Measuring weight without a load cell - Setaram TGA balance mechanism
In this video, I show you a very interesting balance mechanism which was saved from a Setaram TGA that was destined to end up in the trash bin. I thought that this part was valuable enough to save it. I will guide you through the mechanism and its principles. There is also a little bit of circuit design involved because I needed to build a circuit that makes the whole thing work.
What is a TGA? - Information and working principles
Image source: http://dx.doi.org/10.1007/s11663-011-9611-5
TGA or thermogravimetric analysis/analyser is a technique/instrument used to analyze changes in mass as a function of time and temperature, typically in a controlled atmosphere.
It is a very handy tool in materials science, especially where thermal stability is of interest. For example, the steel industry uses it to study the corrosion properties of different steels in different environments and at different temperatures. When the steel suffers from corrosion, weight change occurs. This weight change can be monitored with this equipment.
A simple drawing of the equipment can be seen on the figure next to this text. At the top of the machine, there is a delicate balance mechanism, typically with micrograms sensitivity. I will focus on this part in this article, but let’s quickly go through the whole instrument. One side of the balance arm has a long chain or wire attached to it that allows us to have the sample in a furnace below the balance, while the other side of the balance arm has some counterweights to balance the weight of the chain, the sample holder and the sample.
The sample holder is typically a small basked made out of something inert that tolerates high temperatures and corrosion. Platinum or alumina are often used. Under the sample holder, there’s a thermocouple that monitors the temperature inside the furnace, close to the specimen.
The system is also equipped with a gas or atmosphere-management system that consists of gas inlets and outlets and a vacuum pump. Depending on the test, one can introduce an inert gas, reducing or oxidizing atmosphere, and so on.
As I mentioned above, corrosion studies are common with this type of device, but other processes such as evaporation, desorption, phase transition and others can be studied.
With precise temperature and atmospheric control, and with the accurate logging of the test parameters, one can deeply study material properties with this equipment.
Working principles of the balance mechanism
This is an even more fun part because the device is simply genius.
First, let’s talk about the mechanism as a whole first. The balance mechanism is extremely delicate. The whole balance is sitting on a torsion wire. This allows very fine movements. To allow transportation the mechanism can be locked and fixed. This allows the torsion wire to sit in a “resting position” where it is not exposed to any tension. The arm is also firmly locked in place, so it won’t move at all.
From the centre of the top of the balance arm, there is a thin metal rod that goes upwards. Roughly 5 centimetres from the arm, there is an interesting combination of four solenoids. They are connected in series and what is maybe more interesting is that they are “coupled” to the balance arm by having their “core” connected to the metal rod that comes from the balance arm. So, if we run some current through the coils, they will deflect the four cores in some way which will also move the balance arm. The coil is so sensitive that a few milliamperes are enough to restore the balance of the arm. The question is, how? To answer this question, let’s see what is above these coils.
At the end of the rod, there’s a thin plate with a slit in its centre.
Then there is a super interesting setup here. On one side of the slit, there is a regular red LED. On the other side, there are two photodiodes, deeply embedded in the metal structure. This setup implies that when the balance is balanced, the slit must be in the centre and this would minimize the light coming from the red LED to both photodiodes. They would equally not get exposed to light. Since the whole setup is in a metal enclosure, it would be totally dark for those two photodiodes.
To figure out when the balance is in balance, we should look for the situation when there is some minimum situation with the exposure of those photodiodes buried in those metal channels. The way a photodiode works is that if we apply some bias voltage to them and then expose them to light, they generate a so-called photocurrent. This photocurrent is proportional to the intensity of the light.
When we have two of these photodiodes connected in series, their photocurrent will compete with each other and they can even cancel each other out. Well, this is the exact situation that we are aiming for. When the two photodiodes receive the exact same amount of light, their photocurrent cancels out. The whole mechanism is built in a way that when it is in balance, the light is blocked out by that plate with a slit on it in a way that the photodiodes receive the same amount of light, indicating that the arm is balanced.
This part requires some thinking, so let me show you below how I made it work.
Balance arm. When engaged, the whole balance arm is hanging on a thin torsion wire.
Upper portion of the balance mechanism with the coils and the opto-elements.
Making things work
First, let’s start with the optical part, the LED and the photodiodes.
So, the LED is nothing special. It is a red LED with a 1.65 V forward voltage. I am planning to run it at 5 V, so then the voltage drop across the resistor that I need to add to the circuit is 3.35 V. Using Ohm’s law, and assuming 20 mA current for the LED, this would give us 167.5 Ohm. So, I will use something around 200 Ohm just to be on the safe side.
Then, the photodiodes are a bit more tricky. They are connected to a small circuit board. I had to understand it first. So, there are 3 pins. On the left, the pin is marked with a plus sign. The trace on the PCB goes towards the diode (cathode), and then the other leg of the diode (anode) goes to another trace which then connects it to another diode (cathode). At this junction, there is another pin coming out from the PCB. Finally, the third pin (anode) is connected to the other pin of the second photodiode.
Backside of the PCB that accommodates the two photodiodes
Initially, I had no idea what this was. Then I searched a bit and it turned out that this is a balanced photodetector circuit, or part of it. So, the whole circuit is basically just a resistor and two photodiodes. When the circuit is up and running, the two photodiodes generate photocurrent proportional to the incident light intensity.
So, here one can already see that this setup with the light source, the slit and the two carefully placed photodiodes can be used in a smart way. When the incident light is identical, the photocurrents are also identical. In the ideal case, they can cancel each other out. So, when they cancel each other out, the measured output voltage (measured through a resistor in series with the output) would be 0 V. And when they are off-balance, then the voltage would be anything up towards the bias voltage depending on the distribution of the incident light between the two photodiodes and the rest of the circuit. Also, we usually don’t measure current, especially not with the ADC of a microcontroller. So, we can add a resistor in series with the output pin (centre pin of the photodiode PCB) and measure the voltage. This voltage will be then still zero when there’s a balance, and it will be something non-zero and something noticeably less than the bias voltage when there’s no balance.
Then we can use this voltage to indicate to the control circuitry if there’s a need to apply more, less or no current to the solenoids.
The two photodiodes closely packed on the PCB
Metal plate with a split. This plate moves between the light source and the couplet photodiodes.
So, how is this all related to those four coils?
The essential parts of the balance: balancing coils and “balance-detecting” opto-elements
The above-described balanced photodetector or null detector is the core part of the balancing mechanism. The four coils together in the balance is an electromagnetic servomotor. It generates a force (torque) to counter the weight of the mass being measured. To create this countering force, the coils must be driven with some current. This current can be measured, or more precisely, the current can be driven through a resistor, and we can measure the voltage across this resistor. This voltage (current) is proportional to the weight placed on the balance. Thus, after calibration, the voltage (current) signal can be used to determine the weight placed on the balance.
In this specific equipment the principles of the device are the following:
The balance is initially fully balanced by balancing the weight of the sample side. This is typically done manually by placing weights in a small basket that is hanging on the other side of the balance arm.
At this point, the output of the photodetector should read zero because both photodiodes produce the same photocurrent. Thus they cancel each other out.
When the experiment is running, due to the circumstances that the specimen is exposed to, the specimen suffers from weight change. This throws the balance off which must be compensated.
The photodetector detects the off-balance condition because its output is not zero anymore. Then an appropriate circuit drives more (or less, depending on if it is a weight loss or weight gain) current through the coils until the balance is restored.
This current difference is proportional to the weight change of the specimen. So, with careful calibration, the weight change can be expressed in actual SI units, typically in micrograms.
