Stacking keel backing plates?

[wyb2]

Originally Posted by nickpitcher

Yes I could do that, however that'd take at least 12 days as I'd have to wait 24hrs for the sealant (Sika 291) to cure before torquing each bolt, one after another.

I would not wait to torque. Apply the sealant, torque the hardware, and let it cure.

I know there is this ‘create an o-ring’ theory, but I think it causes way more problems than it solves. Sealant isn’t an o-ring, it’s sealant. It works by bonding to the surfaces as it cures, it doesn’t need additional pressure. By re-torquing after cure, you just risk breaking a good bond that was doing its job.

I’ve always felt this way, but I’ve also read as much somewhere, so I’m not completely making this up. Admittedly I can’t point you to exactly where off the top of my head.

[nickpitcher]

So I've gone ahead and added an additional 5mm plate to each of the existing plates. Pretty happy with the result, looks much more heavy duty.

Thanks everyone for your help

Now to paint it...

[tronic72]

I don’t think those plates are wide enough . They are to spread the load and you’re not spreading it wide enough

At work we have had multiple wake board frame torn out of the boat for this exact reason

[stormalong]

Originally Posted by tronic72

I don’t think those plates are wide enough . They are to spread the load and you’re not spreading it wide enough

At work we have had multiple wake board frame torn out of the boat for this exact reason

Just to reiterate what was said before:

There are 12 bolts holding a keel weighing approximately 1000 pounds. Worst case, for argument sake, let's say that it is 2000 pounds.

2000 divided by 12 bolts is 167 pounds per bolt - Not much of a load for these bolts.

Now add that this is an older boat with no visual history of fiberglass deformation - see OP's picture.

Considering this information the conclusion is that what the OP has done it quite sufficient.

[tronic72]

Originally Posted by stormalong

Just to reiterate what was said before:

There are 12 bolts holding a keel weighing approximately 1000 pounds. Worst case, for argument sake, let's say that it is 2000 pounds.

2000 divided by 12 bolts is 167 pounds per bolt - Not much of a load for these bolts.

Now add that this is an older boat with no visual history of fiberglass deformation - see OP's picture.

Considering this information the conclusion is that what the OP has done it quite sufficient.

Your arrogant single minded reply is the reason many people don’t like forums.

[stormalong]

Originally Posted by tronic72

Your arrogant single minded reply is the reason many people don’t like forums.

Arrogant??? Just because you don't like my analytical response?

I try to help people on the forum with helpful responses, not criticize them with unfounded remarks.

So who is the arrogant one?

[wyb2]

Originally Posted by tronic72

I don’t think those plates are wide enough . They are to spread the load and you’re not spreading it wide enough

At work we have had multiple wake board frame torn out of the boat for this exact reason

How does one determine wide enough? The OP suggested, though I don’t think specifically stated, that the plates were about the same size as the originals from the builder.

The deck of a wakeboard boat is built much differently than the keel stub of a sailboat.

[Nekton73]

Practical Sailor did some testing related to this issue:


https://www.practical-sailor.com/boa...-failure-modes




I'm guesstimating that the backing plates in the picture are 2" squares, so each is covering 4 square inches minus the area taken up by the 3/4" bolt (.44") = 4 - 0.44 = 3.66 square inches of area covered by the plates.


167lbs per bolt distributed over 3.66 square inches = 45.6 lbs per inch of gravitational load on the fibreglass of the keel stub at each bolt.



Assuming the bolts were torqued to 130 ft-lbs which is spec for a 316 Stainless 3/4" - 10 pitch bolt, the compressive force exerted by the nut and bolt is roughly 10 400 lbs force, or approximately 25% of the average minimum tensile strength of said bolt ~ 42 000 lbs.

10 400 divided over 3.66 square inches = 2841.5 lbs per square inch of compressive force on the fibreglass of the keel stub before any shear force is exerted during sailing, or by running aground.



Going back to the Practical Sailor article, they suggest backing plates should be at least 5x the bolt diameter (3/4") which in this case would be 5 x 0.75" = 3.75 square inches. So PS recommendation suggests the backing plates used here are marginally too small.



They also recommend the thickness of the laminate be 1.25 to 1.5x the bolt diameter wrt to potential shear force exerted by the bolt under extreme circumstance. 1.5 x 0.75" = 1.125" Without knowing how thick the keel stub is on the Cavalier 26 this relationship remains an unknown. Given the overkill of early fiberglass layups I wouldn't be surprised if it was at least 1" though.



Despite the unknown of the keel stub thickness, and the marginally too small backing plate, if this were my boat, and there is no known history of failures of this design, and I had no reason to believe there was any damage to the hull, keel, stub or bolts, and I had replaced the old backing plates with ones the same size as the originals and dry torqued to the correct value, I would feel confident that the designers and builders did their job and the keel attachment is safe... and keep sailing.

[RaymondR]

Best results would probably be achieved by using larger and thicker plates with the existing plates placed on top, or just use bigger and thicker plates.

[wyb2]

Originally Posted by Nekton73

…
Assuming the bolts were torqued to 130 ft-lbs which is spec for a 316 Stainless 3/4" - 10 pitch bolt, the compressive force exerted by the nut and bolt is roughly 10 400 lbs force, or approximately 25% of the average minimum tensile strength of said bolt ~ 42 000 lbs.
….

Agree 100% with your thought process and conclusion, but I’d take 1 small exception…

(Fair warning, getting waaay into the weeds here)

Ahhh, torque specs. I would go for a lower torque value for a few reasons:

1) 42,000 lbs tensile strength backs out to ~135,000 psi material ultimate strength (UTS). I’ve seen a pretty wide range for 316 SS, but this is right on the upper end of that range, I wouldn’t rely on it. Additionally, it’s generally a good idea to design to yield strength (YTS) for things that need to stay in service for a long time. For 316, that can be as low as 30,000 psi, which would give a tensile strength of ~9,200 lbs for a 3/4-10. So I wouldn’t want to shoot for a pretension of 10,400, closer to about half that. 50%-60% of YTS is a general rule of thumb, so around 5,000 lbs. This would logically translate to about half the torque, so around 65 ft-lbs.

This might seem low for such a large bolt, but:

2) Most of us are used to working on machines like cars, trucks, marine diesels, etc. These things often use alloy steel (grade 5 or 8) hardware with a higher YTS that is also closer to its UTS. Grade 8 has a YTS of 120,000 psi vs a UTS of 150,000. 316 YTS can be as low as 30,000 psi, vs a UTS of around 80,000, more than double.

3) The K-factor we are using (0.2) assumes dry threads. Any bedding compound in the threads will act as a lubricant. I’ve seen conflicting guidance how much this affects the torque/preload relationship, but it’s definitely not going to reduce the preload.

4) And this is a biggie: fiberglass is not steel. Most hardware torque specs will assume the material being clamped has a similar modulus as the bolt. When the clamped material modulus is significantly lower than the bolt, this changes something important. If they are the same, as additional load is applied in tension, the bolt stretches at the same rate as the clamped material expands, so the tensile stress is basically unchanged until the additional load exceeds the preload. When the materials are different, this does not hold, and basically any additional load should assumed to be added to the preload stress.

[RaymondR]

Excellent analysis and the comments about thread lubricity and it effect upon bolt stress suggest that the angular make up method of achieving proper bolt tension might be more appropriate in this instance.

[Nekton73]

wyb2 - some great additional info and considerations. I will keep this in mind next time I am assisting with a keel re-bed and torque, or when it comes time to do my own.


The weeds are where things get interesting.


RaymondR - what are you referring to when you say the "angular make up method of achieving proper bolt tension"? Googling did not help me, although I did come across this: https://www.boltscience.com/pages/tighten.htm

[RaymondR]

Originally Posted by Nekton73

RaymondR - what are you referring to when you say the "angular make up method of achieving proper bolt tension"? Googling did not help me, although I did come across this: https://www.boltscience.com/pages/tighten.htm

You snug the nut down then tighten it further through a calculated angle of arc to achieve the required preload. It is a more certain method where there is variations of the friction factor of the mated surfaces.

[Lee Jerry]

Originally Posted by wyb2

Agree 100% with your thought process and conclusion, but I’d take 1 small exception…

(Fair warning, getting waaay into the weeds here)

Ahhh, torque specs. I would go for a lower torque value for a few reasons:

1) 42,000 lbs tensile strength backs out to ~135,000 psi material ultimate strength (UTS). I’ve seen a pretty wide range for 316 SS, but this is right on the upper end of that range, I wouldn’t rely on it. Additionally, it’s generally a good idea to design to yield strength (YTS) for things that need to stay in service for a long time. For 316, that can be as low as 30,000 psi, which would give a tensile strength of ~9,200 lbs for a 3/4-10. So I wouldn’t want to shoot for a pretension of 10,400, closer to about half that. 50%-60% of YTS is a general rule of thumb, so around 5,000 lbs. This would logically translate to about half the torque, so around 65 ft-lbs.

This might seem low for such a large bolt, but:

2) Most of us are used to working on machines like cars, trucks, marine diesels, etc. These things often use alloy steel (grade 5 or 8) hardware with a higher YTS that is also closer to its UTS. Grade 8 has a YTS of 120,000 psi vs a UTS of 150,000. 316 YTS can be as low as 30,000 psi, vs a UTS of around 80,000, more than double.

3) The K-factor we are using (0.2) assumes dry threads. Any bedding compound in the threads will act as a lubricant. I’ve seen conflicting guidance how much this affects the torque/preload relationship, but it’s definitely not going to reduce the preload.

4) And this is a biggie: fiberglass is not steel. Most hardware torque specs will assume the material being clamped has a similar modulus as the bolt. When the clamped material modulus is significantly lower than the bolt, this changes something important. If they are the same, as additional load is applied in tension, the bolt stretches at the same rate as the clamped material expands, so the tensile stress is basically unchanged until the additional load exceeds the preload. When the materials are different, this does not hold, and basically any additional load should assumed to be added to the preload stress.

I'm afraid I have to disagree. My $0.015:

1) Are you saying to reduce the torque (preload) from the standard? As you say, the basis should be yield strength. But the bolts should be torqued the standard amount according to the type/size. This is because the purpose of the preload is to clamp the parts together, and you want them to stay there (i.e., touching) or bad sh1t starts to happen. There is nothing to be gained by reducing the preload; it only weakens the joint. (I did not run through the numbers to get what the load in the bolt is, but use the spec torque/preload.)

2) Again, yield is the working metric. Ultimate strength doesn't really matter. Until...

after the joint has failed (yielded), where margin between yield and ultimate is a good thing. For example, if you hit ground hard with the keel, the forward bolts may yield without breaking - keeping the keel attached, even if wobbly, is better than just falling off immediately - but probably don't pick your material based on that.

3) Lubrication doesn't change the torque or preload. As you apply torque to the bolt/nut, that force is split between loading the bolt (stretching) and the friction between the threads. The former is the goal, the latter is a byproduct (which also increases with the load). By adding lubrication, the friction part is reduced and let's you have a better accuracy or correlation between torque and preload. The amount of preload desired in the bolt is unchanged, the amount of torque needed to get it is reduced.

4) I don't think this applies to FRP/steel; the moduluses (moduli?) of the two aren't 'significantly lower.' The compressive load in the clamped material is the same (magnitude) as the preload in the bolt. This is by definition, it is how the preload gets into the bolt. And it occurs even though they have different moduli. (That is the purpose of the large washers (or backing plates) - to enlist more area of the clamped material to compensate for the lower modulus.) It basically occurs the same in reverse.

"...as additional load is applied in tension, the bolt stretches at the same rate as the clamped material expands, so the tensile stress is basically unchanged until the additional load exceeds the preload."
The clamped material will expand with the bolt such that the bolt tension (preload) equals the sum of the compression in the clamped load and the additional load. The load should never exceed the preload, and therefore the tension should never exceed yield.

[wyb2]

Appreciate the thoughtful response

Originally Posted by Lee Jerry

I'm afraid I have to disagree. My $0.015:

1) Are you saying to reduce the torque (preload) from the standard?

What standard?

All torque standards are based an some assumptions about the joint, like base and bolt material. If you were using a nylon bolt because you needed electrical isolation, you wouldn’t just look up the torque spec in some industrial design handbook that assumes steel hardware.

Originally Posted by Lee Jerry

As you say, the basis should be yield strength. But the bolts should be torqued the standard amount according to the type/size. This is because the purpose of the preload is to clamp the parts together, and you want them to stay there (i.e., touching) or bad sh1t starts to happen. There is nothing to be gained by reducing the preload; it only weakens the joint. (I did not run through the numbers to get what the load in the bolt is, but use the spec torque/preload.)

2) Again, yield is the working metric. Ultimate strength doesn't really matter. Until...

after the joint has failed (yielded), where margin between yield and ultimate is a good thing. For example, if you hit ground hard with the keel, the forward bolts may yield without breaking - keeping the keel attached, even if wobbly, is better than just falling off immediately - but probably don't pick your material based on that.

3) Lubrication doesn't change the torque or preload.

It changes the relationship between the two, as you basically go on to explain.

Originally Posted by Lee Jerry

As you apply torque to the bolt/nut, that force is split between loading the bolt (stretching) and the friction between the threads. The former is the goal, the latter is a byproduct (which also increases with the load). By adding lubrication, the friction part is reduced and let's you have a better accuracy or correlation between torque and preload. The amount of preload desired in the bolt is unchanged, the amount of torque needed to get it is reduced.

4) I don't think this applies to FRP/steel; the moduluses (moduli?) of the two aren't 'significantly lower.'

The modulus of FRP is significantly lower than the modulus of steel. FRP is around 17Gpa or 2.5 Mpsi (megapounds per square inch aka million psi). Steel is around 10x that.

Originally Posted by Lee Jerry

The compressive load in the clamped material is the same (magnitude) as the preload in the bolt. This is by definition, it is how the preload gets into the bolt.

Yes. Load and stress and modulus are all different things though.

Originally Posted by Lee Jerry

And it occurs even though they have different moduli. (That is the purpose of the large washers (or backing plates) - to enlist more area of the clamped material to compensate for the lower modulus.) It basically occurs the same in reverse.

Not exactly. The point of the backing plates is to enlist more area of clamped material to compensate for its lower strength, not lower modulus. Strength is how much stress (psi) the material can take without failure. Modulus is a measure of how much the material compresses/deforms under stress. Modulus is basically a measure of softness/hardness that tells you nothing about strength.

Originally Posted by Lee Jerry

"...as additional load is applied in tension, the bolt stretches at the same rate as the clamped material expands, so the tensile stress is basically unchanged until the additional load exceeds the preload."
The clamped material will expand with the bolt such that the bolt tension (preload) equals the sum of the compression in the clamped load and the additional load. The load should never exceed the preload, and therefore the tension should never exceed yield.

My point is basically that those last two sentences operate on the assumption that the bolt and clamped material have a similar modulus. With metal hardware on metal components, this is generally true.

But if the base material is much softer, so relieving stress requires much more expansion, then loads become additive.

For steel bolting steel, if you have 5000 lbs of preload, and you apply 2000 lbs of load in tension, you inherently relieve 2000 lbs of clamping force, so your resulting bolt load is 5000 + 2000 - 2000 = 5000 lbs. I think this is basically what you are saying, and it’s correct

For steel bolting FRP, if you have 5000 lbs of tension, and you apply 2000 lbs of tension load, you may only relieve 200 lbs of clamping force, because the FRP needs more room to expand to relieve stress. In this case your resulting bolt load is 5000 + 2000 - 200 = 6800 lbs. The preload gets exceeded without the parts ever getting unclamped.

My approach is conservative. I’m ignoring a mitigating factor of the FRP relaxing over time (creep), which I think does happen. I’m using a conservative value for SS yield strength. I’m sure tons of people have just looked up 130 ft lbs (or whatever for their bolt size) and never had a problem. But I think it’s more because there is plenty of room for error when you are bolting on a X,000 lbs keel with hardware that could lift XXX,000 lbs, and less because it’s the ‘correct’ torque value.