ISA-88 defines four recipe types that carry a product definition from the corporate level down to a single running batch. The most confused pair is the master recipe and the control recipe, because they look nearly identical yet play very different roles. A master recipe is the approved, equipment-specific template you validate and store; a control recipe is the disposable copy created for one batch and actually executed by the sequencer. Understanding how instantiation converts one into the other is the key to safe, repeatable batch production.
control recipe vs master recipe in one line: A master recipe is the reusable, equipment-specific but not-yet-released template for making a product, held in a recipe library. A control recipe is a copy of that master recipe created for a single batch, populated with a batch ID, a target quantity, scaled formula values, and bound equipment, and it is the thing the batch sequencer runs and records.
ISA-88 arranges recipes into a descending chain of specificity. The general recipe is site-independent and equipment-independent: it states what a product is in generic terms, such as ingredients and process actions, without naming any plant. The site recipe adapts that general recipe to a specific location, substituting local raw materials, units of measure, and language while still not committing to particular equipment.
The master recipe is the next step down and the first that is tied to real equipment. It targets a class of equipment in one process cell, carries the full procedure and formula, and is the version that gets reviewed, approved, and version-controlled. Nothing runs a master recipe directly. It is a template, deliberately held one remove from execution so that approvals stay stable.
The control recipe sits at the bottom. It is generated from a master recipe at the moment a batch is scheduled or started, and it is unique to that one batch. Every value a master recipe left generic gets resolved here: the operator enters or confirms the batch size, the system scales the formula, and specific tanks and reactors are allocated. When the batch finishes, the control recipe becomes part of the historical record and is not reused.
Instantiation is the act of copying a master recipe into a fresh control recipe. The batch engine reads the approved master, stamps the new instance with a unique batch ID and order number, and hands it to the sequencer. From that instant the two objects are independent: the master keeps sitting in the library unchanged, and the control recipe belongs to the batch alone.
Scaling adjusts the formula quantities so the recipe fits the ordered batch size rather than the master's nominal size. If the master is written for a 1000 kg batch and the order is for 1500 kg, scalable charges multiply by 1.5 while fixed parameters, such as a catalyst dose or a jacket temperature setpoint, stay put. Scaling happens as part of instantiation so the control recipe carries the real numbers the sequencer will command.
Equipment binding resolves which physical units the batch will use. The master recipe references an equipment class or requirements; the control recipe names the actual reactor, the actual charge line, and the actual transfer path selected by the scheduler or the operator. Once a control recipe is bound and scaled, it is fully concrete: every phase knows its equipment, every setpoint has a number, and the sequencer can execute without any further interpretation.
During a live batch, operators frequently adjust a control recipe. They may change a hold time to recover from an upset, override a setpoint after a quality reading, or accept a prompt that alters a charge. These edits are legitimate and are captured in the batch record, but they apply only to that one batch. If they silently propagated back into the master recipe, the next batch would inherit an unreviewed change, and the whole point of an approved template would collapse.
This is why batch systems treat the master recipe as read-only during execution and route any changes to the control recipe copy. A genuine improvement to the product definition should go through a controlled change: edit the master recipe, put it through review and approval, and version it. The transient control recipe is the correct place for one-off, batch-specific deviations; the master recipe is the correct place for durable, governed changes.
Keeping the boundary sharp also protects traceability. Because each control recipe records the master version it was instantiated from plus any deviations applied, an investigator can always answer two separate questions: what did the approved recipe say, and what actually happened on this batch. Blurring control-recipe edits into the master would erase that distinction and make it impossible to prove which batches ran the validated procedure.
No. A control recipe is created for exactly one batch and carries that batch's unique ID, scaled quantities, and bound equipment. Two batches of the same product are each given their own control recipe instantiated from the same master recipe, so each one records its own deviations and results independently.
No. The master recipe is a template that must first be instantiated into a control recipe before anything runs. The sequencer only ever executes a control recipe, because only the control recipe has a concrete batch size, resolved setpoints, and specific allocated equipment.
The general recipe is the most abstract, equipment-independent and site-independent product definition, and the site recipe adapts it to one location's materials and standards. Both sit above the master recipe. The master recipe is the first level bound to real equipment, and the control recipe is the instantiated copy below it that actually runs a batch.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.