Wednesday, February 3, 2010

How Strong is She, Really?

rev. 2-19-2010

From time to time every sailor wonders how strong a given part of their boat is. Magazines and books suggest that everything should be reinforced, because there are under-built boats out there to which bad things have happened. The PDQ 32 is among the more carefully built and conservative designs; little upgrading for strength is needed. What follows is a collection of rough calculations I have gone through; I will add to this post over time as I collect more. So far I have found that the PDQ engineers seem to have checked their details.

Note: I am a chemical engineer, not a structural engineer, so I have made only rough calculations and kept my assumptions to myself. There are always guesses regarding lay up details. But I have applied practical expereince gained on this and other more fragile boats - I have been sailing long enough to have broken and repaired many things - and I think I am reasonably close. It would have been nice to have some of this information in the manual, but then I would have nothing to puzzle through during the winter (20 inches of snow are predicted to fall tomorrow).

The davits are a simple design; two 1 1/2-inch stainless pipes cantilevered out through the transom and bolted to an inner bulkhead. I used a simple pipe cantilever beam formula from Mark's Handbook of Mechanical Engineering. Because they have a twist bent into the design and because pipe is subject to buckling failure and not simple fiber over-stress, the calculation is complex. However, assuming a 4:1 safety factor (conservative, to allow for buckling), ignoring the slight inward curvature, I came up with a safe working load (SWL) of 188 pounds. The heavy end of the tender, with motor and accessories, is about 80 pounds. If we allow for 50 pounds for water and a 140% dynamic load factor, we get a load of 182 pounds. Very close. If a sailor leans hard on the davit while entering a boat in the water, perhaps a 190-pound force is applied. I have leaned hard while digging out an incredibly heavy snow load and it felt very solid, though it certainly flexed. So, it seems PDQ engineers designed for all common use scenarios and used a conservative design factor.

What if we add a wind generator? The generator plus mounting pole add ~ 30 pounds. The downward thrust generated by the wind (as transferred through the braces) might be about 40 pounds, but this depends on the exact arraignment. Now we have a maximum load on the davit of 264 pounds. We are beyond the design limits; it will probably be fine, but it is going to flex significantly and possibly crack some day. Taking the outboard off would solve the dilemma, but I like the convenience of leaving it mounted. Choices. Adding solar panels present similar stresses - more weight but less wind load.

What if we add seats, such as those on the Gemini 105Mc? That would be about 480 pounds (people plus structure), plus a 140% dynamic load factor but supported a bit closer in, so it's equivalent to a 220-pound load addition to each end. Added to the dingy load, we get about 484 pounds, not counting falling into the seat or crowding one end, which could easily push the load right to the breaking point. Additional support is defiantly needed. Better, or at least simpler, would be to fabricate a replacement davit from 2-inch stainless tube. Strength goes up roughly with the cube of diameter, and so the SWL should be about 450 pounds. If you have a big family, perhaps 2 1/2-inch tubing is in order. Beefing up the glass in the area with a few more layers might be a good call, too, and going with a slightly larger flange.

The deck in the area of the cleats is 5/16" solid glass (I drilled a hole near the bow cleat while mounting a chain lock). They are mounted with two 3/8" bolts and large fender washers. If we assume a shear strength of 30,000 psi (conservative - 45,000 psi would be more typical of a vacuum bagged laminate), assume that a 2-inches length of laminate would need to fail to pull a cleat (generally a failed cleat pulls a large hole in the deck because of the fender washer on the nut), assume the stress is horizontal and athwartships, and that the leverage is applied 2 rope diameters from the deck, we get a failure strength of about 17,000 pounds. In fact, we may see crushing under the cleat before that, but I would be surprised if the strength was less than 10,000 pounds. There is no mooring or sea anchor load on a bridle that is going to come close. Typically a 1/2-inch line is recommended for each bridle leg, which has ~ 8,000 pound failure strength and a working load of about 1,800 pounds. For comparison, a single 5/16-inch rock climbing anchor bolt is known to hold 5,500 pounds in shear.

Winch Base
After I pulled a winch out while sailing hard on the wind in a sustained 20-knot breeze with a 130% genoa, I considered adding backing plates to all of them. After confirming that the failed winch was incorrectly owner installed in a cored deck section, I realized the factory winch installations in solid glass were fine.

The winch that pulled out had been mounted to a cored deck without the benefit of fender washers or any core replacement. The failure was shearing of the skin and crushing of the core around and under three 5/16-inch nuts. There was no water damage to the core, although there were some signs of minor leakage (perhaps the leakage all occurred during the 2 weeks we sailed with a floppy winch sealed with duct tape). The mounting failed at full load (just less than 1,000-pound line pull, based upon the gear ratio, handle length, and assumed level of one-hand effort). If we give no value to the core resistance, figure the laminate to be 1/32-inch thick (not counting gelcoat), and count 2 bolts, failure would have been expected at 2000 pounds of line pull, but given some unequal loading between bolts and fatigue, a failure at 1000 pounds after11 years of service seems very reasonable. With either an oversized solid glass backing plate, or if installed in solid glass (1/4-inch in this area), the pull-out strain will be similar to a deck cleat (more leverage, but more laminate area) or at least 10,000 pounds and 2,500 pounds SWL, which is more than enough to provide a long no-flex and no-crack service. Failure would likely be where the solid glass tapers to cored construction.

Note regarding bedding of winches. The existing bedding was silicone, it was not stiff, and probably had not failed. The winch pealed up very easily - actually, a 1000-pound line pull removed it, so I'm guessing. Still, I like butyl rubber better for mounting things that may move a bit under extreme load, I know I will remove some day, and are well secured with bolts. Examples are stanchions bases, cam-cleats, traveler and genoa tracks, windlasses, and fresh air vents.
  • Butyl bonds to the deck and never gaps. Even so, it is easily removed with mineral spirits.
  • Installation is neat. Simply trim the excess with a dullish knife a few minutes after you tighten the bolts. A bit more will squeeze out over the next hour, but I generally leave that as an extra gasket. It could be trimmed. No cleaning excess with big messy wads of paper towel, or pre-taping to limit the spread.
  • Butyl doesn't go bad in storage. Keep a roll on the boat.
  • No drying time.
  • Silicon is very difficult to re-seal. Even silicone won't stick to it.
Use the gray material, available from RV parts stores. The black material, more often used for windows, can stain porous gel coat and clothing. Do not use it below the waterline. Do not use near heater vents or where gasoiline or oil exposure are probable; it will turn into goo.

I use 3M 4200 or or even 3M 5200 for things that I won't be moving or that are mounted with screws and could use some extra bonding. But I warn you; 3M 5200 is about as easy to remove as epoxy and getting the hardware off can be ugly. They do sell a de-bonding product, which helps good deal. Trichlor works too.

Windlass Mounting
The deck several feet back from the bow is cored, and there was no evidence on my boat that core had been replaced with solid material. The line pull on the windlass and the winch are similar - 1,000 pounds. However, the windlass came with an integral flange which distributes the load over the full circumference. Additionally,  the horizontal pull on the windlass is much closer to the deck than it is on the winches. This combination of factors gives a much stronger mounting, with failure expected at nearly 10,000 pounds and with a SWL of 2,500 pounds. The winch will stall well before damage is done. What a difference a change in angle and a simple backing plate make, when compared to a sheet winch.

Attaching a Drogue
The Jordan Series Drogue advise is to attach the bridle to the transom with dedicated plates; sound general advise when the construction of a specific boat is not known, for mono-hulls where the cleat spacing is often too narrow to cause a sufficient turning moment (catamarans are wider - this is not a PDQ issue), and to provide for low chafe. Sound practice for serious storm application. Can you attach a drogue line to the winches? I would not. Although there is little chance that even a Jordon Series Drogue could pull them out of a PDQ deck, I wouldn't be surprised if the winch mechanism was damaged, and the strain would be too great to adjust the bridle under load. A tugboat hitch would take the stain of the palls, but the beam is still less.

Will the stern cleats do? Certainly. Attach the drogue to the stern cleats (a single round turn to insure load sharing), backed-up by the mid-ship cleats, allowing substantial extra line in the bridle so that it can be eased in case of chafe. By backing-up the bridle with the mid-ship cleats, the tension on the mid-ship cleat hitch will remain moderate and can be released under load; this is a common construction riggers procedure. It is also simple enough to slack one leg and transfer it to a winch for recovery.

Chafe protection should be provided through the use of anti-chafe gear where the rope moves against the deck under the stern railing. Watch out for sharp spots on the stanchion bases and deck bolts. Examine every connection point for sharp edges, loose thimbles (the above webbing can sometimes make a very serviceable thimble if threaded onto the rope before splicing and it is not prone to working loose), and movement when under load; although I haven't enjoyed a major storm off-shore, I have seen plenty of chafe failures in industrial applications. Just as it is easier to cat a line with a knife when it is under load, chafe accelerates under high load factors.

1 comment: