Final Design

UCSD Mouse Time–Controlled Feeding Cage

Figure 1 Feeder System

Figure 2 Feeder Cage System

The linear slider was designed to hang from the top of the feeder so that falling food particles would not cause jamming. The load cell was attached to the slider and the hopper hung from the load cell so that the weight of food could be measured. The motor was placed behind the hopper and drove the system backwards and forwards in a linear motion parallel to the gear track.

The system was programmed using LabVIEW that communicated with a National Instruments data acquisition instrument (DAQ). The LabVIEW was programmed such that it would drive the motor at user defined intervals to either inhibit or allow the mice to have access to food. In addition, both before and after the hopper was moved, the DAQ read the voltage of the load cell. This voltage was converted into an approximate weight and could then be used to see how much food the mice had eaten.

Mouse Cage

The UCSD mouse cage was slightly modified. A hole was cut in the back of the mouse cage and a wire grating was installed that was wide enough for a mouse to stick its mouth through. The grating position was made adjustable with slots cut in the back of the mouse cage. L-brackets above the grating allowed the feeder to detach in order to allow the mouse cage or the food hopper to be easily cleaned. All the modifications on the cage are compatible with existing lid. This allows the cage to fit within the current UCSD cage setup.

Figure 3 Mouse Cage After Modification

Linear Slider

Figure 4 Linear Slider

The slider is used carry the combined weight of the hopper, motor, and load cell. The weight of these pieces together is 238 grams. In order to reduce the friction between the stroke and the rail, it was desired for the slider to have a small friction coefficient. In addition, the slider needed to have a rail length of at least 75mm so that the stroke could move the hopper a reasonable distance. The model of KSRLST16 was chosen because it has half of the friction as the SROMST25 model that was also being considered.

Load Cell

Figure 5 Load Cell

The load cell was used to measure the amount of food consumed by the mouse. The hopper was attached to one end of the load cell, and the other end of the load cell was fixed. The load cell outputs the difference in weights as a signal of difference in voltages, and the signal is converted to a weight by LabVIEW after amplifying and filtering. It was desired for the load cell to have a high accuracy so that small amounts of food could be measured. It was important that the dimensions of the load cell fit in the system. Furthermore the load cells could not be expensive since eight of them were needed within a two thousand dollars budget. Since the weight of a full food hopper is within 100g, the 0-100g load cell can be used that gives ± 0.0866g accuracy. Micro Load Cell (0-100g) - CZL639HD from Phidgets was chosen for this design.

Printed Circuit Board (PCB)

Figure 6 Low-pass filtering and amplifying circuit

The amplifier is needed to increase the magnitude of the output signal of the load cell for the data acquisition system (DAQ). The current output from the DAQ is 4 mA, and the voltage supply is 5 volts. Therefore, the required gain for the system is 1200. It is difficult to build such a big gain with a single amplifier. The solution was to combine a 200 gain instrumentation amplifier with a gain of 6.4 from the non- inverting amplifier. The non-inverting amplifier, shown in Figure 3.4, can be simply built with using a 3.7 KΩ (R1) and a 20 KΩ (R2) resistor. A gain of 200 was used with the instrumentation amplifier by using a resistor of 499 Ω (Rg). In addition, a low pass filter with a cutoff frequency of 16 Hz was incorporated to reduce the noise for the output of the amplification circuit.

Food Hopper

Figure 7 Food Hopper

The hopper holds the food for the mice. At feeding times, the motor drives the hopper to the cage. The hopper then stops when the two gratings come together, shown in Figure 2. In order for the mice to reach the food, the gratings were desired to be as thin and sturdy as possible. In addition, a lighter hopper would allow for a more accurate load cell to be used. Based on these criteria, there were two designs considered. After iterating over many design attempts, the design shown above was the most optimal.

Motor

Figure 8 High-torque motor

The motor is used to drive the food hopper. The motor must have enough stall torque to drive the hopper with a fully filled food pellet hopper. Also, the motor had to be small enough so that it could be installed behind the hopper. A motor with a large gear ratio, slow speed, and large torque was desired. The 20mm diameter gear motor (#1109) was chosen because of its dimensions (20 x 47 mm) and gear ratio (154:1).

Data Acquisition (DAQ)

Figure 8 NI USB-6001 DAQ

The DAQ is needed to communicate between the computer and the physical system. The system will consist of up to 8 cages controlled at once by a laptop running Labview. Therefore it is necessary for the DAQ to have sufficient number of digital and analog input/output pins. The DAQ should have at least 16 digital and 8 analog pins since each feeder needs 2 digital and 1 analog pin. In addition, it was desired for the DAQ to be compatible with Labview. The system also included a weight sensor in each cage. Therefore a DAQ with a higher resolution would result in more precise weight measurements. Ultimately, the NI USB-6001 was the most efficient and suitable choice for our design.

Summary

The project was ultimately a success. The automated feeding system worked effectively to limit the supply of food to mice and was significantly less expensive than the current commercial systems. The LabVIEW interface was simple enough so that even those without any familiarity of the software could program the cages. In addition, the feeder was removable so that the cage could go through cleaning in the autoclave oven. Further, the load cell was able to keep track of the amount of food consumed by the mice. Overall, every design criteria was met.