We are a small craft business and have an ancient Jacquard loom, approximately 100 years old. Unfortunately, nowadays, machines that would suit our needs are either not available or incredibly expensive. The issue with modern machines is that the weaving of ribbons is not symmetrical, as they use an 'arm' system rather than a 'boat' system.
I want to build a system with 200 motors, each of which can move independently. The speed of the Jacquard is 60-100 picks per minute. That's why I can't buy a modern controller for it, as it requires a minimum speed of 500 picks per minute. The motors would need to move back and forth, essentially working similar to how the Jacquard punchcard operates, with a 0 position and a 1 position. The system needs to be expandable to accommodate up to 500 motors.
Briefly, I read through the entire forum post from start to finish:
After researching, the setup would look like this:
I would use a three-phase power supply for this
200 units of Steren Micro servomotor
13 units of Adafruit 16-Channel 12-bit PWM/Servo Driver - I2C interface - can control 16 motors
or 10 units of SparkFun Ardumoto - Motor Driver Shield - can control 20 motors
2 units of Mean Well LRS-350-36 adapter
1 unit of Magnetic contact switch (door sensor) - (to signal the motors to transition to the next state)
I welcome any comments and advice, and I would like to ask which Arduino might be the appropriate choice?
Do you have the mechanical/physical design all done? The three-phase power supply is irrelevant as ALL the other stuff is DC powered. Don't forget when you move motors, you are also moving a lot of wire, both signal and power which must be kept apart as far as possible.
Let's assume for the moment that your parts choices are correct. Now, let's impose a reliability requirement. The electronics will likely be fine or can be easily upgraded, but it's the electromechanical bits that I'd worry about.
How far do those servos have to move with each operation, and at 100 operations/minute (I assume that's a "pick"), how many picks per day do you expect?
Now, how long should this machine work without failing and what's the cost of failure?
If I were automating a machine like this, the first thing I'd do is buy 5 representative servos and run them at the expected rate (or maybe 20% faster) and see how long before one fails. It's not statistically perfect, but that failure rate will be 40 times higher with 200 servos.
The alternative is purchasing servos or other actuators that have specified lifetimes.
@cedarlakeinstruments has an excellent point. Most servos are for infrequently used toys, most likely contain plastic gears, and will not last long on a machine that runs for hours on end, day after day. Be sure to do the recommended lifetime test before committing to this design.
Thank you and jremington for mentioning this valid problem. As you advised I will try several motor types between 2 - 7 dollars. The motors normal working period will be 60-100 state changes (picks) every minute maximum. The set up of the original machine now is 60 picks / minute and it works about 7 hours a day.
I also worried about the heat these cheap type of servomotors make, as they will be very close to each other, but of course not in contact with each other.
If I replace the servo motors with stepper motors it will probably triple my budget. Also the step motors are much bigger then servo motors, but they are more robust and better able to withstand frequent position changes maybe.
Can you (or anyone) recommend a servo or stepper motor that would be suitable for this task?
I searched for solenoids and linear actuators also, but their prices are higher then step motors, as I found. However, yes, it is possible that almost no service is required due to the lack of wear parts.
If you expect to make serious use of the revised machine, you will probably have to revise your budget. Cheap junk is not a bargain -- it is just cheap junk.
If a solenoid will give you sufficient range of motion, then I'd go with that. There's not much that will destroy a solenoid short of overheating.
I once asked an engineer why he specified a small DC motor that cost $300 to move a load that was only a few grams when there were so many cheaper choices (any motor from a toy would have worked). His answer was simple: this motor will run hundreds of cycles every day for close to two decades. It's cheaper in the long run to use a motor that won't fail than to make a single service call to replace it.
I love all the cheap boards and electronics we can buy these days. They enable many applications that were simply too expensive before, but you have to make sure that using them doesn't introduce a weak link.
Do the intend to use the servo motors to replace the punch card itself, by selectively blocking the pins that go through the punch card, or are you intending to actuate the pins themselves?
You mentioned that a modern controller is not capable of operating slowly enough for your use, but could you obtain the actual actuators used by the modern machines and drive those with your own controller?
Scaling up anything requires an extra level of skill over and above that required to get a simple one to five system going. An example I often use is scaling up making bread in your kitchen to making bread to sell in a shop.
The big big problem with a stepping motor is setting the initial position. Normally this is done by using a feed back switch, mechanical or optical, to allow you to find an initial position. After this you can use them using only the number of pulses required backwards and forwards.
This worries me as to your experience level with electronics. This is a project that requires a lot of skill.
and you gave the link.
The big difference with that link, and what you want to do is the speed of operation. That was for a hand loom where the speed and frequency of the lifting of the threads is quite slow.
I doubt you will find a servo that can lift that fast, even if you supply them with 6.5V instead of 5V. Also most servos will not move through an angle of more that 180˚. Then you have the problem of converting that rotary movement into a lateral movement. I have lately been doing a project that involves this problem. I solved it by tying waxed twine to the pull down mechanism I required, and this required some calibration to take up the slack on the thread loops that I tied. The return mechanism was built into the object I was trying to pull down, so there was no control over the return speed.
Continuous servo motors have no way of controlling the position of the motor only the direction and speed of it. So that is not an option for your project.
The motors would replace the hooks. So instead of the machine hits the punchcards to the head to select the hooks, the motors/solenoids would move the harness cords.
Thank you for your comments. I don't try at all costs to get the most out of everything at the cheapest price, because obviously it won't work out.
You're right, it's a bigger project than I can jump into in my experience, but it has to be done, and the worst that can happen is that I give up trying.
Maybe then I could use electromagnets to hold the hooks that are needed. These hooks are currently raised by the machine in its own rhythm according to the punch card. If I pick up all the cards in each beat, but only keep the right ones up, then maybe I have eliminated the motor and speed problems.
Thus, perhaps no major transformation is necessary. The question is how much of a N-strength magnet is needed to securely hold the hook. But this can be roughly calculated and easily tested.
I would think solenoids would be much better than any type of motor.
Directly lifting the harness cords would take a lot of force. Driving the selector mechanism that is currently being operated by the pins that read the punch card would not take nearly as much force, and would not requires as extensive a modification to the machine.
Do you want to preserve the ability to use punch cards?
Could you post some pictures of the actual jacquard mechanism on the machine?
We had to move the machine and this was made then. The other picture is already in place, but it might not look so good there.
I'm not sure if the solenoids can provide enough pressure to replace the punch card.
The best solution would be to somehow replace the punched cards, because then the machine will obviously do the rest. It is currently a 100 hook Jacquard, and we can upgrade it to a 200 hook Jacquard or more. Fortunately, we have more experience in the construction of metal and wooden structures.
A link about the operation in the previous place. I'm not in workshop to make better video, but I will.
