How do I model a supercharger rotor?

[lohring.miller]

This builder did a lot of experimenting on a screw compressor. The files to print it are available. His conclusion was at that scale it wouldn't work. There are some turbochargers that can work on 50 cc two strokes, but either a vane or a liquid ring pump would be better for really small scales.

Lohring Miller

 

[Richard Hed]

There is free spreadsheet software available, which works very well.
I use it for spreadsheet and word processing.
It replaces office.

Called LibreOffice.
Very good product, and totally free.
I download it to my local computer, so there is none of the cloud nonsense to deal with.

.

Ha, I have a computer loaner for where I work. They have installed msux office and it would not work when I am away from my place of work. So I downloaded Libre and it installed. NOW, NO PROBLEM. msux just can't let people alone, has to be poking and prodding and forcing one to use their shittly product.

 

[gbritnell]

Pictures of the GMC 671 blower.

 

[Bentwings]

I just finished modeling a 6-71 blower in Solidworks. It's modeled at full size then I printed it out at 1/3 to go with my 302 printed engine. It has 3 lobe rotors. I did the preliminary layouts in Autocad then modeled it. I used the helix tool in Solidworks. Let me know and I will post pictures.

Thanks I’ll down load the file I have both fusion 360 and SW maker. Not had good experience with maker it’s hard to justvgetvitvto open. I’ll have to see if I can get my other sn in Minn to help with blower case. He has an extended case super blower. He may not be very excited to take it apart as it is nearly brand new. I’d like to make both 8-71 and a super blower model. Even at 1/4 scale either should be able to blow up a garbage bag. I’ll use real bearings rather than pure plastic so I can spin them as fast as I want.

 

[gbritnell]

Solidworks has a feature called twist. Like I said I did the layouts in Autocad to get the dimensions then created the model with straight lobes then twisted it 20 degrees. I have a Creality Ender 3 Neo and it will print with about a .002-.003 tolerance so although my fits are pretty good I don't know about pumping air with it. You should also drive the rotors with helical gears. Solidworks can create them but you'll have to go to a site called The Engineers Edge/ helical gear calculations to come up with the numbers to get the proper center to center dimensions. This will involve starting with the approximate gear diameter, diametral pitch and some arbitrary helix angle like 20 degrees. Now you have to adjust tooth count and helical angles until you get the desired center spacing.

 

[GreenTwin]

Pictures of the GMC 671 blower.

That's just awesome !
Thanks for the pictures.
.

 

[gbritnell]

Here are the steps for creating the blower lobes
#1 sketch, #2 trim, #3 extrude, #4 flex/twist
The biggest thing is getting the tangencies of the rotors correct as they rotate. Kind of like the involute curve of gear teeth.

 

[mrehmus]

Good work!
If I recall correctly, Gearotic will design superchargers. I printed one from files sent from a friend. 3 rotor version.

 

[JimSvarv]

This builder did a lot of experimenting on a screw compressor. The files to print it are available. His conclusion was at that scale it wouldn't work. There are some turbochargers that can work on 50 cc two strokes, but either a vane or a liquid ring pump would be better for really small scales.

Lohring Miller

I read a magazine article of a company that made screw compressors.
And they made the screws/rotors in short sections, maybe 2" long.
Then they glued them together. This way they could make them hollow and lighter.
Don't remember what the smallest size was, but the bigger ones was for engines with hp I the thousands.


For roots to calculate the displacement of the supercharger per revolution is it the volume of all the rotor tips that one need to calculate?

 

[Jasonb]

George, with your "twist" do you get the correct cross section

If I did similar in alibre it would not give the correct cross section unless the sketch is done at right angle sto the angle of "twist" You can see this when creating helical gears which if done by simply twisting a spur gear profile loose their shape more as the helix angle is increased.

 

[gbritnell]

Hi Jason,
Actually I did it both ways. Here's how I did the solid model for the printed part.
In Solidworks they give you several options for creating the helix. The one I used was length and amount of turns in that length.
I calculated this information then went to the helix toolbar
When the tool opens it puts you in the drawing environment where you create a circle equal to the pitch diameter.
When you exit the drawing the helix work window opens and using the circle you create a helix. The length for one turn was extremely long so I put a quarter turn and the proper distances (it was still longer than needed but I'll get to that)
When you accept the equation a helix is created.
Now you create a work plane on the helix and at the end of it. This work plane will be perpendicular to the helix so not in alignment with any of the three major axes (correct spelling)
Now you draw the rotor shape. You can do this from the origin point or just anywhere on the work plane. Once done you exit the drawing environment
Once created you go to the loft tool. You select the rotor sketch and the helical profile. The helical rotor is created.
The created rotor's ends are not square to the major axes so this is where the extra length comes into play.
You make a trimming plane and cut the rotor to the proper length.
You now have a rotor with the proper 'twist' and the ends square to the major axis.
There might be an easier way but my Solidworks is self taught.
I did solid modeling when I worked but those programs were very different from Solidworks. I know the program fairly well but new things I have to learn on the fly.
I hope this all made sense!

 

[Peter Twissell]

Does that method give correct 'meshing' between the rotors?

I have recently been modelling helical gears, using the sketch perpendicular to the helix method. For gears, this is correct as it represents the way the gears will be manufactured (by hobbing) and maintains the correct pressure angle. However, the gear form has to be 'adjusted'. In my case, I used a 14 DP form to generate a 12T gear with 45 deg helix on a 1.2122" pitch circle.

Intuitively it would seem to me that if the rotor form has been generated normal to the rotor axis as shown on post 87, the helical rotors should be in effect a series of thin slices at that generated form, with each pair of slices rotated slightly to give the required helix angle.
To model that, the sketch normal to the axis would be projected along the helical path.
Such a design would lend itself to vertical milling with a form cutter, with the rotor blank supported on a dividing head geared to the table feed.

Incidentally, none of the superchargers I have encountered have two rotors with forms similar to that shown in post 87. Some have one rotor of a form similar to that, with the other rotor significantly different. Others with two similar rotors have more of a pointed tip form, as shown here in the attached photograph.

Regarding calculation of displacement, as pumps of this kind are scaled down, internal leakage becomes the dominating factor and any attempt to calculate displacement from geometry becomes futile. Add to that the changes in leakage with fluid viscosity, local pressures and distortion/ movement of components and experimentation is the only way to determine any kind of valid results. I have recently developed an automotive oil pump and the range of test results under various conditions is quite staggering.

Edit: I'm sure Bentwings will quickly spot exactly what the pump is for, just from the drive gear form!

 

[gbritnell]

Even with the somewhat pointed rotor tips there is only one contact point along the length of the lobe helix. The one I modeled was of a GMC type with rond lobes. In my CAD layout asthe rotors torn there is still just one point of contact (minus clearance) along the length of the lobes as the mesh.