ctrl32

Maths

Maths blocks

Arithmetic, and the assignment block. Nothing surprising here — the value is in knowing what happens at the edges.

The arithmetic

Block Inputs Output Result
ADD IN1…IN4 OUT IN1 + IN2 + IN3 + IN4
SUB IN1, IN2 OUT IN1 − IN2
MUL IN1, IN2 OUT IN1 × IN2
DIV IN1, IN2 OUT IN1 ÷ IN2
MIN IN1, IN2 OUT the smaller of the two
MAX IN1, IN2 OUT the larger of the two
ABS IN OUT the value without its sign

ADD takes four inputs, and unconnected ones are 0 — so it doubles as a two- or three-input adder without any special handling.

Every other input is 0 when unconnected too, and that is worth a thought on MUL (result 0) and DIV (see below).

Division by zero gives zero

DIV with IN2 = 0 produces 0. It does not fault, does not produce infinity, and does not stop the machine.

This is a deliberate choice: a divisor that is briefly zero is normal on a real machine — a flow that has not started, a length not yet measured, a sensor still warming up. A controller that halts the cycle over it would be useless.

Zero is a plausible-looking wrong answer

The danger is not the crash that does not happen, it is the 0 that flows onward and looks like a real measurement. Where the result drives something that matters, check the divisor:

flow ──IN1─┐
time ──IN2─┤ DIV ├── OUT ──> rate
time ──[GT 0]── Q ──> rate_valid

and let rate_valid gate whatever uses rate.

MIN and MAX as guards

The obvious use is picking the larger of two measurements. The more common one is putting a floor or a ceiling on a value:

setpoint ──IN1─┐
      50 ──IN2─┤ MAX ├── OUT ──> setpoint_guarded

Now the setpoint can never go below 50, whatever a screen or a fieldbus writes into it.

For a floor and a ceiling in one block, use LIMIT instead — that is what it is for.

MOVE — assignment

OUT = IN. The IEC assignment, as a block.

Pin Type Meaning
IN input, number the source
OUT output, number the destination

It looks like it does nothing, and in a diagram it nearly does — its job is to get a value from one tag into another. Where it earns its keep is in a ladder rung, as the plain result box: while the rung conducts, the value is committed; while it does not, the destination holds.

recipe_ready ──[ MOVE IN=recipe_temp ]── OUT ──> setpoint

The setpoint takes the recipe value the moment the recipe is ready, and keeps it afterwards. That "commit a value when a condition is met" is one of the most common things a machine program does.

Numbers on ctrl32

Values are floating point, which means:

  • Fractions are fine. 0.1 °C, 2.5 mm, no scaling tricks needed.
  • Big integers lose precision. Past about 16 million, consecutive whole numbers stop being distinguishable. A part counter that must be exact into the millions belongs in an integer tag.
  • Equality is unreliable. 1.0 - 0.9 - 0.1 is not exactly 0. Never test two calculated values with =; use EQ with a tolerance, which exists for exactly this reason.

Order matters

Blocks execute in program order and pass values through tags, so a chain of calculations has to be written in the order it flows. A block reading a tag that a later block writes gets last cycle's value.

For a long calculation this is where Structured Text becomes the better tool: expressions evaluate in the order you write them, on one line, with no intermediate tags at all.

Tags