With Ethernet everywhere you look, it isn’t surprising that the days of dominance for the IEEE 488 bus could be approaching the end.
Today, we buy a PC from one manufacturer and attach a monitor from another, a printer from yet another, a scanner, a camera…and magically—but not necessarily effortlessly—they all play together. We assume that devices from multiple manufacturers will work together as long as they share the same interface. This happens because the interfaces have been well defined, usually by a standards committee with the objective of interoperability among similar devices from different suppliers.
Twenty-some years ago, the instrumentation world benefited from the work of an Institute of Electrical and Electronic Engineers (IEEE) standards committee. The IEEE 488 standard defines how instruments communicate with computers and with each other. It was a direct result of an interface called the HP-IB developed by Hewlett-Packard.
The adoption of the interface by the IEEE 488 committee opened the door for test-system designers to depart from proprietary buses and single-supplier test systems. Long before PC users enjoyed the ability to effortlessly connect together devices from different suppliers, engineers were networking instruments from dozens of instrument suppliers.
The IEEE 488 bus had a long and successful run and will likely be sustained for another decade before it becomes difficult to find instrumentation to support it. However, there are several reasons why the IEEE 488 bus finally may have seen its best days and why it may become increasingly difficult to find newly designed instrumentation with built-in IEEE 488.
The primary reason: There finally may be an interface that offers enough advantages over IEEE 488 for a shift to take hold. Surprisingly, that interface is not one of the new, high-profile interfaces such as the universal serial bus (USB) or FireWire (IEEE 1394). While these new standards have entrenched themselves in the PC and peripheral world, their adoption for use in general-purpose instrumentation has been very limited.
The interface with the highest likelihood of replacing IEEE 488 is tried-and-true Ethernet. Recently, some instruments have been introduced that offer Ethernet as their only link to a PC, and there are several reasons why this may be the start of an industry-wide trend.
Market Conditions
The last two years have been very difficult for most instrumentation suppliers. Sales of IEEE 488-based products have declined by 50% or more, as evidenced by the revenue drop of the largest suppliers of IEEE 488 instrumentation.
This is compounded by the fact that a large quantity of IEEE 488 instrumentation is on the shelves in companies or available from second-party suppliers for a fraction of its original cost. This means that the investment in developing new IEEE 488 instrumentation is down since revenues to the instrument developers are down substantially.
The flow of new products using the IEEE 488 bus assuredly will be negatively impacted by the industry downturn. While the market for instrumentation in general will likely improve over time, IEEE 488 may not be the primary interface when it does.
Cabling
Adherence to the IEEE 488 specification dictates that the maximum length of the cable be limited to 6 ft. To go much beyond 6 ft, it is necessary to purchase IEEE 488 bus extenders, a pair that will cost from $1,000 to $3,000.
IEEE 488 bus extenders are not 100% transparent because the specification has very tight timing restrictions that make it electrically impossible to extend beyond 6 ft without introducing exceptions. Usually, some minor programming changes, such as replacing parallel polls with serial polls, will get around the restrictions.
In contrast, Ethernet allows up to 200 m between devices (100 m radius from a hub) and greater than 2,000 m if switches or fiber converters are used. These may be purchased for less than $200 and less than $1,000, respectively.
Unlike IEEE 488, Ethernet extenders are 100% transparent and do not require program changes. The main advantage of Ethernet in this regard is the capability to remotely locate instrumentation, for example, on a production line or on multiple pieces of machinery located throughout a facility. In the case of IEEE 488, this becomes very expensive and cumbersome; in the case of Ethernet, this is trivial.
Ethernet’s twisted-pair cables also are significantly easier to work with than bulky IEEE 488 cables. Ethernet also provides the options of fiber-optic cabling or wireless operation if desired.
IEEE 488 supports daisy-chain or parallel-cabling topologies, while today’s Ethernet devices typically sustain a hub-and-spoke technology, similar to a parallel topology but the devices are electrically isolated from one another in the hub. Because of the limited cable length of IEEE 488, a daisy- chain connection is needed to increase the total distance devices can be placed from a controller. If the daisy chain fails, it will disrupt all the devices after the failure.
IEEE 488 cables also are expensive, typically $100 each for a 6-ft cable. In contrast, a similar length of Cat-5e Ethernet cable can be purchased for $10 or less. So from a cabling perspective, Ethernet has substantial cost and performance advantages over IEEE 488. For distributed applications, often the Ethernet cabling already is installed in the facility so the incremental cost of cabling is zero.
Interface Cost
When PCs became available along with plug-in IEEE 488 cards in the early 1980s, the cost of an IEEE 488 controller went from $5k to $10k for a dedicated controller from HP or Tektronix to $3,000 for a PC with a $400 plug-in card. With today’s PCs costing less than $1,000 including a built-in Ethernet port at no extra cost, the $400 required to add an IEEE 488 card to a PC is significant. Couple this with the cost of cabling, and IEEE 488 becomes increasingly more expensive than an Ethernet-based solution.
Electronics
The demand for bus interface devices and bus drivers that interface to the IEEE 488 bus has been going down steadily over the past several years. As a result, the availability of these key devices is problematic, and many IEEE 488 instrument suppliers are faced with one of three alternatives:
- Make lifetime buys, which involve tying up capital in devices and trying to predict the life of their product.
- Design their own IEEE 488 devices, which often are made to be compatible with the existing devices so they can continue to build existing instruments without having to redesign the interface portion.
- Purchase the devices from their competitors who have designed replacement devices.
Not one of these alternatives is particularly attractive, especially to a small or medium-size user of devices who doesn’t have the volume to amortize the expense.
In contrast, the devices to implement Ethernet are plentiful. The electronics cost to implement a 10/100Base-T interface is very low, and the parts are available from at least a dozen suppliers. An Ethernet interface board for a PC, for example, can cost as little as $15.
Speed
The IEEE 488 bus is an 8-b bus, and 99% of the devices in existence are limited to 1-MB/s transfer rates. In reality, most devices transfer at a much slower rate since most of the instruments are based on slower microprocessors that cannot handle the 1-MB/s rate.
An enhancement to the original bus, called HS488, has been around for several years, although there is little evidence that many companies are using it. The HS488 spec calls for a maximum rate of 8 MB/s.
Ethernet can transfer data at much faster rates than standard IEEE 488. A PC with Ethernet built in, presuming that it is not bogged down with network traffic, can transfer data at upwards of 8 MB/s. From a speed perspective, Ethernet and the new 802.3z, 802.3ab Gigabyte Ethernet standard will far surpass the speed capabilities of IEEE 488.
Networking
In 99% of the applications, communications in an IEEE 488 bus application are between the PC and a single instrument. While the IEEE 488 specification does allow peer-to-peer communications, the protocols are not well defined, and it is seldom used in actual test applications.
In the case of Ethernet, peer-to-peer communications are supported along with broadcast and multicast mechanisms that would allow for simultaneous instrument triggering and other control scenarios to be implemented. In addition, multiple controllers can be on the same wire talking to different sets of instruments, further reducing cabling issues.
Possible Caveats
There are aspects of early Ethernet technology that could have been viewed as negatives with regard to instrumentation control. The Ethernet protocol allows for collisions to occur between devices transmitting on the net, and these collisions degrade overall network performance and make the protocol nondeterministic. However, the introduction of switch technology to Ethernet has mitigated these concerns to the point that some industrial-control manufacturers are developing protocols on Ethernet for real-time control of the factory floor.
In facilities with existing networks, studies must be performed to determine if the additional load of instrument control can be supported. The judicious use of subnets and load balancing should make the use of existing nets feasible. Since communications on Ethernet typically are peer-to-peer, existing traffic should not be affected beyond overall network performance.
Outlook for the Near Future
Where they would have had IEEE 488 in the past, an increasing number of instruments are becoming available with Ethernet as the built-in interface. This trend will very likely continue. The adoption of Ethernet for test and measurement applications also will be impacted by the rate at which the market for test and measurement equipment improves.
About the Author
Tom DeSantis is the founder and president of IOtech. He holds a bachelor’s degree in electrical engineering and has more than 20 years of experience in the test and measurement industry. IOtech, 25971 Cannon Rd., Cleveland, OH 44146, 440-439-4091, e-mail: [email protected]
Return to EE Home Page
Published by EE-Evaluation Engineering
All contents © 2002 Nelson Publishing Inc.
No reprint, distribution, or reuse in any medium is permitted
without the express written consent of the publisher.
November 2002