Can We Put Other Instruments on a Chip? Part 1

The article series on Z meters on a chip (Z Meter on a Chip? Impedance Meter Integration and Readout, et al.) illustrates the more general idea of integrating essentially all of the analog circuitry of a medium-performance measurement instrument in one IC. What of other instruments?

When one begins to study test and measurement instrument circuits in some detail, an overlap of functional blocks among instruments is found. For instance, the output amplifiers of pulse generators (PGs) are not much different from those of function generators (FGs) or their digital counterparts, waveform generators (WGs). Precision rectifiers in DMMs can also be found in sine-wave generators and Z meters.

Power supplies are found in all instruments, and while they can differ widely in supply requirements, many have the same voltages: ±5V and ±12V or ±15V are common. Some instruments require isolated supplies, such as those involving drive-sense or stimulate-measure functions or “units” (SMUs). These functions are found in curve tracers, which have developed into TPAs (transistor parameter analyzers), or in a more general and preferred sense (using the same abbreviation) are two-port analyzers. Each port is driven by a SMU and not uncommonly, the SMUs are powered by floating or isolated supplies so that current sensing is low-side (ground-referenced) and hence easier.

Range-switching circuits are common throughout instrument types. Output ranging is by attenuators or can be implicit, as in segmented amplifiers; successive stages of greater power capability turn on as output voltage increases. Segmented amplifier output stages have more parts, needed for the segmenting function, but with a sufficient number of segments, will dissipate far less power than a class-AB output stage. Integration benefits in adding transistors to reduce on-chip power dissipation, and this is indeed possible without going to switching amplifiers. In instruments, additional switching noise reduces effective range and is eagerly avoided, if possible.

Ranging is also in the form of PGAs for sensing high- or low-side currents and voltages. Range switches can often be analog CMOS switches that are already made as ICs. Maxim (the sponsor of this site) and TI sell analog switches capable of an ohm or less of resistance and can conduct 0.2A. This is sufficient for many instrument designs, including Z meter bridge switching and lower-power TPAs. Consequently, an IC such as the ADI AD5933 Z analyzer, which has a ×5 current ranging PGA, no sense-resistor switching, and also no ranging of voltagem could be enhanced to become the analog core of a medium-performance and commercially viable Z-analyzer instrument.

This IC is about eight years old, and in the interim, some of the monolithic ranging capability has emerged. ADI has had nichrome resistor trimming capability for decades, and this could be used to trim the range resistors for a precision current-sense circuit with four decades or more of range. This is sufficient to cover the wider Z range.

Much of the chip would be occupied by the large current and power transistor components for a maximum output current of 100mA to 200mA. And by also increasing the voltage range downward, so that the component under test (Zx ) can have a minimum full-scale (fs) voltage of 100mV, then the impedance range has been extended downward to 100mV/100mA = 1Ω instead of 1kΩ. At 1Ω fs, the Z analyzer is competitive with various lower-cost Z meters. At the high end of the Z range — at 10MΩ — the AD5933 is adequate for ordinary bench use.

Another example is the Maxim MAX038, a 20MHz FG. This shining star of historic instrument integration was preceded by the first-generation 1MHz FGs from Intersil (8038) and Exar (2206). (Interdesign also had a “waveform generator” FG in Monochip Application Note APN-8 by James H. Knapton.) The second-generation MAX038 (for which I would have preferred the designation MAX2206) was clearly superior but was never (to my knowledge) made the core of a commercial FG. Like the XR2206, it has some provision for current ranging (2 pins). It also included a phase detector for phase-locking the triangle-wave generator (TWG) to an external frequency reference.

In part 2, we will continue looking at existing ways to put an analog FG on a chip. We will also continue the examination of the WG to see if it's better or worse than the FG.

Related posts:

0 comments on “Can We Put Other Instruments on a Chip? Part 1

Leave a Reply