Hello everyone, i have this idea of making rugs using a Corexy machine, the idea is to use a tufting gun instead of the nozzle so it can move in 3 direction X Y and Z, the only problem that i have right now is that the tufting gun need to be oriented in the direction that the carriage is going for example if i wanted to make a square the tufting gun need to turn forward then turn to the right then turn backward and finally to the left. i had this idea of using rotary encoders that are connected to an arduino board that control a stepper moteur that turns the tufting gun, the idea is that the encoders will read the rotation of the two steppers on the corexy the orient the tufting gun in the right direction , i don't know if it will work or not, and there's another idea that the motherboard of the corexy will directly rotate the tufting gun without any need of rotary encoders by extracting the x and y coordinates directly from the file. I hope this make sens, i explained it the best i could and hope someone can help.
What you want is a C axis (a rotary axis parallel to the Z axis). For this you will need a more advanced controller than the BlackBox, and some way to generate appropriate Gcode.
My rugbot uses Linuxcnc for the controls, and I do my designs in Inkscape which has an extension, Gcodetools, which has a setting for “ tangent knife” that calculates the rotation. There is an old video @ which shows some of the idea although I have made some improvements. My RugBot project was intended to see if I could successfully build an automated rug/carpet tufting machine for personal use, using as much locally available materials as possible and using a minimum of power tools. And it works. However, there were never any drawings and due to life changes it no longer exists, but I learned a lot and am willing to share. There are some YouTube videos under the name of Badger Bond that show some of my various fiber projects which are all of fairly early stages and not final versions, not that anything is every really a final version. I researched online videos of industrial machines and tried to scale them down to what I could achieve. My RugBot is mounted to the wall of a garden shed, in a relatively arid region, and nothing about the shed is square or level. It goes from floor to ceiling and was extended to about 12 feet wide and I successfully made a bunch of 4X7 foot rugs and some 3X10 foot rugs, along with many smaller ones. It works at about 240 stiches per minute, which is limited by the electronics, and is fast enough for me since I need to constantly check the yarn for defects. I spin my yarn on the machine also shown in a YouTube video. The gantry is belt driven by nema 23 motors, a 3:1 belt reduction, 20 tooth drive pulley, and HTD5 15mm wide belts. I tried loop belts driven by a motor on the wall, fixed belts driven by 2 motors on the gantry, and one motor on the gantry with a shaft connecting top and bottom drive pulleys, they all worked adequately. The motors are maybe 280 in-oz motors and are plenty powerful enough and can accelerate the gantry fast enough to self-destruct. Slow and steady is much better. Much of my time was spent on trying to get a grip on how much accuracy was needed. Tufting rugs does not require great accuracy since you are working on the back of the rug, on flexible fabric, and the important side is the finished side which fluffs up and moves around. I use a heavy weight single ply wool yarn that is about 340 yards per pound, that I spin, but I have also tried different yarns and they all work. The stiches are about 4.5 mm long, and the rows are spaced 4.5mm apart, but you can use what ever works for you. The original guide rollers on the version shown in the video were made from neoprene rubber tube slipped over bronze bushings turning on shoulder bolts. The punch needle in the video is from an Oxford Punch Needle and worked reasonably well. While they are excellent hand punch needles the geometry of them on a machine causes a force in the opposite direction of the travel, which combined with the soft rubber wheels, did not always provide adequate accuracy. Before the motorized movement shown in the video, I tried a pneumatic movement that while it was entertaining to watch it did not work well. When the needle tries to enter the fabric, it sometimes snags a tread before working into a “hole”, which creates an opposite force on the “Z” axis. Due to the flex of the fabric the Z axis needs to be “spring loaded” to stay in contact with the fabric and the first versions used some elastic cord which didn’t always work well and sometimes led to incomplete depth of the stitches. I later switched to a pneumatic cylinder which was able to apply a much more consistent and adjustable force, which could have made the punch needle viable, but I had also made other changes by then. My last version of the needle mechanism was a partially deconstructed electric tufting gun that is now available online and generally has the blue handle. I like “loop” pile, but I would think the “cut” pile guns would also work fine. Due to the design of my rugs, I wanted the direction of the stitches to be done in a continuous spiral rather than a back-and-forth raster pattern. This meant that the yarn had to go through center of rotation to avoid tangling, and that if the needle motor was stationary. I had to figure out how to have two oscillating motions, for the outer needle that opens up a “hole” in the backing fabric, and the flat blade that pushes the yarn through the hole, to work and have unlimited rotation. I only used the front half of the gun and mounted it in a way that it would rotate around the axis of the needle. This was accomplished by having the drive motor move a pair of sliding carriages, through a pair of “Scotch Yoke” mechanisms, that pushed a pair of concentric tubes that could also rotate with the needle. The tubes were connected to push rods that replaced the original ones on the tufting gun and the yarn passed through the center tube. After much fussing about with my relatively primitive skills this actually worked and was fun to watch and listen to. The front of the gun was mounted to a disc that was constrained by a ”lazy susan/turntable” bearing which while not very precision was close enough. It was tricky trying to get the gun mounted in the proper location to rotate correctly but with trial and error it got fairly precise. If I were going to do this again, I would keep the tufting gun more intact and use a “slip ring” for the power and encoder to enable continuous rotation with the yarn passing through the center of the slip ring. My plans on the next version, if there is one, would be to use OpenBuilds C-Beam and wheels for much of the construction. When I started there were not as many options, and I was concerned about the axial forces on v-wheels so I opted to use wheels on three sides of the linear rails (wood strips) so all the bearing forces would be radial. I have since used C-Beam and the black wheels on an upgraded spinning mill and think they are up to the task. The vertical axis is longer than available C-Beam, but I think it could be spliced with pins in the interior holes and maybe some plates on the outside. The version in the video was cantilevered off the wall and this did not work as the plywood sagged and it jammed up. Later versions had wheels under the bottom and ran on the floor which worked much better. Guide rollers still were uses on the wood rail on the wall, but I would now use v-slot on the floor, and I feel confident that using a bunch of v-wheels would adequately support and constrain the bottom. The top would be constrained by v-wheels running on a smooth piece of vertical metal, probably a piece of angle, so that it would keep it aligned but could wander vertically if the base was not perfectly flat. Due to the flexible nature of a rug this would have no impact on the finished product. The “Z” axis would be another section of C-beam to allow for an air cylinder “spring” in the center. The carriage for the tufting mechanism would then ride on the second C-Beam. While OpenBuilds does have an excellent selection of plates available, I have had some custom steel plates laser cut by a local shop when I needed larger ones. There is a fair amount of friction at the contact point of the tufting mechanism and the fabric due to the amount of force needed to keep the needle from bouncing back out of instead of going into the fabric. This can pass a fair amount of torque into the vertical axis so I would also mount that axis closer to the wall and allow for more horizonal carriage motion. My wooden carriage and mechanism are fairly heavy at 30 – 40 pounds (?) so it is counter balanced by an equal weight on the back side of the column. My first version of the vertical movement had the belt looping over the top and bottom of the column and including the counterweight in the loop. This worked fine but for some reason I moved the drive belt to the carriage side of the column and the counterweight is just suspended by a cord going over a pulley at the top and it works equally well. On the theoretical second build I would try to have the belt loop around on the inside of the C-Beam channel. The total weight of the vertical column, carriage, and counterweight is well over 100 pounds and the nema 23 motor with reduction has no issues in moving this at the fairly slow speeds and accelerations. So that is a fairly general overview of some of the mechanical concepts. I may try to find some still photos that may be useful and post them. I have found that OpenBuilds to be a very reliable and more sturdy system than I expected. They are very helpful and will even cut extrusions to length and tap ends. For larger pulleys, shafts, and other custom parts I have found Misumi to be very useful since they will custom bore to your needs. I used HTD5 15mm belts after researching backlash and such, but I think most any profile would work fine with enough accuracy for tufting. It is more a matter of trying to find a matching set of open-end timing belt and coordinating pulleys that you can get bored to the motor shaft and belt pulleys as needed. In the US McMaster-Carr is good for a lot of fasteners, and I also find ServoCity, VEX Robotics, and SDP-SI to have useful bits of kit. I found that the 3D printer can help in many cases and parts can be quite strong and durable if sized properly with the right materials. For the RugBot there are three general parts to the project. There is the mechanical part of the machine itself, the design process, and the computer control part of it. For design and g-code generation I use Inkscape with the extension GcodeTools which can figure out the rotation of the tufting mechanism to follow the direction of travel. Details of my convoluted method and hacks of Inkscape maybe another post here or somewhere else. For computer control I use LinuxCNC which is the only control software that I found that was capable of coordinating all the movement and the stitching in a workable manner, and their community support was critical in making this happen. My files and some conversations are in thier forum somewhere and probably searchable under Rugbot. Badger, Sep 27, 2022
Hello badger, my name is Gleidson, I'm from Brazil, from the state of Minas Gerais. I'm following what you posted about a carpet making machine, I'm trying to assemble a machine using linux cnc to operate this machine, to try to produce one like your RugBot that has a video on youtube. Hello badger, my name is Gleidson, I'm from Brazil, from the state of Minas Gerais. I'm following what you posted about a carpet making machine, I'm trying to assemble a machine using linux cnc to operate this machine, to try to produce one like your RugBot that has a video on youtube. thanks for sharing your knowledge. I saw that you mentioned that you made improvements to the machine, would you have pictures on any forum?
Gleidson - I have relocated and now have no space for large equipment so the rugbot no longer exists. The punch needle in the video was replaced with parts from a tufting gun where I used the same servo motor in a fixed position and the needle parts could still rotate. My upgrade plans were to use a more intact tufting gun in a rotating carriage and either replace the motor with a servo, or hack an encoder system. This would also probably mean using a slip ring for all of the electrical connections to avoid twisting the cables. The wood and rubber rollers didn’t end up being as rigid as hoped so I also thought of using some OpenBuild parts that seem sturdier.
Thank you very much Could you help me understand 2 things: 1- When the machine is doing the sewing, there is a motor with a reducer that makes the needle rotate in the direction that will be sewn, how did you make this gcode for axis A to follow the direction of Z? 2 - When reaching a tight curve, the needle stops working and only turns on again when the needle rotates in the right position, and it turns on again. How did you do that?
hello friend, how did you turn on the servo motor to make the needle pulsate, in a few moments it stops and then returns, how did you turn it on the board? was it output to a spindle type tool or as a 5 axis? I didn't understand how you did it
The blackbox controller (or at least the one I have) allows you to control a servo separately from the Z-axis. And Badger mentions that GCodeTools outputs rotation angle (for a 'tangent knife'), presumably as part of the gcode file. You may be able to use that information to convert the rotation value to an instruction for a servo which could then control the angle of the needle. Alternatively you could use 'the Z-axis' to control rotation and use the servo to move the needle up and down. Regardless, that is the direction I would take if I were using a Blackbox controller.
