By Ehud Shany WinSoft Inc.

With new PXI modules introduced to the market in growing numbers, the replacement of the legacy instruments is available without the modification of the legacy TPSs or their software drivers.

Legacy test and measurement instruments face end-of-life problems, lack of replacement parts, and require on going maintenance and calibration. In most cases, said instruments reside in three-bay Automatic Test Equipment (ATE) that occupies significant floor space. Their dimensions and technical conditions require valuable storage space and cause countless project delays.The replacement of these instruments with PXI based modules requires modifications and revalidation of old Test Programs Sets (TPS), which have been developed under old programming languages such as HP-Basic, Assembly, FORTRAN, and ATLAS. Changes to these software applications can be impractical due to the magnitude (budget and schedule) of such an effort. Until recently, said replacement was almost impossible from a technical aspect, due to the lack of modules with similar hardware capabilities, such as high density switching or high frequency RF modules.

Current status description
The footprint of legacy testers became a major problem in places where these testers were being used. A vast number of instruments residing in these testers are frequently found to be obsolete, and replacement spare parts become unobtainable.These instruments are integrated with other types of equipment (electrical or mechanical) and share a common communication protocol such as GPIB, RS-232, and 1553. This collection of instruments form what is usually known as a test and measurement system or ATE.As part of the need to replace these old instruments due to obsolescence issues, the goal of current ongoing ATE upgrade projects (mainly in the military community) is to reduce the size of the legacy testers by replacing the instruments with a new, smaller version. One of the best options for such an upgrade process is with PXI modules. In those cases that PXI modules are not available as a replacement, a combination of rack-mount instruments and PXI instruments may be found to be best option. In summary, the key factor of said ATE upgrade process is based on cost and schedule requirements. The cost of replacing an obsolete piece of hardware often implies changes in software (part or all of it) and potential communication protocols, to accommodate different functionality. Additionally, changes to the software require system recertification.An ideal ATE upgrade process will include the following elements:

• Support of PXI modules or a mix of PXI and rack-mount instruments with a variety of protocols (such as GPIB, RS-232)
• Reduction of the ATE size from three-bay to one-bay.
• No additional space is required for the implementation of the solution.
• A scalable solution will enable a modular approach in which the ATE will be upgraded over time, while keeping some of the old instruments active until a new budget is allocated.
• No changes to the software will be required regardless of the programming language, communication protocol, timing differences, mix of old/new instruments and quality of signals.

Technology description
The requirements for an almost ideal ATE upgrade/extension process is fulfilled by the presented technology. This technology will enable replacement of common instruments such as DMMs, spectrum analyzers, and counters. It supports common interfaces such as GPIB, USB, and TCP/IP and is not affected by the TPS programming language [2]. Pass-through mode enables commands that are not translated to pass directly to the old instruments. In those cases, the upgrade process is done in steps by which the ATE mixes old and new instruments.

Since the technology does not affect the source code, it does not require its availability. This is a very important feature for classified military projects that cannot expose their source code.

Figure 1 illustrates a traditional tester in which an old instrument is connected to a CPU through GPIB bus.

Figure 1

In Figure 2, the new instrument is a PXI chassis with modules that are able to perform and provide the same functions as the old instrument. The old CPU still sends and receives the same commands as it did in the past. The communication between the CPU controller and the instruments translator/emulator module is done via the existing GPIB bus and the communication between the translator/emulator and the PXI modules by utilizing a TCP/IP protocol. Each command and/or address sent by the CPU controller is analyzed and translated in real-time in to one or more commands of the new PXI modules that is connected to the bus. The new instrument’s reply is treated in the same way.

Figure 2

It is important to mention that TCP/IP is mentioned as one option among several. In other words, the communication can be USB, LXI, or any other communication type.

The number of instruments supported by one instrument translator/emulator module is limited only by the number of PXI slots. If required, a second PXI chassis can be added to provide more modules. As more and more chassis are added, it is easy to convert a three-bay ATE to one- or half-bay ATE that can host the 1U instrument translator/emulator and several PXI chasses.

Figure 3 shows 2 x PXI chassis with the following main modules:

The instrument translator/emulator is installed under the monitor (on top of the PXI chassis) and takes only 1U of space. A Virginia Panel interface serves as new interfaces for the UUT (Unit Under Test)

Figure 3 (click to zoom)

The original ATE is three-bay in size and is used for radio testing by the US Navy.

As the presented technical solution deals with the communication level only (GPIB or others), it is not required to have the TPS source code available, and the programming language of the TPS does not affect this technical solution.

The instrument translator/emulator module, as part of its functionality, resolves all GPIB (or other protocol) addressing and other special communication packages and messages. It supports several modes, including transparent mode, in which it passes the commands through without any action. This setup allows new instruments to be added to a legacy ATE system while the ATE is undergoing several steps of upgrades.

As mentioned before, the PXI based solution can be combined with rack-mounted instruments. In Figure 4, the instrument translator/emulator communicates with both PXI based instruments and GPIB rack-mounted instruments. The GPIB addresses of the old instruments are mapped in such a way that the instrument translator/emulator knows if it is a TCP/IP (PXI) instrument or GPIB (rack mount) instruments.

As the instrument translator/emulator will need to load new emulators/translator along the life of the refurbished ATE, a web server was added for loading new instruments. This server is also used in cases that undocumented commands discovered during the execution of new TPS (which have not been run yet on the upgraded ATE).

Figure 4

Additionally, the instrument translator/emulator can support those cases in which instrument manufacturers released new instruments with a built in emulation module.

While testing several instruments, we discovered the following problems:

• The built in emulator covers 60%-80% of the command set.

• Changes to the emulation were released per firmware release schedule that was not acceptable (2-4 months on average).

• Illegal commands that worked on legacy ATE did not work on these emulators and the manufacturers had no intention to support them.

• ATE local problems such as delay and quality of signal are not being addressed.

The technology described herein has built in tools that handle these and other integration problems. Its tracing tool can collect and analyze the protocol low-level commands in both directions, and the web server capability enables remote support. An internal utility can provide the configuration of all the instruments that are attached to the instrument translator/emulator module, as this information is provided by the instrument manufacturer. This information is accessible via the internet.

ITAs (Interface Test Adapters) and custom-made instruments have the most complex replacement objectives. The new device often needs to preserve the old signals and their characteristics. Again, in this situation, the instrument translator/emulator provides a replacement alternative.

Figure 5 shows a one-bay PXI based ATE in its final configuration. The instrument translator/emulator is hidden and at least half of the rack is empty and ready for future expansion. New TPSs are written in LabVIEW and the old TPS remains untouched.

Figure 5 (click to zoom)

Figure 6 shows a mobile rack in which two rack-mount legacy instruments, DMM and counter from Racal-Dana, are being replaced with two PXI modules. Although the size of the PXI chassis is identical to these two instruments, most of the chassis is empty and may host 16 more PXI modules. The old DMM and counter were left in the rack for demonstration purpose only.  In real life they are not required for the ATE functionality. 

Figure 6 (click to zoom)

Conclusions
The presented technology extends the life of legacy ATEs by replacing the old instruments with PXI based instruments. The old TPS are preserved. New TPS can be written in modern programming languages such as LabVIEW, while the old TPS stays untouched. All future known technologies are supported, such as:

• IVI (IVI foundation)
• Synthetic Instruments (Agilent Technology)
• LXI (LXI consortium)

The cost associated with this technology is significantly lower than with existing methods by which the software or driver is rewritten and then revalidated.

REFERENCES
1. National Instrument Corporation, www.ni.com
2. E. Shany, WinSoft Corporation, “Method and system for software preservation and communication protocol expansion through hardware replacement,” U.S Patent Pending, 2003, Ref # 004966.P002

Ehud Shany (B.S. in computer science) has been the President/CEO of WinSoft Corporation since 1994. For the past twelve years, his company has delivered hundreds of successful software and hardware projects to other companies such as Boeing, Raytheon, Northrop Grumman, the US Navy and many others. Ehud is an expert in test and measurement applications and has extensive experience in automation, robotics, networking, real-time, and database applications.

WinSoft Inc.
1932 E. Deere Ave, Suite 110
Santa Ana, CA 92705
Tel: 949-428-4844 x 202
Fax: 949-428-4842
E-mail: ehud@winsoft.com
Website: www.winsoft.com