Hello, complete Arduino beginner here. I am a hobby weaver, I would like to learn Arduino so that I can develop some tools to speed up the weaving process.
I'm attaching a picture of my loom. Each of the 16 handles operates a 'shaft', which lifts a different set of threads so that I can pass the weft thread through them. I would like to create a device that can be programmed to push down and then release different combinations of levers; the idea would be that I could methodically move through those combinations in a pre-defined sequence by pressing a single button – similar to how you click to progress through slides on Powerpoint.
I'd begin by seeing what is out there as a product that already does what you want. If there is an existing product, find out as much as you can about the device. Consider that taking this project on will, most likely, cost more than buying a manufactured product. I am sure there is much more to consider. This project is doable.
You'll probably start off with buying another loom for use in development.
High speed actuators , possibly air operated, comes first to mind, one for each handle
Each would need to be sized to operate the handle load and fast enough as above.
Major mechanical alteration to the machine would be required to allow the above.
A decent compressor to operate the actuators.
Then you would need pilot valves operated by an interface board to connect to the Arduino.
Arduino choice would largely depend on the eventual program you would need to write.
Attack writing the program with just one or two simulated (LEDs) valves to begin with and work your way up to the final 16 controls plus whatever else would be required.
This is before you get to the shuttle which would be yet another hurdle.
I'd look at servo motors first. Try to work out some way to link them to the levers so they get actuated by the servo rotations. Hopefully they're strong enough, depends on how much force it takes to move those levers.
bluejets:
This is before you get to the shuttle which would be yet another hurdle.
Doesn't look like automating the shuttle is part of the specs to me. I think it's just automating the pattern.
You basically want what is known as a Jacquard mechanism, except driven by an arduino instead of punched cards.
Trying to design a mechanism to actually flip the handles would be the hardest part. After finding a video with a similar loom, it appears the handle pulls on a cord that lifts a horizontal bar. A vertical pull mechanism would be easier to implement, possibly an add-on mechanism that places a hook under each cord to allow for lifting it, while allowing easy removal for manual operation.
In general what you want is a mechanism that uses a single actuator (air cylinder, motor, or even a single handle that you pull yourself) that is capable of lifting all the cords simultaneously, then let the arduino selectively engage or disengage the desired cord from that mechanism. Solenoids are a natural choice for this type operation, although small hobby servos may be more practical, very little torque is required because the servo would be operated at a point where there is no tension on the mechanism. I would probably go for the manually operated handle, with switches for the arduino to detect the engaged and disengaged positions of the handle so it could change the pattern at the appropriate time.
Do you need to keep the lever handles? Does the loom need to go back and forth between manual and automatic modes or can it be fully automatic?
What length of string is actually pulled by the fully rotated lever? Is there a 4 or 6 inch rise to the shafts?
If you replaced the levers with linear actuators and tied the strings to the actuator shaft it becomes mechanically simple with linear actuators replacing the levers but it may not allow for manual operation.
Certainly feasible. I would think that all the mechanical stuff will be the most challenging.
It would probably help to get an Arduino or three before you start the project and begin working through the examples provided with the IDE to get some familiarity with how to program them. Get a few solenoids and/or linear actuators and find out how to wire them up and control them.
david_2018:
You basically want what is known as a Jacquard mechanism, except driven by an arduino instead of punched cards.
Trying to design a mechanism to actually flip the handles would be the hardest part. After finding a video with a similar loom, it appears the handle pulls on a cord that lifts a horizontal bar. A vertical pull mechanism would be easier to implement, possibly an add-on mechanism that places a hook under each cord to allow for lifting it, while allowing easy removal for manual operation.
Thank you for this helpful input, and for your research – you have grasped how the loom works and what I'm after! In fact I think technically what I am trying to design is what's known as a 'dobby' mechanism, where the shaft lifts are pre-programmed and the weaver only needs to press a single button (usually a manual pedal) to move through the sequence of shaft lifts that creates the pattern. The only difference between this and a jacquard mechanism is that jacquard loom allow for lifts of individual threads - whereas every thread on my loom is connected to one of 16 shafts, so I only have 16 inputs to play with.
david_2018:
In general what you want is a mechanism that uses a single actuator (air cylinder, motor, or even a single handle that you pull yourself) that is capable of lifting all the cords simultaneously, then let the arduino selectively engage or disengage the desired cord from that mechanism. Solenoids are a natural choice for this type operation, although small hobby servos may be more practical, very little torque is required because the servo would be operated at a point where there is no tension on the mechanism. I would probably go for the manually operated handle, with switches for the arduino to detect the engaged and disengaged positions of the handle so it could change the pattern at the appropriate time.
I definitely need an actuator (a rotary one I think?) to operate the handles that move the cords that hold the shafts - doing so manually defeats the purpose, as the idea is for a single press to alleviate the need to pull down and release several of the 16 handles each time I want to do a pass of my shuttle. In fact the ideal would be for this to be doable via a foot press of some kind, as that will leave my hands free to move my shuttle (I don't need that part to be automated). I'm open to simply replacing the levers with a motor-operated servo or rotary actuator (I don't yet understand the difference...) but I would prefer to be able to keep the manual option if possible. Are there are actuators that can also be 'switched' to work as manual levers?
The 'next level' up would be to somehow integrate this system with the file format used by weavers (WIF) that encodes both way threads are tied to the shafts (known as the 'threading') and the sequence of lifts (known as 'treadling') for any given pattern. What I'm envisaging is that I could upload a WIF file to the arduino system, and then assuming I've threaded my shafts correctly, all I'd need to do is perform a single foot press to create the pattern.
cattledog:
Do you need to keep the lever handles? Does the loom need to go back and forth between manual and automatic modes or can it be fully automatic?
What length of string is actually pulled by the fully rotated lever? Is there a 4 or 6 inch rise to the shafts?
If you replaced the levers with linear actuators and tied the strings to the actuator shaft it becomes mechanically simple with linear actuators replacing the levers but it may not allow for manual operation.
I don't 100% need to keep the lever handles but I'd prefer to retain the option to manually operate the loom. As long as the 'new' automatic actuators can also work as manual levers, that's okay with me.
I think there is about 2.5-3 inch lift to each shaft when the lever is pulled.
david_2018:
You basically want what is known as a Jacquard mechanism, except driven by an arduino instead of punched cards.
Trying to design a mechanism to actually flip the handles would be the hardest part. After finding a video with a similar loom, it appears the handle pulls on a cord that lifts a horizontal bar. A vertical pull mechanism would be easier to implement, possibly an add-on mechanism that places a hook under each cord to allow for lifting it, while allowing easy removal for manual operation.
In general what you want is a mechanism that uses a single actuator (air cylinder, motor, or even a single handle that you pull yourself) that is capable of lifting all the cords simultaneously, then let the arduino selectively engage or disengage the desired cord from that mechanism. Solenoids are a natural choice for this type operation, although small hobby servos may be more practical, very little torque is required because the servo would be operated at a point where there is no tension on the mechanism. I would probably go for the manually operated handle, with switches for the arduino to detect the engaged and disengaged positions of the handle so it could change the pattern at the appropriate time.
My apologies David - in my earlier post I misunderstood what you were suggesting, and now I think I have it - you're basically saying I should 'work around' the levers by instead creating a vertical lift that selectively engages or disengages the cords, rather than pulling the cords via the levers. That makes perfect sense. So in terms of equipment, where would I begin? I am just slightly overwhelmed at how I'd go about understanding what I need in terms of hardware for this.
For further info, you can see the cords that lift the shafts and are connected to the front levers on page 14, image 27 here:
The cool thing about these cords is that they are made with lots of little connecting 'gaps', so it's easy to put a hook (or another cord) through them. The more I think on it, the better I like David's idea of creating a lift mechanism that selectively engages different sets of cords corresponding to each of the 16 shafts.
Idahowalker:
I'd begin by seeing what is out there as a product that already does what you want. If there is an existing product, find out as much as you can about the device. Consider that taking this project on will, most likely, cost more than buying a manufactured product. I am sure there is much more to consider. This project is doable.
You'll probably start off with buying another loom for use in development.
These mechanisms do exist but such computer operated dobby looms cost thousands and thousands of dollars, much more than I can spend! I think with my existing loom I should be able to upgrade it at a lesser cost.
Projects like this tend to be iterative, i.e. they evolve as you learn what does and doesn't work. It'll almost certainly take longer and cost more than you expect. As already mentioned, you'll really need another loom unless you want to take a months long break from weaving while you work on the project.
I'd start by trying to figure out a mechanism to work one or two levers. That way when you discover that the actuator you chose is unsuitable in some way, you're not out so much cash.
Here is image 27 from the pdf document. It shows the strings behind the levers. Clearly they can be removed from the levers and raised from behind. Whether there should 16 separate lifters, one for each string, or one lifter with 16 separate engagement mechanisms for each string to be lifted is not clear to me.
Thank you all so much for the input so far, it's incredible helpful in terms of crystallising how I might go about building this.
Here is my design thus far. From what I understand, figuring out which linear actuator or solenoid would work would be a matter of using a fishhook to measure the force required to lift all 16 shafts simultaneously (the most the system will have to handle), and then ensuring it can meet that threshold. Anything I need to look out for here?
My main puzzle now (aside from, well, figuring out the whole thing) is how I would connect each one of the 16 pairs of cords that lift the shafts to the linear actuator or solenoid, such that the arduino could 'choose' between them each time the foot pedal was pressed. Do you have any ideas about what such a selective engagement mechanism could look like?
Have a look at how the traditional loom worked, although in this case there would only be sixteen lifters operating the frames instead of individual threads.
Basically a single lifter bar at the top. Metal rod that attaches to the lifter cord at the bottom, with a hook at the top to catch onto the lifter bar. A solenoid or servo can easily push the hook back away from the lifter bar, selectively disabling it and preventing the frame from being lifted. I did run across one example of a dobby mechanism using wooden bars with protruding screw heads to catch onto the lifter bar, but cannot find any great pictures or videos of that working (the company does supply this with either a computer mechanism, or a chain of cards with protruding pins that is totally mechanical).
@david_2018
Thanks for the video of the Jacquard mechanism
The pushing of the hooks onto the lifter bar is going to require 16 movers and a lifting bar.
My thinking is leaning towards 16 direct lifters using small servos like shown by JCA34F. I'm not clear that two wire are required since the lever pulls the string from a single point, and the lifters can be offset. Using servo lifters similar to what people do with rise and fall Christmas decorations and indeed some of the large kinetic sculptures. I would use a small servo turning a grooved wheel to which which the lifting cord is attached. This lifting cord can be attached to the loom cord which goes to the lever. Depending of the radius of the wheel and the lift required the angular rotation can be set. The servos are in parallel so the driving is common. In code, the appropriate servos will be attached for each move.
