece is placed on top of the thin material being cut to level the surface of the guide plate for the Dremel router.

The first test of the vert cutter jig was on an assembly hole pattern jig for 6 inch turbine discs. As you can see, counter clockwise from the bottom, the quality of the cut holes improves greatly. The bottom vent hole is particularly rough as on that pass both the cut pocket in the cutting board and the vent hole were being produced. With no good path to escape wood chips caused a lot of chatter and heating (which makes the material turn "gummy"). The remainder are just improvement with practice.

Hole pattern jig:

Consistency counts; particularly when building something with the balance requirements of a turbine. So, to facilitate drilling holes in runner discs in something like the same place for every copy, a hole pattern jig was made from a disc of 0.223 inch clear polycarbonate material. The disc was cut roughly with a jigsaw to slightly oversized for a 6 inch turbine, and a paper pattern then taped to the disc. The hole positions were transferred from the pattern to the disc with a spring loaded metal punch, and the holes drilled appropriately, 0.375 inch for the axel shaft and 0.125 inch for all the assembly rod holes. The disc was then mounted in a lathe and turned to the correct diameter. In use, a 0.375 inch hole is drilled in the material a disc is to be produced from, and that material is set on the guide peg in the arc cutter cutting board. The hole pattern jig is set over the disc material and a c-clamp used to hold the pattern and disc material in place. Then a 0.125 inch transfer punch and a small hammer are used to mark all the assembly rod holes relative to the 0.375 inch axel hole. With the holes marked the disc is unclamped, the transfer marks made into true punch marks using the spring loaded metal punch, and the punch marks are used as a guide to drill the 0.125 inch assembly rod hole with a drill press.

Transparent runner discs:

Finally! With all the required tools, now we can make some runner discs. Two 0.223 inch thick discs for the runner ends, and four 0.091 inch discs for the inner runner, with five 0.031 inch thick star washers makes a complete runner assembly stack 0.865 inches thick. This is plenty close enough to the original pump style runner's 0.960 inch thickness to use in the new turbine housing with no modifications to the housing. So, these discs, plus a couple of others, were cut using the new jigs.

Transparent runner star washers:

(c.a., April 2004)

That's transparent-runner star washers, not transparent runner-star-washers. Back the better part of a year ago I drilled and cut discs from 0.031 inch white polystyrene sheet to use making star washers for the transparent turbine runner. (It really only matters that the runner discs are transparent to see dye flow. Nothing goes between the star washers.) Just a matter of using the vent cutter jig and trimming off the excess with an X-Acto blade knife to finish them.

Transparent runner assembly:

With everything fitting together nicely, time to relocate to a more dust free envrionment, wash all the greasy finger prints off the runner discs and star washers, and let them dry.

With everything clean, and a board with a 0.375 inch hole drilled squarely through it as well as holes partially through for all the assembly rods, using the hole drilling jig as a guide, we're ready to start sticking things together. Literally, with polystyrene plastic cement. First thing, and if you already don't know this trick, you'll thank me for it later, do NOT try and put the runner together with assembly rods of the same length! Rather, cut them so that they can be placed in the assembly board in a spiral from tall rod to short rod with an eighth to a quarter inch difference in length between each rod in the spiral sequence. This way you can insert the first assembly rod into its hole, and slide it down a bit before you have to insert the next rod into its hole, which means the first rod will not jump out of its hole when you try and insert the second, etc., etc. Really. Do it this way. You'll save yourself a trip to a pretty white room in a nice jacket with very long sleeves. (This trick also works well for things like hand inserting can style op-amps into printed circut boards, by the way. Just cut the leads in a similar spiral pattern.)

Once the first outer disc is in place flat on the assembly board, take the glue brush and daub a bit around each assembly rod, and around the axle rod. Do this for each disc once it is in place to lock the assembly rods at the proper distance for disc separation. The stacking sequence is outer disc, star washer, inner dsic, star washer, inner disc, star washer, inner disc, start washer, inner disc, star washer, outer disc. Before placing a star washer first coat the spokes of the runner disc it is to mate with with cement, and also its spokes on the side it is to mate with. Then place the star washer and press it down firmly. Similarly when you place a runner disc, coat the spokes of the star washer already in place it is to mate with, and its own spokes on the mating side, then work it down the assembly rods and press it frimly in place. To maintain the proper runner disc spacing, place a small piece of 0.031 inch thick polystyrene near each assembly rod and then press the runner disc spokes firmly down and more gently press down near the 0.031 inch polystyrene pieces. Again, daub a little cement around the assembly rods and the axle with each level assembled. The inner runner discs are fairly stiff, so it isn't necessary to keep stacking 0.031 inch polystyrene pieces all the way up the assembly at each assembly rod location. They just need to be in place to set the distance between layers as they are assembled. Of course, it isn't a bad idea to place a 0.031 inch piece into each level at the time you press the runner disc into its final position before applying a daub of glue around the closest assembly rod, just in case the glue is still soft in layers below, but you can do that on a rod by rod basis so long as the runner is held separate from the one below by 0.031 pieces all the way around.

Speaking of all the way around, note the alignmarks on each runner disc in the assembled runner stack in the photo below. If you look closely, you'll see a matching set on the assembly board. I said consistency counts. The hole drilling jig is very consistent, but, it is not perfectly symmetric. Unless your's is perfect, there will be a few tiny wobbles in the hole spacing, so you'll need to mark your runner discs for proper alignment at assembly.

Once the stack is assembled, place 0.031 inch thick pieces of polystyrene in each gap in alignment with two opposite spokes, then gently clamp those two spokes at the hub and leave it overnight for the glue to cure. After that the completed runner can be removed from the assembly board and the assembly rods trimmed off with an X-Acto knife. (Be careful to not scratch up your nice transparent surfaces!)

As with the pump style turbine runner, a length of 0.25 inch diameter cold rolled steel rod was driven through the 0.375 inch OD polystyrene axle tube to provide extra rigidity. After that the runner was chucked in a lathe for a bit of truing. You'll want to make shallow, slow passes with a very sharp and fine point cutter. Otherwise you run the risk of sealing your gaps with rolled over material. A water cooling mist might be appropriate, but not absolutely necessary. Even being careful and slow, the gap edges are likely to flare slightly. This can be fixed by turning the lathe chuck by hand while scraping the gap edges with an X-Acto knife blade. For a finish, you can take a piece of fine wet-dry sanding paper and insert it between and hold it on the gap edges while the chuck is spun at a few hundred rpm.

Transparent runner installation:

Simple. Remove the old pump style opaque runner and replace it with the new transparent runner. A while back I decided it would be simpler to place a turbine in something like an aquarium and inject vegetable dye into a water stream, rather than try and figure out how to generate fine particle smoke on demand and inject that into an air stream (and probably a lot easier on the environment); and, hence, the link to the solar powered fluid mechanics lab at the top of this page. To that end I drilled a small hole at an angle towards the outlet of the inlet nozzle and inserted it a hypodermic needle for dye injection. This can be seen in some of the photos below. A 0.031 inch thick white polystrene piece was cut to fit under one of the circluar vent pieces on the turbine houseing to act as a background for viewing injected dye.

Take it for a spin!

(c.a., August 2004)

With the transparent runner turbine finished a few months ago, I haven't quit on this project, but, I've been working mainly on adjunct portions, in particular, the Solar Powered Fluid Mechanics Lab (SPFML). That has its own web page, which can be accessed from its link at the top of this page.

Turbine testing via the SPFML will be documented on this page, but detailing the SPFML itself will be left to the SPFML web page. Testing to date consists mainly of fixing leaks. However, I did make a few attempts at dye injection for flow visualization. And, I've decided that it is difficult to get it right, and very very messy. I'm giving up on dye injection and will be giving hydrogen bubble flow visualization a try in the near future.

Below is a link to video of the transparent runner turbine spinning away via water flow in the SPFML.

Visualization Upgrades

A few simple changes to the turbine were necessary to implement hydrogen bubbling visualization, and expedite switching runners for investigating the effects of changing vent shapes and vent-to-runner size ratios.

Bubbler electrodes:

Hydrogen bubble flow visualization works through a controlled electrical dissociation of water into oxygen bubbles outside of the visualization region, and a very fine, pulsating stream of hydrogen bubbles injected into the region of the flow to be visualized. The path followed by the hydrogen bubble stream shows the flow streamlines.

The circuit used here for applying power to the water dissociation anode and cathode electrodes is described on the Solar Powered Fluid Mechanics Laboratory page. Construction of the anode and cathode electrodes and their attachment to the turbine housing is described below. During water dissociation, oxygen is produced at the anode, and hydrogen is produced at the cathode.

anode:

(c.a., February 2005)

A circular piece about three inches in diameter punched from a thin sheet of copper was used for the anode. The center of the disc was marked through a pattern drawn on a piece of tracing paper, and an entry hole drilled then cut and filed large enough for the runner shaft bearing retaining collars to slip through. The disc was then mounted on one of the turbine bearing stands using brass tubing standoffs, 4-40 brass screws, and nylon insert lock nuts. The electrical connection to the disc is made via a wire eyelet trapped under a wing nut on one of the mounting screws.

cathode:

(c.a., October 2005)

The cathode was made from a 0.003 inch diameter piece of platinum wire threaded through two holes drilled near the mouth of the nozzle so that the wire runs parallel to the runner shaft in the center of the nozzle output stream.

Quick removal runner:

(c.a., October 2005)

The major modifications to the turbine housing of two-and-a-half years ago, (hard to believe it has been that long!), did make it relatively easy to change runners. But, it still required removing a lot of screws and a good idea of how all the bracing pieces are mounted to know which screws to remove. That's OK when you're just fiddling about, but, particularly for demonstration purposes, a bit slow and inconvenient. So, I made some changes that allow swapping runners without need of a screwdriver.

Two 1-inch square blocks were cut from 1-by-3/4 inch aluminum bar. These were drilled and tapped so they could be attached on top of the existing aluminum strip used to mount the bearing support on the opposite side of the turbine housing from the hydrogen bubbler anode. The outer threaded holes in the existing aluminum strip were aligned in the drill press using a transfer punch in the drill chuck, then drilled out to allow 10-24 screws to pass through the strip and screw into the threaded holes in the newly cut 1-inch square aluminum blocks.

The 1-inch square blocks provide sufficient backing for the bearing mount lexan pieces so that two 6-32 thread thumbscrews on each end of the assembly can be used to retain the bearing mount; allowing the 2-56 screws previously used to attach the bearing mount to the original aluminum strip to be permanantly removed.

Now, by removing the 4 thumbscrews and pushing the bearing out of its pocket in the bearing mount while tipping the mount away from the turbine housing, the bearing mount is easily removed for access to the runner.

With the bearing mount out of the way, the housing vent plate still blocks the turbine runner from being removed. To eliminate the need to use a screwdriver to remove the vent plate, I disassembled the side of the turbine housing away from the hydrogen bubbler anode, and rethreaded the 4-40 holes in the housing side panel to size 6-32. I then inserted 1/2 inch long 6-32 screws from the inside out of the side panel so they could be used as studs for mounting the vent plate with 6-32 wingnuts.

The lower screw required a little special attention. Since it backs up against the lower turbine housing spacer, its hole had to be counter sunk, and a 6-32 flat head screw used in place of the pan head screws used in the other holes.

With the turbine side panel remounted, and the holes in the vent plate resized for a loose fit on 6-32 screws, the vent plate is easily installed on or removed from the side panel via 6-32 wingnuts on the side panel 6-32 studs. (The lower screw is only used for alignment. Without trimming the wings, a wingnut won't quite turn there, and I didn't feel like customizing just one nut.)

Although not absolutely necessary, I chamfered the 1-inch square backing blocks to allow pulling the runner straight out of the turbing housing without having to tilt the shaft in the opposite bearing to get the runner past the backing blocks.

Now, with the thumbscrew and wingnut arrangement, runners can be swapped in a matter of a minute of so.

Portable Air Power:

(c.a., October 2005)

While all aspects of turbine testing could be performed in the Solar Powered Fluid Mechanics Lab, some things might be done much more simply by just dumping air from a fixed volume tank at a known starting pressure into a turbine and watching what happens. For example, comparing runner design efficiencies can be done this way. With proper consideration of mass differences, whichever runner spins the longest for the same application of air through the turbine is the most efficient.

For a small pressure vessel I used something that may be near impossible to come up with these days, a designated 30-pound freon canister originally used in automotive repair shops for recharging air conditioner systems some decades ago, back when chlorofluorocarbons weren't known to be quite so dangerous to the ozone and such.

My father used to own a repair shop, and, a long long time ago, I squirreled a couple of canisters away once they were emptied. (You never know when something might be useful!) If you aren't as big a packrat as me, and you choose to do something like I did here, you should be able to come up with some kind of small tank at a surplus yard, a hardware store, or even an autoparts store.

Basically, I just hung a quick action ball valve, a tire inflator valve, and a pressure guage off the output fitting of the tank. The tank has a 1/4-inch male flare tube outlet. For that connection I used a 1/4-inch flare swivel into a flare to 1/4-inch male NPT (pipe thread) adapter. The NPT adapter screws into an aluminum manifold which has a 1/4-inch NPT threaded hole on each end, and three 1/8-inch NPT threaded holes across its top. On top of the manifold I installed the pressure guage, the inflator valve and an 1/8-inch NPT plug to close the third hole. The guage with 1/8-inch NPT connection I had laying around in a junk box for years. The aluminum manifold and inflator valve with 1/8-inch NPT connection I ordered on-line from McMaster-Carr for about $12.00. The rest of the parts I picked up at a local hardware store for about $10.00 more. The hardware store ball valve has 1/4-inch NPT male thread on one end and 1/4-inch NPT female thread on the other. The male thread was screwed into the end of the manifold opposite the tank connection, and a 3/8-inch hose barb adapter with 1/4-inch male NPT was screwed into the female threads on the ball valve. Except for the flare tube swivel, teflon thread tape was used to seal all connections.

As it sits in the last picture of the three immediately above, the tank and valve assembly works fine. Close the ball valve, open the tank valve, pump up the tank via the inflator valve, connect a hose from the barb connection to the turbine nozzle inlet tube, quickly open the ball valve, and watch the turbine spin. No problem. However, as you may have noticed by now, I'm kind of anal. So, I decided to hack up a few pieces of aluminum to make a platform to support the manifold and not take the chance of at some point getting overly excited and cracking the tank's flare fitting by slamming the ball valve open too hard. The manifold has predrilled holes suitable for 8-32 mounting screws.

While the tank and valve assembly does stand up on its own, it would still be easy to pull or kick over. To reduce those possibilities, I slotted a few boards and made a cradle to keep things a bit more stable.

Pumping it up with a high-volume bicycle pump isn't the worst workout I've ever had. But, I plan to use a 12-volt automobile tire inflation compressor for repeated use. Those things can get a little warm. So, as temperature of the air in the tank is a factor in the supplied volume, a thermal sensor on the tank may be in order.

References:

(Clicking reference numbers here takes you to the text location of the reference.)

[1] Tahil, William, 1999. Theoretical Analysis of a Disk Turbine (2). Tesla Engine Builders Association News (16):15-16.

Last updated 03Sept2006
Alan Swithenbank, alans@cuervo.stanford.edu

Contents: