[Part 1] From Concept to Reality : Project Initiation

"A Coolant Flow Rate and Temperature Modulated Automatic Diffusion Pump Safety Shut-off Relay" or as I call it, "Pete-O-Meter"

It dates back to two weeks ago when I was talking to my PI. Since the diffusion pump has been a trouble maker for a while, the conversation naturally led to the safety features of our diffusion pump. As I have changed the chiller motor few weeks ago, it is uncertain whether the system is reliable under continuous use even though I have test run it for four days straight. 

Then I asked him, just out of curiosity, what will happen in case the chiller fails and the diffusion pump continuous to heat up? Now, we are not sure if the diffusion pump is thermally protected. That is, if any kind of safety relay exists. Due to this uncertainty, a great concern is brewing over a continuous use of the diffusion pump. The consequence of failure of pump is deadly explosion.

Be that as it may, should I put myself in a constant danger of the pump failing and a potential say-goodbye to all my friends or do something about it to fix the problem. The small project, therefore, has begun as a part of my motivation to stop this madness and fix things once and for all.

The project is simple, yet complicated in its core as it requires knowledge in, computer science, mechanical engineering, and electrical engineering. I started to draw a blue print and an electric schematic for what may be the most expensive flow meter ever.

Now the picture below shows up the electric schematic of a Coolant Flow Rate and Temperature Modulated Automatic Diffusion Pump Safety Shut-off Relay, which I will simply call from now on as, "Peteometer" (Pete-O-Meter). 

The brain of the Peteometer is a Parallax BS2pe Microcontroller chip which provides a very simple interface yet very powerful in its customizability and reliability. It is interfaced with the physical flow meter (the one from my previous blog posts about the Opti-Temp Chiller) using two photo transistors coupled with one 650nm diode laser and a infrared LED. In addition, a separate thermocouple module will be connected to the microcontroller which then will be used to directly measure the temperature of the water.

As you can see Pin 15 through Pin 8 are reserved for the 2 consecutive 7-segment display which will display both the temperature reading and the flow rate. Now, some may wonder why I am only using 8 pins for two 7-segment displays, one of which requires at least 9 separate pins, it is one of electricians trick that if I pair up corresponding cathodes from each segment display and hook them up using regular pin assignment to microcontroller pins with two anodes assigned to different pins, say pin A and pin B, then I can selectively choose which segment display to be excited by letting the microcontroller ground only one of the anode (A or B). Therefore reducing the number of required pins from 18 to 10. 

That's that and moving on to the relay circuit. The relay is mechanical enclosed in a plastic cover about a size of kidney beans. Since the diffusion pump master relay uses 125VAC 0.150A power, I had to make sure that the relay is capable of switching on/off this amount of power ratings. The relay chip is bought from a Radioshack and is rated at 125VAC Max, 500mA.

The coil in the relay is not directly excited by the microcontroller pin as I realized that the power requirement of the coil barely meets what the microcontroller can provide. I got around this problem by using a transistor linking an independent high voltage source to the coil where a base of the transistor is set to be triggered by the microcontroller. Therefore, I can still tell the microcontroller when to turn on the relay but I the energy required to turn it on is drawn from a separate energy source. An additional capacitor that is connected in parallel with the coil is to stabilize the voltage across the coil since I noticed a fluctuating voltage from a prototype development stage.

Now, moving on to the phototransistor. The principle of getting light measurements from any phototransistors in general utilizes a typical RC circuit design where the resistor is replaced by a phototransistor. As incoming light excites a phototransistor, the so-created reverse current from the phototransistor prevents how fast a capacitor drains a charge once an external voltage potential is removed. Therefore, if we measure the RC time of the circuit, the result reflects the amount of photons that hit the transistor at the time of measurement. (This is a very crude explanation, but works.) The sensitivity of light measurement depends on the value of capacitance coupled with a phototransistor. Notice that I'm using two capacitors with different capacitance values. The phototransistor monitoring the diode laser beam needs not be very sensitive since the power output from the laser is already pretty decent for this purpose (0.5mW).

The thermocouple module is pretty much all fully built as I'm simply going to buy a ready-to-go module from Parallax, Inc. So nothing really go through there. Although one thing about the types of thermocouple wires, the thing is the microcontroller can be programmed for a specific type of thermocouple. Therefore, it really comes down to which thermocouple table is the easiest to put into the program. I am yet to know, but we will see as I progress.

No comments: