engine and drivetrain Kugelfischer Pump Cones

[Guest Anonymous]

As a follow up to my post last week, Jeff sent along the images of his pump cone collection. Pretty cool, they came from Preston Miller the engine guru of Miller & Norburn fame apparently. Stuff like this goes into the category of "unobtanium". Thanks to Jeff for the pics and info.

www.alpinabmw2002.com

[c.d.iesel]

unoptainium indeed

all you need is a engine or chassis dyno, a lathe,

a chunk of metal, and you to can be a 'tuner' for speed

[PatAllen]

i might be not of help at all, but someone with non negligeable knowledge, a data logger with a wide band unit and record afr data on a known tuned 02 at specific thotle openings and rpm, could end up collecting enough data to make a 3d graph, roll it on itself, translate to flow values to scale, and have it reproduced on a 4 axis cnc...

[dubois]

most of them - except for one - look "pretty simple". Surface finish looks pretty rough for being such "precise" piece of machinery. I am sure it was pretty difficult to produce in the days before CNC, now they can be had cheap - only if you had the drawing. I am sure there was an algorithm to relate engine specs, to fuel curve, to cone dimensions, and that part probably remains a "secret"

[amdavidson]

Someone with access to a 3D scanner should do a scan of the cone for posterity's sake and so that reproductions can be manufactured.

I changed jobs and no longer can get at equipment like that, but I know I'm not the only one on this list who's worked with CMMs and/or stereoscopic 3D scanners.

[Richard]

Someone near ten years ago but the blueprint on line for the racing cone. It could have been Roadfly. It may take awhile but I know I printed it....Would that help??

[amdavidson]

If the drawing is good enough, there are a million little job shops that could knock them out.

If there's enough interest I might even be able to machine a few at http://techshop.ws

[Swiftus]

Hold on a minute! The cone is a spinning part, correct? So you only need 2 axes to reproduce one, ie. a lathe.

A far as getting the measurements of the cone, why not simply take a photograph of it? From dead-on perpendicular to the cone and from a distance away to reduce warp-age, you would be able to measure both dimensions required for reproducing the cone, diameter and where that diameter occurs along the cone's axis.

3D scanning is neat and 4 axis milling is impressive (albeit that video is only demonstrating 3 axis milling), reproducing this part just requires some graph paper, a camera, and a CNC lathe (no need for an indexed spindle either).

Another method for recreating the cone would be to use a nice set of hole standards and measure the depth at which each diameter sits on the cone. With enough measurement, one could possibly extrapolate the desired profile.

HTH,

Jay

[PatAllen]

Hold on a minute! The cone is a spinning part, correct? So you only need 2 axes to reproduce one, ie. a lathe.

A far as getting the measurements of the cone, why not simply take a photograph of it?

not exactly. it doesnt "spin" all the time but rotates on itself some specific degrees related to specific rpm, and moves inward/outward depending on throtle position. so you couldnt just take a single picture of it, but several pictures of it at each few degress increment to capture the chaging profile, hence the requirement to have a 3d profile of it, then maped on a lathe. the video i posted, altough not a true 4axis cnc, is exactly what would be needed.

[amdavidson]

Hold on a minute! The cone is a spinning part, correct? So you only need 2 axes to reproduce one, ie. a lathe.

A far as getting the measurements of the cone, why not simply take a photograph of it? From dead-on perpendicular to the cone and from a distance away to reduce warp-age, you would be able to measure both dimensions required for reproducing the cone, diameter and where that diameter occurs along the cone's axis.

3D scanning is neat and 4 axis milling is impressive (albeit that video is only demonstrating 3 axis milling), reproducing this part just requires some graph paper, a camera, and a CNC lathe (no need for an indexed spindle either).

Another method for recreating the cone would be to use a nice set of hole standards and measure the depth at which each diameter sits on the cone. With enough measurement, one could possibly extrapolate the desired profile.

HTH,

Jay

Take a look at the race cone.

It's not as simple as a basic lathe. There are some lathes with advanced carriage movement that can do asymmetric turning, and they're getting more common but they're hardly run of the mill.

That being said, if I had all the money in the world, there are some shops that specialize in asymmetric turning with PCD tooling that get nearly nano-meter tolerances. You could get some sweet custom cones from those guys.

[dubois]

Hold on a minute! The cone is a spinning part, correct? So you only need 2 axes to reproduce one, ie. a lathe.

A far as getting the measurements of the cone, why not simply take a photograph of it? From dead-on perpendicular to the cone and from a distance away to reduce warp-age, you would be able to measure both dimensions required for reproducing the cone, diameter and where that diameter occurs along the cone's axis.

3D scanning is neat and 4 axis milling is impressive (albeit that video is only demonstrating 3 axis milling), reproducing this part just requires some graph paper, a camera, and a CNC lathe (no need for an indexed spindle either).

Another method for recreating the cone would be to use a nice set of hole standards and measure the depth at which each diameter sits on the cone. With enough measurement, one could possibly extrapolate the desired profile.

HTH,

Jay

Take a look at the race cone.

It's not as simple as a basic lathe. There are some lathes with advanced carriage movement that can do asymmetric turning, and they're getting more common but they're hardly run of the mill.

That being said, if I had all the money in the world, there are some shops that specialize in asymmetric turning with PCD tooling that get nearly nano-meter tolerances. You could get some sweet custom cones from those guys.

Having worked at GE-Evendale in the fabrication shop, planning and making more complex parts than this cone, my first instinct is to design a very special cutter (we had a department just for that), then CNC mill the part (yes, that's when I call my programmer). A lathe, I agree would be very challenging.

[PatAllen]

Having worked at GE-Evendale in the fabrication shop, planning and making more complex parts than this cone, my first instinct is to design a very special cutter (we had a department just for that), then CNC mill the part (yes, that's when I call my programmer)

i beg to differ, if you look at it, the surface finish migh not be super important, as long as the lift of the mettering arm (lever pin) is correct and that it is capable of sliding over the cone...probably a 1/8 ball nose end mill in a 10 000 rpm cnc and a indexed chuck (like in the video) would do very fine imho. i think the mettering arm has a ball bearing at its end so it may "roll" over the cone instead of rubbing on it...anyway, if it where just me i would do it this way. i have a friend with all cnc goodies you can think of...i just dont have a A4 Tii to tune...lol...

[Guest Anonymous]

For anyone seriously thinking about trying to replicate one (or more of these) its not just the cone that is different, the pistons inside are 1.0mm larger at 7.5mm and my recollection is that the head is also that much thicker (my pump is at my engine builders so I can't measure it easily), Jeff may know. I don't believe they needed to uprate the piston size in the A4 pump as the A4 doesn't make that much power (as compared with a full injected race M10, or DOHC M10 (schnitzer) or M12/7.

www.alpinabmw2002.com

[Swiftus]

Ahhh, I see now. So the 'cone' isn't an ideal cone, but more of a cam-shaped cone, if you will.

I still think that you could get away with using a 3 axis mill (A-X-Z or A-Y-Z), but most 3 axis mills are actually 4 axis mills.

Now the measuring and replicating of the part becomes more difficult. I would think that if one had access to a 4 axis mill, they could use a probe and plot a point cloud of the cone. you know, the old skool way of digitizing a 3D part before the 3D scanning phenomenon. Either way would work.

Now I see why Alpinas cost so much.

[dubois]

Having worked at GE-Evendale in the fabrication shop, planning and making more complex parts than this cone, my first instinct is to design a very special cutter (we had a department just for that), then CNC mill the part (yes, that's when I call my programmer)

i beg to differ, if you look at it, the surface finish migh not be super important, as long as the lift of the mettering arm (lever pin) is correct and that it is capable of sliding over the cone...probably a 1/8 ball nose end mill in a 10 000 rpm cnc and a indexed chuck (like in the video) would do very fine imho. i think the mettering arm has a ball bearing at its end so it may "roll" over the cone instead of rubbing on it...anyway, if it where just me i would do it this way. i have a friend with all cnc goodies you can think of...i just dont have a A4 Tii to tune...lol...

Pat, you are right, the special cutter might be just as simple as a nose end mill, and that may explain the rough surface finish on those parts. There is at least a couple ways to make the same thing. I haven't been in a machine shop for more than 20 years, but the principles are the same.