Search
Menu
Cognex Corp. - Smart Sensor 3-24 GIF LB

Choosing the Better Bus Standard

Facebook X LinkedIn Email
Study compares IEEE 1394 and GigE Vision.

Hank Hogan

Which bus standard is better for machine vision? A recently released study by Fraunhofer Institute for Photonic Microsystems (Fraunhofer IPMS) in Dresden, Germany, tries to answer that question. Commissioned by Sony Image Sensing Solutions, the study compares the GigE Vision/GenICam combo with the IEEE 1394 Instrumentation and Industrial Digital Camera interconnection standard.

It explores issues such as usability, setup and development overhead. It looks at plug-and-play operation, as well as at the separation between transport and protocol layer functions. The study also examines data rates, in theory and in practice, along with determinism and latency. And it looks at the load on the CPU in single and multiple camera settings.

Along with a suitably equipped PC, a total of four GigE Vision cameras, from Dalsa, Basler, JAI and Prosilica, were used. Sony 1394 and 1394b cameras comprised the other bus contingent. The cameras all were set to produce the same data rate, with a typical configuration of 640 × 480 pixels at 30 fps with an 8-bit depth per pixel.

In general, GigE showed promise but fell short on implementation. Some of these shortcomings were seen more readily when the technology was stressed. Others, such as those involving interoperability, were evident in even relatively simple settings.

As for GigE’s advantages, one is a matter of network wiring. The copper wire it uses can work over runs up to 100 m, far longer than the 4.5 m of IEEE 1394. However, with optical fiber, 1394b also could reach 100-m length.

The authors conclude that GigE has promise, with higher raw data rates and longer maximum cable lengths than IEEE 1394. However, the lack of quality of service guarantees in the transmission of data and greater loading on the host processor could lead to performance problems for GigE compared with IEEE 1394. In addition, it is sometimes necessary to interact with GigE hardware at a very basic level.

“If you really want to get the maximum performance, you have to deal with very low level timing characteristics and timing constants,” said Michael Scholles, business unit manager for sensor and actuator systems at Fraunhofer IPMS.

Hamamatsu Corp. - Earth Innovations MR 2/24

As a result, the study contends, hardware integration, optimization and interoperability are more challenging for GigE. Because of these factors, the study concludes that system designers might decide to stay with IEEE 1394 because it is the established technology, while they wait for GigE, which is more rapidly evolving, to mature.

However, Sony currently does not make any GigE cameras and is a strong supporter of IEEE 1394. Therefore, it’s natural to wonder about bias in the study. In response to such concerns, Scholles said that the study was designed to objectively compare the specification and performance of GigE Vision with IEEE 1394. It was published in a Sony white paper called “Can GigE Vision deliver on its promise?”

Others, however, have come to different conclusions about the best bus. For example, Basler AG of Ahrensburg, Germany, makes cameras using both types of bus standards. Friedrich Dierks, head of software development for components, said the company uses this approach because both standards offer pros and cons.

He said that if a low load on the host CPU is important, the company suggests using IEEE 1394 because the load will be about one-fifth that for GigE on a fairly standard current machine. In addition, if jitter in the image transfer time — as measured from start to end of transmission — is important, IEEE 1394 is the better choice because it has a value of about 0.2 ms in contrast to roughly 0.8 ms for GigE.

On the other hand, he noted that the company suggests GigE if requirements include cable runs up to 100 m, data rates up to 100 MB/s or when image delivery must be guaranteed. That is because the built-in packet resend mechanism of GigE ensures that data will arrive at its destination, Dierks explained. “Thus, for critical applications where losing an image is not an option, GigE Vision is the better choice.”

Basler’s stance is that there is no right answer to the question of which bus is better. The choice depends on the desired destination.

Published: May 2008
Glossary
digital camera
A camera that converts a collected image into pixels that are black or white digital or shades of gray. The digital data may then be manipulated to enhance or otherwise modify the resulting viewed image.
machine vision
Machine vision, also known as computer vision or computer sight, refers to the technology that enables machines, typically computers, to interpret and understand visual information from the world, much like the human visual system. It involves the development and application of algorithms and systems that allow machines to acquire, process, analyze, and make decisions based on visual data. Key aspects of machine vision include: Image acquisition: Machine vision systems use various...
CPUdigital cameraindustrialmachine visionResearch & TechnologySensors & DetectorsTech Pulse

We use cookies to improve user experience and analyze our website traffic as stated in our Privacy Policy. By using this website, you agree to the use of cookies unless you have disabled them.