Notes on designing curtains with rings

This is a post where I point out things to future Alice which should have been obvious to past Alice when she bought curtain fabric for her spare room.

For context, I am making curtains with 40mm metal eyelets at the top.  I need metal eyelets because I am interlining with blackout fabric, and making a polycotton lining too, plus buckram.  This is quite a thickness to expect shonky plastic rings to withstand.  Therefore, I purchased some badass tools to insert metal eyelets from Hanolex.  It involves the biggest hammer I may ever own.  The break even point in tool investment seems to be when you are making more than one set of eyelet curtains, especially when you take into account the recorded delivery postage in sending your beautifully handmade curtains off to someone else to butcher in the eyelets for you.

I like to use the blackout as an interlining rather than a lining fabric.  When making curtains, one tends not to take the lining all the way to the curtain edge.  I find the blackout fabric to be quite stiff compared to some fashion fabrics used for the front of the curtain, so when using it as a simple lining, you end up with floppy leading edges.  I've even seen this done on sample curtains in a well-to-do posh national chain of stores.  Eugh!  Therefore, I like to use the blackout as an interlining, cut to the finished curtain size and tacked inside, making all the curtain the same stiffness.  I then use a thin cotton or polycotton to line the curtain.

Designing your curtain

Here follow some complicated trade-offs and calculations that I won't repeat here, but consider the following constraints:

1. For a pair of curtains, each curtain needs to be at least twice the width of the window area it has to cover.  E.g. for a 2m wide window with a pair of curtains, each curtain needs to be 2m wide or more.
2. You need an even number of rings, and Mr Google suggests typical spacings for 40mm rings are 16-18cm.  The spacing you choose is to some extent determined by how far your curtain pole stands off the wall, which should be just a little more than half your ring spacing.  Clearly if you have M rings, your total curtain width will need to be M x ring spacing.  This allows you a bit of leeway to play around with the number of rings, the ring spacing and the total curtain width, subject to the next constraint.
3. Typical fabric widths might be 140mm, so you may need several drops of fabric in order to get your total curtain width.  Consider where the join should be between two drops: it ought to fall in a furrow between two rings, i.e. with an even number of rings each side of it.  Make sure your ring spacing and number of rings allows you to get this seam positioned correctly, bearing in mind the width of your fabric (which is fixed).  It's not as easy as you'd think.
4. With all of the above, make sure you're allowing for turnings at the edges on the front fabric!  On top of the finished curtain size, I allow 5" at the top, 3.5" at the bottom, and 2.5" at each side.  I allow a 1/2" seam to join drops.

Having solved that headache, you might want to consider where you place the seams between drops of blackout interlining and polycotton lining.  I didn't want a stiff section in my curtain, so I plan to stagger the seams.  Since it's the next stiffest fabric, I'll place the join in the blackout lining in a different furrow from the front fabric (with an even number of rings each side).  The polycotton lining is relatively thin and floppy, and in any case is on the reverse, so I'll place the join mid-way between that of the front fabric and blackout, even though this is on a projecting fold.

The narrowest drop of front fabric probably looks best toward the outer edge of your window on each curtain.

For lining, I plan for it to appear on the reverse of the curtain as a neat rectangle coming 4.5" down from the top, 2" in on each side, and 2" up from the bottom.  I allow 1/2" for turning on the top and sides, but 2.5" for creating a deep 2" machined hem along the bottom.

I will use 5" buckram along the top edge for stiffening, and I expect to place my eyelets with their midpoints 2.25" from the top of the curtain.  The outer two eyelets are spaced by half the eyelet spacing from the curtain edges.

Matching obvious patterns

I should have unrolled several meters of the fabric from the roll and stood well back before buying.  What I thought was a lovely swirly pattern turns out to have VERY obvious repeats which really ought to be matched.  What is more, my guestimate of yardage in the shop was pretty good at first sight: I bought 6.5m, and my later calculations show I need four drops of 62" each, which is 6.3m with no pattern matching.  Clearly this is very frugal of me, but I am now cursing myself for not allowing enough for pattern matching.

When you are standing in the shop with a winging toddler, how much extra fabric should you buy for pattern matching?  Turns out to be quite simple.  Firstly, measure the vertical pattern repeat.  Then buy enough for the number of drops (N), plus N times the pattern repeat.  So for my fabric, I should have bought 4 x 62", plus 4 x 24" (because 24" is the enormous pattern repeat distance).  So I ought to have purchased just over 8.8m to be safe.  That's quite a difference from the 6.3m!  I now need to attempt to buy another 2.2m+ of the same fabric, six months after my original purchase, if they still have it.  ARGH.

Continuous bias tape: a magic formula

I'm not going to start this post with an apology for not posting, because I don't have to!  Mwa ha ha!  I've been too busy with paid work and chasing a toddler around.

I have been knitting (as always) and I'm still sewing, but I can't post anything yet.

Nevertheless, I thought I'd spend a few moments in thought about continuous bias tape.  The Coletterie has a lovely tutorial on how to make continuous bias tape, which I highly recommend.  I'm not going to try and give you such a tutorial here when a good one already exists.  However, having followed their tutorial measurements to the letter a few times, I note that I do not get the whole number of 1" strips marked across the fabric as they show for their step 4.  The result is that I get less than the 100" of bias tape they propose.  This is probably because I do not edge-stitch as they suggest in step 2: instead, I use a 1/4" seam to match the one they suggest in step 6. 

I like the 1/4" seam and I do not want to do their flimsy edge-stitch.  Fortunately maths comes to the rescue and I bring you the magic formulas for cutting continuous bias tape.

For this method, follow the Colette tutorial, but substitute my measurements.  Also, use a 1/4" seam in steps 2 and 6.

I want to make a total length (T) of bias strip of width (w).

When I come to draw my lines on the fabric (see tutorial step 4), I'll need to draw enough to separate my fabric into n strips, where n = sqrt(T/(2w).  I'll have to round up to the nearest integer to make sure I get at least length T and I don't end up with half a strip.

I therefore need to start in step 1 with a fabric square of side D = n*w*sqrt(2) + 1/2".

So, to put numbers in my example, say I'd like to make 98" of bias tape of width 1".
w = 1"
T = 98"
So I need to draw lines in step 4 to separate my fabric into n = sqrt(98"/2") = sqrt(49) = 7.  I need to draw 7 strips.
n = 7
Now I need to start in step 1 with a square of D by D, where D = 7*1"*sqrt(2) + 1/2" = 10.4"
D = 10.4".

Therefore, I ought to have started the Colette tutorial with approximately a 10 3/8" square, not a 10" square.  No wonder it went wrong!

Don't stop there: use the magic to find the square size D for however much tape of whatever width you want!  THE POWER IS IN YOUR HANDS!

Cascading ruffles: a happy how-to

Ruffles remind me of bad 80s wedding dresses.  Especially the kind of ruffles that are gathered where they attach to the seam: eugh.  They stick out all stiff and bulky.  I do not like this kind of ruffle.

I do like the sort of ruffle that attaches to the seam smoothly, but then flares out along the loose edge.  That feels a bit less granny and a bit more modern.  These ruffles cascade beautifully, instead of sticking out nastily.  See the neckline of this pretty dress by Kathleen at Grosgrain Fabulous:

How to make such elegant ruffles?  We can deploy the magic of maths*!

For the ruffles to fall as they do, the outer (free) edge must be longer than the inner edge (attached to the garment).  Cutting a strip of fabric on a curve gives you an outer edge longer than the inner edge, so Bob is your uncle (or in my case, my mother's aunt). 

For extreme ruffles like the above dress, it turns out that the outer edge of the ruffle must be about 5 times longer than the inner edge (go on, measure it on your screen with string!).  The curve to do this will be so extreme that each end of the strip will meet, i.e. you'll cut out a circle with a hole in it.  Actually, whatever length of ruffle you need, it's easier to calculate as a section of a circle.  I like to calculate what I need to make, rather than trial-and-error cutting.  Fabric is expensive but a little time spent in thought before you cut is free.

First, decide on the width of your ruffle, we'll call it "W".  I want my ruffles to lie on my shoulders like the dress pictured above, so I'm going for a width of W = 8cm.

Next, decide on how much longer you want the outer edge to be compared to the inner edge, call this "X".  For extreme ruffles, I like X = 5, which means your outer edge is 5 times longer than your inner edge.  Lower values of X will give smaller ruffles, while bigger values will ramp up the ruffle mania.

Now imagine drawing two concentric circles (a.k.a. a picture of a doughnut).  The inner circle will become the edge you attach your ruffle on, while the outer circle defines the edge that waves free in the breeze.  Okay, you need to slice through your doughnut at the end to get your curvy strip, but that's fine.

Your generic ruffle pattern.  You're welcome.

The inner circle has a circumference of C = 2 x pi x r, where r is the radius of the doughnut hole.  We'll work out what C and r are going to be in a minute.  The outer circle has a circumference X times larger, because that's the type of ruffle we've chosen.  So, it'll be "X x C".  The radius needs to be r+W, so that when we cut our doughnut open, we have a curvy strip of width W.  So, the equation for this is (X x C) = 2 x pi x (r+W).

Like me, you were probably forced** to take a GCSE (or equivalent) in mathematics, so you can solve your two simultaneous equations for r:

C = 2 x pi x r
X x C = 2 x pi x (r + W)

Full marks if you got the solution r = W / (X-1).

For my happy case of X = 5 and W = 8cm, you get r = 2cm.  So, you draw a doughnut with the hole having radius r = 2cm and the outer edge having radius r + W = 10 cm.  Easy!  (Don't forget to add seam allowances, and finally cut though that circle to get your strip in the manner suggested on the diagram.)

How much ruffle does one circle buy you?  For this you'll be wanting the circumference of the inner circle, which is the edge you'll use to attach your ruffle.  That'll be 2 x pi x r, which is about 6 x r.  For me, that worked out at just over 12cm.  To get a longer bit of ruffle, you'll have to cut out many circles and join the strips together along the straight-cut edges in a tedious manner.

At the end of your ruffle, cut the edge into a pretty curve (see dashed line).

Hemming your ruffles is a pest because of the curve.  You can use a rolled hem on an overlocker in matching (or contrast) thread for minimum hassle, but you have to be okay with that look.  Otherwise, enjoy your painful rolled-hemming!

* I am prepared for more ridicule by posting this.  It's not as bad as the peg bag tutorial though.
** mandatory but it was a pleasure!  ;-)

What's (geometrically) wrong with my peg bag.

I should be cleaning the kitchen, but instead I bring you: what is wrong with my peg bag (geometry edition).

From this ugly photo of the ugly bag, I think part of the wrongness arises because the sides of the bag taper inwards towards the bottom.  I'm not talking about the round-bottom shaping, but the slightly sloping sides.  They're not sloping enough to be style, just enough to be looking like a mistake.  And that's because they are.

I put two darts in the bottom of the bag to give volume for pegs.  I failed to calculate that this would make the slides slope inwards and make the bottom of the bag narrower than the top (well, duh).  In the interests of making my pattern correct next time, I thought I'd calculate what I was supposed to do when I drew those darts on my pattern.

Towards the bottom of the bag, my pattern should look like the image above with darts of angle theta and side l, and dart points separated by a distance x.  The total width of this pattern piece is z.  I also marked distance d and delta, which come in handy later.

 

When I come to sew up the darts, the bag will appear to have a width w when viewed from above.  This is narrower than z marked on my pattern piece, because the darts have given the piece some three-dimensional shape.

When viewed in cross-section, lying on a flat surface, the middle of the fabric will rise above the surface by height h (at least, if it were made of stiff card, not floppy cotton).

The interesting quantities to me when I design the bag are theta and h.  The dart angle lets me choose how "boxy" I want the bag to be:  theta = 0 is no shaping at all, theta = 90 degrees is a square bottomed bag, and intermediate values give gentle shaping.  The quantity h allows me to choose how wide back-to-front my bag will be inside.

From my choice of theta and h, I can calculate some useful design parameters which help me to draw my darts in later.  Time to get out my calculator and calculate these numbers:

One can also find a few more useful numbers, in particular x and l which tell me the separation of my dart points, and how long the dart sides should be.

Finally, the total width of my pattern piece at the lower end should be z:

... while at the top by the hanger (where there are no darts), it should just be width w.  So, there we have my mistake, my friends: it's equal to z-w, if you cared.

I can now re-visit my wrong pattern with real numbers.  I'd just drawn darts in at random, but it looks like I used an angle of theta = 28 degrees, with a dart length of l = 4.5cm.  The top of my bag has width w = 34cm.  I find that tan(xi) = 0.601 (that's tan(squiggle), where "squiggle" is clearly the best greek letter), and  d = 3.86cm.  It means the inside of my bag has h = 3.1cm, but really it's 6.2cm across because I used darts in the back and front bag pieces.  Now for the interesting bit: the bottom of my flat pattern (near the darts) should have been z = 37cm across, not the 34cm that I used.  That's an 8% error: enough to make my bag look crap.

Time to re-draw!

I'm also going to change the mouth of the bag, add elastic loops inside to hold the hanger up, and a small fabric loop on the outside at the bottom to pass the hook through when you fold the bag in half for storage.

Sorry for the algebra explosion: if anyone thinks this sort of thing will actually be useful for them then drop me a comment and I could make a calculator-type thing, if maths isn't your forte...