dosed agent – powder dye or any other powdered substance
the dispenser must be able to dose in a wide range from 3 to 100 grams with an error of no more than 0.2 grams per dose
Dosing speed must be at least 1 gram per second
Indication of the operation process and device settings should be carried out using the built-in touch screen of sufficient size for comfortable operation
The dosing process must be started using a pedal in manual mode, or by a signal at a discrete input in automatic mode.
The design of the dispenser should allow partial disassembly and cleaning (washing) of the hopper and dosing module when changing the color of the dye.
The dispenser must have a self-diagnosis system and report malfunctions of the main components to the user using messages on the built-in display
The design of the dispenser should allow placing containers of various configurations and heights in the dispensing area. Setting up for a new type of container should be quick and easy
The hopper for storing the dosed agent must have a volume of at least 2 liters.
The dispenser control system must keep a count of the remaining doses in the hopper or display the absolute balance in grams
The dispenser control system must be able to calibrate strain gauges
Access to service controls must be restricted with a password
The dispenser control unit must be able to update the firmware from a usb drive
And the most, probably, the main criterion is the cheapness of the design. Those. the design should be as simple as possible, because we naturally could not compete in price with our “brothers from China”.
This, of course, is not a full-fledged technical task, but it was already clear that there was something to build on.
We started development by studying ready-made designs on the market and searching for articles and books on the topic. Many solutions were based on banal sluice valves and those that opened stupidly in time. Of course, there were also screw solutions, but they all rested on dosing due to the rotation of the screw mechanism in a predetermined time interval, but this did not suit us, since it was impossible to achieve the necessary accuracy without feedback. We also found such solutions extremely inconvenient, because it is necessary to calibrate and adjust the opening time every time the dosed agent is changed. As a result of the analysis, we came to the conclusion that a screw feeder with feedback in the form of electronic scales would be the most optimal solution to obtain the required characteristics. Fortunately, there is enough literature for calculating screw mechanisms and the theory as a whole is not complicated.
After our research work, the design concept began to take shape.
Eugene started making the first model in 3D. The calculation of the screw was made:
The bunker was installed on a pair of strain gauges. It was necessary to install an agitator inside the hopper for constant mixing of the dosed substance, since the dye was very light and sticky, and did not always want to crumble under its own weight, which could stop the feeder from working. To install the container, a weighing platform was developed, and a funnel with an additional vibrator was installed to form the flow to prevent the substance from hanging in it. We had practically no restrictions on the length and width for the installed container, but to adjust the height of the container so that nothing would wake up by, a simple mechanism was developed for lifting the entire dosing unit along with the hopper in a fairly wide range. This solution allowed to install containers of almost any configuration. It is based on a lead screw with multi-start trapezoidal thread and a standard lead nut with a backlash removal mechanism. This ensured the smooth running of the entire structure and did not require additional height locks after adjustment. Standard calibrated shafts with a diameter of 8 mm and mating linear bearings were used as guides. It turned out to be very reliable and at the same time simple.
Evgeny used an aluminum structural profile as the basis of the entire structure, it is very convenient to work with this material, it is rigid and easily connected using standard fasteners. In general, the design involved the maximum of standard components that could be purchased, but still many parts had to be designed and manufactured individually. One of the most interesting components was the screw feeder, which feeds portions of the material from the hopper into the funnel. All the details of this assembly were designed as simply as we could and allowed everything to be disassembled and assembled back without much difficulty. Below are photos of the main nodes.
The auger was driven by a stepper motor through a V-belt transmission with a toothed belt. The pulley diameter ratio was chosen to form a reduction gear. This solution made it possible to increase the moment on the screw shaft, as well as to increase the smoothness of the ride. They decided to print the first, experimental versions of the bunker on a 3D printer, because. this allowed us to understand as cheaply as possible what shape of the bunker we would need, and in the future to make it from stainless steel. Printing was carried out from PLA plastic and in the first iterations it was a regular cylinder, and the formation of a funnel was done on the bottom cover of the hopper, at that moment we thought that this would be enough, but how wrong we were …
On the top cover of the bunker, a turner drive motor with a gearbox was installed.
When the design of all the main units was completed, Evgeny proceeded to manufacture them, and also purchased and ordered all the ready-made units that we used in the design.
Evgeny made some of the parts himself on his CNC milling machine, but we ordered some screw feeder components and the screw itself from friends at the factory. The rattle of the machine cutting aluminum haunted Zhenya for a long time after the completion of all milling work. Most of the parts were designed in such a way that they were made from one machine on the CNC, but for the back cover of the feeder, to which the stepper drive motor was still attached, I had to additionally make a reversal template, since it was necessary to mill on both sides and not spoil anything. Despite the automation of the process of milling parts, in some of them, we made holes and threads by hand, because it was cheaper, although it was not always so easy.
And just a month later, a decent amount of parts ready for pre-assembly formed on the table. I still remember my surprise at this photo, I couldn’t believe that not so long ago it was only on a computer screen. The build process has begun. First, Eugene assembled the dosing unit:
Then I proceeded to the main frame and the lifting mechanism:
Then we “married” the feeder with the frame:
We added a guide funnel and got half of the finished dispenser:
The printing of the bunker took more than 8 hours, and in the end the bearings in the printer began to die, they thought that they would not finish printing, but everything turned out, albeit with a terrible rattle.
The first version of the agitator blades, funnel fastening elements, as well as the agitator motor-reducer fastening case were also sent for printing.
One of the complex structural elements was the weight platform. Because weighing should take place in dynamic, not static mode, we needed to get the most “clean” signal from the sensor, which turned out to be quite difficult on a platform with 2 rotating motors, and even a vibration motor on the funnel. To try to combat vibration, we installed the developed platform on rubber vibration pads and through them we fastened the platform to the dispenser frame. The platform itself was milled out of aluminum and hid the sensor itself inside. From above, it was planned to install an additional guide for convenient and accurate positioning of the container. This lining was supposed to be removable for quick installation of the necessary lining for different types of containers. But in the first iteration, we did not make this overlay. You can see the finished platform below:
After the final assembly of all mechanical parts into one, the first version of our dispenser was obtained, ready for real tests with control electronics and real dosed substance, as well as several iterations to improve the design and correct errors, but more on that in the following parts.
Thanks to everyone who read to the end, I hope it was interesting. To be continued…