The Offshore Voyaging Reference Site

Planning a Refit— Keel Removal and Inspection

Can you believe that we are now up to four articles on keels and keel bolts? No? Me neither. I have wondered if I was getting carried away here but, when I really thought about it, I realized that this work needs doing for the following reasons:

  • A keel loss, while comparatively rare, is often (mostly?) fatal.
  • Roughly 35 years ago the industry abandoned bronze and Monel keel bolts, which were pretty much immune to deterioration (when into lead), in favour of stainless and mild steel, both of which can waste in just a few years.
  • It’s impossible for a surveyor to check keel bolts visually with any certainty since, as Bill from Mars Metals tells me, the wasting of bolts often is in the form of waisting in the joint between the keel and hull, so there can be a big problem even when the stud heads and nuts or bolt heads inside the boat don’t look too bad.
  • It’s relatively easy for an unscrupulous seller of a boat to hide telltale signs like cracking in the keel/hull joint and attendant rust stains.
  • The industry, like with rudders, has its head firmly up its ass about the whole keel issue.
  • Ditto a lot of owners.

More Frequent?

Given all that, it’s clear that going to sea in these boats, that came out of the flurry of fibreglass production sailboat building in the seventies, eighties and nineties, without adequately checking and, if required, replacing the keel bolts, would be foolhardy.

Or, to put it another way, as these boats age further, I think that there will be a tipping point where keel losses suddenly get more common, and maybe a lot more common.

Consequences Suck

And even if that’s not true, and keel losses remain rare, it’s still a high consequence failure—the most important part of risk management is understanding the relationship between frequency and consequence.

A good metaphor is that high speed automobile accidents are pretty rare (when measured against miles driven), but we still have airbags and buckle up because the consequences are so horrible.

Understanding is Better Than Delusion

And let’s not forget that the whole idea of this series is to develop a framework (including a budget spreadsheet) to assist those considering refitting an old boat in really understanding what the costs will be before committing. And, most important of all, in being aware of the real costs of worst cases, like replacing keel bolts, that could blow a huge hole in the finances of the unrealistic.

Enough, back to the nuts and bolts of this (ouch).

Do We Need to Remove The Keel?

In the last two parts we looked at testing bolt integrity without taking the keel off. So the question becomes, can we stop there if:

  1. There are no signs of bad stuff happening like:
    1. rust streaks from the keel to hull joint,
    2. hull damage from groundings,
    3. or corroded bolt heads in the boat;
  2. We have had the bolts ultrasonic tested; and/or (preferably and),
  3. the bolts have withstood being torqued.

That’s up to each of us to decide depending on our own risk tolerance. What would I do? Probably stop there.

But let’s assume that we have bad, or even just inconclusive, results from the above. Or we decide out of an abundance of caution that we are going to take the keel off anyhow—the best and safest choice.

What then? 

Login to continue reading (scroll down)

Inline Feedbacks
View all comments
Philip Wilkie

A few months back we had a perfectly nice 36ft coastal sail boat in the yard that had been run aground on a sandbar, but when it was towed off they badly twisted the keel about 20 degrees.

End result; insurance writeoff.

This may well be anecdata (n=1), but surely has to speak very strongly against the vulnerability of bolt on fin keels and their costs to repair. In the same yard (with 15 pens) I’ve seen at least 20 boats with keels removed for one reason or another in the past year. The costs are never discussed, but they must be eyewatering.

And all this risk for a performance gain that for cruisers doesn’t really add much value. After all how often and how long do we like to sail hard on the wind? And when we do, point that high, how much leeway are we losing? And how much does the relatively flat forefoot pound? And the hull flex?

Maybe for cruising boats we should just own up to the reality that bolt-on fin keels are a bad idea period.

Alex Borodin

Hi Philip,

Even if we all agree that bolt-on keels are a bad idea, it won’t change the fact that when looking for a cruising boat, it is almost inevitable that most boats on the market will have just those keels. If I am not mistaken, most builders abandoned encapsulated keels in late 70s and 80s. Hence, the younger the boat, the smaller the chance of it having an encapsulated keel.

Also, some people do like to cruise with speed (John included, as known from other articles here). Even if they are looking just for downwind speed potential in a boat, there is a good chance designers of performance-oriented boats have put bolt-on keels on them since deep into the 70s.


Bill Attwood

Hi Philip
I am another of the guilty ones who raise this idea. ? I agree totally with your thesis. John is correct to point out that encapsulated full or short keels can also have their problems, but if the capsule is lead, and if the boat is from a respected builder, both easily verifiable facts, then one can be free from worry. There are still a number of designs, mostly from the 70‘s and 80‘s, but being built into the late 90‘s and even early 2000‘s, which have this configuration. They may be a little more expensive than French-built cruisers of a similar vintage, but are still affordable. Having followed this series of articles from John and the resulting comments, it seems a no-brainer to me. If you have the money, or are interested in regatta sailing, then go for a well-built new fin keeler, if you are interested in cruising and less concerned (not disinterested!) about speed, then find a full keeler with encapsulated lead keel. And then enjoy the easier motion in heavy weather, able to make way to windward when the speed merchants are suffering. John will probably accuse me of cultism. Guilty, your honour.
Yours aye,

Ernest E Vogelsinger

Guilty as charged, too. My ideal boat is of steel, with a welded hollow keel structure that has been filled with molten lead after sandblasting, and when the lead had cooled tightly glassed over.
If there would only be enough on the market, though…


In 2004 some small cracks between the fore and aft ends of the keel-hull joint on my 1974 Islander 36 became apparent. The iron keel is attached to the hull with (eleven) one inch (8 TPI) threaded rods, which terminate in the bilge, held in place with 1-1/2 inch nuts on fender washers. The nuts and washers are on top of a 1/2 inch thick steel plate that runs the length and width of the flat bilge, essentially a long, wide backing plate. The old threaded rods (aka keel bolts) were removed with little drama (thanks to the yard’s pneumatic tools) and looked good, but they were replaced with new mild steel ones, torqued to 300 ft-lb (as I recall). Before installing the new keel ‘bolts,’ I took the backing plate to bare metal, and then, acting on the yard’s suggestion, coated it with Bar-Rust, then paint. I ground down to the keel-hull interface and found no anomalous gaps between the iron and fiberglass, and no signs of movement, and only a very little superficial rust. So, after a lot more grinding, the yard glassed over that interface, all the way around. Since then, part of the ~annual haul out ritual is to, while on the hard, try and budge the keel bolt nuts – I had a 1-1/2 inch deep socket cut in half and each end welded to a 4 ft long pipe producing a very long socket with the T-bar handle at a convenient height when the socket is set on the nuts in the bilge. I’ve never been able to make the nuts move. While in the yard, I also coat the bilge with Ospho and then paint. And, after almost 20 thousand miles of sailing on the Pacific (CA-MX-HI), including contacts with the sea bottom (a couple soft, and one hard), the hull-keel joint has been completely stable.


Hi John,
I haven’t been able to find that old yard bill, but as I recall it was about $5kUSD, about 1k of which was for the bottom paint, including prep and application. So about 4k for the keel work in 2004. I did as much as I could, mainly grinding and scraping, but I don’t think I lowered their labor that much as I’m not nearly as efficient or skilled as they are, especially at reconstruction. Regarding the glassing of the joint, the glass was not applied directly to the iron (apologies for giving that misimpression), rather the iron was barrier coated (Bar-Rust as I recall), and then filler was applied to fair the transition , and then fiberglass bridging the coated keel to the hull was applied over that. That fiberglass bridge is not structural, it just fairs the hull, and keeps water away from the keel-hull joint. I assume that if the keel moved wrt the hull, the coatings would give and crack. Since that rework the only keel damage I’ve experienced has been on the bottom due to plowing through shallow marina entrances (La Paz, MX and San Francisco bay).

Philip Wilkie

Apologies. I certainly had no intention of irritating you or derailing the thread. It was just difficult not to react as I did when your own article makes such a comprehensive case on the costs involved to repair a damaged fin keel.

I’ve read this series with considerable interest from an engineering perspective, and learnt a great deal. You are right in that bolt-on’s are the reality most people are going to live with. If for the sake of argument we take this as a given, then surely a case can be made for the industry to re-think it’s approach. Clearly the move away from bronze and monel was a bad idea, just for a start.

But if stainless bolts are what we are going to get, then perhaps a rethink on how to design out crevice corrosion and improve inspect-ability would be in order. Or design the web platform and mechanical layout so that hydraulic bolt tensioners are the preferred tool. Or just stop glassing over the bolt heads. I accept there is nothing we can do about the several decades of boat builds out there which have this vulnerability, but what’s the way forward on this? Even the simple measures Steve H outlines above seem really worthwhile.

Because honestly you’ve convinced me that a non-inspectable, hard to repair design for a safety critical design item like the keel to hull connection is a real liability. Certainly if I ever find myself looking at buying another boat this consideration would be very high on my list right now.

Bill Attwood

Sorry John. Look on it as a form of Tourette‘s and have sympathy for the afflicted.?

Andrew Craig-Bennett

I am a recent convert to GRP after 45 years of wood, 40 of them with teak built boats, with, obviously, external bolted keels. It was standard practice in the British isles, and maybe in North America, to pull a keelbolt be it bronze, wrought iron or stainless steel, every ten years, and, if there was any sign of wastage, to replace the lot.

Iain Dell

John – just to confirm your price estimates are pretty near the real-world mark. Friend of mine bounced his fin-keeler off a rock in the UK two years ago but, in his own words, “didn’t see any damage to the keel bolts nor the join when the hull was lifted” (routinely at end-year) so he did nothing and told no-one. End of last season he was wondering why water was coming into his bilge in increasing amounts…….. He’d changed his insurance company in that period and, understandably, neither would pay out. He learned the hard way not be so completely stupid when he forked out just over £12,000 to have the keel removed, repaired and refitted. The other lesson hidden there is how an otherwise intelligent man, an experienced and (usually!) careful sailor, can be so inexplicably negligent towards such an obvious danger.

Eric Klem

Hi John,

What you propose is a very reasonable sounding process to me but I have only removed bolt-on keels on wooden full keelers which is a bit of a different story and has already been discussed.

It is interesting that Mars recommended some form of NDT on the bolts.  MPI requires specialized equipment so you can count on people having to hire someone for that.  Dye testing doesn’t have the capital equipment barrier but most people will have no idea how to do it properly.  The part needs to be really clean which it should be anyways for reinstallation but to make this worthwhile, plan to spend many hours cleaning up the threads.  To be honest, if it were my keel, I wouldn’t feel the need to do this if I had calculated the max expected load in the bolts and it was well below the fatigue limit of the material.  Some materials, including steel and stainless steel but not aluminum, have a stress level below which they simply do not fatigue regardless of the number of cycles.  Handily, the safety factor that most people put on designs means that even if the preload is lost (preload on a stable joint is the best way to prevent fatigue in bolts), the stress is below the fatigue limit.  This isn’t always true, I have actually designed many things where it was not true, but knowing how to do this calculation, I would not feel the need to check for cracks if my calculations came out saying it were true.  Probably easier to hire someone for NDT for most people but I would hope that the majority of cruising boat keels actually have sufficient safety factor to skip it but you don’t know without some engineering.  By the way, most people will make overly optimistic assumptions for analyzing this type of joint if they don’t have proper training.  I would still do a very good visual inspection to check for wastage and I would put a thread gauge on it to make sure they were cut right in the first place (incorrectly cut threads are a fatigue issue).

I have seen a few keels where I would want to really level the top of the keel before using it as a mold for the stub.  There can be advantages of actually having non-flat matings surfaces (breaking the big end of a connecting rod is one example) but in this case I think you really do want a flat joint as the preload is quite low for the area and your locational accuracy is poor.  If you were really sure you had gotten the keel top flat, I suppose just making the stub flat would be better than nothing but it is still better to do as you propose and use the keel as the mold.


Eric Klem

Hi John,

Grounding is definitely a concern as you point out and likely represents the highest loads many keel bolts ever see.  The trick here is that the failure mode is different as there are not the cycles that fatigue requires resulting in small hard to detect cracks that slowly propagate.  With lead keels, you may actually see the lead around the stud pulled up somewhat, that would definitely be a red flag to look further.  The bolts will tend to elongate in the thread area.  It is somewhat material dependent but if the pitch is off in any section, that would be a red flag.  The crudest way to check that is to run a nut down after cleaning the threads and if it starts to bind anywhere, you have issues, unfortunately, if it doesn’t bind, this won’t tell you there hasn’t been a small amount of yielding.  A better way to check is with a thread gauge.  Assuming the studs used are not a material that is likely to be subject to brittle fracture, it is likely that you will pick up the elongation of the thread pitch before you could detect a crack.

All of that doesn’t change your main point though which is that for most people, hiring someone to do NDT is a good idea.  Even if dye testing doesn’t pick up the problem, those people have training and experience that will hopefully allow them to pick up issues that others would not.

And yes, I suspect very few boats actually have their keel bolts properly preloaded.


Stein Varjord

Hi Eric,
As a non engineer who by necessity has had to understand some engineering mysteries, 🙂 maybe my way of understanding preload could be useful? I don’t know if it’s a completely correct comparison, but I see a mast and two shrouds as an illustration:

We first set the shrouds so they are just tight when the system has no load. With wind loads, the windward shroud will get increased load and stretch so the mast leans over and the leeward shroud gets slack. If we then rather preload both shrouds to the load they will have at the maximum side load we want the rig to be able to take, it might seem as this would increase the loads in the system. We might think that the wind load comes on top of the heavy preload.

In reality that’s completely wrong. The windward shroud stays at exactly the same load all the way up until max load is reached. The only effect from wind load on the system is that the leeward shroud tension decreases by the same amount as the wind load. The consequence is that the mast will not move sideways at all, as the only variation in tension happens in the leeward shroud.

With a low wind load, most of the tension on the windward shroud is from the leeward shroud. With a high wind load, most of the tension will be the wind load, but the leeward shroud will not get slack until the maximum tolerable rig load has been reached. From that moment only, will the load on the windward shroud increase above the pre tension it has at rest.

This is how I explain why I put very high tension on rigs. Some bolt situations seem similar, but I assume the whole picture is more complicated.

Eric Klem

Hi Stein,

You are correct that tensioning a rig is not dissimilar to tensioning a bolt.  The area where we have to be careful is the relative stiffness of the bolt (shroud) to what it is clamping (mast section and hull).  Ideally, the thing being clamped is way stiffer as it makes the analysis really easy but with an aluminum mast, I strongly suspect this is not the case.  When it is not the case, the load in the bolt does not necessarily equal the static preload once the system is loaded up.  Depending on the exact scenario, you can start to get stress cycling or opening of a joint so care needs to be taken in design and assembly.

The easiest way to prove this is if your lee rigging goes slack.  Since everything has to deflect to carry a load and it now has no load, that tells you that something has moved as the lee rigging is now shorter than it was before you started sailing.

In general, preloading bolts and shrouds is definitely a good thing unless you are designing a pure shear joint but that tends to be the realm of civil engineering.  Just for most people, you should make sure that what you are clamping is very stiff whenever possible.


Colin Speedie

Hi John
on some older (and arguably better built) British designs, hull-to-keel joints were sealed with a 2 pack putty called Ralli Bondite that had the most tenacious grip imaginable. Taking a keel off that had been bedded with the product was the stuff of nightmares. You could remove all of the bolts and lift the boat and nothing would happen. By the end of the day you’d be using wedges and a sledge hammer to try and break the joint. Which inevitably had the secondary effect that damage to the GRP might be caused during the process. That will have to be made good before any replacement of the keel – so add that into the cost and time equation.
Using wooden wedges and steadily, gently driving them in and then leaving them for a while to ‘prise’ the joint apart can work well and cause only minor damage, especially if the seal is made with a structural adhesive such as Sikaflex, I’ve found.
Best wishes

Stein Varjord

Hi John,
I have only two comments:
The tool you recommend is awesome. I have it and use it a lot. However, it’s called Fein, not Fien. Typo, I guess. Swiss made, but now a part of Bosch. Fein tools are expensive, but are built to last forever. Many other brands make similar tools much cheaper, copied from Fein, the inventor, but they are not at all the same quality.

To slow down epoxy cure time, there are several tricks. You mention some. I’d add these:
– Make sure there is shade from the sun on all items related to the job, including all of the keel and hull, a long time before work starts.
– Be careful about actively cooling down the mating surfaces, as that could create condensation, which will give poor adhesion.
– Cool down the epoxy ingredients before mixing by keeping the closed container in cooled water. Keeping the mixed epoxy tray in water also helps increasing pot life, but is often impractical. Excessive cooling in moist surroundings could create condensation problems.

Robert Berlinquette

Great article, the list of jobs is getting high and will be checking the bolts on the new to us Sceptre 43. But I did come across another Sceptre owner who had removed their keel. It was interesting to see that the bolts are beveled at the base for sealing compound and that keel also has alignment notches on it.