Machining the "rotor" of a PatWankel rotary engine - Page 2
Close
Login to Your Account
Page 2 of 3 FirstFirst 123 LastLast
Results 21 to 40 of 46
  1. #21
    Join Date
    Mar 2009
    Country
    UNITED STATES
    State/Province
    Washington
    Posts
    7,857
    Post Thanks / Like
    Likes (Given)
    5773
    Likes (Received)
    10155

    Default

    how are you supposed to machine the outer cases and slots for the seals?

  2. Likes Limy Sami, adama liked this post
  3. #22
    Join Date
    Feb 2009
    Location
    Colchester UK (where the lathes used to be made)
    Posts
    1,918
    Post Thanks / Like
    Likes (Given)
    1320
    Likes (Received)
    557

    Default

    The outer cases are relatively easy. As previously stated the only precision surface inside the case, is that of the seal slot (which could be handled by a 9 axis mill ;-))

    Almost the entire surface of the rotor is a sealing surface , machining it will/would be like die making.

    Bill

  4. Likes John Welden, adama liked this post
  5. #23
    Join Date
    Mar 2009
    Country
    UNITED STATES
    State/Province
    Washington
    Posts
    7,857
    Post Thanks / Like
    Likes (Given)
    5773
    Likes (Received)
    10155

    Default

    Quote Originally Posted by Billtodd View Post
    The outer cases are relatively easy. As previously stated to only precision surface inside the case, is that of the seal slot (which could be handled by a 9 axis mill ;-))

    Almost the entire surface of the rotor is a sealing surface , machining it will/would be like die making.

    Bill
    Everything only needs to be absolutely perfect for it to work. A whole bunch of spherical stuff and slots and seals will probably all line up no prob.

  6. Likes Billtodd, Limy Sami, adama, Garwood liked this post
  7. #24
    Join Date
    Feb 2009
    Location
    Colchester UK (where the lathes used to be made)
    Posts
    1,918
    Post Thanks / Like
    Likes (Given)
    1320
    Likes (Received)
    557

    Default

    No one said it would be easy ;-)

  8. Likes adama liked this post
  9. #25
    Join Date
    Jan 2007
    Location
    Norfolk, UK
    Posts
    18,726
    Post Thanks / Like
    Likes (Given)
    14249
    Likes (Received)
    14268

    Default

    Quote Originally Posted by Billtodd View Post
    No one said it would be easy ;-)
    The OP did !

  10. Likes adama liked this post
  11. #26
    Join Date
    Nov 2006
    Location
    Athens Greece
    Posts
    125
    Post Thanks / Like
    Likes (Given)
    0
    Likes (Received)
    9

    Default

    Hello BillTodd


    By wire-EDM they were made the eight finger followers (the yellow parts in the animation) used in the first pattakon VVA prototype (more at http://www.pattakon.com/pattakonRod.htm )









    The VVA proved on the roads more reliable than the rest Renault19Energy engine :



    Form the ten finger follower prepared, the two remain still unused.

    It was a simple, accurate and not expensive machining.
    The ten finger followers were cut together (they were designed in a polar array) from a steel disk:



    Their working surfaces (the semi-circular cam-follower-surfaces) were ready for use without other machining.



    You write:
    ”Making a prototype using 5axis cnc may be a better approach. Surface finishing could be problematic!”

    The initial thinking was to make it with a 4-axis cnc.

    Then a CNC lathe (with the eccentric shaft secured on its chock) appeared preferable.

    Now a conventional / modified (by adding a U holder for the cutting-tool) lathe seems even better in terms of surface quality, of accuracy of dimensions, of time required and of cost.



    You write:
    ”Have you an idea of the size of your prototype? (LP prototype was perhaps 200-300mm diameter)”

    It is close to 350mm in diameter.
    We think of a per working chamber capacity between 500cc (as in the single rotor Spinder NSU Wankel) to 650cc (as in the Mazda RX8 Renesis).


    Thanks
    Manolis Pattakos

  12. #27
    Join Date
    Nov 2006
    Location
    Athens Greece
    Posts
    125
    Post Thanks / Like
    Likes (Given)
    0
    Likes (Received)
    9

    Default

    Hello John Welden

    You write:
    “how are you supposed to machine the outer cases and slots for the seals?”


    In a similar way with the inner body.

    Each half of the casing is located on the eccentric shaft



    The red part shows the path of the cutting tool.

    For each recess / chamber in the casing, the “white” gearwheel” is properly secured on the half casing (at an eccentricity equal to one quarter of the pitch circle diameter of the white gearwheel):



    In the animation it is shown only a part on one rotation. It is supposed the rotation of the lathe chock is continues.

    During the first six slides of the animation, the cutting tool is in the air (it cuts nothing).
    For the following 19 slides the cutting tool removes material from the one chamber, doing nothing for the rest four chambers.
    For the last five slides the cutting tool is in the air again (it cuts nothing).


    With the previous machining work, the clearance between the inner body and the outer body can be quite small at operation (like, say, 0.2mm), which means almost all the gas at the end of the compression is concentrated in the “combustion cavity” formed at the “top” of each working chamber, on the outer body.


    After machining all five chambers in each half of the casing, the grooves for the seals are machined on a CNC mill. For the grooves the significant dimension is the width (the distance between the two flanks).

    If the previous are confusing, please let me know to further explain.

    Thanks
    Manolis Pattakos

  13. #28
    Join Date
    Mar 2007
    Location
    On Elk Mountain, West Virginia, USA
    Posts
    2,091
    Post Thanks / Like
    Likes (Given)
    899
    Likes (Received)
    713

    Default

    I am having trouble visualizing your seals, and see a little concern about effective cooling and lubrication of the sealing surfaces...but a more fundamental problem I see is geometric. Your combustion chambers have which surface-area to volume ratio, which equates to loss of heat energy, and large chill zones (boundary layers) of incomplete combustion.

  14. #29
    Join Date
    Jun 2013
    Country
    UNITED STATES
    State/Province
    California
    Posts
    1,036
    Post Thanks / Like
    Likes (Given)
    713
    Likes (Received)
    453

    Default

    Dumb question, but if the rotor AND housing spin, how do you wire for spark?


    BTW, I love your body of work.
    Last edited by mkd; 01-26-2017 at 06:36 PM.

  15. Likes Limy Sami liked this post
  16. #30
    Join Date
    Nov 2006
    Location
    Athens Greece
    Posts
    125
    Post Thanks / Like
    Likes (Given)
    0
    Likes (Received)
    9

    Default

    Hello Magneticanomaly.

    You write:
    “I am having trouble visualizing your seals, and see a little concern about effective cooling and lubrication of the sealing surfaces...but a more fundamental problem I see is geometric. Your combustion chambers have which surface-area to volume ratio, which equates to loss of heat energy, and large chill zones (boundary layers) of incomplete combustion.”

    Cooling:

    In the following air-cooled three-chamber PatWankel:



    the ambient air around the cooling fins of the spinning outer body does the cooling.

    The inner body (red) is cooled, among others, by the three seals (white) that sweep its external surface being inside grooves made on the cool outer body. It resembles to the cooling of a reciprocating piston by its piston rings that slide onto the cool cylinder liner.
    The inner body (red) is also cooled by the fresh charge in a way similar to the cooling of the reciprocating piston in a 2-stroke engine.



    Lubrication:

    The lubrication is easier than in a Wankel wherein at high revs the seals undergo heavy centrifugal forces and wherein each apex seals requires a significant “preloading” by springs inside its groove in order to remain in contact with the epitrochoid working surface when it passes from the area between the “spark plugs”.
    Worth to mention: the seals in a Wankel perform a combined motion: a spinning around the center of the rotor on the eccentric shaft pin, and an orbit around the center of the engine:



    In comparison, the seals in the PatWankel of the above animation is pure spinning, which means constant centrifugal force all around the cycle.
    At high revs this means less friction between the seals and the working surface (because the centrifugal force decreases the force the seal applies on the working surface).
    At lower revs the preloading of the seals keeps them in tight contact with the working surface, minimizing the leakage.
    At high and top revs (i.e. where ther is not time for significant leakage) the seals abut softy on the working surface minimizing the friction in the engine and increasing the power output (worth to mention: in a rotary, the main source of friction is the sliding of the seals over the working surface).

    So, in a PatWankel like the above, the seals need less lubricant.



    Regarding the high surface to volume ratio:

    In the conventional Wankel Rotary the flame does see way bigger surfaces (and has to travel along way longer distances) than in a conventional engine.
    This causes an excessive thermal loss towards the walls of the chamber.


    According the “Liquid Piston” at LiquidPiston, their reverse-Wankel engine:



    which has even higher overall surface-to-volume ratio than the Wankel, runs closer to a constant volume combustion than the reciprocating piston engines; they also claim the combustion ends inside the small cavity at the top of each working chamber.
    The rate of thermal loss is several times higher during the progression of the flame along the compressed fuel-air mixture, i.e. during the combustion.
    During this period, the actual “surface-to-volume” ratio in the LiquidPiston Reverse Wankel (and in the PatWankel) is many times lower than the overall “surface-to-volume” ratio (it is lower than in the best reciprocating engines), which means minimisation of the thermal loss.
    The combustion cavity is so compact / concentrated, that the flame has to travel a very short distance (a few times shorter than in a Wankel or in an over-square reciprocating piston engine) to burn all the air-fuel mixture.
    During the following expansion the high “surface-to-volume ratio” increases the relative thermal loss, but the overall thermal loss is lower than in good reciprocating piston engines.

    In comparison, think what happens in the Wankel during the combustion: the flame propagates slowly in an attenuated (anything but compact) chamber, the burned gas radiates towards the big walls (here the overall surface-to-volume ratio is also the surface-to-volume ratio during the combustion), the combustion last till the opening of the exhaust ports with a part of the mixture remaining unburned etc, etc.


    Take the PatWankel in the animation.

    At TDC (i.e. wherein the volume of the working chamber is minimised) almost all the air / mixture entered into the working chamber is concentrated in a compact cavity (spherical or semi-spherical etc).

    As compared with a Ducati Panigale 1299cc reciprocating piston engine running at the same revs with the revs of the inner body of the PatWankel, the PatWankel provides more than 70% additional time around the TDC (1.15*1.50=1.725)
    The 15% longer dwell at the TDC comes from the harmonic (i.e. pure sinusoidal) variation of the combustion chamber volume:



    the 1.50 comes from the 270 degrees required in order the chamber to go from its TDC to its BDC (wherein the volume is maximised):

    There is plenty of time, lots of squeeze and a very short distance for the flame to travel, enabling the combustion to actually complete inside the compact cavity.

    Most of the thermal loss happens during the combustion.
    The thermal loss continuous during the expansion, however the rate of thermal loss during the combustion is way higher.

    Compare it with the thermal loss in a Ducati Panigale 1299 wherein the flame, during the combustion, sweeps the inner surfaces of a wide (116mm diameter) short (about 5mm height) cylinder (like a coin) having abnormal bottom and top surfaces (valve pockets etc).
    Reasonably, the thermal loss towards the walls during the combustion will be substantially more than in the above PatWankel.

    So, there are reasons for lower thermal loss in the PatWankel, despite the higher (than in a reciprocating piston over-square engine) overall surface-to-volume ratio.

    In previous posts it was described the way of machining the inner and the outer bodies of the PatWankel (it was for a 5-“cylinder” PatWankel, but things are similar in three-“cylinder”, in seven-“cylinder” etc PatWankels).

    The machining allows a very small clearance (say, 0.2mm) between the inner and outer bodies when a working chamber is at its TDC (i.e. wherein the volume in the chamber is minimized).

    The pure spinning of the inner and outer bodies (each about its own fixed axis), allows such small clearance even at extreme revs (in comparison, in a reciprocating engine the elasticity of the connecting rod requires a significant clearance between the piston crown and the cylinder head (like, say, 1mm), otherwise the piston will hit on the cylinder head during the overlap at top revs).
    Small clearance means that during the combustion almost all the air-fuel mixture is in the combustion cavity, the squeeze is strong, the spark plug is at the very-centre or the cavity, the flame has a short distance to cover till the ends of the cavity, the surface on which thermal loss happens is too small, and there is plenty of time as explained above.


    The previous are strictly technical.

    On the other hand,
    if you look at the US3.5$ millions DARPA gave so far to LiquidPiston, at the prize and the support of the MIT (isn’t Sloan Automotive Laboratory part of MIT?) to LiquidPiston and the prize / support of Shikorsky company to LiquidPiston, it seems they see a promising engine design.

    Thanks
    Manolis Pattakos
    Last edited by manolis; 01-27-2017 at 04:27 AM.

  17. #31
    Join Date
    Jun 2011
    Country
    SWITZERLAND
    Posts
    2,867
    Post Thanks / Like
    Likes (Given)
    649
    Likes (Received)
    1216

    Default

    Quote Originally Posted by manolis View Post
    if you look at the US3.5$ millions DARPA gave so far to LiquidPiston, at the prize and the support of the MIT (isn’t Sloan Automotive Laboratory part of MIT?) to LiquidPiston and the prize / support of Shikorsky company to LiquidPiston, it seems they see a promising engine design.
    For use on a weaponized drone it needs at most to survive the duration of the consumption of a single tank of fuel.

    And, the LiquidPiston doesn't pass the hot exhaust gasses though the rotors central shaft.

  18. #32
    Join Date
    Nov 2006
    Location
    Athens Greece
    Posts
    125
    Post Thanks / Like
    Likes (Given)
    0
    Likes (Received)
    9

    Default

    Hello Mkd

    You write:
    “Dumb question, but if the rotor AND housing spin, how do you wire for spark?”

    As you correctly write, both parts (the inner body with the working surface on it, and the outer body with the grooves and the seals on it) spin, each about its own fixed axis.

    Here it is shown the inner body of a five-chamber (five-“cylinder”?) PatWankel:



    There is plenty of space enabling the installation of the complete ignition system (including the generation of the electric current required for the ignition) inside the inner body, near the intake ducts (to avoid overheating).



    The spark plug holes on the working surface of the inner body pass over the seals of the chamber at the end of the expansion (i.e. way later than in the Wankel rotary engine) reducing several times the relative leakage.

    Thanks
    Manolis Pattakos

  19. #33
    Join Date
    Feb 2005
    Location
    Akron, OH
    Posts
    1,834
    Post Thanks / Like
    Likes (Given)
    279
    Likes (Received)
    1376

    Default

    Maybe you want A CNC lathe with a cam turning option such as this:

    Okuma cam turning - YouTube

    Okuma | Cam Options for CNC Lathes

  20. #34
    Join Date
    Jun 2011
    Country
    SWITZERLAND
    Posts
    2,867
    Post Thanks / Like
    Likes (Given)
    649
    Likes (Received)
    1216

    Default

    Quote Originally Posted by manolis View Post
    Here it is shown the inner body of a five-chamber (five-“cylinder”?) PatWankel:



    There is plenty of space enabling the installation of the complete ignition system (including the generation of the electric current required for the ignition) inside the inner body, near the intake ducts (to avoid overheating).
    Figure (c) demonstrates the impossibility of getting a wrench on a sparkplug.

    Drawing a surreal picture doesn't make it so.

    Are you trolling us?

  21. #35
    Join Date
    Nov 2006
    Location
    Athens Greece
    Posts
    125
    Post Thanks / Like
    Likes (Given)
    0
    Likes (Received)
    9

    Default

    Hello jCandlish

    You write:
    “For use on a weaponized drone it needs at most to survive the duration of the consumption of a single tank of fuel.”


    The “Sealing Grid” of a Mazda RX-8 Wankel Renesis has a “time” between overhauls (TBO) between 50,000 and 100,000 miles.

    The apex seals of the RX8-Wankel have a hard life (combustion at both sides, heavy preloading in order to maintain contact with the working surface when they pass through the area between the spark plugs (see the plot with the acceleration of an apex seal as it moves around the epitrochoid working surface), heavy gas leakage at their ends (on the corner seals), etc.


    The seals od a PatWankel like:



    have a, by far, easier life.

    They spin (they do not orbit), which means constant centrifugal acceleration that, at high revs (wherein there is not much time for leakage), alleviates / softens the preloading, reduces the friction, increases the power output and reduces the wear.

    The number of combustions they participate is half (because they belong to one only combustion chamber; they are not shared between neighbor combustion chambers).

    Only the one face comes in contact with hot gasses. The other face abuts on the cold flank of its groove.

    They are neither bouncing between the flanks of their grooves, nor they are slapping the flanks of their grooves.

    They keep the leakage as low as in the reciprocating piston engines.

    Etc, etc.

    Reasonably, the seals of the PatWankel will last a few times longer than the seals of the Wankel, which means a reliability / TBO similar to the reciprocating engines.



    You also write:
    “And, the LiquidPiston doesn't pass the hot exhaust gasses though the rotors central shaft.”

    As explained in previous posts, the exhaust gas needs not to contact the walls of the passageways inside the inner body.
    The exhaust gas passes through a pipe, having a small clearance from the walls and the shaft.

    In the LiquidPiston MXv3, the cooling air passes though the rotor and takes with it the exhaust gas. Not good in case of a catalytic converted.

    Thanks
    Manolis Pattakos

  22. #36
    Join Date
    Jun 2011
    Country
    SWITZERLAND
    Posts
    2,867
    Post Thanks / Like
    Likes (Given)
    649
    Likes (Received)
    1216

    Default

    Quote Originally Posted by manolis View Post
    As explained in previous posts, the exhaust gas needs not to contact the walls of the passageways inside the inner body.
    The exhaust gas passes through a pipe, having a small clearance from the walls and the shaft.
    This design is the perfect equivalent of a ship in a bottle.

    Easy! You just build everything inside-out.

  23. #37
    Join Date
    Nov 2006
    Location
    Athens Greece
    Posts
    125
    Post Thanks / Like
    Likes (Given)
    0
    Likes (Received)
    9

    Default

    Quote Originally Posted by jCandlish View Post
    Figure (c) demonstrates the impossibility of getting a wrench on a sparkplug.

    Drawing a surreal picture doesn't make it so.

    Are you trolling us?

    Hello jCandlish



    Manolis Pattakos

  24. #38
    Join Date
    Nov 2006
    Location
    Athens Greece
    Posts
    125
    Post Thanks / Like
    Likes (Given)
    0
    Likes (Received)
    9

    Default

    Hello Comatose.

    You write:
    “Maybe you want A CNC lathe with a cam turning option such as this:
    Okuma cam turning - YouTube
    Okuma | Cam Options for CNC Lathes”

    Thank you for the links.

    In a normal size 5-“cylinder” PatWankel it is required an about 50mm inwards motion of the cutting tool followed by a 50mm outwards motion of the cutting tool, and this four times into one rotation of the chock.

    In order to work, it needs very low revs of the chock.
    Is somebody in this forum having such a device? What is his estimation for the time required to machine a, say, 350mm maximum diameter / 125mm wide working surface?


    With the eccentric shaft on the chock and the gearwheels, the cutting tool “sees” a rotating part. I think it is preferable, especially for machining a prototype engine.

    Thanks
    Manolis Pattakos

  25. #39
    Join Date
    Feb 2009
    Location
    Colchester UK (where the lathes used to be made)
    Posts
    1,918
    Post Thanks / Like
    Likes (Given)
    1320
    Likes (Received)
    557

    Default

    Manolis,


    Spark plugs: Why not just put one or two* spark plugs on the out side in each combustion chamber (time the spark as necessary) , It'll be one fewer hole for the seal to traverse, give you a broader flame front (with two plugs) and make maintenance possible with normal sized hands.

    *(are doubled ignition circuits still mandated for aircraft engines?)

    Rotor machining: I would bet I am the only one on this forum with a lathe with an easily fitted internal gear* on the nose (made for an experimental speed reducer).

    clamped-nose.jpg

    As such, I gave serious thought to your eccentric turning operation:

    To get it to work requires a spindle mounted crank which, on my lathe could not be too big so would require at similar crank at the tail-stock end for support (not difficult , just extra pain in the arse) .

    My long radius attachment could move the tool (assuming to curve of the Pat rotor is actually radius).

    The gear offsets might be too large for the size of rotor I could turn , but again, I could get around that if I had to.

    The big killer , as pointed out by Garwood earlier in the thread, is that the angle of the tool varies as the stock rotates (just as the angle of the seals change as the rotor slide past) . To fix that would require somekind of cam and pivoting tool....

    At which point the thought experiment ended.

    * I believe the ratios of your gears shown are 5:4 my internal gear is 55T. Making a 44T gear would be a trivial operation.

    In short, I doubt anyone with a cnc lathe will be willing to modify it to make your prototype.

    5 axis milling looks like the only way (BTW you need 5 axes to keep the cutting tool at the optimum angle)

    For the same reason, to keep the tool angle correct, you'll need something more exotic than a standard 5 axis mill to do the cases of your engine. The machine will, at least need an angled spindle nose (maybe even an additional axis to control it) . I believe the motor-sport division of the company I work with, has something like this (they are used to doing impossible things for crazy F1 and car designers)

    A plea: . You are a clever guy, very inventive but, clearly, your experience of machining is limited. If you post on a professional machinists forum , have the good grace to listen to the professionals (do not immediately start an argument - do not be a Troll!).

    Bill

  26. Likes digger doug, Limy Sami liked this post
  27. #40
    Join Date
    Mar 2004
    Location
    Edison Washington USA
    Posts
    10,342
    Post Thanks / Like
    Likes (Given)
    964
    Likes (Received)
    5344

    Default

    I dont doubt that this "could" be built.
    but not, as the OP is hoping, cheaply.

    Remember, Mazda spent something like 40 years, hundreds of millions of dollars, and over a million "working prototypes" in the form of early rotary cars, to get to the RX 8 engine.
    many of those 1 million plus prototypes which were sold at retail to customers, failed prematurely.

    my guess is that 2.5 $ million would barely pay for the machine tools to build the first iteration of a prototype, and the next 100 or so iterations would require roughly 100 times that much money, and 5 years or so, before we get heavily into convincing industries to provide your new materials in affordable mass produced volumes...

    if it can kill somebody, as mentioned earlier, the chances of funding increase.

  28. Likes digger doug, JRIowa, Limy Sami liked this post

Tags for this Thread

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •