Propeller watch. POV - propeller clock Propeller clock on atmega
Hi all! I would like to bring to your attention a simple propeller clock that I assembled on the Atmega8 controller. They are made from readily available parts and are easy to replicate and manufacture. The only thing is that you need a programmer to flash the clock controller and control panel.
A regular 120 mm fan (cooler) was used for the base of the clock. You can use any fans for this clock, both clockwise and counter-clockwise, because while I was assembling this clock, I modified the program a little and switched the display of symbols from the remote control programmatically.
The circuit of the clock itself is quite simple and is assembled on an Atmega8 microcontroller, to synchronize its operation a clock quartz with a frequency of 32768 Hz is used.
The clock is powered by a receiving coil, the energy to which is transferred from a generator with a transmitting coil. Both of these coils make up an air transformer.
There were no particular problems with the circuit and design of the generator, since a generator from a plasma ball was used.
The generator is assembled on the common TL494 microcircuit and allows you to change the width and frequency of the output pulses over a wide range.
Even with a gap of a centimeter between the coils, the voltage is quite enough to start the clock. Just take into account that the larger the gap between the coils, the larger the pulse width needs to be made and, accordingly, the current consumption from the source increases.
When turning on the generator for the first time, set the pulse width (duty factor) to a minimum (the control knob is in the upper position according to the diagram, that is, leg 4 is pulled through resistor R7 to leg 14, 15, 2 of the TL-494). We turn the generator frequency until the squeak disappears, this is approximately 18-20 KHz (tuning by ear), and if there is something to measure the frequency, then we adjust it accordingly within these limits.
The generator board also contains an additional voltage regulator on LM317, designed to regulate the fan speed.
It’s not on the diagram, I didn’t draw it
. Watch a demo video of the clock in action.
Video.
The clock board itself is attached to the base of the fan. I secured it with double-sided tape.
Then I slightly modified the clock circuit from a photoresistor to an infrared photodiode (picture below).
Instead of a simple LED in the transmitter, I now have an infrared one.
The resistor was set to 100k instead of 2k.
The critical moments in the manufacture of a clock are the manufacture of an air transformer and alignment (or rather balancing) of the clock board on the base of the fan.
Take these moments more seriously.
Air transformer.
It was based on a regular 120 mm cooler with bronze bushings. The clock board is glued to the base with double-sided tape.
We bite off the blades from the cooler and grind and level them with a file and sandpaper. The coils are made on a frame made of cable duct. I didn’t come up with this design, I just took this idea from the Internet. To wind the transformer, a base is made from a cable channel. Every 5 mm we make a cut on the sides of the channel and carefully roll it into a circle; select the diameter so that it fits tightly on the plastic base of the fan.
Next, we wind 100 turns of enameled wire with a diameter of 0.25 onto the mandrel from the cable channel.
The current consumption of the assembled transformer turned out to be 200 mA (this is with a fairly noticeable gap between the coils).
In general, together with the fan motor, the current consumption is around 0.4-0.5A.
We do the same for the primary (transmitting) coil, but we try to make a minimum gap between the coils. The transmitting coil also contains 100 turns of 0.3 wire (or 0.25).
In the diagram I have slightly different winding data for these coils.
Hours fee.
The strip with LEDs is made on fiberglass. A hole is drilled in it, a piece of tube from a telescopic antenna is inserted into this hole and soldered to the board (the antenna tube must be cleaned of the shiny coating). You can use any suitable tube, or attach the board in another way, for example using a screw with nuts.
I connected the board with LEDs to the clock board with a regular enameled (winding) wire; it is more rigid than the mounting wire and does not fray when rotated.
To balance the entire board, on the other side we glue a screw with a diameter of 3-4 mm with hot glue, screwing various nuts onto the screw on the other side - we achieve minimal vibration.
To check the functionality of the clock board, we short-circuit the photoresistor with a screwdriver or tweezers; the LEDs should blink.
The clock starts working when 5V (logical unit) appears on the 5th leg of the atmega. That is, when the photoresistor is illuminated, there should be 5V on the 5th leg,
When the photoresistor is not illuminated, there should be a logical 0 (about 0V) on the 5th leg of the atmega, for this we select a resistor to ground from the 5th leg. The diagram shows 2 kOhm, I got 2.5 Kohm.
At the bottom of the fan base we glue an LED so that with each revolution of the fan motor, the photoresistor passes as close as possible to the light source (LED).
Control panel.
The control panel is designed to control the operation of the clock, switch display modes (change the direction of fan rotation), and set the clock time.
The remote control circuit is assembled on an ATTINY2313 microcontroller. The board contains the MK itself with a harness and six buttons designed to control the clock.
I didn’t assemble the housing for the remote control, so only a photo of the board itself.
Information on the purpose of the remote control buttons;
H+ and H- clock settings
M+ and M- minutes setting
R/L change of direction (for screws rotating clockwise and counterclockwise)
font change font (thin, bold and website inscription)
When writing a site, use the H+ and H - buttons to adjust the width of the inscription.
The attached archive contains all the necessary files for assembling the watch;
Archive for the article
If you have any questions about the design of the watch, ask them on the forum, I will try to help and answer your questions as much as possible.
This article is about making unusual watches. They have many names - propeller watches, Bob Blick watches. The screen of this watch is not like any of the watches we are used to. A mechanical display is used to display the time. It is a rapidly rotating lever with LEDs installed on it, which form the image.
The lever rotates at a frequency of about 1500 rpm and the diodes light up and turn off for a strictly defined time. Since the lever rotates at high speed, it is almost invisible, and we only see flashes of LEDs. In each position of the lever, the LEDs light up in a certain combination, which allows you to generate graphic and text information.
Depending on the shape of the lever, the display can be in the form of a cylinder or a disk. The straight lever allows you to imitate a clock.
It is believed that Bob Blick was the first to make such a watch. On the Internet you can find a large number of different options for such watches. This clock was modeled after Henk Sotheby's.
Basic functions
Below are the main functions of the watch:
Time and date display
Setting all parameters from the RC-5 type remote control
Time display in digital and dial modes without date and with date
Displaying five-minute divisions
Uses 5mm super bright LEDs
Creeping line with character generator.
A running line with a length of 128 characters is written to the EEPROM.
Demo mode. Cyclic switching between ticker, analogue and digital display.
Setting the time
Since all the electronics are on a rotating lever, the question arises: How to set the time? In many models, the time is set on the lever itself using special buttons. With this design, you will be able to see the set time only after the lever is activated. If the setting is incorrect, you will have to stop the lever again and again set the time blindly. In this watch, setting is done from the remote control. Setting the time in dial mode looks especially impressive.
Mechanics
Let's move on to the most difficult stage of watchmaking - mechanics. First, you need a fan from the computer's power supply. It is highly advisable to use a high-quality fan with ball bearings; this will significantly extend the life of your watch. As a rule, the rotation speed of computer fans is 3000 rpm or 50 revolutions per second. This rotation speed allows for a very stable image. But a lever rotating at such a speed creates a lot of noise. So I lowered the speed to an acceptable noise level.
Energy can be transferred from a stationary part to a rotating part in different ways. The most common is sliding contact. This method has many disadvantages - contact instability, noise, mechanical wear. The watch I made used a more elegant method. A transformer consisting of moving and stationary work. Its production is perhaps the most important stage in the manufacture of watches. First of all, you need to carefully disassemble the fan. To do this, you need to peel off the sticker from the back. And carefully pull out the retaining ring. After which you can remove the impeller and rotor. We don't need the plastic impeller anymore either. We remove it from the metal base and wind the secondary winding onto it. The winding contains about 150 turns of winding wire with a diameter of 0.3 mm. This is approximately 5 layers. Each layer was coated with silicone sealant (available on any construction market) and dried.
I highly recommend using wire in silk insulation - this will make it easier to fix the turns. A regular wire will slide off the metal base.
To attach the lever, several holes are drilled in the rotor.
Most of the plastic is removed from the stationary part of the fan, leaving only the bottom frame.
The gap between the primary and secondary windings should be minimal. In reality it turns out somewhere between 0.3 – 0.7 mm. To make the primary winding, it is necessary to make a mandrel. To do this, take any cylinder of a suitable size (I used an old capacitor) on which the required amount of paper is tightly wound until the desired diameter is reached. Next, about 100 turns of wire are wound around this mandrel, similar to the secondary winding. After the sealant has dried, the mandrel is carefully pulled out. The resulting wire ring is centered and fixed with sealant to the base of the fan. Thus we received a transformer for transmitting energy to the rotating parts.
Next you need to make a rotor position sensor. For this, any infrared LED and phototransistor are used. The LED is installed on a fixed base. Phototransistor on the rotating part at the same radius. Thus, the phototransistor would light up once per revolution. It is convenient to use a cut optocoupler.
Electronics
The watch electronics consists of two parts - rotating and stationary.
Fixed part
Schematic diagram of the fixed part
It is implemented on the pic16f628 microcontroller, which decodes commands from the IR receiver. This allows you to turn the clock rotor on and off. In the on mode, the microcontroller supplies a PWM signal to the gate of the transistor, which modulates the voltage in the primary winding of the transformer. You will have to select the PWM frequency yourself. For each transformer it has its own optimal value. In my version it had a value of about 7 KHz. The disadvantage of this is a slight whistling of the engine rotor. It is better if it is more than 16 kHz.
In off mode, the engine turns off. Then, after a few seconds, the duty cycle of the pulses in the primary winding decreases. In this mode, energy is needed only to keep the clock running.
To adjust the engine speed, an LM317 microcircuit is used, which is turned on by a key on a field-effect transistor.
Rotating part
Schematic diagram of the rotating part
Energy to the rotating part comes from the winding on the rotor. The voltage from the rotating part is supplied to a rectifier and stabilizer providing 5 V to power the microcontroller. At the input of the microcontroller there will be signals from the IR sensor from the remote control and the lever position sensor.
All LEDs are connected through transistors turned on in current source mode. Thus, the LEDs are protected from overvoltage, which can reach 40 volts. This voltage may vary depending on the LEDs turned on at the same time. The diode current can be taken equal to 50 mA, since the diodes operate in pulse mode.
Finally, I realized my long-time dream - I made a propeller watch! I got this idea a few years ago when I saw this watch in action on You Tube.
The implementation of the idea was complicated by the fact that all the circuits, and there are simply tons of them on the Internet, are implemented on PIC controllers, and I still haven’t been able to flash it. I tried a bunch of programmers, but either my hands are crooked, or the stars were aligned at that time, but all my attempts were unsuccessful. But I haven’t found any circuits on Atmel microcontrollers, the programming of which I have no problems with. I tried to encourage programmers I knew to write a program for AVR, but they didn’t find a response in their souls. Maybe the idea would have remained buried under the rubble of collapsed hope, but recently I started looking through my collection of various circuits on disks that I bought at a flea market...
Small update . The watches made above turned out to be difficult for our readers to replicate. Therefore, a simplified version was made, without the use of machines. Detailed
Many outlandish electronic projects can be found on the Internet, which gives the inquisitive mind no rest.
And even though the “propeller clock” is far from a new thing on the big Web, when I one day came across a diagram of a clock with a stroboscopic effect, I couldn’t pass it by.
A little theory
The main idea of the device is microcontroller control of a group of LEDs mounted on a rapidly rotating base.
The code specifies a loop that repeats from an external interrupt. Let's say the length of the total burst is 15 ms. During this period of time, each LED lights up n-number of times. At low rotation speeds, the human eye will only detect a single switching on of all the LEDs at once. But, as soon as the rotation speed is increased, small intervals of the overall burst will begin to stretch along the X axis, and the eye will begin to detect non-simultaneous triggering. This will continue until a certain limiting rotation speed, at which the 15 ms interval will be rotated to a certain length along the X axis, at which the blinking intervals within the overall burst will be clearly distinguishable and the numbers will be drawn that will add up to the overall picture. A further increase in the rotation speed will lead to a stretching of the total packet of pulses and the numbers will become unreadable.
The board was redesigned for SMD components, because the lighter the board, the less the load on the fan.
The rotating part consists of a main board and an indication board on which LEDs are installed.
I used SS12 Schottky diodes as rectifier diodes. I soldered an 18-pin socket under the microcontroller, since an “idle start” was necessary.
The length of the arm can be adjusted to taste, taking into account the comfortable viewing of the luminous part. In my opinion, a 90-110 degree scan is optimal. A scan option of less than 90 degrees will confuse the numbers, and more than 110 degrees will stretch the image too much in diameter.
Initially, I chose a shoulder length of 65 mm, but the experience was unsuccessful and I sawed off the finished board to 45 mm.
The LED board looks like this:
It has 7 main LEDs and 2 backlight LEDs. All LEDs are 5 mm in diameter.
Connections between two boards are made by soldering the connecting pads. I etched the boards, carried out installation, and connected them. Now you need to place them on the fan rotor.
To do this, I drilled 3 holes with a spread of 120 degrees.
I inserted countersunk screws with a diameter of 3 mm and a length of 20 mm into them. I secured it with nuts and secured the boards to them.
The ends of the secondary winding were soldered to the board. I installed a compensating counterweight on the opposite side of the display board to reduce runout during rotation.
The time has come for an idle run without a microcontroller. I placed the rotor with the circuit boards in its place on the fan and supplied power to the RF generator, the fan is still motionless. The backlight LEDs came on. I checked the input voltage, it dropped to 10 Volts, this is normal. It remains to install a synchronizing optocoupler consisting of an infrared photodiode and an infrared LED. An IR LED was glued to the base of the fan and powered from the main +12 V power supply through a 470 Ohm resistor. A regular IR photodiode is soldered onto the board.
I installed the optocoupler so that when rotating the photodiode would fly over the LED as close as possible.
I programmed it.
I installed the controller in the socket and secured the rotor with a retaining ring.
It's time to launch!
The first inclusion made me happy and sad at the same time. The circuit worked, the LEDs showed the time 12:00, as they were supposed to, but the image was blurry along the X axis. I began a “debriefing”, as a result, I came to the conclusion that it was necessary to replace the photodiode. The spread of the response area from the external interruption of the MK turned out to be too large.
I decided to install a photodiode with a narrower radiation pattern, and also covered the LED with black electrical tape.
The triggering area decreased 2-3 times, and the subsequent activation was pleasing: the blur completely disappeared.
Let me note once again that low-power fans will not accelerate this design to the required rotation speed, and the picture will flash before your eyes. I reworked the project three times, and only the version on a fan with parameters of 0.4 A; 4.8 W; 3200 rpm worked fine.
An obvious disadvantage of the design is the lack of a backup controller power supply. Yes, yes, the time will be reset every time the main +12V power supply is removed.
Unusual dynamic LED clock powered by a motor from a hard drive.
Device diagram:
Well, when all doubts are put aside, we can begin...
To make a propeller watch we will need:
* 2 sheets of fiberglass, one is double-sided (45*120mm), and the second is single-sided (35*60mm).
* Iron and Ferric Chloride (for etching boards).
* Motor from HDD drive.
* Soldering iron with a thin tip, mini-drill.
For watches:
* LED driver MBI5170CD(SOP16, 8 bit) - 4 pieces.
* Real time clock DS1307Z/ZN(SMD, SO8) - 1 piece.
* Microcontroller ATmega32-16AU (32K Flash, TQFP44, 16MH) - 1 piece.
* Quartz resonators 16MHz - 1 piece.
* Quartz resonators 32kHz - 1 piece.
* Ker. capacitor 100nF (0603 SMD) - 6 pieces.
* Ker. capacitor 22pF (0603 SMD) - 2 pieces.
* Ker. capacitor 10mF*10v (0603 SMD) - 2 pieces.
* Resistor 10kOm (0603 SMD) - 5 pieces.
* Resistor 200Om (0603 SMD) - 1 piece.
* Resistor 270Om (0603 SMD) - 1 piece.
* 2kOm resistor (0603 SMD) - 4 pieces.
* Clock battery and holder for it
* IR LED
* IR transistor
* LEDs (0850) 33 pieces (one of them (the last one) can be of a different color)
For the motor driver:
* TDA5140A motor driver - 1 piece.
* Linear stabilizer 78M05CDT - 1 piece.
* Capacitor 100 mF polar (0603 SMD) - 1 piece.
* Ker. capacitor 100 nF (0603 SMD) - 1 piece.
* Capacitor 10 mF polar (0603 SMD) - 2 pieces.
* Ker. capacitor 10 nF (0603 SMD) - 1 piece.
* Ker. capacitor 220 nF (0603 SMD) - 1 piece.
* 20 nF - 2 pieces.
* Resistor 10 kOm (0603 SMD) - 1 piece.
1) First we need to make 2 boards.
2) We are looking for an old unnecessary hard drive to remove the motor from it, in some hard drives the motor is not attached with bolts, but is pressed into the case, pay attention to this when choosing a hard drive, otherwise you will have to cut it out :)