A Brother 910_ Ayab diary/ EMS kit

This is a work in progress post. I have now edited and included its previous version so the most recent observations appear first, the oldest last.

9/8/18 I have received a replacement unit from EMS. I got this far testing a large scale pattern, no fault of the program at my stopping point. I did not notice until I got across the last pattern row that I had actually knit with both strands of yarn. Rather than try to fix the issue, I decided to scrap off. This gave me a chance to test new yarns. I had never used chenille in DBJ on Passap, it sheds too much, resulting in error messages. I like to have my scarves at least 60 inches in length, and this combo looks as though it would reach that length when completed. I also did a bit more editing of the repeat as shown here. We shall see if the equipment and I both last to a whole 1200 rows or so in one sitting anytime soon 😉

6/25/18  the kit apparently has now gone from out of stock to officially no longer in production 

5/28/18 replacement kit never received, it’s back to mylars for me to actually try to get any significant knitting done 

4/27/18 I have been able to develop some fabrics using two knit carriages selecting needles from opposite sides, though the lace carriage will not select from the right if the knit carriage is also selecting from the left. The posts include altering a knit carriage to use as the second one to select needles as an alternative to changing ribber settings between normal and pattern knitting, and addition to posting on combining KC patterning with racking

I had been told my kit would be replaced due to persistent patterning error issues. Sales are presently on hold while the hardware is reviewed, and work is being done on a new version of the software, so any changes in status are to be determined.

3/ 31/ 18 At present there is no English text manual for using the software or a quick start guide. I am sharing some of the differences and similarities in actions I have noted between the unaltered 910 and those of the hack, beginning with settings, and including a few tidbits gleaned from support as I encountered patterning problems. They may be of interest to those making the transition after previous knitting experience and familiarity with the machine, or to anyone curious about the program and some of its features and differences.

Yellow text in this first image from a publication on the 910 indicates matching functions I have identified so far in ayab software.

the full illustration from the 910 manualactions on the unaltered 910

When the pattern selector is down, the pattern is centered on green #1 (right of  0). This means if the pattern is an even number, say 24 stitches, then the pattern limits are green 12 and yellow 12. If the pattern is an odd number of stitches the pattern limits are yellow 12 and green 13, the center being green #1. With an odd number of needles, Ayab places the larger number of needles on the left, its orange (yellow) side.

AYAB

I purchased my hardware kit (now off the market) from the developer : to purchase , wiki , interface , install. Online discussion groups:  Ravelry , Facebook, and GitHub.

At present, there is no cover readily available for the machine once the unit is installed, so the left side of the main bed remains exposed.

Though beeps heard upon preparing to preselect the first row from left to right are described as a triple beep, I hear them as a triple set of multiple beeps with pauses between. Not taking the time to “hear” them all will result in patterning errors from the get-go. Be sure to hear the beep and see the flashing of the unit before reversing carriage direction. Clear the left turn mark at least periodically while knitting. If you have experience with any production knitting on the unaltered machine you may feel the knitting speed is slowed when using the software and cable for downloading each row as opposed to that with which the mylar scanner accomplished the same task.

Ayab can be set for an infinite repeat in length, the function was automatic on the mylar unless one made the choice to stop at any point. A single repeat on the mylar was enough to program across the width of the needle bed, with ayab the full width of the pattern must be programmed, so designs need to be tiled accordingly.

Installation of software can be achingly slow on Mac desktop (I am using High Sierra OS). SiLabsUSBDriver  is also required. Security settings need to be checked/changed even if no warning is given by the computer leading one to assume there have been no issues with the download or the install. The driver will reside in the library/ extensions.

The hardware install is fairly straightforward.

The USB connection needs to be made prior to launching the software for the port to be recognized. If port questions come up, click the “Refresh” button on the AYAB screen. When things are working correctly, you should be able to see a port name from the pop-up menu. The plug-in power supply supplies power only to the needle selection.

The popping noise or clunk is heard usually but not always when you cancel AYAB and restart and is produced by the solenoids. I have been reassured it “does no harm” by support.

The launcher box remains open throughout knitting.

To start knitting:
1. Load your pattern in the width of the planned repeat
2. Set all the AYAB settings at default (2-color, start at row 1, single, center). Check infinite repeat if you want to knit more than one vertical repeat
3. Knit carriage is on the left, outside the turn mark
4. In the AYAB software, Configure/Knit
5. Wait until the software says Please Init carriage at bottom of the screen
6. Bring the carriage across the left turn mark and stop before crossing any needles, wait for the triple set of beeps.
8. Continue across the first row and stop, wait for flash and beep, then proceed with movements to and from each side, again waiting for flash and beep that indicate each row download is complete.

If knitting a repeat in single height, when you are one row ahead of the last selection the row numbers disappear as “Image Transmission is finished. Please knit until you hear the double-beep sound”. The marker on the image jumps down to the beginning. You knit across and it does the selection for the last row and gives the long beep sound. Knit one more row to actually knit that selection. When all needles are selected to B, you cancel your settings and you’re done. 

In some instances clearing both turns marks seems to help avoid errors and ensure consistent patterning. Beware of moving the carriage back and forth for any reason, ie to fix a yarn loop, dropped stitch, etc. The movements in my experience have been counted as rows, resulting in patterning errors as would happen when using a punchcard machine. If working on a desktop as opposed to a nearby laptop you may be missing any visual cues for row advancement provided by the ayab screen.

The image displayed on the screen may be magnified and reduced to fixed amounts, but at present, there are no programmable cues as might have been used on mylar ie. for color changes, lace vs knit carriage actions, etc.

When exploring the menus, creating an easy to view repeat helps to sort out available actions

The single setting is used for all single bed fabrics. The row numbering starts from 1, not 0, and focuses consistently on the row whose pattern is being selected, not the row being knitted; design row start line placement

start needle design stitch position on needle bed 

single design alignment test: left, center, right of the same repeat on an alternate start stop needle configuration 

As mentioned, with an odd number of needles, Ayab places the larger number of needles on the left, its orange (yellow) side. It also automatically mirrors all patterns, so lettering and images appear on the knit side as drawn. The mirroring, however,  is a problem with lace. If working from an established pattern, the lace repeat must be mirrored again either before loading the pattern or prior to knitting it or lace transfers will be made in the wrong direction. It is possible to operate the LC from the right side using Ayab, but if  knit carriage is also set to selecting needles ie. if set to slip stitch, and is used in combination with the lace carriage, the LC will not select when operated from the right, and the pattern will not advance properly (documented by others as well).

I used to constantly use 2 knit carriages in a lot of my accessories. At this point, that function is not possible due to patterning errors or loss of patterning when the second carriage operates from the right.

The first preselection row in Ayab is made from left to right. For any pattern that requires a start from right to left, ie one with color changes every 2 rows, there are 2 options. One is to place the design row start line at the very top row of the repeat. That row will be selected left to right, the first design row will then be preselected right to left, and the 2 color rotations using the color changer on the left can proceed accurately.  The other option is to shift the last row in the programmed repeat down to row 1 position prior to downloading the pattern and saving that as the working repeat.  I prefer the latter since it eliminates having to recall the change in the start line. An example of the shift is seen as applied to the automatically shaped lace trim

The ribber setting: a reminder, later color changers were also capable of being used on the bulky machine

Brother DBJ KRC setting including on the 910: white is the background (color 1), punched holes, black squares or pixels are considered contrast (color 2)

This is the KRC color separation described for DIY use on a punchcard machine: 

some of the possible backings (vertical striper requires extra steps)

Previous related posts: https://alessandrina.com/2015/04/18/a-simple-shape-an-exercise-in-dbj-brother-km/ and https://alessandrina.com/2017/10/26/dbj-and-color-separations-some-previous-posts-links/

AYAB color sequence is reversed from the Brother convention of white pixels being color #1, and black pixels being color #2. It chooses black as color #1, white as color #2. The first pass to the right is set up with the “black” yarn. The actual position in the color changer can vary depending on how the yarn is threaded in the mast. The first pass to the right preselects for the first row of black squares, which will be knit on the first pass from right to left. In one option, the ribber needs to knit the yarn in order to get the carriages threaded with yarn to the other side, the KC makes a free pass, with nothing knit. This is achieved by having the main bed set to slip <– –>, and the ribber to selection to be used in the remaining DBJ fabric. After knitting across and waiting for beep, knit back to color changer, the beep happens again as you reach the opposite side. Press the button to change yarn to “white” and continue in pattern, changing colors every 2 rows. Depending on your choice of start to the piece, another option may be to have the ribber set to slip as well on the first preselection row from left to right, thread the “black” with the carriages on the right, using it to knit that first design row, moving toward its empty slot in the yarn changer, where you will then pick up the “white”. Check cam settings on both carriages before moving back toward the right. Thinking of the pattern in terms of black and white and matching the sequence as described frees one from the factory KRC convention.

KC I or II may be used, needle arrangements can vary depending on the look one prefers on the edges. It is helpful to have some previous experience with ribber use and understanding of how stitches are formed. Patterning resulting from choices on end needle selection on either bed may be considered a boon or a distraction, at times even create a smaller, secondary pattern (arrows)

The circular setting: was intended for tubular fair isle, some of my experiments to produce other fabrics may be found in earlier posts. The ravelry group discussion on the topic including using the setting in knitting socks, with an extensive tutorial by Adrienne Hunter: https://www.ravelry.com/discuss/ayab/3346844/26-50#38

3/15/18 After more testing I have come to the conclusion that end needle selection cancellation on my 910 LC is working properly, but is not operative when I am working with my Ayab interface.

3/5/18 While working on a post on shaped lace edgings on Brother machines  I encountered 2 new issues. One appeared to be I was unable to restore end needle selection on my LC while sorting out the repeats. The other was that if 2 carriages are in use, the lace carriage will not select needles if operated from the right, while if is the only carriage selecting needles, it will work from either side. This is a property that is not isolated only to my set up.

2/28/18 I have continued to intermittently knit swatches on my hacked 910. The software continues to be a boon in terms of avoiding the mylar and producing test fabrics quickly. At one point it was suggested I flash the Arduinogo to the Tools menu, Load AYAB firmware. Choose 910, Uno, 0.9. It takes a few seconds to reload the firmware (the part of the program that’s in the actual Arduino.  The changes from 0.8 to 0.9 were minimal, but it’s best to make sure you are using the latest.  EMSL preflashes them in testing so it’s not normally necessary for the user to do it.” Mine had not been. Doing so has not changed any of the behaviors that make me reluctant to even attempt to knit anything other than short fabric tests.  The trick now has become to keep enough patience and distance to be able to not assume that any patterning issues or fabric inconsistency are due to the software and not my own inattention or faulty notes. A recent observation: if an odd number of needles are being used the rule is that the larger number will be on the left.

1/21/17 I have been reworking some of my previous posts to accommodate for the fact that Ayab preselects the first knit row only moving from left to right. After being contacted by a design school student in Europe with respect to creating specific designs in 2 color drop stitch using ayab, there is now a dedicated post on the topic. Nothing has changed in terms of my having reliable and consistent patterning for any significant knit lengths using Ayab. In addition, the knit carriage movements in the 910 when using the software now behave like the Brother punchcard knit carriages. Any change in movement in the opposite direction or jiggling of the KC for any reason can make the design advance pattern rows. This was never true for the unaltered 910 machines, which were wonderfully reliable if the knitting was interrupted, if carriage were moved outside the knit edges for any reason, did not need to clear end marks for any reason other than on the first row of knit, were able to preselect first row from either side, making patterns that required use of the color changer usable without design first row adjustments, and the list goes on. The Facebook group is an active one, and worth joining for anyone seeking more observations, advice, and inspiration.

1/7/18 Over the weekend it appears my software has now begun pre selecting  first with errors, and now correctly on that first pass from left to right, so I will now be reviewing any instructions I have posted this past month in case settings need to be changed for anyone trying to execute the same fabrics, beginning with  lace.

1/4/18 My Ayab software set up rows work this way: the first pass from left to right only gets the carriage to the right, selecting only first and last needle if change knob is set to KC I. The second pass from right to left preselects the first row of pattern. The third pass (from left to right), knits the first row of pattern, selects the second, and so on; the subsequent selection is correct, no rows of the design are skipped. It has been pointed out to me that this may be a unique feature to mine or one unreported by others. “First preselection row should occur on that first pass from left to right”. My posts on using the software are based on the first pattern row knitting from left to right, not right to left as would happen if the very first pass from left to right preselected for the first row of knitting. At this point in time, it is also not possible for the first preselection row to occur moving from right to left. This, in turn, needs adjustments if patterning needs to occur in 2-row sequences from the left ie. in knitting mosaics and mazes unless preselection happens from right to left,  a result of my version’s “quirk”. Since I also knit on an E6000 the first 2 rows feature to complete preselection for the first row of pattern is a familiar one, and from the beginning, I assumed it was an intentional feature in the Ayab software as well. It actually solves the problem if the first selection row needs to happen toward the color changer. Fixed starting sides can be problematic depending on the type of fabric being knit.  Lots of options to explore.

12/28/17: tuck lace meets hand technique 

12/25/17 my first try at lace: the LC is the one to select for transfers, the KC knits to complete the formation of stitches. Brother punchcards for lace usually begin with selection rows, end with 2 blank rows at their top. Using Ayab the repeats need to repeat across the width of the piece. One nice added feature is that when doing so, blank rows may be planned and left at both sides, creating a knit stitch border and eliminating the problem of paying attention as to whether end needles are selected or not, or what other measures to take. This fabric creates large eyelets, there will be 2 empty needles side by side for the duration. Some of the old pattern books referred to it as one of the “mock crochet” ones.

the resulting fabric, knit and purl sides

one more to try: a large diagonal eyelet lace combining lace and tuck 

1/22/18 swatches with preselection from left, similar results for shapes that are shaped with single row increments in height. Double height, the fabric creates tiny pintucks, not blisters as can begin to be seen at top of swatch; KC slip <—, ribber slip —>;

shapes as seen in test swatchverifying presence of pockets 

For a full post on quilting in one or 2 colors please see later post 

12/23/17 The “quilted” fabric produced below is different than the one achieved by specifically designed color separations, the fabric has an interesting blister like effect, the knit areas have more horizontal texture. Because of limitations due to the eventual needle selection errors, I am experiencing with the software, my swatches are limited in length or have interrupted patterns. My new repeat two of the “pockets” have beads dropped into them to highlight their location 

12/22/17: settings given below are for first preselection row from right to left, not for left to right,  testing out the waters with the circular settings, using the repeat. Please check later posts for reviewed content 

I tried to go for “quilting, but think that may require switching levers on the ribber, so to start with, this was knit in an every needle rib, with the ribber set to knit <—->, main bed set to slip —>, producing a fabric with very subtle pintucks, which may be more interesting in 2 colors
Quilting: KC slip <—, ribber slip —>

The pink and white swatch is in “drop stitch lace”,  was also knit using the circular setting. Instructions for color separations for fabric, how-tos, tools, tips, and samples may be found in my posts on the topic, with the same repeat used in https://alessandrina.com/2015/06/16/geometric-shapes-in-drop-stitch-lace-2-brother-km/ 

the image resulting from my own past color separation its purl side, turned sideways

Previous posts have covered manual color separations for fair isle, quilting 1, quilting2, and a summary of series on drop stitch lace. A later post on drop stitch lace using ayab .

One of the differences in using Ayab, is that knitting set up begins on the left, with first preselection row happening on the second carriage pass from right to left. This makes sense if the goal is to set up patterning for moving toward the color changer. Passap E6000 uses the SX/GX (2 Rows) for set up, with its color changer on the right. The knitter may set the carriages or locks for any chosen setting (knit for knit rows, slip if the extra 2 rows knit are not desired as part of the pattern or those suggested by console or other instructions). In the original state 910, the starting side is of no consequence as long as one is outside the set line, and preselection happens on a first, single pass. The same is true of preselection in the punchcard, one may start on either side. In some fabrics such as slip stitch worked holding techniques, or if the color changer is in use and making moves in even numbers of rows toward and away from it are required, starting side matters. The Ayab circular setting does the color separation for the fabrics above, but other cam and lever settings may need adjustment, based on sorting out what selected needles are doing when knitting starts. For the “quilted” swatch above the left part button on the KC, and the right one on the ribber were used. Sometimes the guide to color selection or cam settings is whether the first square in the imported image for download is black or white. With the loss of the built-in color reverse on a hacked machine, it is easy to work when swatching and testing to color reverse the repeat by using invert, available in the color menu, providing the alternative repeat. In Gimp the grid color may be adjusted as well if you choose to work in black and white as your starting palette. For the individual image, Image/ configure grid changes the preference for single use. 

12/19/17 I am continuing to really appreciate the speed and ease for sampling pattern ideas. That said, I  have also again experienced selection issues varying from lost selection to repetition of the same pattern row indefinitely, frequently after more than 50 rows are knit, and in spite of increasing attention being paid to beeps and flashes. Short pauses also seem to put the software into time out mode. I have not done enough knitting to evaluate whether moving the carriage past both turn marks with any frequency prevents those selection issues. Support is responsive, with new tools come new learning curves for everyone involved.

There is a popping noise /clunk that I have now learned is routine and  “likely from the solenoids being engaged or disengaged, when they’re first powered up, or released at the end.” It is a noise I have never heard in standard 910 use. I usually power down equipment when not in use, and have been told: “For the sake of longevity on the knitting machine, it’s probably better to err on the side of disconnecting the USB cable when you power down the machines.”

My present knitting efforts are to come up with some sort of variation of this fabric, published in an early Empisal Ribber Pattern Book, shared in a Facebook forum post.  I have used racking in a variety of experiments but never traveling over a MB pattern as in this instance. There is now a dedicated post on the topic 

12/17/17 I am brand new at using the system, my comments here are logging my personal experience and observation as I am learning a new tool, not intended as a guide. Please follow forums for advice from folks with more experience and knowledge.

The ayab interface from EMS is now up and running on my 910. The process was easy and straightforward in terms of hardware. Software install on the Mac is very slow initially, but the launcher operates quickly after that.  I am using the system connected to an iMac presently running OS 10.13.2. There initially were issues with port failure when attempting to use the program to download. No security warning was received when the driver download was completed, but rather, the image below, which I took at face value, so I did not immediately access the preferences security panel to make any changes thereIt took some sorting out with EMS support and finally accessing preferences security settings and “allowing” the install to solve the problem. Apparently, there may be some change in the message received within a half hour of the install. So, easy fix, check system preferences security settings regardless of pop up window even if lacking  familiar and obvious warnings such as

At present, there is no cover available to fit over the kit and its wires, so the left-hand side of the machine remains exposed. There is a row counter built into the software, but I tend to rely on the built-in on the 910. Any DIY cover should take clearing the row counter and the knit leader trippers into consideration. There is some conversation about 3D printed ones or even metal. Tiny detail: the kit comes with electrical zip ties to help secure the plastic shields on both sides of the board.  Clip the ends, and twist them so join sits either on top or under the interface, thus keeping them clear of KM parts, and, particularly, the belt drive on the left.

There are a series of beeps and flashes that indicate that the pattern row has indeed been downloaded and is ready to be knit. 910 mylar users are familiar with the sound coming from the reader as each row is scanned. Prior to using the Ayab system, I was initially concerned about beeps for each row adding to machine operating noise, since I am familiar with the Passap screech and at one point was also with the Studio electronic version. The Ayab beeps are actually on the soft side, and background noise or inattention may actually lead one to miss the cues. LED flashes occur in pairs indicating the row is ready to knit as well. Again, any cover may hide the light or soften the beeps, even more, making the cues available via the screen more essential.

The kit comes with a 3 ft. cable. I have rearranged placement of my machines to make using both the Croucher cable, switch box, and wincrea on the Passap, and the Ayab system on my 910. My initial wish was for a longer cable, which is possible to use. That said, there are several knit from screen clues that if you are operating the machine any distance from the computer get lost and become hard to see, including indicators for row numbers and the pattern advancement lines on your motif.

So far I have small FI samples and DBJ  samples. Ran into a problem after 60+ rows of each with mis patterning occurring, but that may well have been operator error. I have used my 910 for production accessories and was very familiar with a personal optimum speed. I believe I almost unconsciously with the initial samples, may have picked up speed and interfered with the accurate download of pattern rows. When I slowed down and was much more deliberate, waited for each flash and sound, a test got me to RC 110 with no issues before I quit for the day.

11/27/17 This punchcard was shared by someone trying to reproduce the implied fabric on their 910. It is intended to be used as a combination lace and weaving card, so on the punchcard machine, 2 carriages would be locked on the belt. If even number of passes are made with each carriage, there is the issue of the card not advancing each time when the opposite carriage makes its first pass. So far I have not been able to get this repeat to work for me as drawn, to produce results anything even resembling the “finished swatch” published photo. It did work for me using the LC carriage for 3 passes, releasing it, and following it by 2 rows knit with the KC set to tuck in both directions, a very different fabric. If the LC  is released after an odd number of rows, then technically when the KC is first in use,  the right to left and back to right selection is not interrupted, it is moving in the direction the LC carriage would have moved, and the punchcard row is not repeated. The first row with LC selects, second row transfers, selects next row. Third row transfers, selects every other needle for the first tuck row on the next pass of KC, and LC is released. The fabric then tucks for 2 rows as punched, on every other needle. The second tucked row is completed with KC on right. The first LC row has now been preselected, as the *LC heads back to the right (1) it transfers the selected needles, selects them again, but since they are now already empty, nothing happens to any yarn, and the next row of transfers is selected (2). The second set of transfers is made to the right (3) as the first tuck row is preselected. LC is removed. KC tucks for 2 rows (4, 5), selecting first LC transfer row with the second pass.* Repeat. The marks on the left should indicate carriage movements for each carriage, but they are off as well in my version of the fabric. My final repeats were for use on the 910 with a mylar.

the pattern book swatch image tuck swatch

Lace is a challenging enough fabric without adding weaving yarn floats, and definitely combining them requires a clear understanding of what the yarn is doing on the needle hooks as one progresses through the repeat. I will be starting a separate post reviewing card markings in punchcard pattern books, and translating them for electronic use. For further information please see post on 589

I have always been interested in machine knitting hacks and began sharing information from the internet on the topic as early as 2013. I own the 910 Brother model, and when the ayab hack first became available, assembling electronic components was beyond my skill and interest. I have a good supply of mylars, but have always knit more complex fabrics on my E6000, which has the Croucher cable and switch box for download to and from a Windows laptop, using Wincrea .  I had known a preassembled kit for the 910 hack was in development from a trip to CA, which included a visit to the Bay Area Machine Knitting Guild  back in April 2016.  A recent post on Ravelry alerted me to the fact that Ayab was “alive and well”, and that an assembled kit is now available and on the market. The hardware kit developer , to purchase  , wiki , interface , install . My understanding is that the software is still being developed in Germany.
Online discussion groups:  Ravelry , Facebook.  My compiled list of online pattern generators, hacks, free KM manuals, and more: 5435

I have purchased a kit and will share my experiences with it as time passes. In addition to bypassing the mylar use, the idea of having software that will render color separations that would normally have to be hand drawn for larger repeats than would be practical, or for complex fabrics, is an exciting one.

An artist whose work I admire, using the interface early on is Claire Williams , her tutorial  updated 2016 with DIY assembly of the original “kit”.

Taking knitting machines apart

Once in a while, beyond deep cleaning, even taking the machine apart may be needed. I tackled my own punchcard machine recently, here are photos of the action in progress
Needles now back in, needle retainer bar restoration left to go. In case you choose to tackle the process on your own, some clear instructions may be found at Knittsings for Brother machines , and on Susan Guagliumi’s blog for Studio brand

For folks that prefer videos Roberta Rose Kelly has a series of maintenance related ones on youtube.

Lace edgings on Brother machines- automated with slip stitch

August 2020: I wrote this post (unaltered) in March 2018. I returned to the topic of automated lace edgings on Brother machines in July 2020. As often happens after some distance, details may be freshly observed or seen in a different light, and approaches may then in turn vary. Some of these edgings have been reworked, and new ones will be added. My blog posts have evolved into “works in progress”, with my adding information as I have time. The latest swatches are knit on a 930 using img2track, but the same repeats should work in the 910 with ayab, after mirroring them.

A reminder: although I make an effort to include punchcard knitters in my writings, the full repeats as given and used for my test swatches, knit on an electronic, are not suitable for use on punchcard machines as they are. They would need to be adjusted in height to allow for the fact that when 2 carriages engage the belt selecting needles as the carriage from the opposite side first comes into use, the punchcard does not advance, selecting the same design row once more. The full repeat would need to be shortened accordingly. Many early published trim repeats were intended for use with hand transfers combined with plain knit rows. Such designs are easily translated for use on electronic models with this method.

The initial goal here is to produce a knit fabric using lace transfers in a familiar way, but the knit carriage will now be set to slip <–>, selecting needles with each pass. In routine lace patterning, the KC is set for plain knit, does not preselect needles, advances the mylar, punchcard, or another electronic patterning.
This image, where all blue squares represent knit stitches and white squares unworked ones, is provided with the intent of its being produced as a piece of knitting on the machine. The initial hand technique approach might be to use a single prong tool to move the edge of the blue in one stitch to the right for a decrease and out one needle to the left for an increase. When the decrease is made on the bottom curve, the needle from which the stitch was taken is put back out of work. When an increase is made, a stitch is moved to work on left, and the last stitch on that side is transferred onto it. As knitting continues, that now empty needle location will create an eyelet.

I decided to “wing it” for the first swatches. This illustrates the same knit shape. Blue rows represent all knit rows and stitches, the empty pairs of rows leave areas to insert lace eyelets in a pattern.

Arrows indicate the movement of the lace carriage on left, knit carriage on right. Red squares now represent transfers to right, green ones transfer to left

A second transfer to the right is included in the pattern close to the left edge of the trim to create eyelets next to the decreases, matching those eyelets created by stitches moving out in the increase segment of the chart

A theoretical series of right and left transfers is then inserted, creating a shape in the center of the trim’s width

Now reducing it all to black and white squares or pixels for first preselection from right-hand side using an electronic. This repeat is not suitable for use in a punchcard machine even though it is 24 stitches wide. Electronics advance a design row with each pass of the carriage when cam buttons are pushed in and with both carriages set to needle select. Punchcard machines in those circumstances do not advance when the alternate carriage is brought into work from the opposite side. This charted repeat is usable as is on an unaltered 910, with the first preselection row from right to left. LC operates from the left, KC from the right.

For use with Ayab software two other things need consideration. Ayab mirrors lace repeats, so either mirror the repeat of the original design or choose action mirror in the software prior to knitting. This is a screengrab of the repeat used for my initial swatch test. I later changed the section where the diamonds cross in at the center of the eyelet repeats In addition: those first 2 rows need to be all knit, so the top row of the design was moved down to row 1 position.
KC preselects left to right, knits black squares to left, preselects the same needle selection again to knit those same stitches traveling back to its home on the right, creating the two knit rows that in “normal lace” would be worked with the KC simply set to knit.  As the KC now knits the second row, it preselects needles for the first row of transfers, rests on the right
LC transfers to the right as it travels to the left preselects the next row of transfers. As it moves back to left it transfers to left, preselects for first of 2 all knit rows, rests on left
The sequence is repeated until the piece reaches the required length.
End needle selection is off on both carriages. As LC travels to left, because of no end needle selection, the first needle on the right is not selected. An option is to manually pull that needle out to insure it knits. I decided I actually preferred the chain created by those single stitches knitting only every other row but was not happy with the elongated stitches on alternate rows edge stitches on the increase side. These would be the carriage actions To review: end needle selection is off on both carriages. As LC travels to left, because of no end needle selection, the first needle on the right is not selected. An option is to manually pull that needle out to insure it knits. I decided I actually preferred the chain created by those single stitches knitting only every other row. I had a serious yarn tangle on the right about halfway up the swatch, with some of the evidence visible in the approximate center of the swatch photo 

changing that crossing of the diamond outlines, still winging it

My fiber is now crisp cotton, unmarked weight, tension 8, and it is much easier to observe areas that may still be a problem for me as the designer. Ideally, I would prefer the lines created by the transfers marked in cyan to match the quality of those immediately below or to the ascending part of the diamond on their opposite side, which would require changes in the transfer sequences and space between knit rows. The LC nonselected stitches on the straight creating every other row slipped stitch is still something I like.  The difference on the edge stitches in the ascending angle is happening because with each transfer out, there is a single thickness of yarn on the new edge stitch, and moving the stitch over a needles space elongates it. Longer loops btw are also created when single increases are made by bringing needles into work on the carriage side prior to knitting a row across all needles. In the trim’s decrease edges, the transfers in each new edge stitch have double the yarn thickness contributing to a different appearance, highlighted with magenta arrows. There are 2 rows that have no edge transfers programmed intentionally in the center before reversing direction, so that area has no resulting edge eyelet.  The last shot at adjusting the repeat, with improvement in those cyan, marked areas. Plain knit rows may be added between each repeat, the edge of the knit will be different than in the remaining trim on each side

Going back to the wheel that has already been invented, how can pre-drawn published MK lace repeats be used instead of “winging it” endlessly or not having the confidence to perform the necessary assigning symbols and proceeding with the required separation? Still trying to work with a pattern that knits for 2 rows, transfers for 2, a chart from Stitch World:marking the 2 all knit rows that will be plain knit by programming black squares The pattern is actually for an all over lace. All the transfers in the central diamond actually are happening in the same direction. If the repeat was programmed in the usual manner, beginning with LC on the left, those transfers on the first design row and in the whole of the center diamond shape would all be to the left. I can live with them all being to the right, and would be interested in more eyelets at the edge of the trim. Here is a new repeat, with the increased eyelets at the curved edge Remember if using ayab to mirror the above repeat (seen on right) prior to preparing to knit it There is a disruption for 2 rows in the very center of the diamond, where no increases or decreases occur on the shaped edge prior to reversing the shaping. It was there in the earlier sample marked with arrows as well, but not as noticeable where the edge was created with the addition of single eyelets on increase and decrease rows.

Not finding any other candidates to alter for this purpose in stitch world at the moment I am left with the option of going back to the Brother punchcard book collections, and reviewing what needs to be changed for those same patterns that work so well with the KC set to plain knit but are now to be knit with the KC set to slip <—–>……

3/7/18

a lace punchcard repeat with transfers in opposite directions Assigning colors to transfers: red to left, blue to right. In theory the same repeat could be used flipped horizontally for operating the lace carriage from the right. Yellow lines highlight the 2 blank rows in card that will be replaced by black squares/ pixels (remember rules are different for slip and lace in punchcard knitting)Since the goal is a trim, things are rearranged for knitting to begin on the full 24 stitches There are 2 ways to get the above repeat to work, one is by adding extra blank rows. I chose to reverse positions for transfers, moving left transfers upon row, and red down one row. With lace carriage operating from the left this is my new repeat, with edge shaping added for trim edgeKC is set to slip after the first preselection row, first set of transfers is selected on its second pass to the right. The first transfers with the LC are made to the right, the second set to the left. If knitting the repeat using Ayab remember the above is mirrored by the software, so choose action mirror prior to configuring

I knit a swatch using the thinner yarn again. The difference in increasing angle can be seen here as well, the pattern is short, so the outer curve of the trim reflects that 

This would be my test repeat for knitting the same trim in ayab with the LC operating from the right. The software would mirror it, no action needed. The last row is shifted to the bottom to allow for the knit carriage to find a home on the left. The first row knit would preselect 2 needles only, with KC then set to slip in both directions, the next two rows should knit, preselecting the first row of transfers to be made by the LC operating from the right.

I am unable to test the repeat. I have had intermittent problems with patterning in the software from the time I installed the kit. In multiple efforts and restarts, I am now getting no selection at all or wrong selection with LC operating from the right, while rows advance in the software and beeps continue. I have knit lace with the LC on the right before, but not with the KC selecting needles as well. Possible solutions and causes to be determined. 3/8/18 I have been told upon testing by others the same issues have been encountered that I did when using the KC to select from left and the LC from the right and the problem appears to be the result of a bug in the software to be addressed in future updates.

3/10/18: an adaptation of Susanna’s automatically shaped trim for Brother KM, p. 223 of “A Machine Knitter’s Guide to Creating Fabrics”. This chart shows the first rows of the published punchcard. Please note: using my own lace carriage, when I tested canceling end needle selection, the problem was not solved. Brother setting recommendations made for a similar transfer pattern in an edging published by them, are given and repeat is pictured at the end of my post.


Row 1: LC moves to the right, no needles selected
Row 2: LC moves to left, preselects next row of transfers
Row 3: LC moves to the right, transfers preselected needles to right (red dots), selects needles (if any) for next pass to left
Row 4: LC travels to and transfers to left (green dots), preselects for the first row to be knit by KC In this particular pattern the direction of the arrows match the direction of the transfers with movements of LC operating from left
Row 5: KC from right, set to slip <–  –> knits all needles in D position, repeats the same selection as it travels to left (yellow marks)
Row 6: KC travels back to right, in the second shaped knit row, preselects for the next transfers to be made by the LC
Row 7: LC travels to and transfers to right, preselects for next row of transfers
Row 8: LC travels to left, transferring to left if any needles have been preselected on the previous pass, and the process continues

The numbers on the punchcard chart do not reflect actual row numbers in knitting because when row 4, 10, etc is reached, on the next pass (a blank row in the card), the card does not advance and the previous selection is repeated. For each carriage to make an even number of passes to travel to and from its original position, the total length of the repeat must be an even number of rows.

Here are the 24 stitch repeats adjusted for knitting on the electronic. A for an unaltered 910, B indicating the direction of transfers (red to right, green to left). Arrows mark the problem row, and my solution to it C on the far right, along with the repeat adjusted for use with Ayab.the ayab screen image for working the edging  Both KC and LC are set not to select first and last needles in my directions. When the decreases start to happen on the left of the chart because the end needle selection for transfer to the right again does not happen in areas marked with magenta arrows,  there is an extra stitch that remains on the left that does not get transferred (orange dot) so it is not knit off and simply get held as subsequent rows of knitting take place. Restoring needle selection in the LC is not a solution for the problem, so the final repeat has been amended by me to get that edge transfer. The pattern starts on cast on 20, not the full 24 stitches. Ayab knitters: use the LC to begin selection from left. The first pass will select for a knit row, push those needles back to B, and the LC will select for transfers to right on the next row as it moves to the left. When knitting is to take place there is a clear distinction in the number of selected needles (black squares). The KC will be set to slip <–  –>. Make certain to remember to return the remaining 4 of the 24 total stitches back to B so the full repeat is in work on the needle bed before continuing in the pattern, otherwise there will be needles there to accept stitches moved over for increases to left.

The two extra passes of the LC in this method result in a 2 stitch border on the increasing and decreasing angles of the piece, creating a much nicer edge than that in the samples at the top of the post. The eyelet so close to the right edge stitch which also winds up being slipped every other row made for very messy loose stitches in the thicker cotton that I could not control. The sample did better when I pulled the non selected needles on knit rows out to hold prior to moving across them with KC. 

When working on electronics the 24 stitch limit in width for the repeat is no longer there, extra knit stitches may be easily added to the right of the pairs of eyelets along the non-shaped edge. Having those extra stitches knit on the right side of the repeat made it possible for me to use my cotton again, giving me a controllable edge on that side without having to pull stitches out to make them knit. To my eye, I find the extra passes with the LC and those extra knit stitches on the right are well worth the effort and planning in the finished piece

the ayab repeat  

Something to try: I found instructions ascribed to Brother
for another version of edging using a pattern that has the same eyelet sequences along the shaped edgings. Their recommendation for a 950i is to remove the non-selection mechanism on the lace carriage, and mention was made of the “rubber wheels on the carriage” being uppermost, allowing the end needles to be selected. My lace carriage has fixed, old fashioned brushes, not rubber wheels, and my eliminating end needle selection did not solve the problem with that single stitch in the center of the repeat. And if there is not enough to keep track of, this also combines fine lace and traditional transfer lace, would work fine as just lace.

3/15/18 after more testing I have come to the conclusion that end needle selection cancellation on my 910 LC is working properly, but is not operative when I am working with my Ayab interface.

Double jacquard using punchcard machines

Each row of double jacquard consists of at least 2 rows of slip stitch patterning, one with ground yarn, the other with contrast. The rows knit per design row vary with the number of colors used in the design. Punchcard knitters are not excluded from producing such fabrics, but the color separation is done by hand or using software, and then the cards are punched. The size of the repeat is less important in creating interesting knits than the choice of yarns, colors, and technique choices. In early art-to-wear days, pre-electronic home knitting machines, large, nonrepetitive images were created by breaking down the large images into several vertical panels the width of the punchcard repeat plus seaming allowance. Upon completion, they were joined side by side to create the final image. Ribber settings for DBJ apply across models in principle and may need adjustment depending on the age and brand of the machine. If working on a bulky machine, consider using the ladder-back method.

The common method suggested in instruction manuals for Japanese knitting machines can be viewed in this punch card, laid over a colored template. The limit in repeat width is the fixed 24 stitch. To fit the design on a single 60-row factory punchcard, the maximum is 30 rows in motif height, because one row of pattern requires 2 rows of punching in the card. The step-by-step separation for the design on the left repeated across the 24-stitch card width follows.

Helpful tools: on one blank card color each section of the card, representing color row sequences. This may be used later to check color separation. Box the edge numbers in pairs, beginning with numbers 56 and 57, skip two rows to numbers 60 and the factory-marked row 1, the first visual row when the card is in the reader, and so on. as seen on the right. This separation splits the first and second row of color 1 between the first and last row on the punchcard.

On a second blank card, prepare a master. The first row is left blank, then 2 rows are punched out, followed by 2 rows left blank, repeat, ending with a blank row. Again working upward, number each row you have punched out on the left-hand side starting with number 54, then 55-56, 57-58, etc, ending with a single row, 23. This will result in number markings matching both sides at the bottom of the card, below the #1 factory mark. My own master began on the left with the # 53. I also added corresponding design row numbers for each row on the far left. (1 and 2, 3 and 4, and so on). Written notes on right: an even number of rows need to be knit in each color, so in this method, the pattern must be an even number of rows in height, boxed numbers represent rows knit in ground color.

Take a blank card and put the master card on top, fastening the 2 cards together with snaps. You can now mark holes to be punched. Copy the first 2 rows of color 2 on rows marked 54 and 55 (or 53 and 54 in my case), and repeat in pairs for the required height. I like to mark the card to be punched with colored pencils that differ sharply for each color. They have the benefit of being erasable and can help with keeping each row distinct when punching.

When the completed repeat is marked, unfasten the snaps and move the card down by one row. You will now see the marks you have made through the top row of each set on the master card. Mark through each lower row any blank squares in the row above.

Again, after completing the repeat unfasten the snaps and slide the master up by 2 rows. The card will now be back to its original position, and up by one row. Now you will see the marks you have made on the blank card showing through the bottom of each row of the master card. Mark through each upper row on any square opposite the ones marked in the row below. 

After the repeat is completed, make certain to punch across 2 final rows for the card to create the overlaps for a continuous pattern, as with any other stitch type. Punch out all marks, and you are ready to knit. To check your design visually lay the finished card over the colored master, and correct any errors if noted. Your card is now ready to knit DBJ.

The first preselection row is from right to left with color 1, which will knit from right to left, moving toward the color changer where colors are changed every 2 rows. This color separation will not work for fabrics that require the same needles knitting for 2 rows.

A second master that may be used with the same approach, for a different type of color separation: 

This separation works with any number of rows. The design will be elongated because each design row knits twice in each color, with 4 rows completing it. Fabric will be thicker as well, so use thinner yarns when working it. Draw shapes in colors that are chosen for clear contrast, not because they are to be used in the finished piece. Here I am back to those triangles used in earlier posts. The choice of which color you begin with, whether in the separation or in the knit is ultimately up to you.

With the masked card, the first row is marked on the very bottom of your new design, I began with the red, which would produce the equivalent of this

Remove the mask. Beginning with the second row from the bottom, mark any square not colored in the row below in a contrasting color  Punch all colored squares and you are ready to knit. I prefer to not rely on built-in elongation if I am making a garment or a long piece of knitting. If errors are made or correction of dropped stitches, loops, wrong color selection, etc. is required, unraveling rows is complicated enough without having to also sort out exactly where you need to be in relation to where the card did or did not advance. Black squares represent punched holes, a segment one separated by the red line in repeat width could be used on mylar or for electronic download

Another master for the elongated design, advancing every row, numbers represent rows being filled in, beginning with odd in this orientation, and even with the card rotated for marking even # rows in color 2

This is the same repeat, color-separated, and elongated X 2; proceed as with the directions above, but in this instance, each row in each color is marked twice. The expanded repeat:

I have used these images to illustrate color separations in far older posts.

The reason the resulting separation here looks so different from the one above is simply that in this latest series of charts the color selections have essentially been inverted. The inversion creates a mirror mirrored back to the original orientation with the inverted colors in place, and the color reverse is now recognizable

As a matter of course it is helpful to be consistent in the choice of starting color when separating motifs. Any single repeats isolated from above are usable in electronic machines, reduced to black and white squares. Electronics vary in the built separations in terms of choosing black or white squares for the first row knit, whether you download an original image or work with built-in ones. With the preselection of needles in Brother or pushers in Passap, with an easy-to-count color repeat and some air knitting, planning starting color of choice and needle placement can be sorted out for more confidence when actual knitting starts.

If you would prefer to work on graph paper that matches the size of your punchcard, there is a downloadable PDF and word doc with some guidelines in my previous posts more-low-tech/ and creating knit graph paper. The posts were in 2011 and 2014, with illustrations drafted in earlier versions of OS and apps. Working in Pages I was now able to print an image just about to perfect scale to part of a blank factory card. Here is part of the first repeat in this post traced through the punchcard holes onto it:

I then exported 2 documents, one to Word, and the other to PDF. The first printed image and the PDF one (opened in Preview) were printed to an identical scale using my Epson Printer. Here are the new docs for you to try, they may need tweaking. My measurements, adjusted in Pages were 11 cm by 18.2, for 30 rows in card height: PDF no numbers card, Word doc exported from Pages no numbers card. I have found squares of the blank card BMP provided may be filled in either GIMP or Paint, using the paint bucket tool, but the printed size is not in scale to the printout of the original in the pdf.

If the scale does not matter and you prefer convenience, this is a numbered “full” punchcard template as PDF scaled to fit on a single page punchcard numbered, and as an editable Excel document exported from Mac Numbers: punchcard templates_ excel