Triple layers - variation for easier weaving

Yesterday I stopped before doing what I always do: rearranging the treadling for easiest actual weaving:

Unless I have very fat yarns in my double (in this case triple) cloth, I try to arrange the treadling so that I can do two picks per shuttle (layer).
It is so much faster to do two picks before having to change shuttles...

(Yes, it shows in the end product:

These are two not quite focussed pictures of double-layer shawls.
They are approximately 2 x life size, and has the ends one-by-one and the picks two-by-two.
The warps are a combination of cottons, 16/2, 20/2, 22/2 and maybe a few 30/2.
The wefts are of course only one quality per shuttle, but I don't remember which grist. The sett was probably somewhere about 10 ends/cm (25 epi).

And no, I haven't tried it for three layers.)

So: here is what I would do before sitting down to actually weave.

First: use the existing tieup, but rearrange the colour sequence.

Next: rearrange the new treadling to straight:

As I weave from bottom to top, and have an overhead beater, this is the treadling I would use, namely start in the bottom layer, working up to the top layer, as seen in the widest section of the warp.
(Yes, on the loom there will be "gaps" where the layers change. I have never seen these gaps after wet finishing - see pictures above.)

Hm.
Remembering one of my doodlings from yesterday (which did not reach publication) - another way of making more-than-two-different narrow stripes - a shift in the warp sequence can accomplish that:

Of course it depends on the actual colours used etc etc, but something to consider, perhaps?

A challenge? - I always love a challenge...

So, the question was: how to make a three-layer weave, with warp-wise layer changes?
(and preferably on 12 shafts "only")

This is how I approached the problem:

(To all Swedish readers (and Ellen: hi, Ellen!): note that all tieups are for rising shed. This means that the layers/colours will be reversed if the tieup is used "as is" for a CM.)

Started with a three-layer (three independent layers) weave - for plainweave layers, that takes six shafts.
For instance like this:

(To make things clearer in my mind, I threaded the first (turquois) layer on shafts 1-2, the next (purple) on shafts 3-4, the third (red) on shafts 5-6. When the construction is ready, the threading can be rearranged for easier threading.)

I made the top layer turquoise, the middle layer purple, the bottom layer red-orange. The three layers do not interact at any point. (Note that the difference in nuances between warp and weft makes It easy to see that the bottom, red-orange, layer has a correct interlacement, even without using the "back view".)

Now, we wanted a lengthwise (warpwise) layer exchange. I decided the left hand side is a good place. Thus, to start the construction, assume another "block" of the same threading and colour to occur at the left side. Like so (left pic):

The same threading on a new set of shafts (= a new block), with the same colour order. We want to shift the layers, so I let them "cycle": the middle, purple, layer goes on top - the bottom, red, goes in the middle and the top, turquoise, layer has to go to the bottom.

OK, I hear you: how do I do this?
I am using Fiberworks PCW (silver, if that matters).
By clicking in the drawdown, I can get ends/picks to the top (for instance).

Middle pic above: all purple threads, both warp ends and picks, are taken to the top.
Next is to fix the interlacement: right-hand pic above. Note that the interlacement should be a continuation of the purple plainweave in the right-hand (first) three-layer block.


Now to fix the middle layer, the red-orange one. Click all ends and picks so that they are under the purple layer, but on top of the turquoise (left picture below). Fix the interlacement - easy because of the difference between the red and the orange - right-hand picture below.

For the bottom (turquoise) layer, it is easier to use "back view".
As it happens, all turquoise threads are already at the bottom... fix the interlacement, go back to "front" view again:

Unfortunately, all the 12 shafts are now used.
As we want another stripe, we add another block (6 shafts) - now the total is 18.

The same procedure again: make the red-orange layer top, the turquoise will be in the middle and the purple layer will go to the bottom layer:

In the hopes that I had missed something important, I let the software analyse my result - alas, I had not: it really takes 18 shafts to do this.

But...

What if: let one of the layers stay in the same position for two stripes: this will reduce the goal to two blocks. With the three "open" layers on the right-hand side of different widths, and several narrow-ish stripes on the left-hand side... voilà, only 12 shafts. An alternative?

(As I, personally, prefer straight threadings whenever possible, I rearranged it for this final picture, showing both front and back:

(Remember: click pictures to enlargen)

Deconstruction

This is a very old lampshade, it must be at least 30 years old. The construction of the shade itself is very easy: because of the pattern of this fabric, it was made as a cylinder with a channel for a drawstring at each end. The cylinder is slightly bigger than the wire construction, and sewed with attention to the pattern.

The fabric was bought.

The warp is a thin cotton. I think the warp yarn is the same all over, but it is hard to tell. The weft is the same, except for the closer bands: that yarn might be a thin singles linen.

Here is the drawdown:

I gave the presumed linen a different colour, to make reading the construction easier. (The detail picture does obviously not contain a whole repeat - fill it in yourselves)
The crammed warp stripes has 24 ends each; for more open parts I did not even try to count the number...
Also, I did not check the sett, but my guess is that the crammed stripes has double the number of ends per unit (this is the reason I left every other "heddle" empty).

The same goes for the weft: the crammed bands have 12 picks, the two first end the two last in each band go in the same shed. I guess that the linen weft is double the thickness of the other yarn (and I think the other weft is the same as the warp).

Each square is roughly 2 cm - maybe 1/2 cm for the crammed stripe/band, maybe 1 1/2 cm for the open square.

What I have been doing - nerd alert

It all started with the DFW (Double Flyer Wheel).
I wanted to know where it came from (or, at least, from whereabouts in the country this kind of accelerated driving was known). I had seen something in a (borrowed) book - had to buy it for myself. Was sort of disappointed - there was a wheelwright mentioned, that much was right, but, as it were, only in passing. He had a name and a year (Abraham Hedman, 1738), but that was all there was.
I (thought I) knew I had seen something like it, many years ago, in a museum. Thought I knew which museum. Wrote to them. Wrote to several more museums, while I was "at it". Result: nothing, nada, zilch. A month later, ONE of the (by now 6) museums answered: "I don't have time to go out in the storehouse to look".
Meanwhile, I got a couple of obscure books on ILL, one of them very dull-looking: Den Ångermanländska linslöjden, en historik. (The linen works of Ångermanland, a history). It is a dull text, but once I started to really read it I found lots of, hm, nuggets(?). And lots of food for nerding...

So, here, friends, goes:

Definition of "prime linen": as you (probably will not) remember, "prime linens" come in many classes. They were (at least most of them) 1 1/2 aln (3 ft, or 88-89 cm) wide.
For class 1 there were 2720-2920 ends (or 30-32 ends/cm); for class 8 (which was NOT the finest seen) there were 4120-4 320 ends (or 45-48 ends/cm).
The History book contains some well-described examples, complete with old (and obscure) weights-and-measures.

So I got out the calculator...
One example was described as 39 alnar, 4 1/2 skålpund; 3600 ends, woven in 2299-dent reed.(If I did my calculations right, this means the cloth was 23,2 m long, weighed 1,9 kg - which would translate to 81,9 grams/metre fabric).

OK.

3066 ends x 23,2 metres of cloth = 83 520 metres of warp yarn.
But we have a weft, too. For convenience, I assumed 40 picks per cm (warp density comes out as 40,5) - which gave me another 83 500 metres.
Which gives a total of about 167 000 metres per 1,9 kg of weight.

Now, linen is numbered in Nel (Number English Linen, I believe), where the number tells how many skeins of 300 yards is accommodated in 1 imperial pound.
One yard is 0,91 metres, so 300 yards = 274 metres (rounded), which means we have 167 000 / 274 = 609,5 skeins - let's round a bit carelessly, say 600 skeins.
As an imperial lb (of today) is approx. 0,45 kg, we can go on: ( 600 / 1900 grams) x 450 = 142 (rounded)
which means this particular cloth was woven with a linen singles (*all* "lärft" is singles) of a number somewhere in the vicinity of 140/1. Handspun.

I did several calculations, approaching from different angles, 'cos I was thinking I had made a decimal error somewhere.
But - all of them came up more or less the same. There were several examples described, and they all came out with a yarn count between approx. 120 to 150. Handspun!

A picture from the book, divided in two to make it possible (I hope) to read the writings.

The caption says, roughly;
"These two photos show the strip, with thrums, cut off from the cloth for which Catharina Andersdotter got the Illis Quorum [a royal medal]. The cloth is of the seventh class, is woven through a weaving reed of 2 200 dents, contains 40 knots, or 4000 ends of warp. Every square centimetre thus counts to 40 ends of warp and circa 37 picks of weft. We have found this linen strip in the royal archives [... something "administrative" I can't translate] 1808 - 1811."
I wish... Or, on second thought, maybe not :-)

But the DFW?

Nothing much. This book tells about several "spinning schools" and some competitions - it seems that the DFW spinners always won, but that did not convince "the people". One of the arguments was that a DFW is harder to spin on (heavier to treadle, needs more concentration), and therefore, even if you are more productive/efficient for the first hour(-s - the competitions were 2 hrs), you can't spin for as many hours as a "prime spinner" usually did. Right or wrong?

Anyway. I still don't know anything about my DFW, it still doesn't work (too cold to do what needs to be done on it, the workshop is unheated) - and the only accomplishment I have done is to amass a lot more "worthless knowledge".

Here is another picture from the book:

which shows another principle for a DFW: one MOA, two flyers. Therefore it has to have two drive bands, I think - and imagine tying two drive bands so that the tensioning works for both, at the same time...?!?

Waiting for more ILL books, and perhaps I have to do some "hands-on" museum search, and...

Another shirt

At about 55-60 ends per cm (which equals about 135-150 ends per inch), this is a fairly fine shirting material, woven in a colour-and-weave pattern.

The pattern itself is easily reproduced – but to find this fine yarn for a handweaver would be quite another thing.

Here is the pattern – that is, here is the principle of the pattern. I have not tried to count the actual number of ends. The fabric is slightly warp-faced, which I illustrate with somewhat fatter wefts.

In the warp there are odd numbers of warp both for the solid stripes and for the b & w sections. Each b & w has an extra black to begin and end the section.

In the weft there are also odd numbers of each stripe, but the extra black threads are omitted.

Grist-and-weave?

I have always wondered if/how colour-and-weave designs would work with thick-and-thin yarns.

Now I had a good opportunity to test it. I had some cotton 8/2, and some 16/2 that was just a shade darker, so I threaded up a hound's tooth pattern:

This is what I got:

Ho hum.
How fortunate I did not have a thin yarn in the same colour...

So - obviously the difference in size was not enough, thought I.

For the next sample I used 4 ends of a white mercerized cotton for the thick and one end unmerc for the thin, threaded to log cabin.

Okay, this time the pattern shows, but the quality is, er, sturdy. And not interesting enough to merit a re-sleying.
Re-threaded to hound's tooth:

Even more ho-hum (or should that be "less"?)...

For good measure, as there was a bit left of the lilac warp, I re-threaded that to log cabin.
Wove, from the top down,

- log cabin in thick and thin
- plain weave with only thin weft
- 2/2 twill with only thin weft.

Now I have one not-very-interesting (but soft) scarf, and one idea less to bother with.
I think it is called "experience"?

(But we did get a "draft-of-the-month"  for September, at the guild site, so it served a purpose)