System-level DMMs continue their incremental improvement. Some vendors have provided faster reading rates with no loss of resolution or accuracy. On other models, more features have been made available. And, to enable the highest possible throughput within a measurement system, yet other manufacturers have developed fast scripting capabilities.
The comparison chart accompanying this article presents details of several representative products and confirms that both PXI and LXI DMMs have become more prevalent. Agilent Technologies’ recent entrance into the PXI modular instrument space was marked by the introduction of several products, among them the M9183A DMM. More recently, Agilent purchased Signametrics, resulting in that company’s absence from this year’s chart.
Comparison Chart
Click here to view the PXI/VXI/LXI DMM comparison chart.
LXI also is an attractive standard for modular instruments but in contrast to PXI doesn’t restrict the mechanical form. This factor is especially important for manufacturers of large switching systems that can take advantage of the interconnection flexibility afforded by a larger size chassis. DMMs often make use of the space by including more features.
According to Tom Sarfi, vice president of business development at VTI Instruments, “We have emphasized [LXI’s] minimum vertical profile so that our DMMs take up less rack space. Our lowest profile DMM is only 1U high.”
Figure 1. EX1266 Mainframe With DMM
Courtesy of VTI Instruments
LXI DMMs may have an integral switching capability. These mainframe DMMs address a class of applications that needs scanned measurements but either no signal conditioning or conditioning that can be provided in a separate chassis. Whether PXI or LXI is more appropriate often revolves around channel count and the amount of external cabling needed in a PXI solution. Figure 1 shows a 1U mainframe with a DMM and plug-in switch modules.
Dual-Path ADC
Basic physics hasn’t changed, so many of the year-on-year improvements in DMMs have resulted from very clever design. An example is the dual-path ADC architecture patented by Agilent in 2004. Fundamentally, high speed and accuracy are incompatible. Certainly, high-speed ADCs are available but with less accuracy than lower speed ADCs. Patent 6,778,125 Dual Path Analog-to-Digital Conversion Method and System addresses this limitation.
A fast ADC and a slow ADC are used on parallel paths. Both ADCs digitize the input signal, but aliases created by the slow converter are cancelled by part of the fast ADC output. Actually, there are several sections to the overall scheme.
The high-speed ADC signal is protected from aliasing by a filter that restricts the input signal bandwidth. At the high-speed output, the band corresponding to the frequencies that could be aliased by the low-speed ADC is folded into the correct spectral position to cancel the aliases. Before the fast and corrected slow digital streams are merged, the high-speed one may be decimated and the low-speed one interpolated to achieve the desired overall sampling rate.
Although the circuit is complex, the end result has the original low-speed accuracy but with a higher bandwidth. High-frequency components are converted with the accuracy of the faster ADC.
A large number of more recent references to dual-path ADCs can be found, but the meaning of the terminology appears to have broadened since 2004. For example, most enhanced accuracy, high-speed sigma-delta ADCs develop two intermediate outputs: one with the signal plus errors and the other with only the errors. These ADCs also are described as dual path. In contrast, both paths in patent 6,778,125 contribute signal-related information to the output.
Dave Tenney, product manager at Agilent’s systems products division, didn’t comment on details of the ADC technology used in a current product but did say “[improved performance] has been accomplished primarily through innovative instrument architecture. In the case of the new PXI DMM, there is a high-speed version that incorporates a dual ADC signal path.”
Local Algorithm Execution
Speed also figures prominently in the latest products from Keithley Instruments. Chuck Cimino, marketing director for DMM, sensitive, and SourceMeasure instruments, said, “To take full advantage of the faster programming speeds our Test Script Processor (TSP) affords, we’ve made a number of analog subsystem enhancements, including faster A-D converters, faster settling DC and AC front-end amplifier designs, and other system upgrades.
Figure 2. Model 3706 System Switch/Multimeter
Courtesy of Keithley Instruments
“The latest additions to the line of switching cards designed for the Model 3706 System Switch/Multimeter (Figure 2) are optimized for high-speed relay control as well as higher channel densities and simplified mass termination and screw termination options,” he continued. “Our thermocouple temperature accuracy and settling performance have been enhanced by relocating the cold junction sensor outboard of the instrument with the terminal connect module. This also eliminates various sources of thermoelectric errors and provides more stable and consistent readings across the higher channel counts the Model 3706 makes possible.”
TSP executes algorithms totally within an instrument or cluster of instruments in contrast to a PC-centric test system. Eliminating most of the overhead associated with PC operating systems, languages, drivers, and interfaces typically results in a 10x to 100x speed improvement. TSP-based instruments can be operated in the immediate mode just like traditional SCPI instruments with only minor command syntax differences and all use standard IVI drivers.
Speed-Resolution Trade-Off
The critical technology at the heart of National Instruments’ (NI) DMMs is the FlexADC. Travis White, product manager for precision DC and switching, explained, “The FlexADC converter is based on a combination of off-the-shelf high-speed ADC technology and a custom-designed sigma-delta converter. This combination optimizes linearity and noise for up to 7½-digit precision and stability while offering digitizer sampling rates up to 1.8 MS/s.
“Furthermore,” he concluded, “all FlexDMMs employ one of the most stable on-board references available—a thermally stabilized source that provides unmatched performance. The result is a maximum reference temperature coefficient of less than 0.3 ppm/°C and time stability on the order of 8 ppm/year.”
Application-Specific Formats
Mr. White commented on the suitability of LXI- and PXI-based DMMs for different applications. In his experience, a need for remote access or LAN connectivity often drives the selection of an LXI-based benchtop DMM as an alternative to GPIB. On the other hand, PXI has been more widely adopted in ATE systems where throughput and synchronization are important. PXI also addresses situations where complementary mixed-signal and RF instrumentation need to be integrated with DMM measurements.
Mike Dewey, senior product marketing manager at Geotest-Marvin Test Systems, agreed with Mr. White: “We are firmly committed to the PXI modular platform. We see it as the ideal platform for delivering cost-effective and compact functional test solutions. For test applications that require a range of instrumentation and switching as well as a DMM, PXI is the logical choice.
“If you need a distributed measurement system, LXI could be a good choice because of the capability to connect the components as a network,” he continued. “A good example would be some kind of distributed data acquisition system.”
Mr. Dewey commented that the high performance of the 7½-digit DMM the company sells resulted from advances in measurement devices as well as a better understanding and characterization of component drift over time. A further factor was the extensive use of software-based calibration.
As suggested by VTI Instruments’ former name, VXI Technology, the company has developed modular test systems for many years. More recently, and signified by the change of name, the emphasis has changed to include several new LXI products in addition to existing VXI expertise. VTI’s Mr. Sarfi postulated, “If modular instruments can penetrate new markets, then you can expect the modular DMM market to grow as well because it is a core instrument.
“LXI provides a more flexible platform for DMM development because it is not restricted to a specific form factor, and this allows vendors to add more capability into their products,” he said. “For example, we’ve tightly integrated our DMM with the system multiplexers within the EX1200 mainframe. An internal backplane allows measurements to be directly routed to the DMM; however, unlike many competing scanning DMMs, we’ve also made the inputs available through banana jacks on the front panel.”
Embedding Ease of Use
Of course, having a DMM and using it effectively are two different things. To simplify test system integration, the soft front-panel code for the EX1200 DMMs has been embedded in the products so there’s no third-party software required.
In addition, a calibration routine also has been embedded. The user can develop a cal routine based on the embedded standard procedures. And, there are embedded test scripts to support the most widely used cal standards. According to Mr. Sarfi, embedding the relevant cal code means that VTI is controlling the software revision level, which relieves some of the cal code development burden from the customer.
Test efficiency is further improved because “we’ve embedded the capability to add scan lists that can be initiated with a single command to the device,” explained Mr. Sarfi. “The scan list has the intelligence to maintain not only a channel list to be scanned, but also DMM configurations on a per-channel basis so that multiple measurement types can be made, such as temperature, voltage, current, or frequency, without the need for host intervention.”
Common Themes
A few common themes run through this year’s crop of DMMs. One of the more obvious is system-level improvement. In other words, why have a high-accuracy DMM if the rest of the measurement system is below par?
VTI, Keithley, and Agilent have LXI DMM and switching combination mainframes. The inherent advantage of this type of product is the tight coupling between switching or instrument cards and the integral DMM. The vendor controls the interconnections to ensure the least possible signal degradation. The downside is that the plug-in cards have a proprietary form factor and arrangement of electrical connections.
If a vendor’s range of compatible cards can address your application, choosing one of these solutions may have the best combination of performance and cost. However, many of these systems offer limited instrumentation beyond switching. In contrast, a PXI or VXI DMM can be freely mixed with a range of instrument and switch modules. Of course, then you are responsible for ensuring signal integrity across the necessary interconnecting cables.
In a related example of system-level improvements, Keithley relocated the cold junction sensor used with the 3706 System Switch/Multimeter for thermocouple measurements. Without this change, the user could not avoid some of the thermoelectric errors the relocation eliminated.
And, don’t overlook software. How much will you need to provide? How much is built-in? Of the built-in routines, how many are useful in your application? What tests can be accomplished without shuffling data to and from the host PC? The software and firmware associated with each manufacturer’s DMM are handled differently, but all vendors certainly are emphasizing them.
A final theme is the specialized technical expertise required to achieve each extra digit of resolution above about five. Over the years, high-resolution DMMs have used a variety of analog-to-digital conversion schemes with varying degrees of success. The approaches that recently have been discussed in some detail tend to be composite circuits.
For example, a dual-path ADC patented by Agilent combines a slow, accurate ADC with a fast, less accurate one. NI’s FlexADC uses a very high-resolution sigma-delta ADC together with high-speed DSP feedback. The DMMs based on these kinds of ADCs can address more complex applications than ones that use off-the-shelf ADCs. This is the reason the composite approaches have been developed and are yet another part of a company’s competitive position.