Like most owners with a new-to-us boat that is far from new, I spent the first couple of winters focusing on stuff that just had to be fixed.
This winter I have had the luxury of moving on to things that need doing, but are not immediately obvious. In this case torquing the keel bolts.
Test
Out of interest, before removing the nuts, cleaning them up, lubricating (or not1) and then torquing to specification, I immediately torqued each nut with the wrench set to one third of specification.
Hi John,
As you know, I have a catamaran and I’m a multihull fanatic, so you’d expect me to brag about how cats have no keel bolts, but nope. I’m gonna pass on that, since I’m actually in the process of taking ownership of a monohull. Did I loose my conviction? Again nope, but it was a deal too good to pass on. A Nauticat 40 from -91 on an auction. Others got scared by the words “may have some osmosis” in the description. It has 3, three, tiny blisters! 🙂 I’ll fix them and protect the rest of the bottom of course, but people get scared easily…
This one has an external lead keel with either stainless or bronze bolts. Not verified yet. It’s a Sparkman and Stephens designed motorsailer, built in Finland with Swan specs and a fin keel that is very strong, so it’s not a weak point, but I will check the bolts, of course.
I haven’t searched yet, but I think I remember that there are chapters here on how exactly to work on keel bolt checking and tensioning, or reseating the keel, if needed. I’ll look for it, to figure out if the bolts should be greased before torquing or not, but since this tip came up right now, it felt right to confess my half boat sins. 😀
Hi Stein,
Congratulations on the new boat. Hope she works out well. See further reading above for links to torquing keel bolts.
Timely as I am waiting for our 6 keel nuts to be torqued on our J Boat. This after having had to remove the keel and reseal it due to leaks we found after being offshore. That’s after we had the torque checked a year ago (they obviously got it wrong or their wrench got it wrong, we may never know). Lessons learned – ask questions about that wrench!
This time, a different yard is having Caterpillar come in to torque the nuts (450 ft/lb), to have a well-maintained and certified wrench to get it right this time. That amount of torque may even need a torque multiplier to get it right.
I do not see any mention of torquing keel bolts when the vessel in on the hard; surely this ought to be a prerequisite to remove gravity from the consideration
Hi Peter,
Surprisingly, it does not make any difference, see this comment thread: https://www.morganscloud.com/2020/06/17/planning-and-budgeting-a-refit-keels-part-3-torquing-keel-bolts/#comment-313797
Thanks John, useful comments in this thread. I recall many years ago having had a boatwright insist that checking/tightening keel bolts whilst on the hard was preferable to whilst floating and I saw the logic in that. I suspect that if there was a situation where there were multiple loose bolts that the argument for “on the hard” checking might be valid though I accept of course that any tightening would certainly be better than none.
Hi Peter,
Glad that was useful. That said, even though it does not make a difference to tension, I would always wait until hauled before torquing the keel bolts because there is always the risk of something going wrong like breaking a bolt, at least on older boats where the bolts may have been weakened by corrosion.
To put some very approximate numbers to it:
Imagine you have a 4000 kg ballast keel attached by eight M24 class 8.8 studs. (Three lateral pairs of two plus singles at the leading & trailing edges.)
Each stud can withstand an ultimate tensile load of 293 kN, seven times the weight of the keel. Individually. Collectively, the studs can support 239 tonnes in pure vertical tension – sixty times the dead weight of the keel.
Torquing each of those bolts to 282 N.m, assuming a reasonable 40% lubrication factor for the anti-corrosion grease, produces a preload of 98 kN per bolt.
Since sailors find tonnes intuitive but don’t usually think in kilonewtons: that’s ten tonnes of tension per bolt. Or eighty tons of total preload for the set, against four tonnes of keel mass.
So, with the boat on the hard, the bolts start at zero tension and you torque them so the bolt tension is ten tonnes each, all of which is preload.
With the boat afloat and the nuts finger tight, we’re starting at (4 tonnes / 8 bolts) = 0.5 tonnes each from the weight of the keel. You torque them to the same 282 N.m to get the same 10 tonnes of tension, which is now made up of 9.5 tonnes of preload and 0.5 tonnes of supported weight.
There’s not a consumer- or mechanic-grade torque wrench in the world that is accurate to 5%; you need specialized hydraulic tensioners to achieve that. The difference between torquing the nuts on the hard and torquing them afloat is less than the variability in the calibration of your tool.
But….. if you do this in the water, what do you do if something goes wrong? Pray that the Travelift is free?
These bolts are specified for the dynamic loads when the boat is heeled over on its beam ends and being thrown off a wave. The static loads of the keel just hanging there are nearly negligible in comparison. But when you put a big fat torque multiplier wrench on one, you *can* produce tensions similar to the bolts’ design limits, and that can be dangerous if they are defective in some way.
Hi Matt,
Now, that’s useful and directly applicable information. Thanks!
Hi Matt,
Great information. I knew the difference between bolt tension and keel weight was high, but I did not know it was that high! Thanks.
It sounds strange, but it has to be that way.
Say your 4-tonne keel has a centre of mass exactly 1 m below the joint. Consider the case where the boat is heeled over 90° to port. That’s a 39 kN.m moment on the joint. The port-side line of contact between keel and hull is in compression and the port three keel bolts are doing nothing; the starboard three keel bolts are in tension. They are perhaps 20 cm apart laterally – so each of the three starboard keel bolts is taking 65 kN.
The keel’s only four tonnes, but lay the boat on its side and the three (of eight) bolts that are active in this situation need to exert a combined 20 tonnes of tension just to hold everything stationary.
And that’s before allowing for dynamic motions. The boat is just sitting there in a dead calm sea, heeled 90°.
This is how you end up with preload and bolt strength figures that look insane at first glance. The preload has to be high enough to ensure zero movement of the joint under any realistic load case.
Interesting subject. I did mine last year after removing them. I didn’t have access to a torque wrench that went high enough. They needed to be 510nm – assuming ungreased threads. I used a meter long bar and used all my strength, pushing with my body weight and my legs off the insides of the boat, like a rowing machine. I did them in a pattern like cylinder head bolts and repeated the pattern several times until they would turn no more. I’d love to know roughly what tension they are at. I don’t have any experience of what 510nm should feel like at the end of a 1m bar. Do you know Matt?
Newtons times metres. 510 N.m is a one metre long wrench with 510 N, or 115 lb, of force pushing on the end of it. Or a two metre wrench with 255 N / 57 lb of force. That’s really all there is to it, and the fancy “torque wrench” is nothing more than a clever mechanism to directly measure that force.
If you want to play around with real numbers, try https://www.engineersedge.com/calculators/torque_calc.htm
To get the nut torque value I used, I actually did the moment calculation at keel horizontal that Matt describes and then convert it to bolt tension. Rather than just using this tension, I doubled the value which is a relatively standard margin to use in industry. I also checked the calculations to make sure that I wouldn’t crush the FRP laminate under the washer plates at the proposed tension value.
Things I didn’t calculate include thread strength, drive strength or pullout force of the studs. The threads on a full size nut of the same material are full strength. And those nuts can also take full drive torque. On the studs, J bolts can pull up in lead but they should be designed with a large margin not to and there is no way to know the geometry required for calculating.
Some of the values people throw around in the owners group of our boat can be very scary. Most people suggest the value associated with ~75% of shank yield which while generally appropriate, ignores that keel bolts are oversized for corrosion. I have also seen people use torque tables for grade 8 and higher. Overtorquing can range from a nuisance (breaking a stud) to a danger (damaging the laminate making pulling the keel bolts through much more likely). When we bought our boat, I actually asked if the keel bolts had ever been retorqued and then asked the value which thankfully was reasonable, I would have been doing a lot of inspection if not. I definitely would not let a yard retorque without agreeing on a torque and writing it down. In some of the keel losses where you can see the FRP broken right around the edge of a washer and the boat had undergone significant repair before, I have wondered if overtorquing was a contributing factor.
Eric
Thanks Eric and Matt! 115lb isn’t a great deal of force so I’m pretty confident I haven’t under tightened them which makes me feel better. I’ve probably over tightened them if anything which is, I suspect, the lesser of two evils. The bolts go a long way down into the lead according to the drawings. I don’t know if they’re tapped in or cast in, but either way I’d say it’s virtually impossible to draw them out, I would imagine the threads on the nuts would strip before that happened. In regards to compressing the GRP there’s a large steel beam between the nuts and the hull so I’m not too concerned about that. Anyway thanks both for the additional insight.
may want to verify the 450 lb figure. That’s A LOT. Sabre spec is 90 lbs.
Hi David,
I’m guessing the Sabre is using a lot of smaller bolts to have a torque that low? Whereas J/Boat seem to use fewer larger bolts, so 450 might be right. Anyway, I agree, and have written an entire article (link in tip) on the importance of carefully checking torque with a calculator first.
Very interesting firsthand update to your already excellent chapters on keel bolts, John! Thank you for sharing all your commonsense experience accumulated over your many years of boating!
If your findings and consultations with still more experienced gurus come close to reality, loose keel nuts would be common and not exceptional, which seems alarming. No wonder you continue reediting on the subject.
Initially I came to believe keel bolt loosening to be caused by some sort of malpractice, like: grounding or inadequate resistance of the keel bed to compression by the bolts keel (e.g. construction too weak, soft polymers instead of epoxy) or not retorqueing after final curing of the bed or primary inadequate torqueing, etc.
However, bolts seem to be found loose more often, than one would believe malpractice to occur. And checking bolts regularly, even after having done so oneself, suggests spontaneous loosening. Why would that happen?
Would you consider elongation of bolts with loss of their “spring” function so well explained in your texts?
Maybe true loosening (turning) of nuts over time due to the repetitively changing axial and especially transverse forces, exerted by the keel lever? (similar to vibration induced loosening?)
Considering these latter 2 thoughts, undoubtedly the applicable bolt preload on each single bolt would be paramount for whether and when keel nuts come loose. So it breaks down to size and material of bolts. Would then not 4 really heavy bolts seem safer than 10 half as large bolts? I mean: simply multiplying the preload by the no. of bolts pretends an even distribution of the expected keel forces over all bolts.
However, boats come with what the builders put into them, and your site teaches us to question them always. Any rules of the thumb about the correlation between each single bolt/nut preload versus the nut-loosening-burden of a keel (e.g. bolt preload capacity > X * keel weight) ?
Additionally, what are the knowledgeable opinions on attainable securing methods for the nuts:
blob:https://www.morganscloud.com/5aee87b8-83fb-4683-86f9-4a7631bdd88eWiring, like in aerospace construction, of course seems the icing on the cake, but I have never seen it on keel bolts, their nuts lacking the necessary holes. Neither cotter pins/splints or castle nuts.
The common locking methods (serrated lock washers, splint lock washers, nylon lock nuts) have their failure rates, which, depending on the testing models, are alarming.
The very sophisticated and expensive Nord-Lock® wedge washers promise more. However, I have not found comments regarding their use on keel bolts. https://www.nord-lock.com/nord-lock/products/washers/
Over-glassing the bolts/nuts with glass fiber is a practice. But how to check torque and retorque then? Apart from the worries about concealed erosion with oxygen depleted saltwater penetrating under the fiber glass in a wet bilge.
Lubricating the threads for a better gauging of bolt tension by ways of torque would seem to exclude the use of Loctite® and similar. Torque checking on cured loctite®ned nuts would not work. Retorquing would mean removing completely + cleaning + applying anew, without any info on whether a loosening has happened or not. Personally, I tend to doubt about the durability of Loctite® over years on a keel bolt/nut in a humid, dirty bilge. I may be over skeptical there.
I wonder if these are thoughts shared by others and would greatly appreciate comments by the knowledgeable members.
I don’t think the problem is bolts permanently stretching (this doesn’t happen at all unless they are severely overloaded) or nuts loosening. The problem is that you’re making a bolted joint out of *fibreglass-reinforced plastic*. Emphasis on PLASTIC. Fibreglass laminate is highly anisotropic; in the direction it’s being squeezed by the bolts, its conpressive strength comes entirely from the plastic resin, and the glass fibres do almost nothing. If the resin in the laminate or the resin in the gap filler creeps even 1% over a few years… goodbye pre-load tension, even though the nut hasn’t moved at all.
Lock washers can help with this, but the best solution is to properly engineer the joint in the first place.
Loctite has its place but I am not convinced that this is such a place.
Encapsulating keel nuts in fibreglass is unbelievably stupid.
Hi Alberto and Matt,
I agree with Matt that the most likely cause of keel bolts being loose is simply creep of the FRP between the metal keel and the washers/nuts. For those not familiar with creep, it refers to a permenant dimensional change of a part that is under stress over a long period of time and the amount is proportional to the stress. Plastics including resin creeps but metals do not.
Another contributing factor here is that the manufacturers do not do a good job of making the FRP perfectly flat and parallel. What this means is that there are spots with much higher stress and this causes creep to happen faster than a basic calculation would show.
One of the issues with running real large diameter bolts that are not very long is that very small displacement changes have huge tension changes as the bolts are a very stiff spring. In fact, if you look at high vibration applications, the designers actually intentionally put a significant distance between the head of the bolt and the first thread engaged to lower the spring rate. Many companies include this in their internal design standards.
In general, if a joint is well designed, there is no need for anything extra to lock a nut from coming off and that includes greased threads. This means that the preload must be significantly higher than any load ever on the joint and that the joint can’t be shifting around if you are counting on friction to keep things moving in the plane perpendicular to the bolt axis. There are some applications where this is tricky to achieve and then you get loctite, safety wire, nylocks, etc. If you put a torque wrench on the keel boats every few years, I don’t see a need for any of these.
Eric
Eric, John,
This is giving me loose keel anxiety. It’s not your fault. If I did not own a boat, I would have no anxiety… But, let’s be honest, that’s no way to live!
We’ve wondered about tightening our keel bolts but left everything nearly untouched in the end. They are there under numerous layers of paint that show no cracks. We tried tightening one, but that thing was not close to budging. That was a year and a half ago.
Our boat is a Columbia 56, 1977 that hit the water in 1980. It’s a one off, off series, so technically the last hull, no°19. It’s a semi full keel with skeg hug rudder. At some point in the boat’s history, the keel was encapsulated in thick fiberglass to avoid lead painting. It’s not like it can free fall. We thought that thing cannot move, like at all. Can it? I’m now wondering if our reasoning is flawed in any way. Could the vibration of the motor cause some loosening over time? Should we tightened it in order to avoid creep in other materials? like the encapsulation? In my mind, the load of the keel is on the bolts but also supported or laying into the thick encapsulation. I don’t think it can move at all… I’ll sleep very well tonight. But I’ll wonder about our keel bolts before bed, it will be nagging at me.
I’ll go do some more reading as there seems to be a wealth of info I haven’t read yet on bolts and keel.
Best,
Marie
Hi Marie,
This is a a hard one since there are way too many variables in play here for me, or anyone else, to give you a solid recommendation:
This is why my general recommendation is that keels on boats this old should be removed for inspection before venture offshore. https://www.morganscloud.com/2020/06/05/planning-and-budgeting-a-refit-keels-part-2-non-destructive-testing-of-bolts/
https://www.morganscloud.com/2020/07/28/planning-and-budgeting-a-refit-keels-part-4-keel-removal-and-inspection/
Ditto rudder and chain plates. https://www.morganscloud.com/2020/03/08/planning-and-budgeting-a-refit-rudders-part-1-the-problem-defined/
At the very least, before continuing your circumnavigation I would recommend torquing the keel bolts to spec, here’s how: https://www.morganscloud.com/2020/06/17/planning-and-budgeting-a-refit-keels-part-3-torquing-keel-bolts/
The good news is that if anything bad happens, say a broken bolt, you could not be in a better place to get that fixed than NZ.
I would not rely on the glassing holding the keel on unless you know it was done to a very high standard and with epoxy resin to avoid secondary bonding problems, and even then I would be doubtful unless a real engineer was involved in specifying the laminate for that purpose. Also, I would very much doubt that the encapsulation is stiff enough to prevent movement if the keel is indeed loose.
To end with some good news, my guess is that the bolts are bronze and if so the chances of one breaking because of corrosion is much lower than it would be for stainless or galvanized steel.
Hi Marie,
Not much to add to what John said. If there are no cracks whatsoever in the paint, that is a good start as it means that both the keel and the nuts have moved significantly. If it were my boat, I would still do as John suggests.
By the way, to retorque you must remove the nut, clean everything then retorque. If you just put a wrench on and see if you can tighten more, you don’t learn that much as the relationship between torque and tension is not well known but will result in a lower than desired tension.
On the glassing, even if it is quite a thick laminate and done with epoxy, the geometry can make it not very structural. On boats where there is no keel sump and the keel is basically just attached to the bottom of a round hull, any glass added takes a very sharp turn onto the keel and this has very little strength. You need a big radius in there to have strength. On a boat with a significant sump, it is definitely possible to make it structural but it must be done right.
Eric
Hi Eric and Marie,
Thanks for the fill on that. On the subject of keel glassing. A piece of good news, if memory serves those old Columbias have a quite large keel sump. Here’s a diagram I found: http://www.columbia-yachts.com/layout_of_yacht.jpg Hard to see from the diagram but, if my aging memory is serving, there is quite a gentle curve in the sump on those boats. A friend had one hauled right next to us for some years.
On the other hand I would be remiss if I did not say that over all the years I have watched repairs in boat yards, glassing over something like that tended to be an effort to hide a problem or alternatively avoid doing the right thing, therefore I’m totally with Eric on the need to torque the bolts, I would not rely on the glassing unless you know for absolute certain it was done right including being specified by an engineer or naval architect who could do the calculations right.
WOW, Eric and John, thank you for your input. I’m always impressed at the knowledge you’ve cumulated. Indeed, Seamer has quite the keel sump. We torqued and nothing budged, not even a smidge. So, that’s a good news. I did not know this, but my husband had also done it after our Pacific crossing in Tahiti. Still, nothing at all then, The first owner of this boat had this boat commissioned off series by Columbia to sail across the world with his kids. The hull is extra thick. I would not be surprised that any fibreglassing would have been done by Columbia as he lived in Costa Mesa by the yard. We will never know the full story. He sold it when he was 89, and not to us, so we will never know the full story. But we are okay with that part!
Hi Marie,
Great to hear you torqued them and all good. Given you have done it a couple of times and sailed many miles in-between the two. I would guess that the whole joint is pretty stable. Nice contributor to a good nights sleep.
Indeed, if only it were always that simple!
“Plastics including resin creeps but metals do not.”
Lead undergoes significant creep. A lead wire loaded by a spring makes a useful and robust explosives delay timer.
Hi William,
We were discussing bolted joints here where the metals (bolts) are specified so their yield strength is not exceeded. (Matt already coved that.) As I understand it, lead has a very low yield strength, particularly if pure. And Eric did mention (in another comment) a possible issue with J Bolts moving in a lead keel, but as he points out there is no way to quantify that, so we have to trust the original engineering. It’s also worth remembering that lead keels generally have antimony added to make them harder, and I suspect, that so doing would up the yield strength. Point being that to be fair to Eric his statement is, as I understand it, correct in the context of this discussion.
Hi John and William,
Yes, my wording was inexact, I should have said most metals. Technically, even that would need to be clarified to apply at lower temperatures where we only need to worry about things like lead and zinc. When I did steam expanders, creep was something we had to be aware of in all our parts because of the high temperatures which can make things like steel creep noticeably.
Unfortunately I don’t know what the creep numbers are for the typical lead alloy used in keels. If anyone knows it would certainly be interesting. I suppose it is possible that we are seeing a small amount of creep in keels that leads to losing bolt preload but I suspect the bigger contributor is the laminate.
One other source I failed to mention is that too many boats have plywood or some other coring material in the interface which fails.
Eric
https://vulcangms.com/material-creep-mechanical-properties-lead-creep/ gives creep data for pure lead as does Marks’ Handbook (index under creep).
There is some creep data for lead alloys in the blue pamphlet under home/resources/library of the vulcangms website, but I have difficulty in interpreting it.
Riddington and Sahota, “Mechanical Properties of Lead Alloys in Compression”, Journal of Materials in Civil Engineering, July/August 2003 measured the initial and secondary creep of lead and lead alloys in compression at several temperatures. The initial (and short lived) creep period can be ignored, and these results are for the secondary creep period for pure lead and an antimony-lead alloy at one temperature.
Material Stress 20°C Strain Rate
(N/mm²) (sˉ¹)
9.99% 2.5 2.62E-9
lead 5.0 3.53E-8
7.5 6.34E-8
10.0 9.23E-8
12.5 1.43E-7
1.2% 1.0 1.36E-9
antimony 2.5 1.86E-8
lead 4.0 4.90E-8
5.0 8.51E-8
7.5 3.83E-7
The higher strain rate for the antimony alloy might be explained by its lower melting point; 20°C is a higher fraction of the Kelvin melting point of alloy than of the pure lead.
A bolt force of 100 kN over a 200mm by 500mm surface would be 1.0 kN/mm². That stress produced a measured strain rate of 1.36E-9/sec in the antimony alloy or 0.041/yr. Of course, the actual strain rate would fall off rapidly as the lead deformed and the bolt unloaded. Knowledge of the actual keel joint construction would be needed to determine the rate of bolt loading, but lead creep could be a significant contributor to the need to periodically re-torque keel bolts.
Hi William,
A lot of that is past my pay grade, but I think I get the gist of it. It would be disturbing if it meant that J bolts in lead keels were slowly pulling out under the recommended torque and would just keep moving after each tightening. Matt, Eric, and thoughts?
Lead is just different from the other metals we use in boats (iron, aluminum, bronze, …); it has significant creep. While in the short term lead has strength, in the longer term is flows without stopping. The antimony added to lead to increase its Young’s modulus makes it harder and more durable in the short term but also lowers its melting point and increases its creep in the longer term.
The problem with maintaining a high keel bolt load would seem to be worst in deep tip weighted (long heavy cantilever), thin cross section (high stress), external (all load on the keel bolts) keel structures.
Hi William and John,
If I remember right, the alloy used in sailboat keels is generally around 3-4% antimony. As someone who is not a material scientist, one thing I have learned is that different alloys do not necessarily behave the way I would expect and so I either need to talk with a materials scientist or see good test data on the exact alloy.
I don’t think I have ever actually spec’ed lead for anything professionally. One of the reasons is its weird properties. We looked hard at it for counterweighting crankshafts at one point but simply couldn’t figure out a reasonable way to keep it in place, even if we stayed below the yield which was hard enough as it would creep and sooner or later cause real issues. Creep is one of the frustrating material properties I run into professionally, we spend a lot of analysis time with certain materials trying to figure out what to do with it.
On numbers, 100kN seems reasonable for a 1-1/4″ 316 keel bolt as might be found in some of the bigger cruising boats. 200X500mm would be reasonable for the stress to be applied across a little ways away from the J. Local to the J, it will probably be more like 30X200mm or something but there is also the bond along the straight section to take into account. One thing, if you take 100kN/(200mm*500mm) that equals 0.001kN/mm^2 or 1N/mm^2 but the strain rate is still right. This is going to depend a lot on the actual geometry and the actual materials used.
Bottom line to me is that there is certainly a potential for creep in the lead to be a factor but I don’t know enough about the alloy and it is going to be very dependent on the design for each boat. For us boat owners, occasional torquing of keel bolts seems like the right course of action. At least on our boat which is admittedly heavily built, the tension in the bolts does not drop rapidly.
Eric
Hi Eric,
Thanks for the fill and check on that. When I get a moment I might give my contact at Mars Metals a call and see what he says. It will also be interesting to see how much the tension on our J/109 bolts has dropped when I check them again next winter. They are 1-1/4″ into lead.
Hi William and Eric,
I got interested and called my contact at Mars Metals. He said that while theoretically possible, they had never seen a problem with J bolts creeping.
He shared that might be because they don’t just cast the bolts in singularly. What they do is build a framework tying all the bolts together into one structure, and further to that in most cases the bolts themselves are made of threaded rod so the lead fills those threads adding a lot of resistance.
His thinking is that if bolts are loose it’s most likely because laminate has compressed, not the bolts moved in the lead. He also pointed out that as laminates age there is micro-cracking which might result in more compression on old boats.
He did agree that it made sense to re-torque bolts on a new boat after say three months of sailing has made everything settle in, and that after that he would re-torque if there were any signs of movement.
Hi John,
Thanks for asking them. I had forgotten how they do their internal structure and I think it makes a lot of sense.
Eric
Hi Eric,
Glad to hear it works for you. I found the idea that these bolts could be slowly creeping very disturbing! That said, he did caution that all of this depends on the boat designer getting the engineering right, or contracting them to do it for them—they have access to an engineer/naval architect for this sort of thing and other projects.
By the way, another interesting comment he made is that, just as you said, if you are doing engineering with a lead alloy, the only way to be sure of it’s yield and creep characteristics is to have a sample properly tested. That said, they mostly do this for industrial projects, not keels where they don’t think it generally required because with the way they build them experience says they are not close to a creep problem.
Hi Aberto,
You could not have more qualified people to answer you than Matt and Eric (Thanks you two). The only thing I would add is that this issue is typical of the marine business where actions that would be considered standard in the industries Matt and Eric work in just don’t get done.
In this case I would bet that in 95% of cases the keel bolts are torqued once at the boat builder and then never looked at again until a concerned owner has a go years later.
Whereas said bolts should be torqued several times before the boat even leaves the builder and probably a couple more times over the first year of sailing to allow for creep (compression) of the joint.
To that end I will be torquing ours again next winter and will report what I find.
Hi John and all,
I do not know what difference it makes, but a Valiant 42 keel is put on with 26 tubes of 5200 which, according to the factory, is alone enough to keep the keel in place. Creep might be minimized (or non-existent) if the weight is largely supported by the 5200 when the boat is at rest and the keel bolts only come into play when sailing.
Two other considerations:
The majority of Alchemy’s ample number of keel bolts are paired athwartships, which, as I understand it, diminishes the “hinging” effect which can “lever” bolts/keel into loosening.
The other is that Alchemy’s keel is not parallel to the waterline, but rather tilted/angled (higher in the bow) so that when I run aground, the blow is directed into the bulk of the hull. In this way there is less likelihood that the keel bolts will be shock loaded in sheer when the keel hits a rock and that the keel will be less likely to delaminate the fiberglass at the fore and aft ends of the keel as the keel transfers the blow into the hull
My best, Dick Stevenson, s/v Alchemy
Hi Dick,
I know some like the stuff, but I’m not a fan of 5200 in that application. The problem is that contrary to common opinion 5200 is a very poor sealing compound so over time it tends to develop cracks that let in water and that’s bad for keel bolts. As I understand it the issue is that it’s a good adhesive but dries quite hard with very poor elasticity and so any changes in the joint cause it to fail, for example temperature expansion and contraction. I discover this the hard way: https://www.morganscloud.com/2016/08/08/goop-and-goo-and-why-i-hate-5200/
And of course the other issue is that if you ever have to take that keel off it will be a horrible job and probably result in significant damage to the fibreglass. I watched a yard taking a keel off a boat that had been put on with 5200 and the whole bilge water sump that was let into the keel top came with it. A lot of very expensive damage to fix.
I also can’t see how creep would be minimizes by 5200 given that keel weight is a tiny fraction of the loads on keel bolts: see Matt’s comment.
Rather than any sort of goop in the keel to hull joint, other than round the bolts, I like to see the keel beaded on Epoxy resin with the right high density filler. This results in a joint with high compression strength and no voids, and no movement, if done right: https://www.morganscloud.com/2020/07/28/planning-and-budgeting-a-refit-keels-part-4-keel-removal-and-inspection/
And finally the experts at Mars Metals confirmed that they really don’t like seeing 5200 used in this way, for the reasons above. They too like to see the joint trued up with epoxy because it results predictable preloads on the bolts and no movement. (I think I’m right in saying that Epoxy (with the right filler) has low creep.)
I have also talked at length with Jamie, one of the two best composite techs I have ever met, who has repaired countless keel joints, and he concurs with above, and in fact won’t let 5200 darken the door of his shop.
As to two bolts reducing hinging effect. I’m not sure about that either, at least when compared against one larger bolt on centre line, assuming both are properly torqued so preload exceeds loads in use. Could be wrong: Matt, Eric?
Hi Dick and John,
5200 is much less stiff than the keel bolts which are also preloaded and therefore the 5200 will carry no load. When you have 2 items in parallel trying to carry the same load, the proportion of load carried by each item is determined by their relative stiffnesses. This is because for an item to carry a load, it is necessarily deflecting in proportion to the load. Preloaded joints are a bit special as the preload means that there is no change in displacement until the load goes over the preload amount.
Bolt placement is definitely important. Roll moments are reacted across a very narrow attachment point. Ideally, you would have all the bolts along the tension (uphill) edge and you don’t need anything on the compression edge. But then once you tack, the keel attachment will be in real trouble so you need bolts down both edges. In a best case scenario, you react all the way to the edge so the roll torque held by each bolt is double what a centerline bolt would be. This means you need half as many bolts to do one side but then you need to do both sides and you end up back at the same number. And of course, you can’t actually put them at the edge. Spreading them out does do a good job of spreading out loads in the keel and keel sump and makes the joint nice and rigid which is good so if done correctly, it makes sense. Whether you can actually react to the far edge is determined by stiffnesses and geometry if you don’t have a flat interface, sometimes small reliefs are made intentionally so that you don’t react the moment over a short lever arm.
Eric
Hi Eric,
Thanks, I find this joint stuff is fascinating. It did take me a couple of read thoughts (and most of a mug of tea) but I think I get it now. Just working on another tip about joints (fittings, not keels) based on my experience on the J/109 over the last couple of years and a new understanding for me.
I wonder what percentage of gear failures on boats are a result of poor bolted joint understanding and/or execution? Of course we will never know, but I bet it’s a high proportion, and most of it avoidable, so worth writing about, particularly since for us lay people it’s all so counterintuitive.
Thanks again for all your help.
Eric & John, Ditto. Fascinating. I have a list of things for my next lifetime that eluded me in this lifetime: rock climbing, engineering training, on and on.
My best, Dick
excuse me: in the text just posted the first link is a mistake: Here is to be read:
Wiring, like in aerospace construction, of course seems the icing on the cake,