My first wankel engine 'bits'!

[itsme]

Hi Mark,

About those apex seals - do you not mean pleasuretool? I think if I left out apex seals, then I may as well give up before I've even started, because there's not even a chance that it will work.

I've also been thinking about side seals and how I could include them on a small engine. I'm sure there must be advantages in having them, but it just adds a lot of extra complexity when making the engine. I'll have to see.

CJL5585, the 'Excel Process' is very simple. In the case of the epitrochoid, you have got the parametric equations for the shape, so it is easy to plot many different points along the curve.

I have set the spreadsheet up so that there are 721 rows. It starts at 0 degrees and then ends up back at 0 degrees going up in 0.5 degree steps (the spreadsheet uses radians, but it's easier to picture degrees). For each 'step', you get an X and Y coordinate. You need to set it up so that there is a column for the X values and a column for the Y values. In a new column, you can then generate the Gcode using the x and y coords that have been generated.

This is an example of what could be written in the Gcode column:

="G01 X "&ROUND(G2,4)&" Y "&ROUND(H2,4)&" F300"

This would write 'G01 X', followed by the the value in G2 (rounded to 4 decimal places), followed by 'Y', followed by the value in H2 (also rounded to 4 decimal places), followed by 'F300'.

If, for example, we assume that the value in G2 is 4 and the value in H2 is 6, then the code would be 'G01 X 4 Y 6 F300'.

It really is very simple, but you must remember that your zero point will generally be right on the centre of the shape (assuming it is symmtrical) or wherever the (0,0) point is if you were to plot the shape on graph paper.

I hope this helps.

Warren

[RotarySMP]

Hi Warren,

Sorry, you are right I was responding to Pleasuretools post.

I have some contact with the guys at Dimond engines in Wiener Neustadt, and can't help wondering what a Mazda style wankel scaled to about 60 HP would be like. I really don't think the air cooled roller rotor bearing of the Fichtel and Sachs KM 914 family was a good direction for small wankels to take. The rotor bearing stresses end up setting RPM limits which limit the whole concept. I would love to see a small wankel with pressure fed plain bearings, oil cooled rotor and liquid cool housings. I bet you could rev the heck out of the thing.

Have you tried mounting a dremel or such die grinder on your mill and grinding the trochoid to get a better finish?

[itsme]

Hi Mark,

I haven't tried anything yet to get a better finish, but I have thought about it. At the moment my mill is only capable of doing about 2000 RPM. This speed is really the limiting factor on the mill. It has also got those horrible nylon/plastic/rubbish gears in it, which I really don't like too much. I'm therefore planning on getting rid of those gears and replacing them with a nice, smooth, v-belt drive. While I'm at it, I can change the ratios to speed it up a bit. There are guys that have had this mill running at 8000 RPM with the standard motor using a belt drive. I'm not planning on taking the speed that high, but I'd like to try grinding the housing with this setup. I'd use a larger diameter wheel (>40mm), to try and get the tip speed (or whatever it's called on a grinding wheel) up a bit. I have no experience grinding, so I might find that the finish wouldn't be suitable if I did this, but I still need a higher spindle speed (I think I'll aim for 5000-6000 RPM). Does anyone know what kind of finish I could expect with a 40mm grinding wheel at, say, 5000RPM???

Mounting a Dremel type tool on the mill also sounds like an excellent idea. I'd never thought of this. Any ideas on how I could attach one onto a Sieg X1??? I'm quite keen to try this before I do any (more) major mods to my machine.

As for rotor bearings, I've never really paid too much attention to them. At this stage, my goal is to just get an engine running. Once I have got past that (rather large) barrier, then I can pay more attention to things like bearings and how performance can generally be improved.

Mark, just out of curiosity, do you know how they make the housings at Diamond. What sort of machinery are they using?

Regards
Warren

[CJL5585]

Warren,
Appreciate the info on the Excel spreadsheet. Thanks.

I will be following your progress, and am wishing you the best of luck in your endeavour.

Keep us posted WITH some photos.

Jerry

[RotarySMP]

Hi Warren,

Try the grinding wheel and see. In my experience the dremel grinding wheels are not very round, so you will want to dress it with a diamond before use.

How about bolt a plate to the side of the head housing, and make a bracket off the plate like the CNC wood routers use. Your work envelope is so small that you don't need the spindle in the middle of travel.

The SW92 uses short lengths or drill rod as home made needle bearings between e shaft and rotor.

The rotor housings are cast, and used to be machined on the huge (like 40 ton) trochoid grinder that was originally built for Norton. That grinder was left in England. Now they are gound on a CNC, but it is not done in house. The reports I have had are that the CNC grounds trochoids are the most acurate ever.

They have some sort of coating (probably something like Nikisil) on the Al.

[itsme]

Hi Mark,

Thanks for that info. I was thinking about trying a cast iron housing and rotor to start off with - no chrome or any other coatings. I'm thinking about then using aluminium end plates and steel apex seals. I'm fully aware that this combination probably wouldn't be too durable, but if I can just get something to run, I can then put more time into the materials on the next engines. At least with CNC, I can make lots of the same part without boring myself to death.

Regards
Warren

[itsme]

Hello again,

I have just ordered a new rotary tool with a flexible extension to modify for my mill. I am planning on taking the flexible extension apart and building the end of it into a blank 2MT arbor that will then fit in the spindle on my mill with a modified drawbar. I think this will give me greater accuracy and it also gets the (relatively) chunky motor out of the way. Hopefully I'll be machining super-accurate bores in the near future (assuming the stuff arrives soon...).

Just as a matter of interest, does anyone know anything about the cooling system on the DKM 54 engine (the first running wankel). I know that the fuel/air mix passing through the rotor and its shaft did a lot of the cooling, but I'm not too sure on how the trochoidal housing was cooled. I have seen photos of the housing, and it doesn't appear to have any cooling channels in it and it would also seem rather complex to try and get coolant to this moving part. Any ideas?

Regards
Warren

[RotarySMP]

I don't know about the DFM 54, but all the KM914 derivatives use a total loss oil system and air cooling of the rotor. The Wankel GmbH engines of about the same size used a wierd system where they ported the induction charge through the rotor to cool it and lube the bearings, and then back out and around and into the intake. The advantage was that the premix system was the only lubrication system. The down side would be throughing away the wankels inherent superior breathing, and the RPM limits which roller element rotor bearing impose.

For an engine the size you are making, an air cooled housing should be fine. The SW92 and OS/Graupner with their roller bearings have no rotor bearing cooling or lubriction other than blow down premix oil caused by the lack of side seals.

Cast iron housings with spring steel apex seals would probably a good place to start. Look forward to hearing your results.

I am starting to look at buying a house (not easy here in Vienna Austria - A city of around 2 million, but there are only about 50 houses on the market). One I have a garage I'll be able to bring my mill home and CNC it.

[itsme]

Hi there,

I think I have finally decided what engine I am going to have a go at building. I was thinking about designing my own from scratch, and even had thoughts about a DKM type engine, but I have come up with a better idea. I am going to basically build the SW92, BUT with a few changes to the design.

The biggest difference will be that my engine will use a proper epitrochoid rather than 2 circles. The porting may have to be adjusted slightly to suit the epitrochoid. I am going to keep an eccentricity of 3.5mm, however I am going to decrease the rotor radius from 28mm to 25.5mm. 28mm seems too big for this engine. I'm not sure if this has something to do with the housing that was used. The rotor width will stay at 18mm. After these changes have been made, the capacity of the 'new' engine is 8.4cc. I will therefore be making an SW84, not an SW92 (I may as well change the name slightly while I'm at it...). The only other change I'm going to make is to the cooling fins. I prefer the look of the OS style fins to the ones on the plans. Oh, I'll also replace the homemade venturi with a commercially available carburettor.

This should save a lot of time, as the basic design is already in place, so I can concentrate on the more important parts.

Mark, maybe you can help me here on some detail. I was looking at the plans for the SW92 and it looks like the internal gear is cut into the rotor. It seems like quite a good idea to do this. Am I correct in thinking this? Finally, does the author give any detail on how he actually cut that internal gear in any of the text? It's at times like this that I wish I could read German...

Thanks
Warren

[RotarySMP]

Hi Warren,

I speak german, but the notes in the text are very limited. I have read these instructions a few times and am of the opinion that the were written for an early model of this engine, or the drawings are missing some detail.. Note that the photo top right also shows an motor with OS style cooling fins. In the assembly instructions it mentions - "The stationary gear is secured with four srews. It has 14 teeth , the holes are not symetrically distributed. The screws must be peaned to sercure them." This is not illustrated in #8 (the stationary gear) or part four, the side housing, each of which should have four M2 holes not evenly distributed. I think you are going to have to look at how the phasing gears mesh on a Mazda and align this the same.


On the rotor it says:

The cuts in the rotor must be evenly distributed, so that the apex seals function precisely. The given angles must be maintained, and the axis of symmetry must cross the center line. Through lapping of the sides of the rotor, using lapping paste on a asurface plate, you are aiming for the following tolerance: The middle of the rotor should be 4 to 5 microns wider that the corners.

The author gives no info on cutting that internal gear, but you will just have to grind up a form tool (I don't know how to ally the figures, but that box under the rotor defines the gear shape. I'll take my plans to work and ask one of the guys to explain it.)

Then you will need to make a shaper attachment for your lathe, and a way to index 21 on your lathe spindle.

[itsme]

Hi,

Getting the lathe spindle to index could be a bit of a challenge. I was thinking about maybe making a cutter (like an involute cutter) with the correctly shaped teeth for an internal gear. I would then use this cutter to make a broach, which would look similar to a spur gear, but have many steps getting larger on each step. Once this is done, I could then use some kind of press to press the broach through the hole and at the end of it, I should have an internal gear.

I don't think that the shape of an internal gear is the same as the 'negative' of a spur gear, otherwise I could just use a normal involute cutter to make the broach. On the other hand, I don't think the shape is that far off. I reckon a fairly accurate gear can be made at home, but I don't think it will ever be perfect.

Although the plans don't give much detail on the fixing of the spur gear, it does show the screw that is used. I think it is part 26. I've been looking at the plans, and still can't quite figure out why those screws aren't symmetrically placed. Is there something that I'm missing. It would just make sense to place them 90 degrees apart. I guess I can always make it up as I go along...as long as it works in the end.

Regards
Warren

[Mat-C]

Have you checked out available planetary gearboxes? Small/cheap cordless screwdrivers often use these because it is easy to stack multiple stages axially.
You probably know about these, but for those that don't: a planetary gearbox has a driven centre gear with 2 or 3 gears revolving around this, meshing with the driven gear and an inverse outer gear, the (slower) drive being taken off the mounting of the revolving cogs.

http://www.du.edu/~jcalvert/tech/planet.htm (diagram is middle-left).