Hello out there,
I`m currently planning a new long time project of mine. I was inspired by the amazing clock of Christiaan Postma Clock by Christiaan Postma on Vimeo . I translated the layout to German and came to a number of 164 moving elements. I planned to move each element by one stepper motor, but I don't need to control each stepper individually. The 28BYJ-48 stepper driven by ULN2003 controller seems to be the cheapest solution (less than 1$ for each pair). For RTC I would take the DS3231 module. Challenge of this projects seems to me the wiring and controlling of the steppers. The cheapest solution I could imagine is to control the ULN2003 directly from the Arduino. Negative side is the amount of wires. Is there maybe an easier and cheaper way? I made an schematic of the wiring, it would be great if someone could check if I could drive the steppers this way.
Components:
-1x Arduino UNO
-1x DS3231
-164x 28BYJ-45
-164x ULN2003
-1x Button (to set the time in hours)
-wires
-1x external power supply 9v (power will be calculated when first sample stepper arrives to measure the consumption)
Problems to solve(Questions):
-Can the Arduino pin handle x41 ULN2003 without being overpowered?
-Which lengths can I use for the wires
-Save the actual position of the motors independent from power source (in case of an power lost)
Does each of the 41 steppers need to move at a different speed or can a single step signal control all of them?
Looking (briefly) at the video suggests that several of the "hands" are linked together and could be controlled with a single motor.
If you need to move each stepper with its own set of steps you may find that the Arduino cannot calculate fast enough - that depends on how many steps are required each second.
Very cool. All the motors rotate at the identical speed. You don't even need separate ULN modules. You can just parallel steppers on the output of one module. You could even try stringing them in series too. Of course current, voltage and power will need to be calculated for the strings and an appropriate driver selected.
The video description suggests that it was done with clock mechanisms. They are kind of like steppers with a simple one-pulse-per-second input. For the bulk quantity required, it may be more effective to buy cheap clock mechanisms.
@MorganS: Thanks for your great input. I`ll have to take a closer look at those clock mechanisms maybe they could make things quite easier. But as I already saw they are using a single coil with some gears which take more space and needs a housing. The original clock is 1.4m x 1.4m in size but mine will be quite smaller I think those gears could take to much space.
This project is intriguing. But so many parts etc! Could it be better implemented using a LCD screen and software? Also thinking that if the system generated 'five thirty eight' then it would be better than 'five' etc. Maybe your clock is going to do that.
MorganS:
Because it's easier to run wires and clock individual motors than it is to run a belt or chain drive around 144 spindles.
Maybe a single motor for each row or column.
If not, this looks like an opportunity to use the tiny stepper motors that are used for car instruments and can be powered directly from the Arduino I/O pins.
Southpark:
Could it be better implemented using a LCD screen and software? Also thinking that if the system generated 'five thirty eight' then it would be better than 'five' etc. Maybe your clock is going to do that.
An LCD screen is not an option for me I like the look of those mechanic parts. Yes, I startet the project at first with the layout of the well know Wordclock. Bad side is that you need something around 500 to 600 motors to realize this layout. Way to much for me.
Robin2:
Maybe a single motor for each row or column.
It could be an option, but you also have to keep in mind that every Element needs to be absolutely synchronous or after months of spinning you can't read the clock. That is my biggest fear in this project spending hours of work and get a clock which is not working correctly. The only solutions I see to connect the elements with no slip are gears and chains (belts have slip) and both needs definitely more work in assembling and building.
Back to the electric side of the project. How is it possible to string all motors in series? Would there not be a drop of the voltage for each motor? If I only use one ULN2003 (or a bigger one) I need to use extra relays to handle the current? Just a quick calculation:
28BYJ-48 12V Stepper
12V Current per half-winding:125mA
Total PSU current: 250mA
7V Current per half-winding: 70mA
Total PSU current: 140mA
I want to run the stepper on 9V which is interpolated something like 184mA total psu. For 164 Stepper this is 164*184mA=30.176 or 15A for each channel. Thats a lot.
Most of the time the steppers stands still and than they need such a high current. So I maybe should use a Capacitor for those short times of power consumption.
Go back and read Robin's Stepper Motor Basics. Steppers use about the same amount of power when they are standing still. This thing is going to make a lot of heat when it is switched on. I can't imagine the power bill for running it for a year.
A low voltage stepper will work on only a couple of volts, but it needs a proper current driver and not a ULN. These should be possible to put two or three in series on a 9V supply. It's not ideal, the inductance of the windings may not be happy. Parallel should work fine. But you have so very many.
MorganS:
Go back and read Robin's Stepper Motor Basics. Steppers use about the same amount of power when they are standing still. This thing is going to make a lot of heat when it is switched on. I can't imagine the power bill for running it for a year.
I don’t want to power the steppers when they standing still. I thought because of their permanent magnet and the really light weight clock hands that they might hold their position without power. I have ordered 5pcs of smaller micro steppers. As soon as they arrive from China Ill wanted to check If this will work. Did I have missed something and you could already say that its not working this way?
MorganS:
Because it's easier to run wires and clock individual motors than it is to run a belt or chain drive around 144 spindles.
Though I think the mechanism involved would be a work of art in itself - a single belt can drive many shafts
(though the friction forces rapidly get worse), so you might be able to reduce the complexity somewhat. Having
12 motors each driving 12 shafts might be better(*) than 144 motors or 144 shafts from one motor.
