The dance continues

(Warning: this one is very geeky.)

The piece is more than 25% woven now; last night we turned the cloth beam again. Another turn and it will be time to unchain and lengthen the warp.

Draw-in is now visible and obvious, since the recently woven web is now wound over a layer of previously woven web. It's about 1" narrower at each side now. I'm not worried about this from a construction viewpoint, because I figured in a 10% draw-in when I was making calculations. But it does mean the top corners of the chiton will have a slightly trapezoidal effect. Hopefully nobody will be looking at the top edge anyway.

During this recent period when I wasn't weaving on a regular basis, something happened to my understanding of the dance. I've now hit on a sequence that is a great deal faster than the previous ones I tried. I've just about doubled my most recently clocked speed for a "there and back again" pass. Now it only takes two and a half minutes rather than five.

Here's what I'm currently doing. I wrap the left selvedge, then put the penion through the natural shed. I use the comb to loosely beat the weft into place, still on the open shed. Then I wrap the right selvedge, and here's where it gets a little tricky. With my left hand I pull and hold the sinister (rightmost) shed rod. I place the penion on the natural shelf formed by the heddles. With my right hand I take the weaving sword and draw it through the countershed, clearing any sticky yarns. Then I put the sword under my left arm, pick up the penion, and put it through the countershed. I drop the first shed rod, move to my left, and repeat the process with the second shed rod. Then I move back to the sinister edge of the warp, drawing the tip of the sword across the threads as I go as if I were using a pin beater. This helps the countershed fall back into place so the original, natural shed is clear again.

I beat in sequence from the sinister selvedge. Holding the handle of the sword in my right hand, I put the blade in the natural shed at about a 45 degree angle to the fell, with the hilt nearest the fell at the selvedge edge. Then I lever the point of the sword upward with my left hand. When the flat of the sword meets the fell, I push up gently, just until I feel the weights moving. Then I step to my left and repeat the process. It takes from four to six such movements to cross the entire warp, depending on how the weft thread lies.

You'll notice here that in both passes I am beating from the attached toward the free end of the weft. This is a trick I learned while weaving rutevev tapestries. I usually use my free hand to adjust the weft tension while I beat in this fashion. Managing the tension in this fashion helps me produce cleaner, straighter selvedges with less draw-in than any other method I've tried. Beating this way adds some walking to the process, because it means I have to cross the loom (from dexter to sinister) twice: once after passing the second weft and again after I've beaten to leave the sword at the sinister end of the loom. However, the speed I gain by not having to manipulate the sheds as much more than makes up for it.

Besides, the weavers of the Classical literature all walked up and down at their looms. Next I should work on being able to sing as I go.

That was painful

After weaving an inch's worth of picks (24), I took a close hard look at what I was getting. The last inch of warp at the left (dexter) selvedge was weighted twice as heavily as the rest of the warp. This was due to my seat-of-the-pants call that two inches of warp was an appropriate number of ends per warp weight.

However, the texture of the cloth at that more heavily weighted edge was completely different from, and preferable to, what I was getting elsewhere across the loom.

It was a hard decision, and I dithered for a couple of days before actually doing it. But I elected to untie and redistribute the weights across the entire warp. I didn't have enough extra loom weights to do the whole thing at 12 ends per weight, though. I did have enough to use an inch and a half (18 ends) per weight rather than the two inches (24 ends) that I had on the loom originally. So I've re-weighted the entire warp.

Now I'm working on beating that first "inch" of picks more tightly together, because it measures somewhat more than one inch in length. The heavily weighted area measured something less than an inch; I will also try to spread out the weft in that section a bit so the density is the same all the way across the warp. I am hoping I have found a happy medium that will give me a nice texture of (mostly) evenweave.

I'm also happy that the weights are distributed completely evenly across the warp now. Also, because it is more firmly stretched the warp makes a different and more clear sound now when I strum it. I feel much more optimistic about the project now.

Back to work!

Warp is weighted!

I have finished putting the weights on the warp -- all except for the selvedges, that is. Each selvedge is a pair of heavy single-ply wool yarns. I'm not sure they're strong enough to use a whole loomweight for each one, but I'm not sure what other choice makes sense. I want them to hang very straight and stiff so I can reduce the tendency to draw in while I'm weaving.

Here is one of my loomweights. They are terracotta, based on finds from the fourth century BCE in Macedonia.

And here is about half the warp, in the late afternoon sunshine. I opened the window to give more natural light for the photo, and the dramatic contrast made most of the photos useless. I kind of like this one, though.

Next up: heddle knitting!

More on warp stats

Well, I was right that my warping would get faster. I noticed after the first 18" or so that I had noticeably quickened. Near the end of the warp I timed myself again and found that it was only taking me 5:15 to do an inch of warp, as opposed to the 7:45 it had taken earlier.

I picked up speed partly because I found a better way of ascertaining how many loops I'd done. It proved impractical for me to count to twelve while I was doing the loops; instead I would stop and check, either by eye or by actually touching the threads, to see if the proper number were there. After I learned the most efficient place to look, I didn't have to touch the threads to count them, which picked me up some speed. I also got better at the peg-wrapping sequence. But mostly I think it was the mood music that helped, since the improvement in speed happened on the first day I began listening to music while working.

I timed how long it took me to advance the warp of the header band: approximately 1:30. I figure I advanced it about once for every six inches of band woven. So if I advanced the band warp about a dozen times that makes 18 minutes of advancing the warp.

I'm going to call that as an overall time of about seven and a half, maybe eight hours to weave the header band all by myself on equipment I'd never used in that way before.

As I was weaving the header band, my gauge slackened just a tiny bit. Where I should have finished with 72 sets of warp bundles, I have only 69. (Surely this number is the Cypriot's way of laughing at me, or perhaps blessing my patron.) That works out to be a sett of 23 ends per inch rather than the 24 I was aiming for, but I don't think that's a problem.

I'm more concerned about possible take-up, but I factored in a 10% take-up in both directions. My other warp-weighted loom experience suggests that there'll be a greater than 10% take-up in the weftwise direction, but I am hoping it won't be greater than 14% because then the garment might be too small to fit properly. And after all, my experience has been with weft yarns that were woolen spun, not worsted spun. As long as I don't get too much draw-in, the take-up shouldn't be a problem. That's what the big beefy Navajo warp singles yarn is for: chunky selvedge reinforcements to keep the web from drawing in markedly. And if the finished piece doesn't have a lot of draw-in and take-up, so much the better; it'll fit a little more graciously that way.