How could I make this curved conrod?

[SillyOldDufferModerator @sillyoldduffer]

[SillyOldDufferModerator @sillyoldduffer]

This item is the con-rod for a model mill steam engine. I'm trying to make it properly as per plan.

The rod part of the con-rod is of circular section and the whole is curved on a 300mm radius. The rod is ⌀4mm at the ends and it bellies out to ⌀6mm in the middle. In a full size engine this is done to minimise the weight of the rod whilst keeping it stiff enough not to kink under compression.

My problem is that I don't how to turn that R300 radius along the rod. I have a manual lathe. The curve implies that the tool point has to smoothly move in – out – in as well as from right to left. I prefer not to approximate the shape with tapers because that's not what the specification calls for.

Can anyone help please?

Thanks,

Dave

[Phil H1Participant @philh196021]

Dave,

I have done this before (no pictures sorry) as follows (describing the central section only);

1. Turn the centre of the rod to 6mm diameter along its full length.

2. Use the top slide set at a shallow angle to turn a slight taper in one direction.

3. Reverse the rod (assuming you are turning between centres) to turn an identical, slight taper in the opposite direction. The end of the taper would achieve the 4mm diameter – or close to it at both ends.

4. Blend the tapers/ short parallel central section by using flat files.

5. Once the file work has blended the rod to a smooth curve, use progressive grades of emery cloth to complete the rod to a nicely polished finish.

I understand that you are trying to work as accurately as possible (by quoting the 300mm radius) but I doubt if anyone will check it.

Hope that helps.

Phil H

[Jeff DaymanParticipant @jeffdayman43397]

1. I suggest simplifying the design and not perpetuate designer's silliness (rod could be straight or tapered or rectangular etc.) but if you MUST make it spherical r300:

2. make a free floating auxiliary cross slide and have it follow a cut sheetmetal template with curve cut into it

3. calculate coordinates of multiple points diameter @ length, turn in steps at these coordinates, file to finish / remove steps.

4. find a friend with a CNC lathe and make it in 1 minute plus 5 mins programming.

[Clive FosterParticipant @clivefoster55965]

If you have sufficient room behind the lathe and are able to detach the cross-slide feed screw rendering it free moving the curve can be generated using a link rod of appropriate length.  300 mm in this case.  Basically one end of the rod is fixed to the cross-slide and the other end to a strongpoint behind the lathe. The strongpoint needs to be positioned so that the link is in line with the cross-slide travel at the point where the nascent con-rod is of maximum diameter. As you move the saddle along the bed away from the point of maximum thickness the end of the link fixed to the cross-slide describes a circular arc pulling the slide with it.

With the cross slide feed screw disengaged the cut needs to be put on via the top-slide which must therefore be set parallel to the cross-slide. The tool tip needs to be in line with the link fixing if the arc cut is to be accurate and symmetrical.

Much easier to say than to set-up and get working right. After the first three you will know what you are doing but the first one in particular can be more than little fraught. Given the end fittings on that rod I'd definitely budget at least two practical runs on a suitable piece of scrap. Tool needs to be very sharp.

Clive.

Edited By Clive Foster on 06/10/2017 17:25:00

[egaParticipant @ega]

Doesn't Tubal Cain recommend making these by allowing the rod to give under the turning tool in the middle? ie take advantage of the flexibility of the work to achieve the desired result.

I will try to find the reference.

Got it: page 53 of his WPS 28 Simple Workshop Devices.

Edited By ega on 06/10/2017 17:39:04

[Mike PooleParticipant @mikepoole82104]

I would be inclined to make a template and as described above use a combination of turning and file work to match template. For a one off this is probably the simple solution. If the time has come to make a template follower for your crosslide it could be a big time saver on the next job but it might be a long time before you get that time back.

Mike

[JasonBModerator @jasonb]

Have a look at my post part way down this page on Fish Belly rods, will also look better than that one in your drawing with the constant radius

If you do want to keep to the 300 rad then work out a set of co-ordinates at say 1mm steps and machine to that then blend with a file.

Edited By JasonB on 06/10/2017 18:21:13

[Chris Evans 6Participant @chrisevans6]

As Jason alludes to it is simple maths to work out the steps. You know the radius so the calculation is the square root of the radius squared minus the distance you choose to put the step in at say two mm step overs. When you have worked these out more simple maths can give you an angle to turn between steps as you get all info needed for a right angle triangle. Easier to just blend with a file though.

[NobbyParticipant @nobby]

I would say make a template like Mike says set up similar to my effort  The template fits in to tailstock Nobby

Edited By Nobby on 06/10/2017 23:22:51

[HopperParticipant @hopper]

Personally I'd turn the suggested tapers and blend it with a file, per Phil H1 above.

But if I wanted to fiddle about and turn an exact 300mm radius, i would make up simple version of a ball turning tool. Piece of flat plate bolted to the cross slide, extending out the back of the lathe, with a pivot pin sticking up out of it, circa 300mm behind the centreline of the lathe spindle. Then a piece of long flat bar, circa 450mm long, with a hole drilled in one end to fit on the pivot pin in the bottom plate. Then attach a tool post, or any block with a slot to hold a lathe tool bit, positioned so the tool bit tip is exactly 300mm from the pivot pin centre.

You then hold the end of the long flat bar as a handle (you could even fit a nice brass knob to it) and move the cross slide into position to centre the tool along the length of the to-be-curved section of the job. Gently sweep the end of the flat bar back and forth the required amount as the cross slide is fed in to put on the depth of cut. When the cuts at each end of the job meet in the middle, job's done.

Once job is done, throw the whole device in the scrap box never to be used again.

[Raglan LittlejohnParticipant @raglanlittlejohn]

Nobby's idea is ideal for this job. I did a similar set up for a model loco smokebox door. Use slow power feed (sliding), and keep adjusting the cross slide to follow the template. you have to keep watching the dti, and try to keep the needle in the same place. Best to work from the middle of the template, as you have to wind the cross slide in to keep up. If you get behind with the cross slide you just stop cutting, and don't spoil the job.

[Russell EberhardtParticipant @russelleberhardt48058]

Many years ago I had to make some fish bellied pushrods for a 1930 sports car. They were very slender and tapered in a gentle curve at both ends to the ball end to reduce the reciprocating mass. The originals had been made by a knitting needle manufacturer! The method I used was to mount them near the centre in a collet and then support the end in a female centre which was offset to give the correct taper by bending the rod. That resulted in a nice curve in one pass. The ball ends were than case hardened and polished.

The pushrods were about 1 ft long and 5 mm maximum diameter so flexed easily. It might not work on Dave's shorter rods but the method is worth remembering for longer rods.

Russell

[egaParticipant @ega]

Russell Eberhardt:

Your method sounds similar to the Tubal Cain method I mentioned earlier (my description of it was not very good).

The fish-bellied push rods remind me of Bugatti's banana-shaped tappets.

[Harold Hall 1Participant @haroldhall1]

Dave

Have a look at my method shown on my website here **LINK**

Harold

[JasonBModerator @jasonb]

This one is done as a bender, I tend to use that more for rods that taper in one direction rather than two. Small live ctr in a boring head to get the offset.

 

The other thing that is not good in the OP's original drawing is the abrupt junction where rod meets flange and small end, just asking for a fracture, best to have a small internal radius where they meet.

Edited By JasonB on 07/10/2017 13:16:13

[Anonymous]

Hydraulic copy unit and CNC milled pattern.

If I was going to use handraulic methods I'd follow the approach set out by Phil H1. Which is exactly the method I used to form the crown on these pulley wheels:

So the radius isn't exactly 300mm? Who cares, it doesn't matter as long as it looks ok. Remember that the original was probably cast or forged so wasn't likely to be to drawing. If forged it was probably finished with files and emery paper anyway – kept the apprentices occupied.

Andrew

[Neil WyattModerator @neilwyatt]

As Jason has pointed out, if you copy the 300 radius you need to copy the sharp corners to be true to the design

I would ask the original designer how HE did it!

Anything other than blending some taper cuts seems OTT.

I'm sure that's how I did this one…

[Russell EberhardtParticipant @russelleberhardt48058]

Posted by ega on 07/10/2017 11:41:12:

The fish-bellied push rods remind me of Bugatti's banana-shaped tappets.

The Talbot used an interesting rocker design as well:

[egaParticipant @ega]

"The Talbot used an interesting rocker design as well:"

Again, the aim seems to have been to reduce the reciprocating weight; I notice that the valves were waisted, too.

Was the Talbot a Georges Roesch design?

[Sam Longley 1Participant @samlongley1]

Not that I know anything about these things but could one turn to near enough then mount a small grinder on a radiused arm & grind the final shape; or is that not a possible option to finish such things

[Russell EberhardtParticipant @russelleberhardt48058]

Posted by ega on 07/10/2017 17:28:08:

Was the Talbot a Georges Roesch design?

Yes. They were very successful in sports car racing in the early thirties.

Russell

[Roger Provins 2Participant @rogerprovins2]

I've made several of these over the years and have always used a method similar to that laid out above by Phil H1.

Quickest and simplest way.

Roger

[SillyOldDufferModerator @sillyoldduffer]

Brilliant answers covering a wide range of possibilities!

Very helpful to know that the form I'm asking about is 'Fish belly' or 'Barrel Shaped'.

Apologies, if I missed anything but in order of accuracy the techniques you've described are:

1. Approximate the shape by blending tapers. I accept that this is plenty 'good enough' in most cases. (Phil, Jeff)
2. Calculate the diameters at a series of points along the rod. Cut and blend them. This would be much closer to the specification. (Jason & Chris)
3. Bend the rod to the required radius. May not work on a short stubby rod. (Russell. ega & Jason)
4. Free the cross-slide and follow a template or linkage (Clive). I like the template approach because (as shown in Nobby's pictures), it has many applications – different shapes can be made and copied with it, but making one is quite an investment (Jeff, Nobby, Raglan, Mike)
5. A ball turner variant. (Hopper)
6. A CNC or hydraulic lathe programmed to cut curves. If only I had one! (Andrew, Jeff)
7. Grinding. This, I guess, is how such shapes are mass-produced. (Sam)

Neil highlighted a problem with perfectionism. Inspector Meticulous would not be satisfied with just an R300 fish belly: he would insist on sharp ends as well! The more I look at problems like this the more I appreciate what goes into exhibition standard work.

Special thanks to Jason who can remember advice he posted in 2014 (I can't remember what I did yesterday) and also to Harold Hall for the link to his website where he provides much helpful detail on 'barrel shaping', including a table of coordinates.

I have to cut grass next but will report back after I've made a real one.

Many, many thanks,

Dave

Edits, typos…

Edited By SillyOldDuffer on 08/10/2017 11:36:23

Edited By SillyOldDuffer on 08/10/2017 11:38:15

[Anonymous]

Posted by SillyOldDuffer on 08/10/2017 11:35:28:

Neil highlighted a problem with perfectionism. Inspector Meticulous would not be satisfied with just an R300 fish belly: he would insist on sharp ends as well! The more I look at problems like this the more I appreciate what goes into exhibition standard work.

Any Inspector Meticulous worth his salt would sack the designer for creating a poorly engineered and difficult to make part with no fillets.

Andrew

[Author]

Posts