IN REVIEW
Several ribber fabrics are produced by altering the settings on the ribber’s carriage to slip, in this series, for an even number of rows in both directions.
This requires manually changing the ribber setting from slip to knit and back for the length of the piece. In electronic machines, where the pattern advances with every pass of the carriage, there is another option.
For example, to produce DBJ with the backing in one color, settings would be changed manually every 2 rows, using the color changer on the left.
Designs with the deliberate placement of white blocks representing each blister can be created.
It is a good idea to test tolerance for each of the yarns involved as a hand tech or small programmed repeat before planning significantly larger designs.
Some things to remember: when 2 carriages are selecting, each carriage needs to move far enough at the end of the needle bed so as not to be locked onto the belt.
Extension rails are required. There is a limit to the number of needles used since both carriages will have to clear the stitches on both side edges.
On the Brother ribber bed, some stops keep the combined carriages from going off the bed. There is one on each side (magenta arrow on left, the blue arrow on right), and to remove the ribber carriage from its bed, it, in turn, needs to be tilted forward before reaching the stops to clear them on either side. In my own machine, I have removed them. This allows for using the altered carriage from the right in wider pieces of knit, but beware of moving too far with the coupled carriages; the ribber may separate.
Altering the KC sinker plate and arm: remove 6 small screws from the sinker plate, leaving only the arm.
The carriage with the altered sinker plate in place, in turn, will then be used to replace rows that were to be knit with the ribber set to slip in both directions <– –>.
In my sample, it operated from the right, with the combined carriages (KC2), from the left.
For consistency, I am editing the original post and will continue to refer to the coupled carriages = KC2, the altered single bed one = KC1. KC is the abbreviation commonly used in publications for Knit Carriage. The change knob, which affects end needle selection, is marked on the KC as for I (end needle selection, indicated by the black arrow) and II (cancel end needle selection). Initials KR in publications are often used to refer to ribber carriage in setting discussions.
My first experiments were knit using an Ayab interface, abandoned since this post was first published.
Later 2026 experiments were knit on a 930 using img2track.
My Ayab interface required the first preselection row always from the left, img2track preselection for that same row can be from the right, based solely on the design.
With my first try, I made no effort to consider which color gets chosen first in the color separation (ayab = black, 910 = white). There is a limit as to how far the single carriage from the right (KC1) can travel on the needle bed to the left because the combined carriages on the left (KC2) are held in place by the pin.
It would be possible, if needed, to separate KC2 and push further out on the extension rail, but perhaps not practical, so there are some constraints on the fabric width able to be produced.
My first swatch has some manually created pintucks on the knit side (white only, knitting extra rows, joined together by all knit rows with pink).
The reverse, purl side, is knit by the ribber set to N/N, all in one color. The stitches held on the ribber while the white only knits on the top bed are visibly elongated (left swatch bottom). There is some color confusion on the knit side on the first couple of rows of DBJ, solved in the second swatch. My repeat for the planned width using Ayab. 
Knitting on 910, a single repeat may be programmed. Start knitting with the color intended for white squares, which will also serve as the solid backing color.
For more similarities and differences between the original and the altered 910, see the Ayab diary post.
To knit using Ayab: begin the knitting with the color intended for areas marked in black squares, which will also serve as the solid backing color.
Set the change knob to end needle select to ensure the first and last needles in use knit.
Preselect the first row from left to right with the ribber set to knit (N/N), it will remain set that way for the remainder of the process. When on the right, set coupled carriages (KC2) to slip <– –>, knit one row to the left, both carriages stay on the left.
Knit the next 2 rows for DBJ using KC1 with the altered sinker plate operating from the right, using the color intended for areas marked with white squares. The main bed will only knit.
Return the KC1 to the right, on the extension rail, and knit the next two rows with KC2 operating from left. Repeat, changing carriages and consequently colors every 2 rows.
Blisters or pintucks/ purl bumps are created when one bed knits more rows than the other, whether as simple knitting or in the pattern. Periodically, the knitting is sealed by at least one-row knitting across all stitches on both beds.
In this version, sealing rows must occur in pairs to allow for color change.
The first chart shows a tentative repeat, planning for black squares to create the blister shapes, drawn in 2-row blocks to allow for color changes every X even number of rows
The image color inverted, so white areas will create the blisters in slip stitch (col 2) while black squares will knit (col 1)
Most published patterns for these fabrics will also include an all knit rows, containing all black pixels, to seal the shapes knit on the main bed only, with the ribber set to N/N.
By using 2 carriages to select needles, one (KC1) may be set to slip <– –> to knit X number of rows on the top bed only, while the pairs of carriages are set to normal knit on both beds, with the cam button set to end needle select. Selection will continue, but no patterning occurs as a result. KC1 may also be set to slip in both directions if preferred, on black pixel solid rows.
End needle selection helps to anchor the blister color on the top bed.
If end needles are not selected on the top bed on specific design rows, push the needle out to D manually before knitting the next row.
A proof of concept swatch knit on only 24 stitches, the pink knits for 4 rows, the white for 2; the pink yarn is a cotton, the white an acrylic/wool blend.
An expanded pattern repeat is planned for a larger test swatch. Here, there are 2 black squares added at each end of the repeat to ensure that those stitches on the top bed are knit on all slip stitch rows. The new color is wool, this time knit for 6 rows in slip stitch before sealing with 2 passes of the combined carriages with the contrasting color.
Single bed slip stitch rows appear on the knit side in color 2, the reverse, purl side is in a single color, 1, and is formed by all knit rows. Its stitches are, in turn, elongated, since they are held and not knit, while the opposite bed knits for multiple rows.
In the sample, the first and last stitches on each side were on the ribber, creating a single white slip stitch edging.
One can adjust such details to suit.
The first preselection row was made after casting on with both carriages(KC 2) set to knit, moving from left to right with color 1, where they stay. KC 1 with an altered sinker plate was threaded with color 2, and begins from and returns to the left-hand side.
A detailed shot of the edge: note the white, single, elongated stitch uppermost, and the pairs of contrast color ones in the “border”.
2026 edit
Any of the 24-stitch design repeats is suitable for use in punch card machines. However, when using them, the option of using a second carriage selecting needles from the opposite side is not possible since the method of advancement of the card will change the repeat.
The ribber setting also must be changed manually to knit for rows where every needle is selected, to slip in both directions when the variable needle pattern is preselected. This is true when using the paired carriages in electronic models.
The first preselection row is planned toward the side of the first carriage(s) to be used.
Reviewing the design repeat, added variations were planned for automating the design, beginning with a 24X16 file,
tripled in height to 24X48, and added solid double lines for pairs of all knit rows with both knit carriages set to slip in both directions. Here, each slip stitch block repeats for an identical number of rows. 
![]()
Note that the widest float formed as stitches are slipped only on the main bed is 8.

Keeping the slipped stitch rows to no more than 4, returning to the original design and the idea of staggered rather than straight edges, with the last pair of design rows shifted to the bottom of the design,
led to a new 24X12 repeat, 
![]()
and the visually unexpected result.
Potential repeats for electronic machines while seeking symmetry, retaining the positive/ negative elements, 48X84, ![]()
reducing the vertical black line, 44X84.![]()
And a last 48X84 adaptation.
The tension in both beds needed adjustments to form larger stitches, resulting in more bleed-through of the light color. ![]()


I am always interested in automating stitches to facilitate as many details as possible in creating fabrics that imitate hand techniques, without relying on row counts and a lot (in this instance) of “hooking” stitches up manually. The swatches below were part of a series of posts on ruching as a hand technique
The hand tech chart mirrored, as the springboard for my automated fabric.
Both led to my exploring the possibility of a cousin, working on the double bed, and using the 2 carriages. The process began with a proof-of-concept swatch. Two colors were used to highlight what the stitches knit with each of the 2 carriages are doing. Both KC1 and KC2 were set to knit in both directions. I began with KC2, knitting the needle set up for the fabric in a dark color, on a small number of stitches.
KC2 operated from the left, KC1 from the right. With KC2 returned to the left, KC1 knit 8 rows. It is best to work out the limit for how many rows will knit on the top bed without stitch problems prior to any automation of functions. *Two rows were knit with KC2 in the dark color, 8 rows with KC1, ribber racked to proper position**, * to** were repeated. The elongation of the stitches on the purl side results from the fact that the ribber knits far fewer rows than the main bed, and in addition, the stitches on the row knit on its second pass in the pattern are pulled across the bed at distances matching the racking positions.
The plan is to automate the texture, knit it in one color, and find a way to track correct racking positions: cast on may be in any preferred method. With the pitch set to P, it is easier to transfer stitches between beds to the desired configuration. Every needle is in work on the top bed, for a multiple of 10+6
ribber needles with stitches that will be moved with racking![]()
placement of stitches on respective beds
change pitch after transfers, ribber moves slightly to right,
final configuration prior to patterning and racking sequence begins![]()
The starting racking position is 5. Racking handle markings for Brother begin with 0 on the left, 5 at the center, and 10 on the right. The ribber is set to half pitch since part of the needle bed will be knitting every needle rib. An often-overlooked clue as to what is happening or is about to happen is found in the arrows just below the racking position indicator. The red arrow indicates the direction in which the bed was racked on the last move. Since racking for my experiment will only be to 3 different positions, I began by choosing to use 5 pixels center, left, center, right for the full repeat sequence, but later amended the repeat to numbers of pixels equal to the specific racking indicator number. The needles will be selected before the next row is knit, a reminder of the racking selection. Sequence will actually be 5, *0, 5, 10, 5,** 0, 5, etc. The racking indicator
My first Ayab repeat was planned for use with 2 carriages: KC1 (single carriage) is set to knit but not select end needles (KCII), and will be producing the multiple rows gathered to create the blisters/ hems. KC2 (coupled carriages) are also set to select needles; both KC (knit carriage, cam button also on KCII) and KR (ribber carriage) are set for normal knit. They will create the sealed stitches joining up the blisters/ hems. The same color yarn is threaded in both. Blank squares will knit because a carriage set to knit overrides needle selection or lack of it in stitches on needles in the work position. My repeat is 46 stitches wide. Because the knitting is started with the ribber already in racking position 5, the first move in the pattern is turning the handle to the right, toward 0. The concept illustrated
is set up and ready to go, racking position 5
4/19/18 As is true in any knitting, things can go quirky. I began to have a single needle on the main bed, not knitting on rows knit with the combined carriages, then ran into dropped stitches in racked groups. The problem was initially not with the software but appeared to be a ribber issue, which, after checking and balancing, was resolved.
4/22/18 And today’s problem is the software, with persistent, intermittent selection errors. I did achieve a sample by manually pushing wrongly selected needles out to D by hand on problem rows, which tended to be 5 and 7 in all black areas. Some reminders and observations: ayab auto mirrors all images. If this repeat is entered (I added a single square as a marker for racking position 0), the software will actually be knitting this, the “image as it would appear on the knit side of the fabric”. What to program? anticipating the above getting mirrored by the software, I entered this,
resulting in this.

That said, remember that turning the racking handle to the left is toward increasing numbers on the indicator, and to the right is toward decreasing numbers. For me, that is counterintuitive. Mirroring the image again, and working with the repeat below can help with tracking racking movement even more. With the single dot on the right, turn the handle toward it, to the right, and the movement will be towards 0. With the movement of the marking row to the left, turn the racking handle to the left, toward 5, and so on.

Next on the drawing board: a fabric using the same technique, but one that I might like more. This image was a portion of a greyscale pin of a pattern book from a Russian pin.
I tested a concept for recreating it as a hand technique, trying to sort out how many rows I could knit before racking and the racking sequence. The best result was with a single sealing row, which in turn required changing the ribber slip setting for one row only,
so it’s back to the drawing board. I think any automation is best done using the ribber to do the all- knit rows, the main bed to needle select racking positions. Results will be added to the post on combining KC patterning with racking
The present setup with 2 carriages may be used for solid color backed quilting. Using the altered KC1 operating from the right with no yarn in feeder should work to drop stitches in drop stitch lace where the repeat is altered to allow for the knit carriage with no yarn to do the stitch ditching while selecting needles as well. Related color separation and swatch may be found in the last segment of the post: revisiting-drop-release-stitch-lace/