Differential Conductance PCB
Overview
The purpose of the Differential Conductance PCB (DCP) project was to maximize laboratory equipment utilization
I undertook the DCP project as an undergraduate research fellow in Dr. Kenneth Burch's LASE lab. The project lasted from the summer to the autumn of 2016 and was the first significant project I took on in the lab.
​
The overall intent of the project was (i) to increase the utilization of lab equipment by adding automated switching to the magnetotransport setup so measurements could happen overnight, and (ii) to allow remote switching to the setup so that researchers can run experiments from outside the lab (e.g. while traveling)
​
More information on the project can be found in each of the below sections.
Motivation
When conducting magnetotransport studies, many system measurement parameters need to be altered manually. The DCP project aimed to solve this issue to increase utilization.
LASE magnetotransport measurements probe the differential conductance of a sample as a function of applied voltage, temperature, magnetic field strength and angle, and more. In order to take these measurements, a sample is wired into a measurement circuit and placed into a magnet.
FTS on BS
Sample
Because the sample must be integrated into the measurement circuit to take readings, its composition/characterization will affect the circuit (via e.g. its resistivity). In order to compensate for the wide range of samples that are measured, the signals from all of the electrical equipment are fed into the so called "blue box" for processing.
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
On this blue box are knobs that change internal resistances in order to calibrate the circuit for measurement. If these knobs are not calibrated correctly, the data is too noisy to glean any useful information from.
​
​
​
The knobs must be calibrated whenever a measurement condition is changed, such as temperature or probing location on the sample. As a result of this necessity, the magnet is not utilized around the clock because:
​
1. A researcher must be in the lab to determine the proper knob settings at the beginning of each run
​
2. Even if the knob settings are known for each run, a researcher would need to be in the lab to physically change them to the proper settings for each run
​
Therefore, the magnetotransport system can only be utilized when a researcher is in the lab, meaning that experimental throughput could at least double by overcoming these issues
Poor tuning masks Andreev Reflection
Proper tuning reveals Andreev Reflection
Inside the blue box, a knob can be seen connected to many green wires
Measurement equipment with the blue box (top right)
The DCP project was created in order to address these shortcomings. The goal was to modify the blue box so that its "knobs" could be controlled (1) remotely, and (2) automatically. The addition of the DCP alleviates the above challenges by:
​
1. Allowing researchers remote access so that knob settings can be tuned off-site
​
2. Incorporating automatic switching into the measurement program. Therefore many runs with different sample conditions may be queued, and the software can "turn the knobs" to the optimal position for each run, allowing the magnetotransport system to be utilized around the clock.
​
Note that the "knobs" of the blue box are replaced with electrical switches in this design, but will continue to be referred to as "knobs" for simplicity.
Design
The design uses solid-state relays (SSRs) and shift registers, controlled via an Arduino and interfaced through LabVIEW, to control a circuit
The layout of the PCB I designed for this project can be seen below. The circuit works as follows:
​
-
Knob positions between 1 and 8 are input into the accompanying LabVIEW software. One position is selected for each of three knobs.
-
The software processes the inputs and passes the information to an Arduino.
-
The Arduino outputs a sequence of voltage states to its pins
-
This sequence of voltages is fed into a triple-chain of shift registers, one shift register for each knob
-
One of 8 outputs on each shift register now has HIGH voltage state, while the rest have LOW
-
The corresponding SSR is turned on, creating a current path through several resistors in series
-
The selected resistances parametrize a voltage divider.​
​ Since the samples measured in the magnet have such a wide range of resistances, there are a wide range of possible resistances in the divider. This allows for :
​
1. Several division levels
​
2. Many absolute resistance levels for the same division level so that the effect that the sample has on the circuit can be minimized.
Arduino Connections
Shift Registers
Knob 1 SSRs
Knob 2 SSRs
Knob 3 SSRs
Ribbon Cable Connections
Ribbon Cable Connections
Ribbon Cable Connections
Code
The code to control the circuit was written in LabVIEW utilizing a package to interface with an Arduino, which electrically controls the switching in the circuit
LabVIEW code which controls the switching can be seen below:
​
-
Knob positions, either entered by the user or retrieved from a run queue, are input and validated
-
An indexing operation is performed to build the array that will be fed into the Arduino block of code
-
The array is processed, generating an output sequence of voltage states on the Arduino. This sequence sets the shift register pins to the proper voltage states to achieve the input knob positions
Indexing
Write to Arduino
Argument validation
PCB Fabrication
I utilized the department PCB printer in order to print my design. Along the way I optimized the printer, learned the printer software, and wrote a guide for other researchers
I'm a paragraph. Click here to add your own text and edit me. It's easy.
PCB immediately after printing
Arduino wired to shift registers
Hover over an image for more information
PCB Printer Guide
PCB immediately after printing
Shift registers and resistors after layout
Results and Future
After testing and implementing the circuit, unexpected non-linear behavior was observed. The project was therefore put on ice, but future directions are still considered
The circuit was successfully incorporated into the blue box, as seen below. After testing the circuit with some standard "samples" (i.e. resistors) and observing the expected behavior, the old circuit was swapped out with the new PCB. The software worked as expected and integrated well into the general measurement software.
After some time being used to obtain data with samples of interest, unexpected non-linear behavior was observed. After some testing, it was determined that the PCB was the cause, and it was removed from the system. The hypothesis as to why this behavior started is that the voltage limits on the SSRs was exceeded, causing a breakdown of the semiconductors within them.
Re-implementation of a similar circuit would be quite straightforward. The SSRs would simply have to be replaced with better relays. These relays could be either heavier duty SSRs or, ideally, contactors. Space restrictions would make the latter infeasible, so perhaps e.g. a reed relay would be a possibility for a sufficiently small relay that has a physical break in the circuit when in the "off" state, as opposed to SSRs whose semiconductors are always part of the circuit.
​
There are several other possible ways to improve the DCP. One possibility would need no new hardware. This could be that the measurement system determines the optimal knob settings for each run automatically. Currently, these still have to be determined by a researcher, and then queued to allow for 24/7 use of the magnet. The software could be programmed to iterate through all knob combinations, run small test sweeps, and then process the data to determine which has the lowest noise.
DCP Inside the blue box
Note: The burn on the underside of the PCB is due to the substrate flexing when the solder paste was heated and does not affect electrical performance
PCB in the"blue box"