Frequently Asked Questions

Signal Integrity

EMI

Samtec connectors meet FCC Class B EMI Requirements?

Federal requirements (FCC 47 CFR Part 15, EN55022, etc.) require EMI compliance testing of active electronic systems such as computers. While passive components can impact the overall system level EMI performance, it is not possible to directly test a connector, cable assembly, resistor, or bolt for EMI compliance.

Shielded connectors and cable assemblies help reduce emissions and improve immunity of electronic products?

Yes they can help, and in some cases dramatically. Signals that have periodic switching such as clock lines are a classic noise source. If energy from this type of source couples to a cable that exits a shielded enclosure, the cable can radiate the coupled noise much like a cell phone antenna radiates (although not as efficiently). Placing a shield over the cable (and terminating the shield properly) can reduce the radiated emissions dramatically. For a board to board connector inside an enclosure, the situation is similar except the radiated emission occurs inside the enclosure so the radiated emission may or may not be visible during an EMI test.

How much shielding does the Q2™ products have compared to Q Series™ products?

Testing showed an improvement of 10-20 dB over the frequency range of 1-4 GHz and 0-10 dB from 4-10 GHz. This was a comparison test where the radiated field of the shielded connector (Q2™) was compared to the radiated field of an unshielded connector (Q Series™). The key to a meaningful test is that all variables in the test setup and on the test board need to the same for both tests. Full wave simulation showed an improvement of approximately 10 dB over the 1-10 GHz frequency range as well.

What are the best practices to reduce EMI when using a board to board (BTB) connector?

There are a few assumptions that need to be stated up front before we get into answering this question:

• We are addressing open pin field connectors like the Samtec QSE/QTE or SEAM/SEAF products, not a coaxial RF connector configured for BTB applications We are primarily interested in digital applications that span kbps to Gbps data rates. We are not addressing very low frequency analog (audio) type applications. The connectors are part of an overall interconnect system that can include printed circuit boards (PCBs).

It is important to understand the radiation mechanism or dominant EMI antennas in a BTB system. One analogy that works well is to consider the PCB ground planes as elements of a microstrip patch antenna. Any longitudinal voltage potential developed across the connector appears as a voltage source which serves to drive the PCB ground planes, not unlike a feed element driving a microstrip patch antenna. This is a useful low frequency (50 MHz to 100's of MHz) approximation; at higher frequencies the radiating system is more complex but the best practices remain the same.

To minimize the longitudinal voltage potential across a connector, we need to minimize the self-partial inductance of the signal return path. Some brief terminology definitions are needed:

• Loop inductance is the only inductance that is “real" or can be measured. Self-partial inductance is a useful mathematic construct; it can be calculated but not directly measured. “Self" means we are interested in the magnetic flux developed from only one conductor of the loop. “Partial" means we are looking at only a portion of the path, specifically the portion between the PCBs.

To minimize the self-partial inductance across the connector there are some general principles to follow:

• A short BTB stack height is better than a taller BTB stack height. Minimize the loop area in the connector by having ground pins integrated into the signal pin field (1:1). Use broad, flat conductors for signal return, dedicated planes are superior to pins for signal return Make the characteristic impedance of the signal and ground pattern as low as possible.

If EMI were the only issue, the BTB connector should have a very low characteristic impedance (< 1Ω). Planar distribution for signal and return currents would have the lowest self partial inductance across the connector and the lowest EMI. Clearly this is not a practical guideline as data transmission requirements dictate a close match to the system impedance (typically 50Ω).

The same physics that applies to EMI reduction in BTB connectors also apply to PCBs. Disruptions in the signal return path such as splits in the ground plane drastically increases the self partial inductance of the return path which can lead to EMI problems.

The best EMI strategy for electronics is to consider multiple areas such as PCB design (stackup and circuit layout), connector selection, signal return management (grounding), power filtering (decoupling/PI design) and logic selection with an emphasis on spread spectrum clocking and edge shaping.

Power Integrity

If I am using a Samtec ECUE-12-XXX-T1-FF-XX-1-XX FireFly™ cable assembly, can I run auxiliary power through the twinax?

Auxiliary power is usually run through the PCB. The objective of a Flyover® cable assembly (such as the FireFly™ products) is to take the data off of the PCB to lessen the losses associated with the PCB and run the low-speed signals and power through the PCB. It is best practice to avoid running power through the twinax cables. If power must be run though the twinax cables, the power circuits must be very well bypassed and filtered at both ends. For more assistance, contact [email protected].

Additional considerations:

• Even if the power circuit is well bypassed and the current does not exceed the current capacity of the chosen conductor(s), the voltage drop has to be carefully checked. The TTF-36100 micro twinax cable specification says that the center conductor has approximately half an Ohm resistance for each foot length. The shield resistance is approximately three times lower. So for each foot of length and for each ampere of current, the voltage drop is 500 mV along the center conductor and 150 mV along the shield.
• The relatively high resistance of the shield that comes from the very small size means that the twinax cables lose shielding effectiveness up to higher frequencies (as opposed to bigger coax cables with heavy braid). As such, there is strong interaction among the twinax conductors below about one MHz, so if one carries power, any residual power noise at low frequencies will induce common-mode noise on the high-speed signals.

I am considering using a Samtec UDX6 connector. It has a block of high-speed pins and power blades in the same housing. How should I assign the power blades to different nets, what are the considerations?

In addition to the usual signal-integrity considerations, there are two potential interactions to consider related to the power blades: (1) how the pin and blade assignment impacts the power delivery and (2) signal-to-power crosstalk. For power delivery, the DC resistance (which determines the current carrying capacity) and the loop inductance matter most. The resistance also varies slightly with pin assignment. However, the inductance, which influences the transient response of the power path, will change significantly with blade assignment. When the inductance of the power delivery path matters for the application, we usually want to minimize it. The loop inductance formed by the blades is directly proportional to the size of the loop.

If the goal is to minimize loop inductance, we need to assign the power and power-return (or 'GND') blades in a pattern that forms the smallest possible loop size. This can be achieved by selecting adjacent blades for power and ground.

The crosstalk between power blades and high-speed signal pins also depends on the geometry of interacting power and signal loops. The interaction can be minimized by making the loops smaller, by placing them farther apart, or by forming orthogonal loops so that the magnetic coupling is minimized. If, in addition, we can also create the two loops such that the capacitive coupling also cancels, it will minimize the full electromagnetic coupling.

With multiple signal pins and power blades in the connector, there is no universal way to make use of crosstalk cancellation by orthogonal current loops, so we need to focus on minimizing loop size and placing them as far apart as we can.

In the high-speed block of pins, assigning signal and ground pins adjacent to each other is the accepted norm. In the power-blade section, assigning power and ground to adjacent blades will provide two benefits: as we saw above, it minimizes the power delivery inductance and it minimizes the crosstalk between power blades and signal pins.

And one last detail: the above considerations looked at circuit loops, when the signal, whether it is the high-speed signal or power noise, is considered differentially between the pins or blades. Sometimes common-mode noise is also important to minimize. It can be achieved by placing the pins and blades that are 'quieter' with respect to the ambient ground structure next to each other.

To read more on the subject, including pin and blade assignment illustrations, see Pin and Blade Assignment Considerations in Samtec UDX Connectors. For more assistance, contact [email protected].

Does Samtec have testing procedures for their Power Integrity Products?

Yes, you can find a copy of the Power Testing Standards White Paper by clicking here.

Working Voltage is listed in VAC, but what is it in VDC?

The DC rating is the peak of the sine wave from the AC rating, which is 1.414 times the rated VAC. For more information, click here to review our Power and Voltage Ratings White Paper.

The Current Carrying Capacity is based on what voltage?

The CCC is applicable up to the rated working voltage.

What is the difference between Breakdown Voltage, Working Voltage, and DWV?

Breakdown Voltage is the point of failure of the connector. Working Voltage is the maximum continuous voltage that the connector should be used at. DWV is the test voltage of the connector.

Why is Working Voltage 1/3 of the DWV?

It is rated at 1/3 the DWV to take into account typical surges and spikes so the Working Voltage won't exceed the Breakdown Voltage.

What information do you have on voltage surges and current surges?

These questions need to be answered on a per application basis. Please contact our Engineering Support Group.

Why do some Current Carrying Capacity Curves end at 105 degrees C and others at 125 degrees C?

Tin contact interfaces are rated at 105 degrees C, and gold contact interfaces are rated at 125 degrees C.

How does Samtec measure Current Carrying Capacity?

Samtec measures at the centralized hot spot of the contact or contact grouping with a calibrated thermocouple. Click here to view a photo of our lab set-up.

How much will the number of contacts energized affect the Current Carrying Capacity?

This question needs to be answered on a per application basis. Please contact our Engineering Support Group.

Do Samtec's products meet Creepage and Clearance standards?

Creepage and Clearance requirements vary greatly depending on the actual application and as such must be dealt with on an individual basis. Please contact Samtec's Engineering Support Group (ESG) for further information regarding Creepage and Clearance.

Does Samtec have a specific group to contact for questions related to power connectors and/or their testing?

Contact our Engineering Support Group for these types of questions.