Vision Docs - All your documents in one place

Now available on the app!

Imaging & Vision Handbook More

Improve your level of machine vision know how

BETTER TO KNOW MORE

Benefit from the expert´s knowledge

Vision Docs - Handbook available now

VISION EXPERTISE

AT YOUR FINGERTIPS

Get the Imaging & Vision Handbook! More

Cable testing and validation

As vision systems evolve, we tend to see more cameras per system, faster data rates and the use of network technology between the camera and PC. All these factors increase demands on the cabling and poor cabling or incorrect network switches can have serious effects.

Therefore it is important that your cable provider uses the correct equipment to ensure poor cable quality does not cause problems. Our in-house cable manufacturing department uses the Fluke cable testing equipment to ensure our cables perform to the highest levels.

As part of our quality assurance process we issue an individual test certificate for each cable produced. Tests are carried out using a PCbased automatic cable tester which verifies all wires to ensure we detect incorrect wiring or short circuits. The system works based on the principle of Ohmic resistance.

Using a Fluke cable analyser we are not only able to identify a faulty cable, but also to diagnose the type of fault and where the fault lies. Parameters like cable impedance, capacitance, data skew and signal crosstalk are all analysed and can be validated against cable specifications to ensure compliance. If required, we can supply a test certificate for each cable tested and include it with the delivery. Using our cable analyser we not only test copper cables, but also optical fibres where the purity of the fibre ends and the attenuation of optical fibres is documented.

The image above shows cable testing in progress, while the image below shows a diagram of a complex cable wiring set-up.

For testing GigE Vision cable installations we use a Fluke network analyser that allows us to measure the maximum throughput rate of complete network topologies. We use this to validate the combined performance of the Ethernet switches and the cables we supply and also offer this service to our customers to validate their complete network installations. If you have a need for this type of testing, please contact us for support.

Network analyser

CameraLink - test methods

The maximum frequency possible for data transmission depends on a set of parameters, such as wiring material, material batch, cable length and connector type, etc. In addition attenuation, reflection and cross talk might occur on the cable. When the CameraLink standard was first released, all CameraLink devices were equipped with ChannelLink chips, as a mandatory requirement of the standard, however, with the latest revisions of the standard, more and more implementations are based on FPGA implementations of the protocol which enable additional control of transmission parameters such as pre-emphasis and equalisation. Only when these parameters are optimised, can data be transmitted over a cable with maximum or extended length. We also find more and more cameras using the full CameraLink bandwidth or maximum frequency of 85 MHz on the cable. In addition to mechanical and electrical testing, this requires additional tests of the maximum data rate possible. This is undertaken using a PC based test rig which identifies the bit errors on the cable at increasing clock frequencies.

Early CameraLink testing techniques were undertaken using simple grey wedges as image content. Experience has shown however, that the grey wedges were too "soft" and did not test all transition combinations of the data signals so that all crosstalk errors possible in practical use were not detected. To avoid this, we now use a defined pseudo random pattern (LFSR pattern), that is sent over the cable while error bits are counted at the receiving end. The frequency of the data pattern is increased until errors are detected, which then defines the maximum speed the specific cable length can operate at. This method enables the detection and elimination of quality issues due to production and material variations. These tests are practical and in contrast to so-called eyepattern tests, they can also be used for high volumes of cable production.

Compared to tests with grey wedges, tests with LFSR patterns produce inferior results and lower maximum frequencies are to be expected, due to cross talk and result in a reduction of transmission frequency of 5 to 10 MHz in practice. In practice its the application that decides how sensitive an application is against single error bits. Following the current trend to use FPGA-based CameraLink implementations in cameras and frame grabbers, only the complete combination of camera - cable - frame grabber defines the maximum possible cable length. Practical use shows that with careful component selection, CameraLink is still a very powerful and reliable interface.

USB - test methods

USB 3.0 is a relatively young interface in industrial imaging. While there are specification parameters for USB3 cables certified according to the USB IF standard (www.usb.org), the USB3 Vision (U3V) standards groups is working on practical test methods for checking of USB3 cables for use in high-speed imaging. Unlike most consumer applications machine vision can demand data transmission performance close to the maximum and too many errors would overload the interface error correction. Similar to CameraLink, bit error rates should be measured to give evidence of the quality of the cables. In real applications it is quite complicated to qualify cables for image transmission due to the vast hardware variety available for this consumer interface.

STEMMER IMAGING however is working hard to define methods and solutions for those tests. Currently we are able to test and validate system sets of USB3 equipment purchased from us to remove risk for our customers.