Views: 0 Author: Site Editor Publish Time: 2026-11-12 Origin: Site
Most sweater programmes fail before the first panel leaves the machine. A sweater flat knitting machine bought on the wrong gauge, bed width or number of systems is expensive to correct, and the catalogue spans beds from 52 to 100 inches, gauges from 3GG to 18GG and one to three knitting systems, at prices from about 5,000 to above 80,000 USD.
Choose the machine by working backwards from the garment. Gauge follows yarn count and target fabric weight, bed width follows the widest panel, system count follows monthly volume, and architecture follows how you will assemble the sweater. Only then does price become a fair comparison, because a cheap machine that cannot knit your widest panel is not cheap.
The sequence below follows that order, from product profile to acceptance test.
Start with the product range: silhouettes, target fabric weight in grams per square metre, stitch structures, colours per row, widest panel and annual volume per style. Each maps to one machine parameter, and that mapping, not the catalogue, sets the specification.
Most mistakes come from choosing the machine first and fitting the product to it. A 12GG machine bought for chunky cables will not pass the yarn through its latches, and a 52 inch bed cannot knit a 140 cm coat panel. Both were specified against the wrong input.
The right input is the approved technical pack of a garment already knitted somewhere, in-house or by a subcontractor. Where none exists, knit three: the heaviest, the finest and the widest style. Those define an envelope that every machine you shortlist must cover.
Write the profile before approaching suppliers, or they will shape the conversation around their stock. It needs silhouette families and size range, fabric weight per family rather than an average, the structures actually sold, colours per row, trim programme, annual volume and style changes per month.
Once the profile exists, the translation is mechanical.
Garment attribute | Machine parameter it sets |
Target fabric weight | Gauge band, which fixes the usable yarn count |
Widest panel, measured flat | Bed width, a limit no setting can exceed |
Stitch structures | Cam functions: knit, tuck, miss, transfer, racking travel |
Colours per row | Yarn carriers, plus intarsia carriers if needed |
Annual volume and style changes | Number of systems, and number of machines |
Assembly route | Architecture, from blank knitting to knit to shape |
Gauge is the number of needles per English inch, and it fixes both the finest yarn the machine can run and the weight of the fabric. Choose the gauge band from the target fabric weight, then confirm the yarn count inside that band, because the two are one decision.
Coarse gauges leave room for thick yarn and give visible stitch definition; fine gauges form smaller loops for smooth, light fabric. Mismatch is costly both ways: thick yarn on a fine machine breaks and damages needles, while fine yarn on a coarse machine gives loose loops and panels that grow after washing.
Gauge | Yarn count, fabric weight and typical product |
3GG to 5GG | Pitch 8.47 to 5.08 mm. Resultant Nm 3 to 9, fabric 420 to 850 g/m2, garment 400 to 850 g. Chunky cables and fisherman rib |
7GG to 9GG | Pitch 3.63 to 2.8 mm. Resultant Nm 10 to 17, fabric 320 to 450 g/m2, garment 280 to 450 g. Crewnecks and polo knits |
10GG to 12GG | Pitch 2.54 to 2.12 mm. Resultant Nm 16 to 30, fabric 180 to 350 g/m2. The workhorse range |
14GG to 16GG | Pitch 1.81 to 1.41 mm. Resultant Nm 26 to 40, fabric 150 to 220 g/m2. Fine merino and lightweight layers |
18GG | Pitch around 1.27 mm. Resultant Nm 40 and finer, under 160 g/m2. Ultra fine luxury knits |
The orientation rule is that resultant metric count at the feeder falls between 1.5 and 2.5 times the gauge number, so 7GG runs Nm 10 to 17 and 12GG runs Nm 18 to 30. Two identical ends halve the resultant count. Because gauge has been measured on both a 1 inch and a 1.5 inch scale, convert quoted figures to needles per inch and confirm on a knitted sample.
Interchangeable needle sections let one bed run 3GG, 5GG and 7GG, or 8GG to 12GG, covering two seasons with one purchase. Each changeover costs production time and fresh tension settings, doubles the needle inventory and demands operators competent on every configuration. Below a few thousand pieces per gauge per year, a second machine is the better buy.
The workhorse is a computerized V-bed flat machine with two needle beds facing each other. Only a double bed can knit true rib, transfer stitches between beds and shape a panel by widening and narrowing, which is why single bed machines cannot knit rib and circular machines cannot shape a panel at all.
The V-bed arrangement lets the beds hold loops alternately so stitches can be moved, held and returned, which is what makes rib, cables and armhole narrowing possible without cutting.
Mechanical machines select needles with cards or chains. They survive for extreme coarse gauge and craft effects, but pattern changes need physical setup, tension varies between operators and shaping is largely manual. Where a stitch must be reproduced across repeat orders, the computerized pattern file becomes the specification.
Three routes sit on the same machine family, and the choice is commercial as much as technical.
Architecture | Output and trade-offs |
V-bed, blank knitting | Fabric cut to shape and sewn. Fast per panel, but wastes yarn and relies on overlock seams |
V-bed, fully fashioned | Panels knitted to shape and linked. Near zero cutting waste and stronger seams |
Whole garment | Knitted in one piece, no linking. Above 80,000 USD, slower, and demands strong programming |
Fully fashioned knitting is where most programmes land, and the knit to shape logic explains why flat beds dominate heavier knitwear. A circular machine knits a tube instead, so a sweater made on one must be cut, which curls the edges, wastes yarn and puts a seam where a selvedge belongs.
Set bed width from the widest panel plus a margin, and treat it as a hard limit. A wider panel cannot be knitted at any setting, and splitting it adds a visible seam plus a linking operation that costs money on every piece.
Measure the widest panel flat after relaxation and add 5 to 10 cm for selvedge, take-down grip and the outer needles you cannot rely on. Check the trim programme too, since a flat knitted collar can be wider than expected. Bodies usually fall between 50 and 65 cm, so a 52 inch bed covers most of a knitwear range.
Bed width | What it realistically covers |
52 inch (132 cm) | Most body panels, sleeves, collars and trims. The common volume size |
60 inch (152 cm) | Longer bodies, oversized drop shoulders, wide sleeves |
66 to 72 inch (168 to 183 cm) | Longline bodies, coat panels, knit dresses, plus size grading |
80 inch (203 cm) | Very long dresses and coats, wide blanket style panels |
100 inch (254 cm) | Blankets only, rarely garment panels, and the width costs floor space |
A wider bed adds frame mass, price and footprint, and the carriage travels further for every course, which lowers output on narrow panels.
Each system is one set of cams and feeders completing one course per carriage pass. Three systems approach three passes in one, but the real gain is 2 to 2.5 times a single system once shaping and stops are counted, so buy systems for the volume you can keep loaded.
System count | Real output and best use |
Single system | Baseline. Sampling, complex intarsia, coarse gauge where yarn limits speed |
Two system | About 1.8 to 2 times baseline. Mid volume, moderate patterning |
Three system | About 2.2 to 2.5 times baseline on plain and rib. Bulk programmes |
Four system and above | More courses per pass, heavier maintenance, fewer qualified technicians |
Brochure speeds of 1.2 to 1.6 m/s mean little alone, because a machine spends much of its life reversing, transferring, changing yarn and stopped for faults. An estimate built from five steps is more honest:
Take the stroke length, panel width plus carriage overrun
Divide by sustained speed and add 0.3 seconds per end for reversal
Divide courses by the number of systems to get the passes
Multiply passes by time per pass for pure knitting time
Apply an efficiency factor of 0.55 to 0.7 for transfers and stops
A 12GG plain body of 300 courses on a one metre stroke gives about 200 seconds of pure knitting on three systems, or 8 to 12 minutes of real time. A five piece sweater occupies 45 to 60 minutes, so 20 hours yields 20 to 25 garments, while a 3GG cable takes three to four times longer.
Four mechanisms decide whether a machine earns its keep: needle selection, transfer and racking, take-down and sinker control, and yarn carriers. Weakness in any one shows up as a structure you cannot knit, not merely a speed you cannot reach.
Mechanism | What it decides and how to verify |
Needle selection | Colours per row and jacquard resolution. Ask for a fine motif sample |
Transfer and racking | Rib, cables, links and the narrowed armhole edge. Inspect a shaped panel |
Take-down and sinkers | Panel stability and openwork. Measure a full width panel after relaxation |
Yarn carriers | Colour blocks and plating. Count carriers and confirm intarsia carriers are included |
Stop motions | Yarn break, needle break and racking fault must stop the machine, not knit on |
Transfer is mechanical rather than a setting: the cams must move stitches from either bed while another system keeps knitting. Racking travel, usually 2 inches each way, decides which cables are possible. Take-down, controlled in 24 or more levels, holds tension through shaping and determines whether a panel measures correctly after washing. Intarsia also needs dedicated carriers, or colour blocks become jacquard with floats that snag.
The control system decides daily productivity, because a machine can only produce patterns your team can programme. Confirm the software is learnable in weeks, that files exchange with customers and your existing library, and that faults are logged for remote diagnosis.
A sweater flat knitting machine with a large pattern library, USB and network transfer, and parameters displayed in your operators' language is productive from month one; one that needs a single specialist to convert every design idles when that person is away. Test four things: time an import of your own design file, ask how alarms are logged, ask whether a controller board is stocked regionally, and confirm training runs on your patterns.
Purchase price is the smallest part of the decision. Landed cost, installation, programming time, needles, energy and downtime over five years outweigh the gap between a cheap and a well specified machine, and utilisation decides payback more than any specification.
Indicative prices run from 5,000 to 15,000 USD for an entry level 52 inch two system machine, 20,000 to 60,000 USD for multi system with full jacquard, and above 80,000 USD for whole garment. Compare cost per piece at your own volume instead.
Cost element | Basis and typical value |
Depreciation | 30,000 USD over five years at 12,000 pieces a year, about 0.50 USD per piece |
Direct labour | One operator to four machines, about 0.23 USD per piece |
Energy | 3.5 kW average draw over 20 hours a day, about 0.16 USD per piece |
Needles and consumables | Annual replacement budget, about 0.10 USD per piece |
Maintenance | Scheduled service and parts, about 0.13 USD per piece |
Total conversion | About 1.10 USD per piece, excluding yarn and finishing |
If in-house knitting saves two dollars a piece against outsourced conversion, a 30,000 USD machine making 12,000 pieces a year pays back in about 15 months. Below 40 percent utilisation it never does, which is why volume must be settled before the model.
A 52 inch machine needs about 2.85 by 0.95 m of floor, an 80 inch about 4.0 by 1.0 m, and a 100 inch about 4.7 by 1.05 m at close to 1.7 tonnes, with pallet truck access and clearance behind the bed. Smaller machines run on single phase 220V and larger ones need three phase with stable voltage. Needles are consumable, so keep an annual allowance and working stock at the machine.
A machine is only as good as the support behind it. Trial knit with your own yarn on a representative pattern, fix the acceptance test in writing before the deposit, and confirm that needles, cams and control boards are stocked regionally.
Send your own yarn in your own counts and ask for the three styles that define your envelope. The same Nm 2/28 wool that runs cleanly on one bed may break repeatedly on another because of feeder geometry, so the trial must produce panels you can measure, wash and inspect.
Write these into the contract and hold a balance against them:
Witnessed knitting of three representative styles at the ordered gauge and width
Measured output over an eight hour run, with efficiency stated
Dimensional check after relaxation and one wash
Function test of every stop motion
Spare parts list with prices and delivery times
Drawings and a parts manual in your language
Operator and programmer training on your patterns
Warranty stated separately for bed, carriage and controller
Three questions surface once the specification is settled: how many machines the order book needs, whether a used machine is a sensible entry, and what to stock from the first week.
Divide the monthly target by working days, then by realistic daily output for your heaviest style, and add 15 to 20 percent headroom for sampling and changeovers. A target of 8,000 pieces over 22 working days needs about 364 a day, which at 20 to 25 pieces per machine implies 15 to 18 machines. Plan on peak season, because a fleet sized to the annual average will be short when it matters.
Yes, with conditions. Well maintained used machines knit the same panel as new ones at a fraction of the price, but the risks are hidden: beds worn enough to lose gauge across the width, carriage guides with play, unsupported controllers and needle types out of production. Inspect with a technician, knit a full width panel and measure five points across it, and price a re-needle before agreeing. Used equipment suits sampling rooms and coarse gauge basics, not fine gauge work that must hold tolerance.
Hold needles at 2 to 3 percent of the bed needle count, roughly 15 to 20 on a 52 inch 12GG bed of about 620, plus sinkers, one spare cam segment per system, feeder belts, a take-down roller covering, fuses and the sensors most likely to fail. Order them with the machine, because freight on a small parts order often exceeds the parts and a missing needle type can idle a line for a week.
How to Choose a Flat Knitting Machine for Sweater Production
How to Choose a Circular Knitting Machine for Sportswear Fabrics
Terry vs. Fleece Circular Knitting Machines: Key Differences
Rib Knitting Machine vs. Interlock Knitting Machine: Which Should You Choose?
How to Choose Circular Knitting Machine Gauge for Your Fabric