I have admired the old 4-stroke marine engines for a long time, and finally got a short break from work, so I am taking a few days off and trying to hash out a design for a 3 hp Frisco Standard at about 40% scale, which would give a 12" flywheel, and almost a 2" bore.

Lots of tricks and turns in this design, but i really like the side-shaft layout.

There is a twin kit available, but it is quite costly, and I really want a single, so I am making up my own 3D models and then 2D drawings, using photos of old Friscos.

So far so good with the design.

It seems to be coming together.

Edited By PatJ on 11/01/2021 08:38:55

I need help with the helical gears for this engine.

As with all 4-stroke IC engines I am aware of, the camshaft turns at 1/2 the crankshaft speed.

I have verified with videos of running Frisco Standard engines that the vertical camshaft is indeed running at 1/2 of the crankshaft speed.

My problem is that the helical gears at the camshaft-crankshaft junction appear to be the same size, both in photos and on the patent drawings.

So my question is, how do I get a 2:1 gear ratio using cross helical gears that are the same size?

Obviously I am missing somthing.

My understanding of gears is not good, so that is hampering things.

If I increase the diameter of the camshaft gear to twice that of the crankshaft gear, then the camshaft gear does not fit the engine (it strikes one of the vertical supports).

I am stumped on this problem.

Is there an obvious solution that I am overlooking?

Anyone have any experience with side shaft helical gearing?

Thanks in advance.

.

Ahhhhhhhhh…………….thanks very much Jason and Rod.

I knew that knowledge base was here.

So it seems there is a bit of magic to this gear arrangement, at least in my mind.

I can't quite wrap my head around the speed difference caused by a variation in the angles, but I will take your word for it.

My plan for fabricating these gears, which may or may not be viable, is as follows:

1. Draw a standard spur gear in 2D in CAD (or download a profile from a supply house.

2. Pull the profile into Solidworks, and extrude while rotating the necessary degrees.

3. 3D print the gears.

4. Make resin-bound molds around the gears.

5. Heat the 3D printed gears until soft, and carefully extract from the mold. The center of the 3D printed pattern can be mostly left open, such as with the gear on the right in the photo above.

6. Pour the mold (with gray iron), clean up a bit with a file, and then run in the gears with some fine lapping compound.

7. The shrinkage factor will have to be pretty close for the gears to work, but not exact.

8. I use a ceramic mold coat, and so the finish on my iron castings is almost perfectly smooth, and I don't have to wire brush the casting to clean off the sand.

9. The resin-bound sand make a very accurate casting, since unlike greensand, you do not rap the pattern in order to withdraw it from the sand.

I could also use a lost-PLA (or a similar filament that is designed to burn out cleanly), and use a ceramic slurry, although I do not have the slurry, and am not set up for that method at the moment.

I think the resin-bound sand method would work.

The slurry has a fixed shelf life, which is perhaps 1 year?, and so I don't really want to purchase that for just two gears.

The lost PLA method is extremely accurate though, assuming your shrinkage factor is correct.

The profile on the gear on the right above does not look the same as the one on the right though, so that may be a fly in the ointment.

I will ponder this further, but I greatly appreciate this information.

.

Edited By PatJ on 28/01/2021 13:14:10

It seems like for the gears above, I could lay out the profile for the first one, rotate a cut to make the 3D model, and then take a section from an appropriately skewed plane to come up with the section for the other tooth form, and then rotate that.

This way, I know the gear profiles would mate correctly.

.

Edited By PatJ on 28/01/2021 14:28:14

I think that issue is addressed at 9:36 in this video?

I was able to convert a downloaded gear model from a supply house, with teeth at 45 degrees, to a Solidworks part file, and then add two of those into an assembly.

The teeth appear to mesh ok.

I used that same tooth form to create two new gears, with 30 and 60 degree twists, and those gears do not appear to mesh.

I will have to try and find the article that Rod mentioned above, and try to replicate his machining method in Solidworks.

.

Eric-

Thanks for pointing that out. I saw the photo, but did not realize the tooth count varied on each gear.

Very interesting.

The 3D model for this engine was going along so well, and I was saying to myself "this is not going to be too bad at all". I spoke too soon, and this gear thing has turned out to be rather an iceberg strike.

So here is what I am faced with:

1. My machining skills are somewhat limited (improved greatly, but still limited as far as helical gears).

It is safe to say I would not be able to make helical gears in a reasonable amount of time, if ever.

2. I would like to be able to cast this engine in more than one size if desired.

3. I would like to be able to download standard gear 3D files from the supply house, and use those in Solidworks.

4. I would like to be able to 3D print the downloaded gears, and make molds from those, using the methods previously described (heat the 3D part just enough to remove it from the resin-bound mold).

5. I don't want to modify the dimensions of the engine 3D model that I have already created.

So the only solution I can come up with which would ensure that all the gears would fit and mate correctly is as shown below. The intent is to hide the bevel gears inside the engine base, so that they will not be visible.

Sort of a kludge of a solution, and I try really hard to always avoid kludge solutions, but in this case, for me, it is either this layout, or don't build this engine.

I know all these gears will mesh correctly because they are commercial gears.

The angle on the helical gears is 45 degrees.

And one unintended benefit from the layout below is that it gives adjustment between the vertical camshaft and the horizontal crankshaft, ie: the small beveled gear can be slide to the right or left as needed to get a perfect fit.

It would otherwise be somewhat difficult to make two gears so exact that they fit the centers of the two shafts exactly and still meshed correctly.  The lower helical gear could also be shimmed up or down as required.

Edited By PatJ on 30/01/2021 12:24:49

Edited By PatJ on 30/01/2021 12:27:02

Edited By PatJ on 30/01/2021 12:29:45

I attempted to cast a pair of full-sized gears for a windmill restoration, and was not successful due to mold failure, but I believe that if I had heated the 3D printed gear pattern to the point where it was soft, I could have successfully removed it from the bound sand.

This began a discussion on another forum about how it was done in the day, and so I looked at my photos from the Soule Museum in Meridian MS (USA), and there were a great number of gear patterns.

The question was raised , and I think it is a fair question "Why cast a gear? Why not just cut a blank and then use gear cutters to cut the teeth?".

My guess is that back in the day, it was far easier to cut gear teeth in wood, and then cast the gear.

A cast gear would probably only require minor cleanup and cutting to get it the exact size, and so a great deal of machining would be eliminated for every gear produced if a cast gear was used.

Back in 2012, I wanted to make castings in gray iron, and "they" said "you can't do that in a hobby setting".

"They" were wrong. Then they said "Well if you do make an iron casting, it will have inclusions, and will have chilled spots that are too hard to machine". "They" were wrong again.

I must admit I was a member of the "they" group, and I did not believe it could be done on a hobby level either.

But the proof is in the pudding, and as I have mentioned to some non-believers, come over to my house and I will demonstrate first hand how it is done, and you can drill and machine the parts yourself and see how easily they machine with no chills or inclusions. I can repeat that on any given day.

Then someone came up with the "lost-PLA" method, which is to 3D print a pattern using PLA filament, then coat it with multiple layers of ceramic slurry, burner out the PLA in an electric kiln, and then cast the part or parts in the same fashion as the lost-wax method.

I said "there is no way this will work". I was proven wrong, and someone sent me samples of a small part which had the surface finish and accuracy of a die-cast part, but in gray iron. I drilled the thin samples, and they drilled perfectly with no chills or inclusions.

I have not tried the lost-PLA method, since I think I can cast a sufficiently accurate gear with resin-bound sand.

The ceramic slurry has a given shelf life, and so that is another reason I have not tried it yet.

And the lost-filament (ceramic shell) method is a multi-step method (multiple layers of shell, burnout, etc), and so is far more time consuming than making a resin-bound mold, which takes perhaps 15 minutes total.

And there is 3D printer filament that supposedly will burn out very cleanly, and it is described in this video.

The accuracy of parts made using the lost filament method is phenomenal (in my opinion), and infinitely easier than the lost wax method (also in my opinion, not having done either though).

Edited By PatJ on 31/01/2021 06:56:55

Edited By PatJ on 31/01/2021 07:02:26

And here is my failed attempt to cast a pair of full sized windmill gears.

This was not my pattern, and I only had one pattern, so I could not heat it up to get it soft enough to get a clean removal from the sand.

I actually think I could have cast a gear with this mold, but there was some mold cracking, and so I did not feel the casting would be accurate enough, and I abandoned the attempt.

The bound sand does produce a dimensionaly accurate casting, since the sand sets prior to the pattern being extracted, so the mold is a dimensionaly exact replica of the pattern.

The fact that there is no draft angle on the gear pattern is the reason why it would need to be heated to get it soft for removal from the sand.  I would have most of the center part of the gear pattern removable (the center part could be wood), and thus only a thin outer section with the teeth would need to be removed.

I do think this is a viable method though if the pattern can be sacrificed.

Edited By PatJ on 31/01/2021 07:12:12

Edited By PatJ on 31/01/2021 07:14:26