Industrial Oscillator Design Notes For Gateways

An industrial gateway is expected to run in a cabinet that may be warm in summer and cold in winter, to start reliably after a power interruption, and to keep its network link up for years without a clock related error. The oscillator that feeds the processor is a small component with a modest cost, and it is also the component that decides whether the timing budget is met at the extremes of the operating range.

This article works through the parameters that an industrial application actually depends on, and the layout rules that keep the specified performance intact on the board. The emphasis is on the figures that are easy to overlook because they do not appear in the headline specification.

What The Industrial Rating Covers

The industrial temperature range is quoted as minus forty to eighty five degrees Celsius, and a part carrying that rating has been characterised across the whole range rather than at room temperature alone. For a network interface the relevant limit is usually the accuracy required by the protocol: an Ethernet link tolerates a reference error of tens of parts per million, and the oscillator has to stay inside that limit at both ends of the range.

The rating is a statement about the part, not about the board. A part that meets its specification in free air can exceed it inside a sealed enclosure where the local temperature is higher than the ambient. The temperature that matters is the one at the oscillator, and it should be measured on a prototype rather than inferred from the room.

Frequency Stability Over Temperature And Life

Three figures combine. The initial tolerance at the reference temperature is the smallest of them, and it is the one most often quoted. The temperature stability adds the deviation across the operating range, and the ageing adds the drift over time, usually quoted for the first year and for each year afterwards at a lower rate.

For a product with a ten year life the ageing term becomes significant, and it should be added to the temperature term rather than compared with it. A part with a small initial tolerance and a poor ageing figure can end up outside the protocol limit before the end of the design life, which is the kind of failure that appears in the field rather than in the qualification test. The three figures should be summed and compared with the requirement.

Start Up Time And Power Sequencing

The start up time is the interval from the application of the supply to the first valid output edge, and it has to be shorter than the time the processor allows before it abandons its own boot. A typical figure of a few milliseconds is comfortable for most processors, but it becomes a problem where the supply ramps slowly or where the oscillator is enabled by a signal that arrives late in the sequence.

A slow supply ramp is the more common cause of a start up failure. Some oscillators require the supply to reach a minimum level before the internal circuit begins to oscillate, and a supply that lingers below that level delays the start without any visible fault. Where the design includes a controlled ramp or a sequencing circuit, the oscillator should be the first device to be enabled and the last to be disabled.

The Enable Pin And Clock Gating

An enable pin allows the clock to be stopped when the processor is in a low power state, which saves more current than the oscillator itself draws. The logic is not uniform between parts: some devices enable the output when the pin is high or left open, while others require the pin to be tied to the supply. Reading the polarity from the datasheet and not from an assumption is the whole of the precaution.

Leaving the pin floating is a risk rather than a convenience. A floating input can be driven by noise in a cabinet full of switching equipment, and the result is an oscillator that stops intermittently for no apparent reason. Where the enable is not used, the datasheet normally recommends tying it to the supply through a short connection, and that recommendation should be followed even when it costs a resistor position.

Supply Current And Rail Separation

The supply current of these parts is modest, often tens of milliamps, but it is drawn in narrow pulses synchronised with the output transitions. Sharing a rail with an analogue to digital converter or a radio frequency section allows those pulses to appear as noise in the measurement, so the oscillator should have its own branch from the plane with its own decoupling.

The decoupling arrangement is a small ceramic capacitor at the pin plus a larger capacitor nearby, with the ground connection made by a via directly beside the pin. A layout that routes the oscillator supply through a long trace to a distant regulator reintroduces the impedance that the capacitor was placed to remove. The same separation applies to the output: the clock trace should be routed away from the analogue section rather than alongside it.

Footprint And Layout

The output trace should be short, should not pass under the oscillator, and should cross an unbroken reference plane. Where the trace has to reach a processor several centimetres away, it is worth asking whether the oscillator can be moved rather than whether the trace can be extended. A length of fifty millimetres is a practical upper bound for an unterminated single ended clock before the edge quality begins to suffer.

The footprint should be taken from the recommended drawing rather than scaled from a similar package, because the pads of a small oscillator are positioned around its internal lead frame. A pad that is slightly out of position allows the part to rotate during reflow, and on a package that contains a suspended quartz blank the resulting stress can shift the frequency permanently. The standard practice for pad design and the placement rules for component placement apply directly, and the reflow limits should be taken from the datasheet rather than from the paste profile. The defects that appear when a part moves during soldering are described under component shift.

Additional Considerations for This Build

Practical attention to aging pays for itself here, because it decides whether the finished board behaves as the drawing intended. Where the requirement is not stated on the fabrication drawing or in the assembly notes, the shop has to assume a default, and that default is rarely the value the design was simulated with. Stating aging explicitly, together with the tolerance that applies, removes the assumption and keeps the result predictable from batch to batch.

Process Control and Verification

On a design of this kind, enable pin is the item that decides how the rest of the board is arranged. A short note on the drawing about handling, storage or packaging is often worth more than an extra decimal place on a tolerance. Where a value sits close to a process limit, the drawing should say so, since the shop can then open the process window rather than working to a nominal figure that carries no tolerance.

FAQ

Is the industrial temperature range enough for an outdoor cabinet? The range covers the component, not the enclosure. The temperature at the oscillator inside a sealed cabinet can be well above the outside air temperature, so it has to be measured rather than assumed.

Should the enable pin be left floating when it is not used? No. A floating pin can be disturbed by noise. Where the function is not required, the pin should be tied to the level that keeps the output active, as the datasheet specifies.

Does a lower supply current mean a weaker clock output? Not necessarily, but the edge rate and the drive capability are separate figures. A part with a low current may need a shorter trace or a different load, and the output specification should be checked against the receiver.