Chain stitch hooks · Jersey · Straight stitch · Transverse shuttle · Vibrating shuttle

That elusive flexible stitch

Flexible stitch is necessary for sewing jersey so that it would stretch with the material. However, regular straight stitch of a lockstitch machine is usually not flexible. So what to do?

Popular solutions include using elastic thread, sewing with narrow zig-zag instead of straight stitch or sewing with an overlocker. But what if you really just want to have a neat proper straight stitch sewn with regular thread, and have the seam stretch with the jersey? Let’s look at the mechanics of stitches – and the mechanics of sewing machines.

The ability to produce a flexible stitch depends on seam tension which in turn depends on the upper and lower thread tension setting as well as on the strength of the feed – the force with which the fabric is being pulled.

Thread tension

To make a flexible stitch, the machine must be able to sew with low tension, both on the bobbin and on the needle. Lowering tension on the needle is not a problem in any machine (as far as I’m aware), but bobbin tension can be challenging due to differences in shuttle or bobbin case construction.

Each machine has a minimal tension with which it can sew reliably. Reducing tension below the minimum results in untidy and uneven stitches, thread pulls or skipped stitches. Although there are general tendencies in each type of mechanism, the actual minimal tension varies from machine to machine because it is the result of many factors working together.

In round bobbin machines, the bobbin rotates in place, so that the path of the thread from the bobbin to the needle hole remains the same all the time. The bobbin rotates because something pulls on the thread; that “something” is the material being transported by the feed dogs and the lower thread loop being pulled up by the upper thread. Thus bobbin thread tension remains constant at each phase of the stitch (barring jerks in movement).

In long bobbin machines, the bobbin moves with the shuttle, and the thread path from the bobbin to the needle hole varies depending on the phase of the stitch. Here too the bobbin rotates inside the shuttle because something pulls on the thread, that “something” being the shuttle moving away from the needle hole. Of course the material being transported by the feed dogs also adds to the pull, but this addition is very small compared to the shuttle movement. The path of the bobbin thread is longest at either end of the shuttle race, with the needle hole being in the middle.

A transverse shuttle in its left most position
A transverse shuttle in its right most position

Thus a long bobbin shuttle has two distinct points regulating thread tension, and they may deliver the same or different tensions depending on the spring. The length of thread path from the bobbin to the needle hole is also different at either end, and it is the longer path that determines effective thread tension. At least, that’s the theory. In practice, both tensions are important because of jerks and inertia, and I leave it to you as an exercise to figure out. 🙂

Since the bobbin thread is pulled by the shuttle movement rather than by the movement of the fabric, a long bobbin machine can deliver much lower bobbin tension than a round bobbin machine. Shuttle movement is akin to the movement of a take-up lever for the upper thread which pulls some thread off the spool in preparation for looping a stitch which is done under zero tension. A moving shuttle also prepares a bit of thread to loop without tension.

Among the long bobbin machines, bullet shuttle machines are best – vibrating shuttle and cylindrical transverse shuttle – because their shuttles provide a more effective and reliable tension regulation with a greater range of tensions than most boat shuttles. However, there are exceptions in both camps: some bullet shuttles have awkwardly shaped springs that block thread path so the lowest tension that the shuttle can deliver becomes quite high. Likewise, some boat shuttles have improved construction compared to most others so that thread is not blocked anywhere and the resulting lowest tension becomes very light indeed. So always check your shuttle to be sure, and in particular make sure that the block is not caused by a wrong bobbin, century-old dirt or a warped spring rather than poor design.

Feed strength

The strength of the feed is the combination of the feed dog action and foot pressure. Walking foot attachments or integrated needle feed change the overall behaviour of the feed and need to be considered separately.

In a regular lower feed, feed dogs push the fabric up from below, press it against the foot and shift it to the back (or to the front, for reverse sewing). At the same time, the foot is pressing down on the fabric against the feed dogs. Feed strength is reduced with reducing foot pressure, but also with reducing feed dog height. The latter also has the effect that the fabric is being transported for a shorter period of time on each stitch, and travels a shorter distance. This means that the movement of fabric begins later and ends earlier. This generally results in lower bobbin thread tension (I am not entirely sure why, but tests confirm it).

Transverse shuttle machines have a very gentle feed because their feed dogs are being pushed up by a spring instead of being moved by a fixed shaft like in newer machines, where the force of upward movement is much greater. Spring-driven feed of TS machines reacts differently to changes in foot pressure than the familiar axis-driven feed. In a spring-driven feed, the strength of the spring pushing up the feed dogs must be balanced by the strength of the spring pushing down the foot. Too much foot pressure chokes the feed completely and prevents the fabric from moving. Too little foot pressure also makes it difficult to transport the material and shortens the stitches. The balance point depends on the strengths of both springs but also on the properties of the fabric – thickness, stretch, softness. Raising foot pressure also increases bobbin thread tension; I don’t understand why, but testing shows that it does.

The netto result of a spring-driven feed is that it is much gentler on the fabric than axis-driven feed, and that with the right foot pressure it can deliver a better balanced seam at very low thread tension on very fine jersey material.

Stitching done by a TS machine (Singer 48K) on extra light jersey. Can you even see the stitches? They are less than 1mm long – and the tape measure is there to prove it. 🙂 This seam has 60% stretch.

The exact interaction of various factors in thread tension and feed is quite complex and is greatly influenced by the quality of the machine – precision of manufacture makes all the difference. Too much “play” between the components nullifies any finer adjustments you might make.

Chain stitch

Flexible stitch can also be sewn on a single thread chain stitch machine. In contrast, two-thread knotted chain stitch commonly found on coverstitch machines, has hardly any stretch.

Single thread chain stitch looks like a plain straight stitch on the top side and has loops on the underside. When sewn with sufficiently low tension, the loops are round and serve as a “storage” of thread. When the fabric is stretched, the loops stretch too.

It is easy to calculate the amount of stretch that chain stitch can deliver. When the loops on the underside are round, each stitch “stores” 4.14 times stitch length of thread. When stretched, each stitch requires 3 lengths of thread (one on the topside and two on the underside), so excess thread is 0.38 stitch length, or approximately 40% stretch. There is also a little extra thread stored in the thickness of the material. When using short stitches, it can add up to give you up to 50% stretch.

Thread tension determines how much stretch is stored in the stitches:

Increased tension gives elongated stitches like in the top drawing. These stitches store less thread, so stretch is reduced.

Round loops like in the middle drawing give 40%-50% stretch. This is the best tension: it gives a tight seam with a good stretch.

Wide loops like in the bottom drawing give even more stretch, but the seam is less tight and the stitches may show when the seam is opened flat. In addition, most machines struggle with low tension and are not even able to produce such a stitch.

Narrow zig-zag

Another way to sew a flexible seam is with narrow zig-zag. Here too it is easy to calculate the amount of stretch you’re getting: it depends on the stitch length of zig-zag – the tighter, the better.

However, calculated values of stretch do not work out in practice because the stretch of fabric between two corners that doesn’t have thread over it, has to stretch twice as much as the bit of fabric on the other side of the zig-zig stitch. The fabric is being stretched unevenly and the overall seam stretch is greatly reduced compared to the calculated values.

Still, narrow zig-zag offers a very good stretch and is particularly suitable for thicker knits. On fine jersey you notice that it’s zig-zag – the seam becomes too heavy not only because it’s got a width to it but in particular because the fabric twists around the stitches.


Round bobbin machines are generally not capable of delivering flexible seams with straight stitch because their minimal bobbin thread tension is too high. However, lowering feed dogs can help reduce bobbin thread tension, and some machines can produce seams with up to 20% stretch sewn at half height of feed dogs.

Vibrating shuttle machines can deliver flexible seams with up to 40% stretch on all materials except very light jersey. Although the seam is still acceptable on the lightest jersey, it starts getting uneven in tension – it appears that even with zero foot pressure the feed is still too strong and stretches out the jersey.

Transverse shuttle machines can achieve a flexible stitch with up to 60% stretch on very light jersey without stretching out the fabric.

For comparison, chain stitch gives up to 50% stretch and narrow zig-zag can deliver even more, if the fabric allows. However, both these stitches make much heavier seams than straight stitch.

Buy me a coffeeBuy me a coffee 🙂

16 thoughts on “That elusive flexible stitch

  1. Elena,
    A huge amount of strong technical detail in this post Elena, many thanks. I don’t know of anyone else out there who combines the sewing and mechanical understanding you show. You make a strong case for the comeback of transverse shuttle machines too!
    Dan H


    1. Thanks, Dan! Revisiting old technologies can give new insights. I understand now why TS machines remained popular for so long, even though round bobbin and VS machines were really dominating the market. There’s nothing better than a TS for sewing your finery. 🙂 And for free hand quilting and embroidery too, I noticed! I’ve been working with some extra fine viscose jersey with 60% stretch lately, and my usual VS machine proved just a little too rough for it, and not enough stretch, so I took a closer look at the four TS machines I’ve got. There will be some (more) posts on that too!

      But the surprising conclusion for me was that a TS machine outperformed a chain stitcher, so then I did the maths to see why, and it’s all there! As well as ideas to improve the chain stitcher too… 🙂


  2. Dear Elena
    Thank you very much for this article. It is very interesting.
    How many kinds of machines did you try to find the lowest possible bobbin thread tension for?
    Best Regards


    1. I have tried TS, Wittler TS, VS short arm, medium arm, long arm, vertical rotary single cycle, vertical rotary twin cycle, vertical oscillator side and forward facing, horizontal oscillator side and forward facing and horizontal rotary. Each type in at least 3 variants from different manufacturers, so modified compared to the original. Some types in many more variants from many more makes. Just have a look at the makes in the Categories list. Obviously, I haven’t tried all possible machines, but I tried quite a few.


      1. Oh – thats a lot. So in general the TS are the best for low tension?
        I use my spring dynamometers all the time to set bobbin thread tension and needle thread tension. I hope you use something similar – they are cheap and easy to use. This is one supplier:
        When I try to go lower in bobbin tension on a machine like the Singer 201, then I think the relative variation in bobbin thread tension becomes quite big – like 0.1N to 0.25 N when I pull up the thread. So it may also depend of the design of these bobbin tensioners.


      2. I don’t use any measuring device, I go by feeling. It is not just the strength of pull that is important, but also – and may be more so – how smoothly the thread pulls, whether it gets caught and comes out with jerks or not. “Bullet” type shuttles are far superior to “boat” shuttles in this respect, so only cylindrical TS deliver smooth low tension.


  3. You no doubt got good feelings, Elena 🙂 I am an engineer, so I like to put numbers on things. I have just started to read many of your informative articles. I noticed, that you experimented with the bobbin tensioners and it got me inspired.
    So as a start I tried to compare two bobbin tensioners set to about 0.15 N thread tension with the same thread. The variation of the tensioner of a Bernina 910 was about +/- 0.02 N. The variation of my Singer 201 was about +/- 0.05 N. I have removed the tension spring of the 201 tensioner and cleaned it, but it did not help. So perhaps I need to polish or do something else to this 201 bobbin tensioner, because it do not feel right. By design It do have a shorter length of thread below the spring compared to the Bernina.
    But I guess the quality of the thread must also play a role for this variation. What kind of thread do you use, when you test this?


    1. Perhaps the spring is badly shaped and you need to bend it. It is a common problem that springs lose shape with time. Perhaps there are scratches just on the hole – thread does wear out metal over time. Check the bobbin holder and make sure the thread path is smooth and free of catches. I deliberately use fairly low quality thread for testing so as not to make the test too easy. It is common polyester thread, slightly uneven and rather fluffy, so it gets caught easily on surface imperfections. But this thread is adequate for most sewing, so I use a lot of it because it’s cheap.


      1. Thank you for your answers Elena. I got another 201k too, and its tensioner got less variation. So I will try to fix it on this 201k.

        Do you agree, that such variations in bobbin thread tension cause less stitch quality?

        You use quite many of the TS and VS machines. These shuttles cause more vibration, so you cannot sew that fast. Don’t you miss to sew fast?

        Are some of the newer industrial sewing machines able to sew with low bobbin tension?

        I spend some time to make some of my vintage domestic machines sew heavy thread and fabrics. I do not have much space, so I cannot have an industrial sewing machine set-up. And I need to be able to transport such machines too.


      2. Sorry, I cannot say about industrial machines. They are very specialised, but whenever jersey is sewn in industry, it is NOT sewn with a plain lockstitch machine. Other mechanisms are more suitable, or elastic thread is used.

        I do not require high speed in sewing because I make clothes, and seam length averages at 20cm or less. There is not enough length to develop any kind of speed! I do not need speed, I need control and reliable stitching. And of course if I make curtains or do embroidery, I use vertical rotary machines for maximum speed. 🙂


  4. Sorry, I do not know much about these shuttles, so my question can be silly. With the TS, you need the point of the shuttle to catch the loop of thread at the scarf. So I guess this point needs to reach out and away from the center line of bullet. But I do not know that much about the timing of all this, but I assume that the needle will be on the side to this shuttle for some part of the sewing cycle. I know the old needles did not have much scarf. The scarf allows a small hook to pass close by for a short time. I guess this will be different with the old kind of shuttle and point of them.

    When you use the term “boat type”. Is it so, that the this kind of shuttle will normally be used with the VS machines with no long arm?

    I actually got two or three Singer 48k in my area for sale. So perhaps I should get one of them.


    1. Boat shuttle is the original TS shuttle. Bullet shuttle is a VS shuttle. Cylindrical TS is a TS machine with a special straight bullet shuttle, like in Singer 48K. This is a rare variant, most TS machines use a boat shuttle. The shuttle has an open top and actually looks like a boat.

      Regarding timing, please see my post “It’s all in the head” as well as “Adjusting timing on a vibrating shuttle machine”. The distance between the shuttle and the needle in TS and VS machines is larger, so no scarf is required, but it does help. These machines make a much larger and better loop through a different needle motion. Less efficient and rather slow but so much more stable.


    2. Jones TS has a special boat shuttle which does allow for low tension. See “Jones transverse shuttle machine”. The only other boat shuttle with similar performance that I came across, as in a Wittler hybrid.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.