The Offshore Voyaging Reference Site

Series Drogues: Learning From Randall Reeves

As you pretty much all know by now, I believe that the series drogue designed by Don Jordon is by far the best storm survival and anti-capsize gear available.

My belief is based on the solid science done by Don Jordan in cooperation with the US Coast Guard, much of it based on earlier work by the Wolfson Unit of Southampton University.

But, just as importantly, we have verified that science, and learned more, from some of the best and most experienced offshore sailers of our time, including Tony Gooch, Trevor Robertson, and Susanne Huber-Curphey, all of them with multiple successful series drogue deployments during Southern Ocean circumnavigations.

A few weeks ago, I spent a morning continuing that process with Randall Reeves of Figure 8 Voyage fame. Here’s what I learned:


Login to continue reading (scroll down)

94 Comments
Oldest
Newest
Inline Feedbacks
View all comments
Trevor Robertson

Hi John and Randall
The video of the dyneema bridle whipping about is both instructive and disturbing. It is not something I had anticipated when considering a dyneema drogue.

I think the problem is two fold:

1. The light dyneema drogue needs considerably more weight at the end to keep it in tension as the vessel surges back on the back face of a wave, a point you both make. This is easily solved by making the anchor weight heavier (double?), but of course this has to be done before deploying the drogue. I will certainly do this and thank Randall for the warning.

2. The dyneema warp has far less inertia than a heavy, wet piece of 18mm nylon double braid so whips about much more readily. The video makes this really clear. The solution to this may be to make the bridle of nylon double braid and the rest of the warp of dyneema. The heavier nylon section should be much less lively than the light dyneema. It may even be necessary to make the part of the leader section of the warp (between the bridle and the first cone) of nylon but judging by the video, this should not be necessary.

3. Stein Varjord’s planned use of glass fibre battens is ingenious but adds another job to be done in the cockpit under difficult conditions. I would rather avoid it if possible.

Having seen Randall’s video, my thinking is that when/if I put together another Jordan-type series drogue, I will use a nylon bridle and all the rest in dyneema, and double the end weight from Don Jordan’s specification for a nylon warp.

Scott Arenz

Hi John,

Great article summarizing the continual refinement of the SDDDJ. The video clip of the Dyneema bridle whipping around seems a perfect example of the law of unintended consequences, as I don’t recall anyone anticipating such increased flotation or motion. The proof of real world use and testing is invaluable!

Regarding the idea of the retrieval loops, although I appreciate the several descriptions that you and others have written on how the splices would be made and attach, I still can’t visualize them. Can anyone point the way to a diagram of how it would be done?

The evolution of Stein’s proposal for protecting vane gear is fascinating, and I hope that someone will put together a real world example before too long. I think the proposed block and control line at the outboard end is a good one. Of the several methods I can see used to attach blocks to the ends of backstay flickers, it would seem that lashing it in place best preserves the structure of the composite batten for the rigor of offshore use.

As for the attachment of the lightweight control line to the bridle, I wonder if chafe will be a concern at this location? I’m trying to imagine how much movement there will be between the bridle and line attachment as the batten flexes in the innumerable load cycles of a deployment.

Given that the bridle is such a critical part of the drogue, testing of the batten system in benign conditions is definitely needed, both for proving the hardware design as well as the modified deployment procedures. An additional question is whether it will tend to increase or cancel the whipping motion of a slack Dyneema bridle.

Lastly, can anyone comment on the use of special deployment bags such as the ones made by Ocean Brake?

Looking forward to the discussion!

Dave Warnock

I was also wondering about the deployment bags. If there is space on deck to fasten the bag stretched out it looks to me to be a good option. Could it even have lines to the bridle attachment points so that after putting the weight out the deployment bag is dropped off the stern and retrieved after full deployment?

Dave

Dave Warnock

John,
I was making a few connections – which might have only been visible inside my head 🙂
The Ocean Brake deployment bag looks quite large when opened out into a mat for deployment. The Rival 38 cc we are currently buying doesn’t have much clear space on and around the aft cabin, a line snaking out has multiple things to snag on. The article mentioned a fear of being close to the line as it is released.
So I figured that hanging the deployment mat over the stern would keep any danger well out of the way and reduce deck space and potential for snags reduced.
Apologies for lack of clarity.

Dave

Philip Wilkie

Dave,

We only have 2m of transom width and plenty of things to catch onto. Putting the deployment bag overboard and letting the JSD run off totally out of harms way, would mitigate my one big remaining concern. I’ve no idea if there are any unintended consequences, but it’s surely worth a test.

Reading highly experienced people who have watched these things actually deploy is highly cautionary. The goal has to make the JSD idiot-proof, because it’s my sense that by the time a storm compels us to use our it, we’ve all been reduced by fatigue and anxiety pretty much to the level of idiots anyhow.

Having said that, these articles on the JSD have increased my confidence enormously, and has justified our AAC membership purely on their own.

Alan Pottasch

Hey John, thank you for another in a long line of informative essays. I have been following and anxiously waiting for this recent addition. I am still going to try my jumars since this will make quick work of the retrieval. However in a sustained blow or if broaching becomes possible i would like to have some loops in my rode as a second option. I will be re-rigging my JSD with amsteel once i understand the process. Reading past discussions and this one I have remained fogging on the loops and their attachments. I’m hoping you could clarify. I am competent with brummal splices but i am unsure of the rest. Is your idea to create several 1 inch dyneema rings and then to attach them to the rode? (in lengths dependent on vessel length) and if so how would one do this? I am sure there is something I am misunderstanding…

Trevor Robertson

Hi John and Randall
I have done a bit more experimenting with retrieval methods for drogue with dyneema warp and here is what I have found:

1. Andy Schell is entirely correct that a cut spice is unsuitable for making a loop in the warp to allow retrieval. I thought cut splice held in place with a shoemaker splice might work as the shoemaker splice relies more on the friction of being woven through the warp and less on the outer part of the splice being compressed. However experiment shows that a shoemaker splice does not have enough friction when lightly loaded to take any significant pull against the normal direction for loading the splice. Unless someone comes up with a splice for dyneema that does not depend on compressing its outer part, a cut splice is not going to work.

2. A soft shackle of smaller diameter dyneema threaded through the drogue warp works as an attachment for the retrieval line (at least in a fair weather cockpit simulation), but if permanently rove is more likely to foul something as it runs out than a brummel splice loop. I hope/believe it is possible to thread a soft shackle through the unloaded part of the drogue warp forward of the retaining line as the drogue is retrieved, but I have too little experience with using dyneema after it has been under tension to know whether it can be fluffed up enough to pass a fid through it easily. It needs to be easy to do as the cockpit of a small vessel in the aftermath of a major gale/storm is not a place for fiddly work with a fid.

3. The suggestions to use a climber’s ascender (Jumar) have considerable promise. I tried it on undersize dyneema (6mm), all that I had to hand, and it holds well. There was no tendency for the 6mm dyneema to slip through the ascender when it was winched as tight as I could manage using my sheet winches. Presumably 10 or 12mm dyneema would be even more secure. There was no discernable shredding of the dyneema line by the ascender jaws provided the load is entirely relieved before opening the ascender – but that has to be done regardless of the type of rope used. The dyneema was crushed nearly flat by the ascender jaws but could be fluffed up to its usual form when the load was relieved and seemed to have suffered no damage.

Conclusion:
I think it is worth pursuing a retrieval method that does not rely on permanently attached external loops because of the danger of these loops fouling when deploying the drogue. The speed and forces involved as the drogue runs out in force 10 is dramatic; Randall Reeves makes this point in the latest post, as I did in one of my earlier ones. If I put together a dyneema drogue I will initially try to retrieve it using Jumars. You need two Jumars of course, one on the running retrieval line to the bows and the other on a short retaining line aft to hold the drogue while the running line is reattached. If that does not work I will try using soft shackles inserted where required during the retrieval process and removed immediately they have done their job.

Of course I will trial the retrieval system in fair weather, but unfortunately that often misses a problem that only shows when using the gear in earnest.

Alan Pottasch

Trevor, I believe I was the one who mentioned the use of a Jumar, it is a brand made in Sweden but readily available everywhere. The importance of that brand is the aggressiveness of its teeth. Actually the lack there of. Gibbs from England makes a version but it is much too aggressive for this application, high loads would run the risk of tearing the plait of the dyneema. There is no issue with any of these ascenders losing a grip on dyneema. Quite the opposite, they are very tenacious. While Jumars are hugely overbuilt for their intended use (6000lb breaking strength) they are only intended for 1000 lbs or less. Climbing rope breaks at 3000. I have seen a falling body shock load a Gibbs and shred a climbing rope. I don’t know what 10,000 pounds shock loaded would do to dyneema. It wouldn’t look pretty but I believe the dyneema would not fail. Either the jumar or a cleat or the winch would. But..we are talking about retrieval, when loads are much less..if it doesn’t break the winch it wont break the jumar. To be clear, the jumar would be placed in lieu of the loops John was suggesting. Pendants off the jumar could be run direct to a winch or to a block at the bow and back to the winch. You would need two jumars.

Richard Elder

Hi Trevor & Alan
On an earlier discussion about this issue I suggested that a modification of a line clutch that was actually designed to hold high loads on Dynema might be the ultimate solution. The problem with using Jumar jammers is that they are not really designed for that type of line or the loads that will be encountered.

John correctly pointed out that no manufacturer would bother to build such a device for such a small market. However that was not exactly what I had in mind. With CNC milling so common, it seems feasible to design a hinged side plate that would allow side access to a well proven and reliable standard clutch. If done properly it would be faster and more reliable to release and could be attached at any point along the drogue line.

Drew Frye

Just as a matter of mountaineering trivia, Jumars where made in Switzerland by Jümar Pangit. I still have a pair of the 78s. I suppose they could have moved production to Sweden.

I’ve broken several name brand accenders used in test rigs; you are at VERY high risk of ruining them over ~ 1200 pounds. Rope damamge is also very likely, or I should say, probable. So make sure there is a load-limiting mechanism. Anything over 600 pounds is abuse.

There are belaying devices better suited for this, but nothing is designed for use with Dyneema, and they are optimized for specific rope types and diameters.

Although not specified in this way, most single climbing ropes have a breaking strength between 58– and 6500 pounds. They will break at lower stress if knotted or drop-tested (sudden load) but that is not a standard breaking strength for any rope.

Drew Frye

Great video. There is a lot of food for thought in that. The “obvious” suggestion is nylon, but with that much surge it will fatigue. The best answer is to keep the slack out, perhaps by adding weight, but perhaps also by setting some jib, but a jib sheeted tight is actually the wrong shape; you really want something that does not exist, like a tiny V-shaped riding sail a the bow. This would provide forward tension.

Maybe polyester is the best material, with just a little natural stretch and better fatigue properties. Wouldn’t you love to see them tested side-by-side… but that’s not realistic. I’m bettering, though, that a clever engineer could run a simulation based on that video and would tell you that X5 stretch along the entire drogue (not just the bridle) would solve the problem.

Yeah, I’m thinking Dyneema may be the wrong material from a purely engineering perspective. I don’t like it on travelers, for example, where I would rather have some give. I’ve made that mistake. Dyneema is a great material, and I use it for many things, but I am unconvinced it is best for a JSD. Maybe. But maybe not. Maybe polyester instead of nylon. Better fatigue resistance than nylon, virtually the same abrasion resistance as Amsteel (side-to-side this is probably true–Amsteel brags about end-to-end, which only matters on pulleys), and perhaps just enough stretch. But it’s bigger and more expensive.

(And obviously I meant 5800-6500 pounds, above. I’ve pull tested quite a few samples.)

Eric Klem

I wonder how big of a deal the slack is from a loads perspective, while it looks impressive I am not convinced that there is actually much shock loading (it is a big deal from the perspective of vane gear damage obviously and less slack should give a nicer ride). The steady state load on the drogue is limited by the drogue’s resistance in the water so if there is no slack, then there is no need for shock absorption assuming that the system is designed appropriately.

Once you introduce slack into the system, the impact loads are a function of the mass impacting, the speed, and how much give there is in the system. In the case of a drogue made of dyneema, the mass is going to be low and the speeds are not all that high. The only negative is that the give is pretty low too so even dissipating a moderate amount of energy will have a high impact load. Doing the most basic calculation, 100′ of 1/2″ amsteel (6.4 lbs, 2.9kg) which is more than appears to go slack with an impacting speed difference of 5 knots (2.6 m/s) will have an energy of only 0.5*2.9kg*2.6m/s*2.6m/s = 10 joules which is not a lot of energy. The actual energy will be higher as the bridle is not perfectly in line and that has a more than linear effect but this should still get it in the ballpark. If the first 50′ stretch 0.05% (this is the stretch at 1% of BL), then the force is only 1260 newtons or just under 300 lbs average from this impact which would have a peak of around 600 lbs. It isn’t even completely additive to the static load as the impact will be over before the cones are fully loaded up.

If we take the extremely conservative view that the whole drogue stopped and now needs to be accelerated all at the same instant and it is in a straight line, the numbers look more like 0.5*25kg*2.6m/s*2.6m/s=85 J. Still dissipated over 50′, the stretch will now be more like 0.1% which is more like 2.5% of BL, leaving a load of 4450 N or 1000 lbs average and 2000 lbs peak on a line that is rated to 34000 lbs. Given that this represents only 1/17 of the breaking strength of the line and I believe is a very conservative way to look at it, this doesn’t overly concern me. Don’t get me wrong, I don’t like the slack but also don’t think that it necessarily means that dyneema is not a suitable material. If anyone thinks that this is not a good way to look at it, please let me know.

Eric

Drew Frye

There is SO much about JSDs we don’t know, but I do have a few thoughts.

The impact is more than 10 joules, because you have accelerate both the mass of the materials AND the water contained within the cones. This is huge factor on parachutes and should not be ignored here. Not a huge number, but probably more than 50 kg, not 2.9 kg even in the first case. The entire drogue contains more than 100 kg of water. I know this from experience.

The other nasty thing about slack is it’s effect on fatigue. Depending on the fiber, fatigue life is 3-5 times greater if just a little residual tension remains on a rope, rather than going completely slack. It has to do with fibers moving in the bundles and also how the crystals and molecules coil and uncoil. You can Google this. It’s a big deal for nylon, but I’m uncertain about polyester or Dyneema. It is even true of metal rigging, explaining why higher pretension is better than slack rigging. It also sets the maximum length for many bridles where yawing is present; you want to avoid a slack leg (although that is not immediately relevant here, since both legs went slack).

If we are curious about the magnitude of that sort of shock load, that could be measured by load cells in fair weather by gathering some slack and hitting it under power. I doubt it is a big number, but the effect of cyclical slack on fatigue and chafe is what I would investigate.

So complicated and testing is so impractical. This is why sharing experiences is so important.

Eric Klem

Hi Drew,

I had made the assumption that any forces from the water were included in the base load that Jordan calculated but you have me rethinking that now. Since water is dense and incompressible, there must be some amount of water that would need to be accelerated. It would be interesting to know the max acceleration Jordan calculated for just as a function of a wave strike and what it would be. As an aside, prior to good FEA software for impact calculation, we generally found that hand calculations were significantly conservative due to the assumptions that all forces would align, all of the worst loadings would occur simultaneously and that we had failed at the point of plasticity but you did need to get your boundary conditions right in the first place.

While theoretically possible to build a dynamic model to look at this, you would have so many important assumptions about your boundary conditions in the model that you would be forced to validate the model with testing to get any reliability from it.

As to other materials, I agree with you that I don’t like nylon due to loss of strength when wet and poor fatigue characteristics. Dacron does seem like a reasonable material if trying to avoid dyneema.

Thanks for the response.

Eric

Trevor Robertson

Hi John
Re connecting a brummel splice loop to the main rode on a 12 strand dyneema JDS.
First an apology for being verbose, but without diagrams describing the options takes many words.

I have tried and failed to find a reliable way to attach a retrieval loop to the main rode that does not require the main rode to be cut and spliced at each attachment point.

The loop is easy: its tail (with or without a brummel lock) is safely buried in its own standing part, which is under tension and thus compressed when the retrieval line is attached – so far so good.
BUT I cannot find a way of connecting the loop to the main warp that puts the outer covering of the stressed part (the loop) under tension. The warp is has to go inside the loop’s tail. I can see no way of inserting the main warp into the loop’s tail without cutting the warp.

A conventional brummel-and-bury is useless, as is a shoemaker splice – the wrong part (the stressed part) is buried and there is nothing to compress the outer and hold things in place.

A shoemaker splice that was also sewn and seized held when winched hard but it felt unreliable, and those feeling are correct often enough to take notice of them. I don’t trust seizings on something as slippery as dyneema.

Linked or cow hitched eye splices at each point in the main warp that the retrieval line is to be attached would work and not require an external loop. Although I do not like the idea of cutting and splicing the main warp more often than necessary, this is the only option I have come up with so far that works. As there are already at least eight eye splices in the bridle and drogue, all functionally critical and most highly stressed, I suppose a few more make little or no difference to the reliability of the whole. Perhaps it is the way to go.

It is annoying that I have spent so long on such a trivial problem and still not found an elegant solution. Doubtless somewhere in the discussion there is something important that I have missed. I sit holding fid, needle and palm, hoping for enlightenment.

Addenda: for completeness of my previous comment, the testing I did on climber’s ascenders used Petzl Ascension gear, French made and rated for 8 to 13mm lines. The line was unsheathed 12 strand 6mm Dyneema.

Craig Burnside

For retrieval anyone tested something as simple as a short piece of webbing through the dyneema and knotted vaguely similar to the cone connections to give a prussic something to bite onto?

Craig Burnside

Well, with some 12mm D12 max onboard which will be the bridle and having worked extensively with it, some squashing wouldn’t worry me in the slightest. That stuff is near bulletproof, just watch the bend radius, especially on a splice. Much less a worry than the risk of any loops fouling. What sort of loads did Randell et all experience on retrieval? Enough to be a worry for the rope would have me more worried about a winch if it was round one.

Drew Frye

Inserting webbing (suggested below) might be hard on the rope, but what about inserting tapered line inside (a short section of core) to make fat spots? It should not weaken the rope and would be quite easy to do. A few stitches would lock it in place. You would have to test the holding power of a prusic, but it may be one of the very few ways to grab Dyneema 12-plait without damamge. It would be smooth and not damamge the rope.

Just spit balling.

Craig Burnside

“but what about inserting tapered line inside (a short section of core) to make fat spots?”

Just been thinking exactly the same thing. In the case of 12mm d12 it wouldn’t even need to be tapered that much with so much spare capacity. A loop sticking out of the dyneema snagging in those conditions really is scary!

Craig Burnside

Back home on the boat next week so will have a play. Only have a 500Kg lever hoist (so it says but cheap from a market in china) so double reeved could prob get up to 3/4T. Some 5mm d12 max onboard which might be OK for a test then hunt down the USB microscope.
12mm D12 is over 20T MBL, seriously doubt if tonne and a half on a prusik would bother it much. I’d be much more wary to be trying to use a jumar or shunt, something might well go bang with them.
I’ll see if I can break something 🙂

Are those loads measured or theoretical? I used to be able to guess vaguely close up to about 500Kg playing with load cells but above that it’s just crazy tight.

Craig Burnside

Think a single cone at the outboard end of each bridle leg might help a bit to keep tension on the legs?