Silk sails?

[probablynot]

I'm in limbo.
I've had one eye's cataract sorted out, but until the other one's done I just can't focus on my model-building.  However, that doesn't stop me thinking.

And I've been thinking about sailmaking.  Model ship sailmaking, I mean.
Here's the problem I've been looking at:

Scaling down:
Let's assume that a full-size, working ship might have a set of sails made from 10-ounce canvas.  Then scaled down, a 1:12 model (for example) would need sails made from 0.833-ounce fabric.  That's ten, divided by 12 cubed (since weight is 3-dimensional) but re-multiplied by 12 squared (since area is 2-dimensional).
Similarly (and more relevantly to the sort of models we usually seem to make), a 1:50 model would need sails made from 0.2 ounce fabric.

My experience so far is that the fabric supplied in most kits for sailmaking is 3-ounce unbleached calico, and we usually just accept this and hope it'll look OK when it's on the model.  But sadly it doesn't.  The 3-oz sails on my USS Enterprise (see the build log) are equivalent to sails on the real thing weighing nearly nine and a half pounds per square yard.  The mainsail alone would have weighed well over half a ton.  Dry!

Using overweight fabric for a model's sails simply makes them look like lead sheeting.  Yes, you can impregnate them with glue (or whatever) and shape them to look as though a wind is blowing, but it's still difficult to get away from that leaden look.
So how can we get a look that simulates the way a real sail would hang in a real situation?

Seems to me we have to use a fabric that's so light that it would hang like the real thing in a dead calm, and could flap like the real thing in a breeze. So it has to be as close to the scaled-down weight as possible.

I've been looking at model sail alternatives.
Cotton does come in weights lighter than 3 ounces, but this is more often achieved by having less threads per square inch than by using thinner threads.  So the lighter the fabric is, the less likely it is to look like scaled-down sailcloth.  And it's flimsy and transparent, so it's less easy to sew using ordinary sewing cotton in your average household sewing machine.
The synthetic equivalents to cotton seem usually to be as heavy as - or even a bit heavier than - the real thing, so there's little point in going there.

But ...

Silk.

Considering silk brings up an extra problem.  They don't measure its weight in ounces per square yard, or even in gsm (grams per square metre).  It's Mommes.
I came up with these [approximate] constants:

1 ounce per sq. yard    is    33.94 gsm
1 momme            is    3.75 gsm
so
1 ounce per sq. yard    is just over 9 mommes


There are two types of silk that look to me as though they could have a texture and weight that's useful in the model sailmaking context.  One is Fuji silk; the other is Habotai silk.  They're both quite basic weaves.  Not shot silk, or anything too like the exotic stuff loved by dressmakers or handbag makers!  And they seem to come in very light weights.  You just have to select a colour that's close to unbleached calico, or dirty white.
Or maybe your model calls for a lightly-tarred look ...

Today I've sent off for some samples of some Habotai silk.  Light ones, less than 6 mommes.  I want to discover how easy they would be to sew with my Toyota EZ-1 sewing machine.
I think I ought to look for some fine sewing machine needles too ...

[NenadM]

Even I have almost zero experience, I was thinking about silk sails for my CS, so I'll follow this topic with enormous interest.

 

Thank you for posting this!

[probablynot]

The first of my silk samples arrived in the post today.  Tiny samples, about 3 inches square, so I haven't tried stitching them yet, but I've photographed them and done a bit of assessing.  Here they are:



Each picture equates to approximately one square centimetre.

The first picture is for comparison.  It's the sail fabric that came with my Constructo Silhouet kit.  It weighs 2.94 ounces per square yard.  The thread count is approx. 68 threads per inch (tpi).  It is natural, unbleached calico.  It feels rough, looks rough; The colour could possibly be called cream, or off-white.

The second picture is of Habotai silk that the importer describes as "natural, prepared for digital printing".  It contains some dressing which stiffens it slightly.  It can easily be washed out after the printing stage.
Its weight is 9 mommes, equivalent to 1 oz per square yard or 34gsm.  The thread count is 112tpi.
I like the idea that it can be printed on.  Think of sails with motifs on (eg, the big red cross on the Santa Maria, or company badges on VOC ships).  You could print all the stitching lines, even the reef points.
I don't think he stiffening is enough for the fabric to be fed through an ordinary home inkjet printer.  But glued to a sheet of paper at the top edge, it ought to work.
The fabric is smooth but not shiny.  It's a straightforward even weave, looks the same both sides, doesn't snag like some satin weaves will do.
It's slightly translucent, but not unacceptably so.  It hangs and lays beautifully (by which I mean it falls naturally into shapes that nicely simulate the shapes that full-scale fabrics would take).  The fabric can be dyed, but it does come in many colours, several of which are acceptable simulations of sail canvas.  

The third picture is also Habotai silk, similar to the previous one but it's lighter in weight and it isn't stiffened.  Its weight is 6 mommes - ie, a mere 0.6 oz per square yard, or 20gsm.  The thread count is 178tpi.
The colour of my sample is 'called natural', which seems to mean slightly off-white.  Like the previous sample it hangs and lays very well.  Probably even better, as it's about half the weight of the other sample, but with only a 3-inch sample it's hard to judge.
It would be difficult to print on this.  It could probably be done, though, by using an iron-on transfer.

I've ordered a metre of each fabric, and when they arrive I'll be testing them for printability and sew-ability.  I've already learned that I'll need to use special sewing machine needles ("Microtex" needles, size 60/8), so they're on order too.

I hope this will all be of interest to some of you.  I'll let you know in due course how I get on with my testing.

[DBorgens]

Brian,

While I am a long way from making sails again, l am interested even if, or perhaps because, my next sails will likely be furled. Thank you in advance for sharing your rsearch.

Dave B

[Jaager]

Since silk is animal protein and not cellulose, what sort of glue works with it?   The chemical stability is different - chlorine bleach will denature it.  Normal PVA is close to vinegar in pH,  but pH 7 (neutral) PVA for book binders is available.

 

Given your thread counts,  how does a 600 tpi cotton pillow case fabric or a 1000 tpi fabric?  provided you can find something that is not Sateen.

[probablynot]

Those thread counts for bedding are scams. They count 'threads' wound around each other to make real threads, and so a fabric listed as 1000-count can really be about 80.

[bridgman]

You are attempting to solve an age old problem that seems to go on forever. Using weight as a scaling factor 

will not give an accurate comparison  since weight does not scale linearly  as length measurements do.

Your method of comparing actual material samples is probably the best method for comparison to a  particular model.

[Jaager]

I wondered about the bedding, thanks for the info.

 

Bridgman brings up a factor similar to the one that I realized about scale - we are dealing with 3D objects.  Rather than weight, try a volume comparison.  Getting an accurate value for thickness of a compressible material is similar to getting a diameter for scale rope.  But I am guessing that if fabric volume could be scaled down, the material behavior would also scale.  Unfortunately, I suspect that few fibers hold up at 1-2% of real world.  If silk can be woven as its original worm extruded mono filament we might get somewhere.

[probablynot]

Obviously. You can't get total accuracy. You can't scale wood grain down. Thin thread will never hang like heavy rope (look at 99% of the hanks we all hang from our belaying pins).

I'm not looking for total accuracy. I'm looking for better. I'm experimenting, and I'm sharing my experiments with you. Even hoping for feedback and new thinking.

Interesting question about glues on animal protein though, Jaeger. I'll watch out for that.

[Dee_Dee]

Brian,

 

Good stuff!

 

The photos are a bit fuzzy, but the weave patterns on these fabrics look interesting.  

 

The first photo is woven with a plain weave.  The quality of this fabric can be seen in the individual fibers, there's lots of 'fuzzies', an indication of very short fibers and low quality.

The second photo, the weave looks like a form of a basket weave.  There are multiple individual threads going up and down with each pass.  In some spots, it looks like two threads, but on the upper left side of the photo, there appears to be three threads going up and down.  Since these three threads are not twisted together, they count as three individual threads.  Going left to right, there appears to be multiple threads with each line.  Of note, there are no fuzzy ends, indicating a higher quality fiber and fabric.

The third photo, it's a bit blurry, but the weave appears to have a slight diagonal pattern, indicating a twill type weave and the fibers look a bit heavier.  The twill weave will allow a bit of 'billowing', but only in one direction and it will reduce the vertical stiffness.

 

Fabric weave is really difficult to photograph, would it be possible to get better photos?

If it's sunny tomorrow, I'll try to get some photos of my Giza 45 pillow cases and the two pima cotton fabrics I've suggested to compare fiber size.

 

Have you looked at any linen fabric?  I have a few meters of various linen fabric but have yet to find the ideal linen.  I've been looking at the linen used for needle work.  I have a square of 40 count linen, but the weave is a bit open and the fibers have the typical linen nubs.  I've been looking for a sample of 50 count linen. 

 

Another fabric I'm looking for is tencel.  Tencel has properties similar to linen but is stronger and can be woven in a tighter weave.  Due to their properties, tencel and linen would not be good for furled sails.  

 

Interesting stuff! 

 

Here's a screen shot of where I see three threads running on the verticle. 

Edited May 19, 2016 by Dee_Dee

[probablynot]

Interesting, DeeDee.  So I had to set the camera up again and try to get closer, clearer.  Here's my setup.

 

 

The following pictures were taken on a black background, to show the individual threads better.  First pic is of the 1oz 'prepared for printing' silk; the second shows the 0.6oz silk.

Seems to me that what we are seeing isn't separate threads.  It's one, flattened thread, and we're seeing the individual filaments of silk as produced by the industrious little silkworm!

 

 

 

I had a bit of a shock this morning.  My two 1-metre sample lengths are going to cost £36 ($50) including delivery!  I might go ahead and buy them anyway, but I'll check out some alternatives first!

Edited May 19, 2016 by probablynot

[JerseyCity Frankie]

I think the issue is more how you treat the fabric you have. In my opinion there will be very few fabrics that can behave "in scale" without some help from some sort of coating. Diluted white glue or clear acrylic medium or any other transparent coating that will cause the fabric to set in a way the model builder can exert an influence on are going to be the deciding factors. Flat sails, sails set with no belly added to replicate the effect of being filled with wind, are going to be fairly easy to replicate without worrying about coatings starches or stiffening agents. But any sail that has to show scale wrinkles or droops or sags is going to require a coating that will dry hard enough to fix these topographic features into the fabric. Humidity and environmental factors can change the way the fabric behaves over time and you may wind up with sails changing their dimensions in one way or another, this too is reason enough to seal the fabric with a coating. I think its easy to fall into a mode of thinking about sails that is too literal. Replicating realistic sails will require some out-of-the-box thinking. But I think the addition of sails is worth any amount of trouble since the final effect is always better -in my opinion- than a model without sails. Check out my HMS Victory build log for a deeper discussion about how I make sails.

[wefalck]

Just a comment on silk as material: it causes serious preservation problems. Silk sails and flags deteriorate fast, so museums today typically do not accept models with silk fabric anymore.

[robnbill]

Have you looked at cotton lawn cloth? I used lawn cloth to make sails. It is 100% cotton and is made with the finest cotton threads I could find. For flags, I printed them on iron on transfer materials and used a hot iron to transfer the patterns o the cotton.

 

Anyway, as Wefalck mentioned, silk is not stable over time so you probably want to look at cotton. I have looked at Linen but have been unable to find any that was smooth enough to use for rope making. I have not found any linen fabric that is a s light as the cotton lawn.

 

Lawn cloth would have to be treated to hold any shape.

[EdT]

I am not familiar with lawn cloth, but linen cloth that was commonly used for drafting before synthetic film (I'm dating myself.) is quite fine and free of "slubs" - those little bumps in spun linen thread that effect our use for rigging line.  Linen drafting cloth was coated with wax that can be removed by boiling leaving a fine cloth.  Although I have not used it since my last model airplane (1950's) silkspan, a non-woven, fine material, is commonly used for sails.

 

Ed

[dvm27]

Realistic looking sails are always problematic. They are indeed hard to scale down without looking too bulky. Look at the beautiful examples from French contemporary ship models below. David Antscherl has given a very good explanation of sail making in volume four of the Swan series and Admiralty Models workshop http://www.admiraltymodels.com/Workshops.html. While a bit labor intensive, they are very convincing looking sails.

 

 

 

[probablynot]

Thanks for the comments everyone.

JerseyCity Frankie - I did take a look at your Victory build log.  My first reaction was to think I ought to just shut up and let the experts do all the talking about sailcloth and stitching.  Your work is beautiful, so intricate, and you're clearly well experienced in what you do.  Me, I just have a general interest in stitchwork and the fabric it's done on.  Although as a teenager some sixty-plus years ago I did spend a fair bit of time making a nuisance of myself in Gowens sailmaking loft, looking at the tools of the trade and the fabrics used, and learning some techniques.  By the time I was seventeen I had made three different sets of sails for my kayak, each one with the proper gore, battens, bolt ropes and all.

But I think we approach the matter of sails on model ships from slightly differing viewpoints.  I'm less keen to have the sails fixed in a rigid billow, though we do share the desire to use a fabric that can be wrinkled and creased to resemble furled sails.
I would prefer my sails to hang with that same languid slackness that real sails have in a flat calm, but to look as though they would billow if the wind blew.

Wefalck - thanks for looking in.  Those links in your signature are fascinating.  I accept that silk may have a lifespan measured in decades rather than in centuries.  I'm not making museum-quality models, and if any of them were to last beyond the end of my children's lives I would be very surprised.  I suspect the same sort of thing could be said for 99% of the models that feature here in MSW.  So I still think it's valid to look at the lighter grades of silk in this context.
However, Robnbill and EdT are right - I was overlooking the fact that there are grades of lawn and linen that are just as light as some silks.  Maybe I'll look into that later.  I've never worked with silk before, and I'm just curious to try.
---------------------------
My working samples haven't arrived yet.  But I've now got my special, fine sewing machine needles (Microtex 60/8) and I was tempted to see how my initial samples (3-inch silk squares) behaved when I tried sewing them.
First attempts showed me I had to reduce the stitch tension drastically - from a 4 (for ordinary kit-supplied fabric) down to 1.  Stitch length had to be reduced too.
Those Microtex needles have really small eye-holes - my machine's automatic threader refused to work, and I had to do it manually.

In each of these examples I stitched a 2mm-wide crease, then folded it over for a second run of stitches to simulate a flat seam.  Ignore my wavy sewing - I wasn't trying for neatness and I did it all without actually measuring anything.  :

First picture shows Fuji silk, 73gsm (roughly 2.15oz).  Fairly easy to manipulate and sew.



Second and third pictures show Habotai silk, 36gsm and 35gsm respectively (1oz).  Slightly more care required, and a slower stitching speed, but the fabric behaved OK.



Fourth picture shows the best I could do with fine Habotai silk, 20gsm (just over ½-ounce).  My 60/8 needle - the finest I could find - wasn't fine enough for this fabric.  The weave was seriously distorted regardless of sewing stitch length or tension.



When my larger samples arrive I plan to try some more serious cutting/folding/sewing exercises.  I also want to see if I can use "iron-on" fabric transfer paper to (eg) put a pattern of seam lines and/or a motif onto a sail.
 

[Dee_Dee]

Brian,

 

These puckers are normal for sewing single layer silk and other light weight fabrics.  

 

There are numerous tricks to minimize the puckering.  These two tricks will reduce the puckering by maybe 80%.  There are another 5 tricks that will reduce the puckering more, but can cause other problems.   

 

This first trick is easy:  Put a small piece of scotch tape over the needle hole in the machine throat plate.  This will reduce the size of the needle hole and will keep the fabric from being tugged down into the throat plate as the needle goes through the fabric.  The tape will not interferer with the machine operation   

 

This second trick is the best way to reduce most of the puckering, but can cause other problems, especially with silk fabric if not done with a light touch:  Hold the fabric in front of and in back of the needle and hold the fabric taut.  Start sewing and allow the feed dogs to do their job of feeding the fabric.  Before starting to sew, make sure you have hold of both the top and bobbin thread.  This takes some practice to get right, so try it on some inexpensive light weight fabric to get the feel of holding the fabric taut and letting the feed dogs do their job. 

>>>The problem with using this technique with silk is the high possibility of significantly distorting the fabric weave in the spots where you held the fabric.<<<

 

Also, when experimenting, instead of using same color thread on top and bottom, use high contrasting threads, for example a red and yellow thread (in the same weight.)  This will help  to make sure the machine tension is perfect.    

 

Due to working, I've not had the chance to take photos of the various fabrics I have.

[probablynot]

Thanks for that DeeDee. I hadn't thought of the scotch tape idea. Holding both ends of light fabric used to come naturally with my old foot-operated sewing m/c, but my current one's automatic so I've sort of got out of the habit.

And yes, I should've used different colour threads. Next time ...

[Kevin]

just to throw my spanner in the works, I am looking at using Calico, but no doubt that comes in different grades, I have 37 to make on my victory build, and hope to be in a position to start making them before long, so following this thread with intrest