Prop shaft tube, bearings and seal design

[pdwiley]

Another question.

My plans call for a straight piece of 50 NB heavy wall pipe. The prop shaft will be either 1 1/4" to 1 3/8" depending on what I can find at the right price. I don't really care as I have the tools to do the taper, threading and keyways myself. My engine is a Yanmar 3QM30H, 27HP at 2600 rpm, with a 3.5:1 gearbox if it matters.

I seem to recall seeing a prop shaft/tube design here at one time that incorporated thrust bearings in it. Anyone have a reference? I'd prefer to keep the thrust off the g/box, not because it can't take it but because then I can put a universal/CV joint between the box and shaft, which eliminates a lot of the hassles about shaft alignment and movement due to engine torque reactions on anti-vibration mounts.

I have a preference for plastic oilon bushings top & bottom and fill the prop tube with light grease. I know the bushings swell 11% in water and can allow for this in their machining. Is there any reason I shouldn't do either or both? Anyone want to suggest another readily available industrial plastic or similar?

Why should I pay marine chandler prices for dripless ceramic shaft seals etc when industrial pump seals are the same thing at a fraction of the price? We're only talking about 600mm to 900mm of head, after all.

I have a lathe more than big enough to handle the prop tube to machine seats for bushings and to true up a welded flange on the inboard end so actually making it isn't a drama, just getting the design as good as I can.

That's probably enough for the moment. My angle grinder is telling me that it's feeling neglected.

[MikeJohns]

I like grease filled shafts, but use waterproof grease.
Ultra high molecular weight polyethylene is a good marine bearing material.
You can buy all sorts of stock sizes and turn them out yourself, or buy standard bushings.
I think AB plastics is worth a try, There's also vesconite from Vesco plastics ( Both in Victoria ).

I'd use a commercialy available thrust bearing, they are a good idea if you have room.

I sent you a PM

[Frosty]

A universal joint can take thrust, its a constant velocity joint that can not.

[CDK]

I use stainless steel stern tubes, oil filled, integrated thrust bearings and universal joints. There are several posts with pictures and drawings on this forum. Search for DIY surface drives.
The system has been in place for several years now and the oil bottles show the exact same level the day I filled them. No absurdly priced marine seals, just standard industrial items from SKF.

[anthony goodson]

Polyacetal machines well and doesn't absorb water

[pdwiley]

I use stainless steel stern tubes, oil filled, integrated thrust bearings and universal joints. There are several posts with pictures and drawings on this forum. Search for DIY surface drives.
The system has been in place for several years now and the oil bottles show the exact same level the day I filled them. No absurdly priced marine seals, just standard industrial items from SKF.

That's the sort of thing I had in mind, thanks. My tube is 60.3mm OD Sched 80 black pipe. I could swap to s/steel as I haven't welded anything in place yet but to what advantage?

The sizes aren't the same as my shaft will likely be 35mm OD but I have a bearing catalog and I see that your sleeve is 35mm OD anyway.

Questions:

How many hours of running time do you have on this setup?

How did you hold concentricity between the tapered roller bearings and the radial bearing at the prop end, or didn't you worry about it? I know some bearings are self-aligning (I have a number of them).

How did you lock your prop shaft to the thrust sleeve in your drawing? That one, for my setup, has me a bit puzzled.

PDW

[CDK]

Here are the pictures and drawings:
http://www.boatdesign.net/forums/surface-drives/diy-tunnel-drive-20433.html#post172724

I did not keep track of the number of running hours but looking at the date of the original post it must have been several 100's. Occasionally I lift the floor to check the oil level an decide whether or not to make a signaling system. Never noticed a change, so I forgot about it. The trick is the double seal behind the props, the chamber in between is filled with silicone oil, the best water repellent I've ever seen.

While constructing the tubes I used a dummy shaft with discs to simulate the bearings and a threaded end to firmly clamp the flange to the tube. Because I didn't trust my ability to weld the parts together without a flaw I took the whole to an experienced welder who did it for me with a TIG welder.
I let the dummy in place until the construction was cold, then inserted the real shaft with bearings and confirmed the concentricity.

From the end of the prop shaft I removed appox. 2 mm to create a shoulder for the bearing carrier. That takes the forward thrust.
The reverse thrust is taken up by the splined adapter, kept in place by the oversized clamps with 4 of the strongest Allen bolts I could find. That part also has to transmit the torque, so I probably overdid things a bit.
If there was more space for the engine, gearbox and universal shaft, I would have probably used a more orthodox construction, but this has survived lots of abuse like bollard tests, a rope in the prop and changing gear with 3000 rpm because the electronic throttle control got stuck.

[pdwiley]

Thanks, I've downloaded those drawings.

If I turn the inboard end of my prop shaft down from 35mm to (say) 30mm, that gives me a shoulder to handle thrust in forward. If I put a flange on the inboard end of the shaft where the flange face butts against the other bearing cup then this will take care of thrust in reverse. The inboard seal then runs on the outer surface of the flange not the shaft, which is no big deal as long as the surface finish is within specification. The material can be carbon steel as it won't - in theory - be exposed to salt water. I can machine the flange to the SAE specification without any dramas.

I'd do my bearing housing differently to you, as in essence a double flanged bobbin, but that reflects what we have to work with rather than anything fundamental. I'd weld it all up first, then stress relieve it in a nice wood fire, then machine it. Get one flange right, then flip it over in the lathe and machine the rest of it in one setup.

[CDK]

I concur.
My lathe is too short, otherwise I would have followed the same procedure.

The carbon steel flange will get a bit rusty because of the environment, but that ends at the seal. A light coat of silicon oil will keep it shiny for at least a year.
How will you secure the flange to the shaft?

[hartley]

FROSTY i thought better of you .universal joints ,such as the hook type
typified by dana spicer ,hardy spicer etc etc will not take thrust loadings , why do you think we take so much time and effort to incorporate proper thrust bearings ,if you doubt this ,take one apart ,see those little needle rollers ??? some one else made the same statement in another thread ,about
time it was put to rest .Now frosty i like your posts so seasons greetings to you and yours ........cheers Hartley

[Frosty]

It cant take as much thrust as it can take torque!!!!!

A thrust bearing Is a different thing.

Why am I telling you this

[Tigawave]

Beware of grease, grease is a poor conductor of heat, so if you are planning on using a polyacetal or polymer composite bearing (which all absorb water and suffer bore closure) be very careful of thermal expansion of the bearing. I've seen many failed bearings due to too much grease working its way back from the stern gland.

The poly bearings differ widely in their bore closure characteristics due to moisture absorption and thermal expansion. You pays your money and takes your choice.

Phenolics exhibit some of the best bearing properties currently, and are significantly different to other poly bearing materials. There are a few suppliers each with different properties so it can be worth shopping around if you want the best. If it's low power low torque low rpm then it's not such an issue.

[Frosty]

I agree with tigar . I have seen a few boats have to have the shaft cut from the boat.

You would think that a rudder would not suffer the same, as it is minimal rotation, but that also does not seem to be the case.

I used cutlass bearing on my rudder shaft, yes a bit expensive but 7 years down the road they are perfect and can be removed as if they were new.

The rubber of the cutlass also absorbs vibration that high speed rudders can sometimes generate.

[pdwiley]

I concur.
My lathe is too short, otherwise I would have followed the same procedure.

The carbon steel flange will get a bit rusty because of the environment, but that ends at the seal. A light coat of silicon oil will keep it shiny for at least a year.
How will you secure the flange to the shaft?

Assume you mean the inner flange that's going to get connected (somehow) to the g/box via a stub shaft/universal joint setup.

My first thought was a straight sliding fit, keyway and retaining nut as simplest and allowing adjustment of bearing preload via machined fits, spacing rings, what have you. However I see that the recommended way is a taper fit a la the reference to Gerr's book pages on Google. This has the disadvantage of making the fit to the opposed tapered rollers more difficult, assuming the taper engagement can move slightly depending on torque applied to the nut.

So - I don't know, yet. My big lathe has a taper attachment so actually cutting tapers is relatively straightforward (once you do a dummy one or 2 in something sacrificial of course, to check the angle...).

Ditto whether to use 316L or 2205 for the shaft. I need to check out the local prices on both before deciding.

On universal joints, my advice is that they will NOT take end thrust, the little needle bearings will die rapidly. They also need a shaft angle of around 15 degrees so the bearings don't brinell rapidly. A CV joint deals with issue 2, but not (AFAIK) the first one.

PDW

[CDK]

That's what I meant.
Gerr may state that a taper fit is the best way, but you'll never be able to predict the bearing play until you fasten the flange nut. Is there no slack at all you may have already damaged the bearings and if there is, you have to pull it off again which can be pretty hard. At the next fit it will slide a tiny bit further up the taper.

If you have calculated a straight sliding fit can take the torque, I would do it that way. The only other, stronger option I can think of is splines, but you need special equipment for it. I had my SAE spline adapters made by a friend in a large machine factory and still had to spend several hours with a Dremel grinder in the support of my lathe because the splines weren't deep enough.

[hartley]

Frosty I don't know why you are telling me this either:confused: you say and i quote ''a universal joint can take thrust'' I say it cannot ,simple really
.....cheers Hartley

[Frosty]

A UJ in a heavy goods vehicle prop shaft takes many hundreds of ft lbs torque.

The load from the 2 yokes is transfered to the spider and across to the opposing 2 yokes. This load on the needles is no different to load from 90 degrees which would be thrust.

I am sure you confusing this with a constant velocity joint which can NOT take thrust at all and you will no doubt use these on your thrust bearings assembly as do others.

Ask your engineering section, Im sure they will explain to you.

[pdwiley]

That's what I meant.
Gerr may state that a taper fit is the best way, but you'll never be able to predict the bearing play until you fasten the flange nut. Is there no slack at all you may have already damaged the bearings and if there is, you have to pull it off again which can be pretty hard. At the next fit it will slide a tiny bit further up the taper.

If you have calculated a straight sliding fit can take the torque, I would do it that way. The only other, stronger option I can think of is splines, but you need special equipment for it. I had my SAE spline adapters made by a friend in a large machine factory and still had to spend several hours with a Dremel grinder in the support of my lathe because the splines weren't deep enough.

I agree completely. Tapers are fine in theory, but there's a downside.

I am leaning towards a keyway plus a Fenner type taper-lock bushing. You can buy the hubs made to have flanges etc welded to them. I have used these successfully on many pulleys over the years and can't see why one of appropriate size for the torque load shouldn't work. What I'm not sure of, as it's never been an issue for me in the past, is if I can get the exact location on the shaft that I need. If not, same problem in essence as a tapered fit, except you don't need to do any machining.

Suitable inboard tapered roller bearings were $AUD 16.50 each and it took me less than 30 minutes to bore a test piece of Schedule 80 pipe to the necessary 62mm inside diameter.

I actually can cut splines if I had to as I have a horizontal mill with dividing head (not that you need one for 6 or 12 splines) and also a vertical slotter for cutting internal keyways and splines. However, it's a pain unless you have a spline cutter of the correct profile and I couldn't be bothered making one for a single job.

A bushing may be better for me than a bearing in the outboard end because, while I can get the bores machined parallel, after I weld the tube assembly into the keel I can't guarantee that the bore will stay precisely parallel. A self-aligning double row ball bearing would be perfect here, but unfortunately I can't find any of the correct size (approx 52 OD, 35 ID).

PDW

[CDK]

SKF only produces 35/72 and 35/80 double row bearings (1207 & 1307), the others normally follow in their footsteps. Your order will not be large enough to make a special.

[Tigawave]

Why not use an oil lubricated plain composite bearing, if there is small mis alignment in the tube, this could be taken up on epoxy bedding. A simple clearance fit plain bearing as used on commercial vessels, some suppliers will supply spiral grooved bearings to actively circulate the oil which also aids cooling. The side loads are small so oil film support will be adequate. These work well for years and there may be no need to reinvent the wheel.

[hartley]

Frosty re uni joints and thrust I dug out an old book by Dave Gerr titled Boat Mechanical Systems Handbook ,under a section "universal joints '' he says and i quote "neither the standard old fashioned uni joints also called Cardan or Hook joints ,nor modern c v joints can take thrust,at least not much '' he goes on to say you may get away with a uni joint and no thrust bearing with a low powered engine ,say 10 hp but don't try it with 300-hp.
cheers Hartley

[Frosty]

I agree I said that a CV can not take thrust.

A universal joint can take massive torque,-- the twisting force from the side is called torque the force from the front or pushing force would be thrust.

As the bearing does not have a thrust side or any side, being roller bearings in a cup then force can be taken anyway including thrust. The limit being the strength of the spider and not the bearings as you suggested.

Now if you are involving mangle then that is a different conversation, and the UJ must be on the output shaft itself as indeed it is on a vehicles prop shaft. One being on the diff yoke and the other on the gearbox drive spline yoke.

Some have centre UJ's but always have a centre shaft bearing holding its position.

[pdwiley]

Why not use an oil lubricated plain composite bearing, if there is small mis alignment in the tube, this could be taken up on epoxy bedding. A simple clearance fit plain bearing as used on commercial vessels, some suppliers will supply spiral grooved bearings to actively circulate the oil which also aids cooling. The side loads are small so oil film support will be adequate. These work well for years and there may be no need to reinvent the wheel.

Yes, for my lower bearing that's what I plan on doing. I can get ball bearings of the size I need and I can machine the shaft pipe to suit easily enough, but if I get any welding induced distortion there goes the alignment. A bushing is safer.

PDW

[pdwiley]

SKF only produces 35/72 and 35/80 double row bearings (1207 & 1307), the others normally follow in their footsteps. Your order will not be large enough to make a special.

Yeah, funny that....

With respect to your outside bearing assembly, I can't see what holds the entire inner assembly inside the 63.5 pipe except possibly the compressed O ring. Is there a plug weld, inner weld or similar that's not shown on the drawing? I can build the same basic assembly except use a longer bushing in place of your ball bearing.

PDW

[CDK]

No, there is just the neoprene O ring.....

Sealing and holding the assembly in place seems a bit much for just a ring, but if the cavity has the proper shape, so the ring doesn't get pinched on the inner side, it exerts enormous pressure on the inside of the tube.
I got the idea when I made a linear motor for outside use. The groove for the O ring wasn't deep enough but I assembled it anyhow, only to find that the shaft seemed to be welded to the cylinder.

The recess in the bearing carrier is somewhat rounded, the one in the cover is flat. I assembled it with the shaft removed, tested it with a heavy object, disassembled it and took a new O ring for the final installation, although the old one didn't show any signs of damage.

[pdwiley]

No, there is just the neoprene O ring.....

Sealing and holding the assembly in place seems a bit much for just a ring, but if the cavity has the proper shape, so the ring doesn't get pinched on the inner side, it exerts enormous pressure on the inside of the tube.
I got the idea when I made a linear motor for outside use. The groove for the O ring wasn't deep enough but I assembled it anyhow, only to find that the shaft seemed to be welded to the cylinder.

The recess in the bearing carrier is somewhat rounded, the one in the cover is flat. I assembled it with the shaft removed, tested it with a heavy object, disassembled it and took a new O ring for the final installation, although the old one didn't show any signs of damage.

OK, thanks, I just wanted to check that I was interpreting the drawing properly. I need to think this through a little more but as I'm away from my machine shop etc for Christmas, that's no big deal.

Those drawings of yours are very helpful. I've fabricated most of the bits for the upper thrust assembly now. Mine is completely separate from the prop tube. It gives me one extra gasket to leak, in theory, but makes it easier to manufacture, allows minor alignment tweaks and total replacement/rebuild in the field if necessary. I need to get a taper-lock bushing and see how accurately I can control its position on the shaft before deciding what's best but there are a couple of ways of addressing that including a threaded adjustable bushing between the bearing cup and taperlock bushing or shaft flange. I'd rather not have to, on the KISS principle.

PDW