The Jitter Budget as a Fraction of Cycle

Ada re-derives this chapter’s own numbers step by step, at full precision

foundations
math-foundations
calculation-audit
applications
Ada ADA · CALCULATION AUDIT

The Jitter Budget as a Fraction of Cycle

The chapter’s jitter table quotes each application in mixed units — milliseconds in one row, microseconds in the next — hiding what really separates them. Put in one currency, three applications tolerate jitter of a full 10% of their cycle, while a printing press allows only 0.8% and semiconductor handling just 0.16%62.5x tighter. This audit converts the jitter budget to a fraction of cycle to reveal the harder constraint.

Companion to the chapter Real-Time and ISA-95 — every number here comes from that chapter.

Ada: The jitter table gives each application a cycle time and a maximum jitter, but they are quoted in mixed units — milliseconds in one row, microseconds in the next — so what really separates them is hidden. Let me put every row in one currency: jitter as a fraction of its own cycle time.

  • Simple I/O: 1 ms / 10 ms = 1000 us / 10,000 us = 0.10 — 10%, or 1 part in 10.
  • Process control: 10 ms / 100 ms = 0.10 — 10%, 1 part in 10.
  • Packaging machinery: 100 us / 1000 us = 0.10 — 10%, 1 part in 10.
  • Printing press: 1 us / 125 us = 0.008 — 0.8%, 1 part in 125.
  • Semiconductor handling: 0.1 us / 62.5 us = 0.0016 — 0.16%, 1 part in 625.

Three of the five applications share one budget: jitter may be a full tenth of the cycle. The printing press then demands 0.10 / 0.008 = 12.5x tighter timing, and semiconductor handling 0.10 / 0.0016 = 62.5x tighter still.

The design meaning is that the raw cycle time already told you semiconductor handling was fast — 62.5 us — but converting jitter to a fraction of cycle exposes the harder constraint: its timing has to hold to 1 part in 625, not 1 part in 10. That fractional budget, not the absolute speed, is why the chapter routes only these applications to hardware synchronization such as EtherCAT distributed clocks or IEEE 1588, where a bounded clock matters more than a merely fast link.

Every number above is taken from the chapter’s own material and re-derived step by step.