Many SCADA and PLC systems do not ask you for endpoint ranges when scaling a signal; they ask for just two numbers, a scaling factor and an offset. Those two numbers are the slope and intercept of the straight line that turns a raw value into a usable one, and knowing how to compute and check them is a core skill. This guide defines the scaling factor as the gain or slope and the offset as the bias or intercept, shows how to derive both from known endpoints, and walks through the sign conventions and common mistakes that make offsets a frequent source of quiet errors.
Scaling Factor and Offset in one line: A scaling factor and offset are the two constants that convert a raw signal into a usable value using a simple linear equation: multiply the raw value by the scaling factor, then add the offset. The scaling factor, also called gain or slope, sets how much the output changes per unit of input, and the offset, also called bias or intercept, shifts the whole line up or down to fix where it starts.
Any straight-line conversion can be written as an output equal to the input multiplied by one number and then added to another. The multiplier is the scaling factor, and the adder is the offset. In algebra this is the familiar slope-intercept form, where the scaling factor plays the role of the slope and the offset plays the role of the intercept. Every linear scaling task, no matter how it is presented in a given system, reduces to finding those two values, so understanding them lets you read and set up scaling in almost any tool.
The scaling factor, often called gain, answers how much the output moves for each step of the input. A larger scaling factor stretches the input's range into a wider output range; a smaller one compresses it. If the input and output happen to share the same span, the factor is one and the multiplication does nothing, but in most real conversions the raw span and the engineering span differ, so the factor scales between them. It is the number that gives the conversion its steepness.
The offset, often called bias, answers where the line sits vertically. It shifts every output by the same amount, independent of the input. If a raw zero should map to a nonzero engineering value, or a raw value that is not zero should map to engineering zero, the offset is what makes that happen. Because it adds equally to every reading, an offset error does not distort the shape of a trend; it slides the whole thing up or down by a fixed amount, which is exactly why offset mistakes can be easy to overlook.
You rarely know the scaling factor and offset directly, but you almost always know two endpoints, the same anchor pairs used in two-point scaling. The scaling factor is the change in output divided by the change in input between those two points: take the difference of the two engineering values and divide it by the difference of the two raw values. That ratio is how much engineering value each unit of raw value is worth, which is precisely the slope.
With the factor in hand, the offset falls out by working backward from either known point. Multiply that point's raw value by the scaling factor, then subtract the result from that point's engineering value; what remains is the offset. Doing this with either endpoint gives the same offset if the arithmetic is right, which is a handy self-check. Once you have both numbers, the conversion is fixed, and any raw value can be scaled by multiplying and adding, no endpoints required at run time.
A worked shape makes it concrete. Suppose a raw range of 0 to 32000 counts should map to 0 to 100 percent. The output span of 100 over the input span of 32000 gives the scaling factor, a small fraction, and because a raw zero maps to zero percent the offset is zero. Change the mapping so raw zero should read as 50 percent instead, and the factor stays the same while the offset becomes 50, lifting the entire line. Seeing how the two numbers respond, one to the ratio of spans and one to the vertical position, is the heart of using them confidently.
Signs are the first thing to get right. A scaling factor is normally positive, but for an inverted signal, where the engineering value falls as the raw value rises, the factor is negative, and the line slopes downward. The offset can be positive or negative depending on where the line needs to sit. The units of the offset are the units of the output, not the input, which sounds obvious but is a frequent trip-up: adding an offset expressed in raw counts to a value already in engineering units produces nonsense. The offset must live in the same units as the thing it is being added to.
The most damaging mistake in a data pipeline is double-scaling, applying a scaling factor and offset twice to the same signal. This happens when a value is scaled in the PLC and then scaled again in the SCADA or historian by someone who assumed it was still raw. The result is wrong in a way that can be hard to spot, because it still moves with the real measurement. The defense is knowing, for every point, exactly where scaling is applied and applying it once. In a cloud SCADA such as Merobix, part of setting up a point is establishing whether the device already delivers engineering units or a raw value that still needs the factor and offset, so the conversion is neither skipped nor duplicated.
Other common errors are subtler versions of the same theme. A correct scaling factor with a wrong offset yields readings that track changes perfectly but are uniformly high or low, which is why an offset error can hide behind a trend that looks entirely healthy. Confusing gain and offset, or entering one in the other's field, breaks the conversion outright. The reliable check is the same as for all scaling: verify a converted reading against a known physical condition. If the value responds correctly to changes but sits at the wrong absolute level, suspect the offset; if it responds by the wrong amount, suspect the factor.
The scaling factor, or gain, is the multiplier that sets how much the output changes for each unit of input, giving the conversion its slope. The offset, or bias, is the adder that shifts the whole line up or down, setting where it starts. In a linear conversion you multiply the raw value by the factor and then add the offset, so the factor controls steepness and the offset controls vertical position.
Take the difference between the two known engineering values and divide it by the difference between the two known raw values; that ratio is the scaling factor. To get the offset, multiply either point's raw value by the factor and subtract that from the same point's engineering value. Doing the offset calculation with both points is a good check, since it should give the same result either way.
Double-scaling is applying a scaling factor and offset twice to the same signal, usually once in the PLC and again in the SCADA or historian, which produces a wrong value that still moves with the measurement and so is easy to miss. You avoid it by knowing exactly where scaling is applied for each point and applying it only once. Confirming whether a device sends raw counts or already-scaled engineering units is the key step.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.