Arne,

1.      One obvious one is the one Toni pointed at. When sailing inshore, it is more a rule than an exception that wind- and wave-direction do not match. Through all my years of fjord sailing, I have concluded that the good/bad tack varies much with the wave direction.

Agree! Wave direction might be the valid reason for limping at that specific situation on the port tack.

2.      Sheeting angle due to offset of the sheeting point. It is not that straightforward to get this perfectly right. For instance, on my Ingeborg, I often sheet in another 30cm on the sb. tack and then pay out again on the port tack, when tacking.

We made that observation of different sheeting angle for each tack, too, and I trained myself over the last months of almost daily sailing in getting it right - on both tacks. Be assured, while "racing" against Playmobil I definitely sheeted to optimum.

3.      The rather big area of detached airflow on port tack may also have something to do with this, but more sailing is necessary to verify this or that.

That is the question: What is the impact of the detached airflow?

“I’d rather limp on one leg than on both”

Okay, I've been going through all the footage and fotos. It was pretty clear, not a lot of averaging to do. Have a look at the results in the following sketches:

On starboard tack (the "good tack") it was as expected: when sheeted in properly, all telltales were flying - except one. It was not possible to adjust sheet angle so that the most forward telltale was flying in lee and windward. We tried hard, impossible. My conclusion from that is, that the luff is stalled at a small area due to the distorted webbing luff shape. Clearly, the flow reattached quickly. The influence on lift and drag I can not quantify, but it can only be bad.

On port tack (the "bad tack") it got interesting: airflow from the luff to 50% chord is detached, and then reattaches behind. That's a huge detached air bubble!

One further thing we discovered while sailing against the Breehorn 37 Playmobil today (see here): We were able to keep exactly, really exactly the same angle to windward as them on the good tack/ starboard tack. As we tacked simultaneously and very close to each other, a direct comparison could be made on both tacks. On the bad tack/port tack, we went upwind about 5-10° less than them. Now, it could be that it was their boat which went better upwind on port tack, but as the bermuda rig is symmetric it more probable that Ilvy went 5-10° worse on port tack/ bad tack. 

Its easy to get the tacking angle from the gps course. However, it does not tell if the TWA's are the same on both tacks. Obviously, for Ilvy, it is not. We do not have a wind-sensor onboard, as I like the KISS principle also regarding electronics. But without one, I can only speculate on the TWAs for each tack.

I hope my semi-scientific observations can be of any help.

Paul

edit: after discussing with Toni on the topic, we remembered that the waves were more against us on port tack - when Playmobil was pointing higher than Ilvy. As it is 37" against 25", they probably didn't feel the waves that much as we did... Hi, field testing noise, how I missed you!

Short update on the luff lashing: I lashed the luff now also sidewards directly to the battens, as in my sketch below ("better luff lashing"). Today's sail put it to test, and it worked well. The gap is gone, and with it the unnecessary load on the first stitches of the batten pocket.

However, it did not help my distortet luff webbing.

Cheers,

Paul

With Arne’s comment on how I sewed on the small webbing eyes which are used to lash the luff to the batten, and comparing his fotos of Ingeborg and Malena to Ilvy, I realized that I obviously did something stupid: I lashed the sail only in forward direction to the battens, but did not lash the luff “sideways” to the battens. That sketch explains it better:

If you look closely at the following fotos of us beating to windward in an F5-6 yesterday in the Kalmarsund, you might discover what I mean: The luff edge is leaning far away from the battens. This might be reason for my distorted luff… What it definitely does, however, is putting some unnecessary load on the first stitches of the batten pockets. Of course, all this applies only to the starboard tack!

I fixed the lashing this morning and will report back after the next sailing, in a few days (we are waiting for easterly winds right now to carry us down through the Kalmarsund).

Just some comments on Arne's latest, very in depth comment here in this thread:

·         I make the camber curve by shaping the barrel rounds with a bendy wooden spline. For that reason, there is just a moderate curve near the luff. I decided against a more ‘advanced’ foil after learning that the shape would be ruined anyway on the port tack, and the sail still worked well on both tacks, with flying telltales at the leech.

Totally agree on that! I am almost afraid to say the following (and I might be totally wrong), as I am quite new to the junk rig community, compared to the experienced experts like you are, Arne, Slieve, David, Paul T., Paul M., etc... but: It seems to me that is not yet understood why the cambered junk draws as good as it does on the "bad tack", the port tack. After all, it is just physics, no miracle involved here... I didn't find anything like publications on this topic, yet, but have there been wind tunnels tests or CFD simulations conducted with the cambered junk/ SJR? It would be most interesting to not only speculate on why it is working so good, but to verify it!

Paul S. has used CAD to make that barrel shape, and has clearly (from the photos) added a good deal more camber to the luff area than I use.

Agree again. I mean, drawing a spline with CAD is basically the same as using a wooden spline. However, CAD gives you easier possibilities to modify that spline, and I definitely did not use the exact same curvature as you did, Arne. It would be interesting to compare both.

·         At each batten end, there are two loops; a big one for the batten to rest in, and a small horizontal one for tying the sail to the batten. This last one appears to have been made differently from the way I do it. I start fitting the small loops by stitching half of it to the reverse side of the sail. Then I wrap it around the sail’s edge and stitch the other half onto the front side, just leaving a gap, big enough to tie on that lashing. This way, the lashing forces the edge of the sail to sit next to the batten.

Yep, I sewed both parts of the small webbing loop to the same side. However, the stitches lead right to the very front edge of the luff, thus I don't expect a huge difference here.

·         I stitch the boltrope onto the assembled sail. The method lets me keep control of the tension in that boltrope. I wonder if Paul S by wrapping it into a ‘sock’ of sailcloth, may have ended up with a little bit slack boltrope? I could be wrong on this one.

Could be, definitely! Though I remember when sewing I really tried to tension the webbing inside the envelope before sewing it together. When I try to tear the luff by hand, it is absolutely stiff and not moving. The webbing I used was specified for 5-6t load...

Anyway, if we study the big numbers, Ilvy’s sail has been a roaring success. She clearly sails like a witch, both upwind and downwind.
To me, her main success factors are a big sail area, a cambered sail and good helm balance.

[...]

I try to keep focus on the big factors.

Definitely! I am most thankful for you, Arne, to have developed over decades such an amazing sail, which now propells boats like Ilvy in a most astonishing way. More than that, you even published the TCPJR documents, which provide such a low entrance to making your own sail. All for free. An ideal example of Open Source Philosophy. I bow to that!

I also totally agree that the big factors make the most difference. Pareto. The biggest factor is to move from bermudan to junk. Then comes camber in junks, then maybe SJR. Those three factors already achieve about 80-90% of sailing experience optimisation, compared to bermudan sailing (I'm a bit pathetic here, maybe.). It seems to me, that every further development can contribute only little more, as the big factors are already there and in use. However, why not increase the performance by even as low as 1%, if it does not mean increased work, cost, hours, complexity? Why not to improve, even if only a tiny little bit, if the improvement costs next to nothing? That's why I started this thread with the loosely sewn boltrope luff idea, as it costs (money, hours, complexity) no more than the webbing luff but gives better performance.

As long as it is easy to achieve attached airflow behind the sail, verified by the leech telltales, one knows that both sides of the sail are working, unlike with a flat sail.

However, I would like to discuss on this topic! It is an absolute improvement to using leech telltales to using none, and I fully support you repeating yourself in naming the importance of those!

But: I think we are not able to derive the conclusion of attached airflow over the sail just because the leech telltales are fine. The leech telltales only tell us, that airflow is attached at the leech. What happens at the luff or at 30% camber, we can't derive of the leech telltales. What happens at and around the mast, can't be told by the leech telltales. At Reynoldsnumbers/ airspeeds we are talking about when sailing, the airflow is definitely detaching behind any round section like the mast, be it on port or starboard tack. The leech telltales prove that the airflow has to be reattaching at some point before reaching the leech. Okay. But where? With which effect? At what cost, talking about Lift, Drag, alpha tolerance, etc...?

Maybe I should explain one thing: If one wants to sail a good junk, building your sails, Arne, is a very good thing to do! I enjoy sailing with it every day! Not only performance-wise, but also regarding simplicity, safety, handling. I am totally fine with that. However, I, personally, really like to understand things and optimize them - if the improvements are in good relation to the effort. I really am fascinated by fluid dynamics, that's why I like being a shipbuilding engineer: In my career so far, I conducted and organised quite some wind tunnel tests, towing tank tests as well as programming lengthy CFD simulations (compuational fluid dynamics) of different hulls (commercial ships, planing and semi-planing boats, catamarans, etc...), propellers and shapes. A well functioning wing profile is of real beauty to me - and I don't equal well functioning with L/D only, but rather look at the whole system including handling, stalling behaviour, etc...

So, I hope to not offend anybody by asking critical questions or proposing improvements! It is just a real passion of me to understand - especially the physics of such an outstanding concept as the junk rig.