07-07-2015 | | By Marco Enge
The increasing need for speed in high-end consumer electronics such as gaming and entertainment systems shows little sign of slowing down. Faster and faster processing rates in computers and smartphones means the exchange of data is accelerating. Maintaining signal integrity gets harder at high speed, and connectors play a much bigger role. Here are five things you need to know when specifying connectors for a high-end system.
Rise times faster than 1ns generally require high-speed data connectors
First of all, do you need high-speed data connectors, or will any bog standard connector do the job? Generally speaking, signals with a rise time faster than 1ns are defined as high speed. The rise time is the time it takes to change between a logic 0 and a logic 1, though it’s usually specified as the time taken to change between 10% and 90% of the maximum signal level. If the rise time is 1ns or less, specialised high speed data connectors need to be specified, as beyond this level, the connector is the weak link in the chain for signal integrity. Bandwidth and rise time are related, so if you know the bandwidth of a signal, you know its rise time, and vice versa. Bandwidth for square wave signals may be calculated as 0.35 / rise time. As an example, signals with a 1ns rise time will have a bandwidth of 350MHz.
High-speed connectors are designed for either differential or single-ended signals, but can be used in either
Nowadays, the majority of high-speed data connectors are designed for differential pair communication, but it pays to double check. High-speed data connectors may also be designed for single ended applications, or may offer good performance for both single ended and differential. However, just because something is designed for one thing doesn’t mean it can’t be used for another. For example, a single ended optimised connector may be used for a differential signal, but it’s unlikely you’ll get the same speed out of it. Likewise, the impedance of the connector may be substantially different. It’s important to make sure you are working with parameters defined for the signalling scheme you are using.
The -3dB insertion loss point might not tell the whole story
One of the key parameters of a high-speed data connector is its insertion loss, which tells you how much of the signal is lost due to reflections from impedance mismatching, crosstalk to other signal lines, and the resistance of the parts of the connector.
A connector’s bandwidth is generally described as the frequency difference between where the signal drops to 70% of its power at the high end of the spectrum and the low end (the two 3dB points). Insertion loss of individual connectors is often given as ‘-3dB insertion loss’ (insertion loss at the -3dB point) so that different connectors can be compared.
Insertion loss depends on the frequency of the signal, but it’s not a linear relationship; small resonances can occur at certain frequencies depending on the physical proportions of the connector, increasing insertion loss. Depending on the actual frequencies present in your signal, there may be some unforeseen effects if this coincides with a resonance point.
Non-electrical properties also help define a connector
Aside from the electrical properties of the high-speed data connector, there are several other things to consider when specifying, just like with any type of connector. For example, are there any space constraints in your application that would have implications for the connector size or density? If so, you may find that your options are limited.
Have you considered the number of times the connector will be mated? Some connectors (like PCB daughter board connectors) are only expected to be mated once during the whole life of the equipment. Whereas connectors for, say, camera memory cards are expected to be repeatedly mated and unmated in uncontrolled conditions and therefore require entirely different types of contacts that can withstand this sort of treatment.
What environment will the end system be used in? Will it be outdoors? Left on the dashboard of a hot car? Or in a server room with strictly controlled air conditioning? Connectors have different levels of protection from moisture and dust ingress that need to be considered, as well as the operating temperature range.
Cost is often the critical factor
Just about all the electrical and physical properties listed above are traded off with cost. As you’d expect, higher performance connectors are higher cost, but careful consideration of the system requirements should allow you to specify a connector that will work reliably at the lowest possible cost point.View Avnet Abacus connectors here: www.avnet-abacus.eu/connectors