MEASUREMENT DEVELOPMENT
adjust and refine the IBD continually. A complete rethink of design techniques
Designing an instrument capable of measuring very low currents and having an output stage capable of producing 3000V is extremely challenging. Previous SMUs in this particular product family were only capable of sourcing voltages up to 200V. For this design, the level of energy in capacitances in the circuit was multiplied by more than 100 (E=0.5*C*V^2) for the roughly 10X increase in voltage. In addition, components that connected 3000V to the low current section, including the PCB, demanded impedances greater than 3000V/100fA or 3E18 ohm. All this required rethinking layout techniques, guarding and protection techniques, and the selection of isolation components that spanned the output voltage to a low current sensitive node.
In this particular design, the current measurement section incorporated many MOSFET components for bypass and switching of ranges. In addition, we used several compensation capacitors to connect the output voltages to the current measure section. Although these parts worked well in similar low voltage products, the higher voltage of this design meant that the capacitors stored much more energy (E=CV^2) than in the low voltage designs. As a result, when a DUT would short suddenly, this energy would be dumped into the current measure section, damaging these sensitive MOSFET devices, leading to either a complete failure of the MOSFET or increased leakage. That meant we needed to develop some innovative ways of reducing the voltage or rerouting this dumped energy around the circuits to protect these sensitive devices.
Figure 1. Thermal video imaging was used to identify hot or damaged components on PCBs during the evolution of the Model 2657A’s design
These bypass circuits needed either to have high impedance inherently or to be guarded to achieve this high level of impedance. The problem components included the compensation capacitors that connected the high voltage to the current sense circuit. These components could not be guarded, yet needed to achieve high impedance. After a long search, we finally identified a compensation capacitor that could
handle the high voltage while ensuring low leakage, as well as meet other critical parameters, such as for part size and dielectric absorption.
Finding suitable high voltage capacitors was far from the only difficulty we encountered in component sourcing. Connecting the source and sense leads to the DUT required low noise triax cables capable of withstanding 3000V from center to guard and guard to shield. They also had to have low leakage from center to guard to allow sub-picoamp measurements. Before we started this project, such a cable simply didn’t exist, so we began working with cable manufacturers to develop one to our specifications. Once we had the cable defined and verified, we moved on to develop triax connectors that could meet our high voltage and low leakage specs and our safety requirements. This connector design had to be adaptable to a variety of applications, including board mount, panel mount, and panel pass through designs. We contacted half a dozen connector manufacturers, but only one had the capabilities of designing a custom high voltage triax connector.
3D printing to the rescue Ultimately, the most complicated component design task was creating an output relay that could switch the high voltage, withstand twice the maximum voltage, and have low leakage. We wanted to be able to use this relay as an output ON/OFF relay for the product as well as use it in a switch system. This complicated the relay definition because each relay application had a different set of safety spacing requirements. These differing requirements had to be combined into a single relay definition based on the worst-case safety spacing. We started working with engineers at two different vendors to develop possible solutions, approve prototypes, and then test their capabilities. After holding multiple meetings and testing several prototypes, we determined that a split guard reed relay offered the best solution.
Our prototype tests let us determine the best material to provide the high isolation and low current leakage that we required. By this time, we were down to a single relay vendor with the capabilities to provide us with a solution. To further complicate the design, some of the required creepage distances could not be met by using off- the-shelf bobbins. This resulted in some creative bobbin designs by the relay vendors to meet our requirements and still have a manufacturable relay. The vendor used 3D printing, which allowed developing bobbin prototypes quickly for faster reviews and tests. In the end, we have a robust relay that meets our stringent requirements. The design and debugging processes took significantly longer than in earlier SMU development projects because of the complexities inherent in high voltage design prototyping and
32
www.siliconsemiconductor.net Issue I 2013
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42