Search, interview, test task, selection, hiring, adaptation - the path is difficult and understandable to each of us - both the employer and the employee.
A beginner does not have the necessary specialized competencies. Even an experienced specialist has to readjust. The manager is pressured by questions, what tasks to set for a new employee at the start and how much time to allocate for them? While ensuring interest, involvement, drive and integration. But do not risk critical business tasks.
To do this, we launch relay-race internal projects. They consist of independent short stages. The results of such work serve as the foundation for subsequent developments and allow the beginner to show himself, join the team with an interesting task and, without the risk of failing an important project. This includes gaining experience, getting to know colleagues, and the opportunity to show oneself from the best side when there are no strict restrictions from the legacy side.
An example of such a relay development was the theme of a rotational screen based on the stroboscopic effect with the ability to display an arbitrary custom dynamic image made on the phone screen. Prototypes can be found .
The work was carried out sequentially by several employees and will be continued by new employees for the duration of their onboarding (from two weeks to a month, depending on the abilities and level of competencies).
The steps were as follows:
a) think over the design (having studied the available samples, description of analogues, showing creative initiative);
b) develop a circuit diagram, arrange it on the board;
c) develop a protocol for transferring an image from a phone to a device;
d) provide smartphone control via Bluetooth LE.
The starting option was to use something very compact like a three-leaf spinner, which, when manually rotated, began to show inscriptions. The BLE module was located in one lobe, ten RGB LEDs in the second, an optical sensor in the third, and a battery in the center. A circuit diagram was drawn up and the first experiments were carried out. It became clear that the level of picture quality is very low, the resolution is small, the game effect is short, and the possibilities are modest. Yes, and spinners are gone as quickly as they appeared. It was decided to raise the bar and develop a rotatable stroboscopic screen. At least it can be used for practical purposes at exhibitions and conferences, and interest in such solutions will not disappear in the near future.
In terms of design, there were two main questions: how to arrange the LEDs (in a vertical plane, as in the example above or horizontally) and how to power a rotating board with LEDs.
For educational purposes, the LEDs were located only in a horizontal plane. As for the power supply of the board, there was an important choice: either we take a bulky, noisy, but cheap commutator motor, or use a more elegant solution with contactless power transfer using two coils - one on the motor, the other on the board. The solution, of course, is elegant, but more expensive and time consuming. the coils had to be calculated first, and then wound (preferably not on the knee).
This is what the prototype looks like
The specificity of mass production is such that every extra cent in the cost is important. Success can be determined by the cost of a handful of passives. Therefore, it is often necessary to choose a less efficient but cheaper option in order for the manufacturer to remain commercially competitive. Therefore, imagining that the rotary screen would be put into mass production, the developer chose a collector motor.
The resulting prototype, when launched, provocatively sparkled, rustled and shook the table. The design that provided stability turned out to be so heavy and overall that it made no sense to bring it to a serial prototype. Rejoicing at the interim success, the decision was made to replace the motor with an air-gap rotating transformer. Another reason was the inability to power the engine from the computer's USB port.
The LED board is based on our RM10 module and six LED drivers. .
Drivers have 16 channels with the ability to independently control each. Thus, 6 such drivers and 32 RGB LEDs in total have the ability to show 16 million colors.
To synchronize and stabilize the displayed image, two magnetoresistive Hall sensors were used. .
The plan was simple - the sensor gives an interrupt for each turn of the board, the position of the LEDs is determined by the clock between two passes and their azimuth and glow in a 360-degree sweep are calculated.
But something went wrong - regardless of the speed of rotation of the board, the sensor randomly issued either one or two interrupts per pass. Thus, the image turned out to be smeared and folded inward.
Replacing the sensors did not change the situation, so the Hall sensor was replaced with a photoresistor.
If you have any thoughts on why a magnetoresistive sensor could behave this way, please share in the comments.
Top side of the board
With an optical sensor, the image is clear, but stabilizes for about 30 seconds. This happens for a number of reasons, one of which is the discreteness of the timer. This is 4 million ticks per second, divided by 360 degrees with a remainder, which introduces distortion into the output image.
In Chinese stroboscopic watches, the image is set in a couple of seconds at the cost of the fact that a small segment of the circle is simply not displayed: there is an empty space on the circular image, it is imperceptible on the text, but the picture is incomplete.
However, the problems didn't end there. microcontroller cannot provide the necessary data transfer rate for the possible number of shades (approx. 16 MHz) - the screen produces 1 frame per second, this is not enough for the human eye. Obviously, you need to place a separate microcontroller on the board to control the image, but for now it has been decided to replace the MBI5030 with . This is only 7 colors, including white, but this is enough to work out the software part.
Well, and importantly, for the sake of which this educational task was started - to program a microcontroller and control it through an application on a smartphone.
The scan is currently transmitted via Bluetooth directly via nRF Connect, and the application interface is in development.
Thus, the intermediate results of the work of the relay team are as follows:
The rotating screen has a line of 32 LEDs and an image diameter of 150 mm. It displays 7 colors, installs an image or text in 30 seconds (which is not ideal, but acceptable for a start). Through a Bluetooth connection, you can send a command to change the image.
And this is how it looks
And for new young developers to successfully learn, it remains to solve the following tasks:
Overcome the lack of microcontroller RAM for full-color display of the color palette. Refine the application for the formation and transmission of a static or dynamic picture. Give the structure a finished look. We'll keep you posted.
PS Of course, after finishing work on Bluetooth LE () we will design and implement a Wi-Fi / Bluetooth version on ESP32 But this will be a new story.
Source: habr.com