Category: Machine knitting (MK)

More than 6 years ago I produced a series of scarves that were double-sided, reversible, and were considered “manly” by some of the customers at my shows. Some were one color, some in 2. I found an early post with no clear instructions for them, but with this image and that of a punchcard marked for a racking pattern (given below). Coincidentally the question of checkerboard rib knit patterns came to light in a forum, and I found myself reviewing the technique, with part of the intent to reproduce this fabric. I have, over the years, been terrible at keeping good notes (if any). At times what I was working on was so “obvious” I had confidence I could rely on my memory. At others my attitude once the problems were worked out and a limited one of a kind series was produced, was that I was “done” with that particular fabric. Now here I am, years later, with a mystery pattern on my hands and a time-consuming quest, wishing I had documentation for how on earth I achieved it

Definitely not “there” yet:

Early translations from the Japanese or German manuals did not always communicate clearly the meaning of symbols or actions required to be taken by the knitter. A bit clearer meaning may be gleaned from these instructions in Brother Punchcard Pattern Volume 5. The hatch marks on the racking symbol indicate the number of pitches the ribber is moved in either direction. The number of stitches moved corresponds to the number of needles in work on the ribber. Recommended settings for English rib also known as half-cardigan, royal rib, or pearl-stitch, with no patterning or racking, the beds may be reversed. It is a 2-row repeat. One row is tucking on the main bed on every needle in one direction,  knitting on every needle in the same bed upon return to the starting side. The opposite bed knits every row. There is one textured round followed by one of 1X1 ribbing. It is not a balanced structure, looks different on each side. Variure denotes fabrics wherein the same bed is tucked for 2, 3, or even 4 rows followed by one or 2 rows of rib knitting every needle on both beds. The tucking bed may also be set up with needles out of work and with added racking as well.
In fisherman’s rib, the settings are for a tubular tuck, with opposite tuck buttons set in the carriages. While a bed tucks on every needle, the opposite bed knits every needle, reversing the process on the return row. It is also a two-row repeat and referred to as full-cardigan stitch or polka rib.
It has the same appearance on both sides and does not contain any rows in all rib. An additional illustration of both The recommended Brother ribber’s “needle rule”.More tuck rib variations may be found in posts on interlock 1, and 2.
Adjustments to the needle rule may be needed depending on the fabric. If only one bed is knitting while the other is tucking, having the first and the last needle in work on the knitting bed. In English (aka half fisherman’s rib) only one (either) bed tucks.  Directions marked with green #6 on left are from the Brother Ribber techniques. The remaining images are for the single-color racked checkerboard pattern from Brother Punchcard Pattern book #5. No pattern card is involved, every other needle arrangement suggested on the right accommodates slightly thicker yarns. A half-pitch setting is used.

In the above instance, the main bed is not performing any patterning function, it is knitting on every needle in the work position. On a punch card machine, a card may be used to track racking positions. With the carriage set as usual for patterning and needle selection but with no cam buttons pushed in, the main bed will continue to knit stocking stitch. No rows are punched, and the numbers on the card in the colored columns indicate the racking position for corresponding rows. The “card” on the left is designed to match racking positions and carriage travel directions (colored arrows) to mirror those in the publications. Since a 36-row minimum is recommended for continuous punchcard use, the “card” on the right has added a 4-row segment for each segment of racking directions. The full repeat is now 40 rows rather than 32 in height. The numbered columns on the far right are as they would appear on standard blank Brother punchcards. The number one is at the level of the first visible row while the card reader drum is actually selecting the first design row. End needle selection is canceled (KCII on electronics). The first move as indicated by arrows is to the right, so the first row is preselected from right to left. The card is then set to advance normally and released. If any errors are made treat card adjustments as you would in any other fabric.

Adding main bed needle selection for selective patterning: the actual punchcard here includes annotated changes in racking sequences from 5/4 pitch positions at its start to 5/6 racking positions for the top half of the completed repeat on its left side. It may be used as-is or set to double length either for use with a single color or combined with color changes every 2 rows.

The card as punched may be used in many ways. In past experiments, I have shown that not changing the racking pitch for a single row while keeping the two alternating pitches constant created scale-like textures rather than checkerboard patterns

here again for a 16-row sequence

On fabrics with racking enlarge the stitch size by 1/2 to one full number to accommodate the stretch needed in racking the stitches.
Color changes: fabrics made in full fisherman rib are reversible, while those in half fisherman are not. In full fisherman rib in order to knit a specific color, that color must be knit for two rows and tucked for 2 rows alternately. In Passap AX with pushers and arrow keys must be used, in Brother, ribber needles would need to be hand-selected to the proper position on every row. 
The yarns used should be soft and have some stretch. Every other needle patterning may be used with slightly thicker yarns.
Racking in the same pairs of pitches ie. 4-5, 4-5 with no added actions taken, produces vertical columns, akin to results in any fabric that repeats the same functions in the same locations on the needle bed

I knit my first “checks” sample on a 930 electronic programming a single repeat to match the card used double length. The goal: the check sample pictured in punchcard pattern book 5 

In programs or machines that allow for memos that correspond to design rows for each repeat, enter the racking pitch number beginning with design row number 1, and continue to 32 or more as needed. The racking sequence is changed at the halfway point of the full repeat.

Added experiments: using the same electronic repeat above, here I worked *20 rows racking every row between positions 4 and 5. One row was then knit on all stitches on the top bed** (I pushed needles out manually rather than changing cam settings), repeated * to**. The fabric reminds me of racked herringbone, the “checkers” are distorted Changing color every 2 rows shows the same lean in the fabric. I have had intermittent problems with my ribber, stitches begin to simply not be picked up by the main bed and are dropped for no apparent reason A very different fabric is created using the repeat and instructions below*Knit 2 rows, rack 1 pitch to left; knit 2 rows, rack 1 pitch to right to RC 20 (or preferred row count); knit 1 row continuing in pattern to the opposite side without racking**. Repeat * to**. One repeat of the 2 sequences is 42 rows in height. At row 1 of each new (here 21 rows) sequence, the carriage starts on the opposite side Changing colors every 2 rows is possible. The racking will begin with the carriages on alternate sides of the machine after the single-row knit without racking. In segment 1 racking occurs on the left, color changer side, and in segment 2 racking occurs on the right, opposite the color changer. 

“Full” fisherman rib with patterning on both beds: on Passap the back bed is capable of many more patterning choices than in Japanese machines, and strippers help hold loops in place on the needle beds. The Passap “needle rule” places the first needle in work on the front bed, and the last on the back bed. This is also variable depending on the fabric being knitted. Using the repeat

tech 129: (black square tuck for single row) on the front bed. Set up back bed after prep rows, making sure pushers are in the same work/rest position as on the front bed *Knit 2 rows, rack to left, knit 2 rows, rack to right to preferred row count ie RC 20; knit 1 row still in the pattern, without racking**. Repeat * to**. One repeat of the 2 sequences is 42 rows in height. The resulting pattern is reversible.” This swatch was knit with the bottom 3 blocks using 6X6, both arrow keys on the back lock, and N in front. The back bed pusher setup is doing the patterning. The top 3 blocks are set to pattern selection on both beds, using AX, arrow keys, and KX on the front lock. Using technique 130 will double the height of the repeat, working each row twice.

On Brother, the second bed selected needles will face tucking needles on the main bed. This creates a knit stitch on one bed, holding down the tuck loops on the other, allowing for side-by-side tuck loops on opposing beds. Such selections would need to be made on the ribber manually. Both beds are set to tuck, both arrow keys. The needles on the ribber immediately below the ones tucked on the main bed are brought to the E position and face the tuck needles on the main bed, while its non-selected needles will tuck. Rack before pushing those needles (black dots) that will be knit up to the E position. The Brother settings for full fisherman suggested in their Ribber Techniques Book and manuals produce a “circular” tuck stitch, with each bed tucking and alternately knitting on all stitches in opposite directions, so the cam button setup is different than when one is planning textures in varied patterns It is also possible to produce “checks” without any racking at all. On the Passap, this sample was produced eliminating racking completely. The front lock was set to knit throughout / N, the back lock alone did the work. Pushers were selected 6 up, 6 down, the back lock set to AX with left arrow key for even multiple rows divisible by 4. I used 24-32 to get a sense of scale. The arrow key was canceled for the next 2 rows AX 0 to switch the pushers.  The working repeat became *32 rows 6X6, left arrow key, 2 rows AX 0*with the front bed programmed

Here the front bed is programmed for the repeat below, technique 130 (black squares tuck for 2 rows After the initial prep on the front of the bed, prior to knitting the first pattern row, pushers were manually selected on the back bed to match the pusher work/rest position selection on the front bed. Their position will change as the back lock moves to the left. After 12 rows, the arrow key was canceled for the next 2 rows to AX 0 to switch the pushers. The working repeat became *12 rows AX, left arrow key,  2 rows AX 0* with the front bed programmed, and its lock set to KX. The back bed produces a “checkerboard”, the front bed produces checks as well, but in a vertical alignment

A similar half-fisherman (only one-bed tucking) fabric may be produced on Brother machines by automating patterning and switching “beds”. The main bed is set to tuck in both directions, the ribber is set to knit throughout. My sample was pretty much a disaster at the start. After trying different carriages, switching out needle retainer bars on the main bed, checking alignment, and every trick I could think of I was rewarded with stitches simply not being picked up at intervals by the main bed. Time for a break for both the operator and the machine.  

a bit more success:

Below is my electronic repeat, 12 stitches by 56 rows in height. It is intended to mimic the work done by the pushers on the Passap. Alternate groups of 6 stitches will knit (black squares) or tuck (white squares) for 2 consecutive rows. At the center and the top of the full repeat, the two extra rows of squares result in alternate groups of stitches tucking or knitting for 4 rows, contributing to the shift in the color and texture of the checks.
Passap specials: the idea of hand-selecting needles every row while watching multiple loops tucking on both beds and even adding racking is far too daunting to my mind. Highly textured patterns are far more easily produced in machines that allow for a greater range of patterning on both beds.  To review, E6000 tuck settings:
N, EX: same on both locks, may be used without pushers or console
KX and AX: tuck in both directions
OX and DX: tubular tuck. FB: tucks right to left, free pass left to right, BB is opposite
The FX setting is incorporated into several techniques used with KX, 104, 105, 112, 113, 167, 259, and 260. Some techniques for adding back lock settings are 106, 114,145, 158, 167, 168, and 190. Techniques 259, 260, 269, and 270 use racking; 200, and 212 require manual changing of arrow keys; 284 uses the U100 transfer carriage in combination with fisherman rib for an embossed effect.
Using FX settings with pushers full fisherman rib can be combined with full needle rib or half fisherman rib on the opposite bed.

Technique 167: use FX/KX

Front bed pushers are always selected up from right to left by the console independent of pattern, so they will knit. Set up pushers on the back bed in the pattern after the first row of the pattern, and make certain they are in the opposite arrangement of work/rest positions on the front bed. EX knits on all needles from left to right. Black squares represent knit stitches and white tuck ones. Making lock changes at the start of the repeat: knit 14 rows,* Knit 2 rows FX <–/KX, followed by 12 rows FX 0/KX** Repeat* to ** One full repeat of the 2 sequences is 28 rows. The original BW building block is 6 stitches wide, 7 rows high, pairs of each form the unit used to form the larger repeat blocks

My chart for my full working repeat test sample: dots represent pushers, green highlights rows with lock changes for pusher reversal. The latter are made here on RC13 and 14 rather than RC 1 and 2

Getting back to that scarf and reversible checks, I finally sorted out the how-to and a repeat in a different number of stitches and rows. Technique 180: disregard console directions. Set up with 1 extra needle and pusher on the back bed at each side. Pusher selection on the back bed as described below matches half the number of stitches in the full repeat starting on the right side of the back bed. End with a single pusher on the far left in the opposite work/rest position of pushers in the group to its immediate right. Reset the front lock row counter manually at the end of each full repeat (24 for mine) back to 000. As an option one may choose to knit half a repeat at the top and bottom of the piece. My first swatch is testing one full repeat + a few rows. My full checkerboard repeat is 24 stitches wide by 24 rows high, composed in turn of a set of 4 blocks 12 stitches wide by 12 rows high. The AX setting changes pusher selection every 2 rows, and the AX 0 rows reverse pusher selection, resulting in the shift in the patterning on the back bed. Knit the first 24 rows (full repeat #) with no lock change, I found it easier to reset the arrow key at the start of the repeat on RC 1 and 2, rather than RC 23-24. The single BW building units are 6 stitches wide by 6 high with blocks 4 producing the 12W X12H repeat segments A working chart for the full repeat: Black dots = pushers in their work/rest positions, numbers on right = full repeat in rows

April 2025: Some knitters are visual learners and prefer video instructions. Elena Berenghean has created a clear, step-by-step video on knitting the punchcard scales on a Brother machine, shared in her Patreon, https://www.patreon.com/posts/dragon-scales-125827456
on YouTube https://www.youtube.com/watch?v=H_9ntWS3ZUc
and Instagram https://www.instagram.com/elenaberenghean/
This old-timer admires Elena’s work and teaching methods. Her hand techniques are usable on any KM. So many rabbit holes to jump into, continuing interest and learning in our craft for new generations.

I reviewed this post and am amused that I narrated it in the first person as I worked through getting things to work. It is wordy, but it documents the trials, errors, and sometimes headaches before any success in developing DIY patterns and techniques.

Over the years, various fabrics have been named dragon scales or crocodile stitch. Here, dragon scales have referred to shapes created using a lace technique and resulting in a pattern such as this

that was followed by hand-knit samples and an investigation into possibly creating a ribber fabric with auto-shaping resulting in similar protrusion
ribber-pitch-a-bit-on-racking-1-chevrons-horizontal-herringbone/vertical chevrons/ herringbone which eventually led to this, where a reversal in racking periodically shifts the lean in opposite directionsautomating the pattern in half fisherman rib/ mylar repeat tracking shown. Any repeat in a factor of 24 may also be used on punchcard machines. The start of a series in varied colors and fibers: sometimes I enjoy getting back to the simplicity and predictability of punchcard machines, even though punching those cards can be slow and a bit tedious. I am now curious about striping again and forming a larger scale with a crisper fold.
The chart is for the working idea, the punchcard image illustrates how some of my cards begin to look as my work evolves. When marking cards for any action, the fact that the eye is not on the same design row as the reader needs to be considered. Here, racking numbers begin to get marked on what would be row one on a factory punchcard, 7 rows up for Brother KMs or above the two all punched rows on any other brand blank punchcard, or roll # position.
Though the final repeat is an even number of rows in height (42), note that each half repeat is not (21). The color changer sits on the left, so the first preselection row is left to right, with the cam button on KCI to ensure end stitches knit. Any color changes happen every even #X rows, so they will technically be in a slightly different spot on the alternate repeat. Some additional trial and error, random yarns: the white is a 2/15 wool, the yellow a 2/12, the blue an unknown thickness, also wool. The best of the lot, but not “there ” yet, is going back to one-color knitting. So then you go for a yummy alpaca and silk, make a racking error and manage to correct the pattern, and lo and behold, the yarn breaks halfway across the row a repeat up from there! “They” do keep talking about how relaxing knitting is ;-). Yarn specs: Fiber Content: 80% Alpaca/20% Silk; Weight: Lace; Gauge: 8 sts = 1″, 1/2-lb cones/3472 YPP (1736 yards/cone)This yarn is an English import, 2675 yards per pound. It felts into a lovely fabric (not the goal here) and knits tolerably well. The fabric is quite stiff, however, and the surface change is minimal and nearly completely lost.2/18 Jaggerspun wool-silk: worth a shot at a scarf. Starting ribber cast on left, followed by 2 circular rows, one closing row right to left, and first KCI row from left to right will set up patterning in tuck so that the direction of the arrows on the left side of the card, lines up with the racking number appropriate for that row before knitting it. The fabrics below are as they came off the machine, with no subsequent blocking. I have some lovely cash wool in 3 colors, 2/48 weight. Using 3 separate strands fed through the yarn feeder separately resulted in uneven feeding, loops, and too many problems. Using 2 strands “worked” easily, but the fabric was nearly flat. Cautiously winding 3 strands onto a cone before knitting gave far more predictable results; this image shows a scarf in progress. The color difference is due to the lighting at the moment. My best advice to anyone attempting this is to knit slowly.
The most likely spot for errors in my experience is at the point where
1: no action is taken for a row (or more in later swatches), so the racking position remains at 10 for 2 rows, and 2: for racking position 9, the knit carriage position is reversed in each half of the repeat.
One can also reach a left-right rhythm and, without realizing it, begin racking between positions 9 and 8 as opposed to 9 and 10, throwing the pattern off. Another look at racking positions: the numbers reflect the racking position before the carriage moves to the opposite side, and the arrows show the direction in which the carriage will be moving. Once the knit carriage moves, the card advances, so glancing at the card after that move will show the action for the next row at eye level, which can be confusing at times. A finished piece, 9.5″ X 64″, in the coned 3 strands of merino. Occasional single strand caught on ribber gate pegs, no yarn feeding issues. The fabric has not been blocked in any way but allowed to “relax”. I like the larger scale of the “scales” but would still like to introduce striping in a way that pleases my eye. The knitting is slow thanks to all the racking, but it is probably faster than using single-bed holding for similar shapes, with a very different finished look.
8/16: When the fabric relaxed, it became less 3D. Back to introducing stripes in contrasting color, an act of faith after lots of trial, error, and a punchcard redesign, this may have been worth the effort when done. I am choosing to cut the yarn and weave in the ends formed in the longer solid areas, and I am permitting myself to knit only while I feel focused on manual changes in color and racking. It may take a very long time to get to “scarf length,” and here is the fabric in a completed piece, about 54 inches in length when off the machine. The top right photo shows the reverse side of the piece; the bottom right is how it might appear when worn.  A test on how to use 2 carriages or change settings followed, allowing the turning stripes to help the scale shape bend outward into a “point”. I found that to get the width I needed, along with striping, it was simpler to change ribber settings to slip <– –> for all knit rows and retain the use of the color changer on the left.
It is easy to share successes. There are also those days, however, when one should not be anywhere within range of a knitting machine, and perseverance does not lead to anything positive.
The above scarf was knit in charcoal, using 3 strands of cash wool. Two strands of the blue yarn created a nearly flat fabric; 3 strands did the job. I now turn to true black and white. Knitting 3 strands of the black was impossible at any tension for any length. Then I noticed the ribber on the right was lower than it should be. It turned out the bolt used to adjust the height of the ribber was loose, and the slightest turn of it loosened it completely. So then it took way too long to get it back in place. Got things back together and set up, and with each movement of the racking handle, the ribber dropped on the right. After more fiddling that got me nowhere, I retrieved the ribber for another Brother machine that had not been used for years. It was dry, and the grease on it had turned black. Time flew while cleaning, oiling, and waiting. Back on the machine, the right ribber bracket of the alternate ribber will not allow it to drop on that side, so it’s back to grease and patience, and yes, I finally got it up and running, only now the smell of the oil and lubricants makes me want to leave my apartment. Outdoors, the temperature is a dozen degrees warmer than inside it and grossly humid. I don’t want my knitting to smell like the solvents either, so the remainder of the day is called in as a period of rest and recreation mixed with a touch of, hopefully, amnesia.
Moving on to the next day: success in one color with no major problems or errors produced a black scarf, 64 inches long with lovely bumps, shown here as it appears immediately off the machine 8/16: 3D shapes held up very well

So what would that true black in the thinner weight do with those stripes in a true white? I found myself forgetting completely to set the carriage to tuck for several tries, then messing up the color-changing sequence. Time for more R&R.
8/7: After several tests with minor variations in the pattern, sorting out yarn weights preferences, I decided to “go” for a version of the same stitch type as the charcoal and white in true black and white. Again, I am not able to use 3 strands of the black. Got a third of the estimated desired length knit, and whoopee! About 10 stitches dropped off both beds on the color changer side. Oh, the joys of unraveling several rows of sewing thread weight black yarn, in racked tuck stitch, down to an all-knit row in the white to make certain the proper number of stitches are in work on both beds. Got that far and am ready for more R&R.
And 8/8, the last in the series, at least for a while, in true B&W. The 3D texture becomes reduced by the weight of the finished piece as it is worn. The service manual http://machineknittingetc.com/brother-kr120-kr710-kr830-kr850-kr230-kr260-service-manual.html provides information on additional ribber adjustments.
The part in question in my troubleshooting appeared to be #24, the “slide plate guide stud”.
In the image below b/, the bolt that became completely loose. I discovered after getting things back together that a, which secures the ribber bracket, is actually directional with a barely perceptible difference in shape, and if accidentally rotated 180, it will keep the ribber bracket from changing height positions and working properly. Rotating it restored the expected actions, so now I have 2 well-functioning ribbers to work with.      As of 8/16, I now have a lovely, equally bumpy fabric in all 3 colors, using 2 strands each of the cash wool at the same tension. I am now using my alternate KR 850 ribber. The height and other adjustments appear identical to my eye. I am reminded of my teaching days in a Brother punchcard lab, where at times the same model machines might be side by side, and fabric would work perfectly on one machine while not on the other supposedly identical model.
Students were not allowed to swap off machines. The one exception was if that was the only way to produce the stitch types in their final projects after I attempted to isolate any other possible issues.
“They” do keep talking about how relaxing knitting is, but with machine knitting, there are lots of opportunities to wonder about that suggested fact.8/17: Complete a royal blue scarf in the smaller scale repeat, previously executed on my 910. The punchcard below the image may be used to achieve the same fabric. 8/18: Trucking on, planning a couple more pieces with the large-scale repeat. It seems I have been having more drat-it moments than one might ever want, resulting in discarding hundreds of rows of knitting for many reasons, including racking operator errors. I have also encountered another problem. In the past, I have used clear cellophane or a variety of other tapes to seal off holes accidentally punched in the wrong place.
I rarely produce multiples of any of my pieces, and my limited edition items were usually knitted on an electronic due to its increased ease in adjusting the repeat width and height to suit. Transitioning from the solid repeat to the striped one, I decided to punch out holes on my original card to test my ideas, and when returning to the large scales, I was too lazy to punch yet another card and taped over sections I wanted to eliminate from the selection. Hundreds of rows into yet another piece, I began to notice odd behavior in needle selection, which was fully remedied by investing time into punching a new card, and yes, starting over yet again. Note to self: do not use taping over in the future, no matter the tape, especially when knitting multiple pieces thousands of rows in length! 8/21: I am working on a final series of large-scale, single-color scarves. As has often been my experience in knitting long pieces of ribbed fabrics (most of my scarves are 1200 rows or more), I have a talent for developing problems after the ¾ point. Two factors that can affect stitches not forming properly: “suddenly” can be the result of 1: the slide lever setting being changed accidentally when moving ribber sinker plate ie to correct patterning errors and bring it to the opposite side, and 2: the ribber alignment for needle positions relative to each other on opposite beds changing slightly from all the side to side motion in racking nearly every single row.
The slide lever has 3 positions. I have gotten used to simply leaving it in its center setting (lili) for my knitting and used to teach students to keep that constant, if possible. Sometimes, when knitting ribbed cuffs, bands, or collars, the differences in length and width changed in the separate pieces and were not noticed until one was ready to join them.

Adjusting needle bed positions (for more, see https://alessandrina.com/2015/01/13/a-bit-on-ribbers-japanese-kms_-alignment-and-symbols-1/)The last piece produced by me was in a charcoal color, using the same yarn brand and weight as the black. All things being equal, using the same tension (required to avoid knitting problems), the charcoal version stitches were considerably looser and longer due to changes in gauge. This scarf will put this fabric to rest for me for a very long time.
This was my final, pre-punched card, and its markings

Some Passap patterns 2016/01/13/racking-2-vertical-chevrons-herringbone/
and 2016/01/09/ribber-pitch-a-bit-on-racking-1-chevrons-horizontal-herringbone/