Learn how to avoid the pitfalls that can ruin all of the benefits an automated switch can provide.
Budget constraints, personnel cutbacks, and shorter product marketing windows are the signs of the times. This type of working environment demands your operation be lean, mean, and just plain smart with its existing resources. So, how can you justify buying more equipment?
Tremendous thought and planning should go into both in-process and final testing. A key goal should be to use equipment that is already a company asset in a more cost-effective manner. Squeezing the last ounce of efficiency from a test process or actual test equipment is a great step to being more effective in the market place.
Specifying and buying new automated switching to augment your testing may bring your facility to the next level in productivity. Most sophisticated test equipment built in the last 10 years typically has some sort of remote-control interface that can be used to set up the equipment, make measurements, and record the hard data for evaluation.
For communications or telemetry installations, this may be the time to move to a more automated solution, getting away from setting up a test each time it�s needed or using manual patch bays and patch cords.
The Warning Signs
Growing companies go through various cycles. Many firms call these growing pains, but with planning and scheduling, these cycles can be properly addressed. The more sophisticated the product being manufactured, the more testing probably needs to be done.
As products become more complex, you don�t want your testing process to become a bottleneck. Without automating some or all of your testing, you lose repeatability, reliability, and the speed that test automation provides. Once software control is established for the application, connectivity and results are repeatable without human intervention and error. This saves time and money.
Automated switching isn�t rocket science, but it does take some in-depth understanding of the pitfalls that can ruin all of the benefits an automated switch can provide. Poor frequency response, poor isolation, poor switching elements, or even the wrong connector or cable types can contribute to a poorly performing design.
Defining the Real Switching Need
When designing a system where an automated switch is desired, most of the effort is directed toward picking the stimulus and measurement equipment. This is easy to understand because, when designing a test, you start with the product that needs to be tested, figure what kind of tests need to be performed, and then select the equipment to perform the test. Naturally, any automated switching usually is the last thing to be considered, yet it is the single most important requirement for a successful test.
When you begin to define your switching needs, take the time to clearly evaluate what really must be connected. This not only applies to the environment between your UUT and the test equipment, but also to many other applications that can benefit from the addition of an automated switch.
Many older installations around the world still have patch panel bays. When they decide to automate these, an engineer simply might count all the source patch sockets and all the destination sockets. They mistakenly use this number to define the size of the switching configuration.
Using too big a switch is a waste of money and space and usually has lower performance. Typically, not every source needs to connect to any destination. Looking closer, you simply might be able to segment your switching needs into types of signals such as high current, low frequency, data, RF signal, and other similar groupings.
Other parameters to consider when defining your automated switch requirement are how can you expand it and how much future expansion capability do you want. Only you as the system designer can determine all these issues, perhaps with guidance from a quality automated switch manufacturer.
Pulling the Plug on Patch Cords
Automated switching is the core of a good test system or modern communications center. Automated switching reduces the number of connections that must be made manually during the test of a UUT or a mission at a control center.
For ATE, you only have to connect once to the UUT, and then various signals are connected by the automated switch between test points on the UUT and the source and measurement equipment. With even larger demands, more switching can be added to make testing more efficient.
Many times, the source and measurement equipment are the most expensive components in the overall system design. Using this same equipment to test multiple UUTs is the next step after automating. Perhaps rather than testing one UUT at a time, a full tray or cart of devices can be tested sequentially while the test engineer prepares the next tray or cart.
For communications installations, patch bays have been very popular in the past. As time and technology move forward and personnel reductions continue, automated switching is replacing patch bays. Tremendous manpower is relieved for other tasks.
The familiar phrase time is money is the easiest way to justify the implementation of an automated switching solution. The time spent manually setting up equipment to test a product or connecting recorders and receivers at a communications facility is not time well spent. It is time where the valuable asset is sitting idle during the setup procedure. It also is a time when valuable engineering personnel must perform manual connections that are prone to errors and when equipment is subjected to cable wear and damage.
Making the Change to Automation
The key to automating many test processes is the automated switching system. It is the core piece of equipment and usually the least understood. This is demonstrated time and time again with engineers attempting to build their own automated switch. If you intend to design and construct your own automated switch, keep in mind that building a reliable high-performance automated switch is still a fairly complex task.
Building an automated switch involves more than simply coupling various switches together. Impedance matches and loss calculations through the switch paths are keys to the design of the switch.
For higher frequency applications, at each connection there will be reflections and losses. This is why it usually is more cost-effective to choose a supplier with the necessary experience to deal with the unique problems associated with an automated switch design.
A company usually maintains equipment longer than a design engineer stays with a single company, so support of an automated switch from a supplier would continue. Would you design and build your own network analyzer or signal source? Of course you wouldn�t.
During the implementation of automated switching to replace manual patch panels and patch cords, it is typical for system engineers to keep the patch panels for some time. This can come in handy during system integration, software development, and overall system troubleshooting. It also might give a warm feeling to the engineer who is new to specifying an automated switch.
In a typical configuration, the signal sources are routed to the input patch panels and then from the input patch panels to the input of the automated switch. The output of the switch would connect to the output patch panels and continue on to the user�s destination. Only the system designer can weigh the benefit of keeping the I/O patch panels in the overall design, taking into account that additional losses, mismatches, and failure points are introduced.
Semi or Full Automation?
The final system design decision may not be yours at all but determined by your budget. Remember to consider all parameters of your system design. Nothing will happen without software running your test programs, and data will not be stored or evaluated without having system computer hardware somewhere.
If budgets are too tight and you need to scale back your design, take a look at the most time-consuming and repetitive or manual connections you have to make and automate those connections first. You still can experience significant savings in time and money even if some of the connections are made manually.
Many times, switching equipment is delivered with some type of control GUI, and that is a start. You will need a complete software solution to control the test equipment, the automated switch, and possibly the UUT.
There isn�t an off-the-shelf software package that will accomplish this for you, but there are many tools and drivers available to help the software programmer complete the task. Do some research, and stay with a software solution that has some history of use and the best tools for your application.
Technology and Topology
Many articles and papers about the different types of switch topologies are available. There are just as many, if not more, about the technology of switching elements. There is no single ideal type of switch for all applications, just as there isn�t a single ideal connector type or cable for all purposes.
An automated switch usually is designed for a specific purpose or a limited range of applications. The technology of the actual switching element also plays a very large part in the purpose of the automated switch and its performance.
Switching-element technology falls into these basic categories with each having its purpose, strengths, and limits:
Mechanical
Mechanical is the oldest technology and provides the most choices. One of the big advantages of a mechanical element is the tolerance to overstress current or voltage. Also, it does not introduce any distortion of DC or low-frequency signals passing through. The downsides are limited contact life, larger physical size, and higher cost.
Solid-State
With tremendous advances in technology, use of solid-state switching elements has grown. It has its purpose and place. A major advantage is that the contact does not have any mechanical life limits.
Digital
In modern ATE and communications centers, digital control and status signal switching will be required. It only makes sense to switch these signals in their own technology domain.
Optical
Advances in optics are significant for both the ATE and communications industries. Because fiber connections are delicate, they lend themselves to automation if possible. Lower losses in switching light have made this a reality.
The topology of an automated switch is application specific and depends upon what needs to connect. Automated switch suppliers typically offer a host of configurations to meet various demands.
Multiplexers
The multiplexer is the most common configuration and a very useful building block for larger configurations. Also referred to as a 1�N or N�1, the multiplexer will connect a common source (1) to a number of destinations (N). In reverse, the N�1 topology will connect the single destination to one of a number of sources.
Matrices
The much more advanced and flexible matrix configuration has two axes and is referred to as an N�M topology. With this design, multiple inputs can be interconnected to a number of outputs simultaneously.
So-called blocking designs support only 1:1 connections, allowing a given input to connect to only one output at a time. Others are designed so that a given input can connect to multiple outputs at the same time.
This also has the effect of fan out of the input signal. This type of automated switch is not needed for many ATE applications but is very prevalent in the communications industry.
Consider the Future
Due to the ever-increasing demands by industry, the system designer needs to get more out of less and get it quicker. Tomorrow�s demands will be even higher. A good design allows for reasonable expansion in the future.
When researching the types of automated switching available in the marketplace, keep in mind possible future requirements. There are many factors that affect the cost of an automated switch, each having its impact on features and performance specifications.
Thoroughly study your current switching needs and identify possible vendors that can provide solutions to meet your needs. Also, investigate the next higher performance-level automated switch to compare the pricing difference. It may be easy to justify the additional cost difference compared to buying new equipment later. Should additional performance be required in the future, the automated switch will be ready for it.
Conclusions
Automated switching is here to stay. There always will be a place for connecting test equipment and various types of communications devices manually, but as these fields continue to grow, so will the automation of them.
Designing and building your own automated switch are expensive, time-consuming, and fraught with technical pitfalls. Much of the performance and success of implementing an automated solution depends upon the automated switch design since ultimately all signals in the system may pass through it at some point. The correct automated switch is as important as, or even more important than, the peripheral equipment that connects to it. If you select a deficient switch design, your entire system performance most likely will suffer.
About the Author
Norton W. Alderson is the vice president of marketing at Universal Switching. Formerly a senior staff engineer with Matrix Systems, Mr. Alderson has nearly 30 years design experience of switching equipment for the ATE, communications, and broadcast industries. Universal Switching, 7145 Woodley Ave., Van Nuys, CA 91406, 818-785-0200,
e-mail: [email protected]
November 2005