Ratio Trimming: Choosing the Resistor and Trim Sequence

Map the sign of each available adjustment

For the ratio Q equal to R2 divided by R1, increasing R2 raises Q while increasing R1 lowers it. This sign rule is elementary but easily lost when a tester reports gain, attenuation or percent error instead of the resistor ratio itself. A passive divider output, an inverting amplifier gain and a bridge imbalance each need their own signed mapping.

Make a small adjustment map from every accessible resistor to every monitored output. Mark an element as unavailable if its trim region is covered, cannot be measured in circuit or has already reached its qualified geometry limit. The theoretical ability to increase a resistance does not establish a usable manufacturing adjustment. Probe access, cut quality and the surrounding circuit determine whether that degree of freedom is genuinely available.

Calculate the reachable ratio interval

If the current values are R1s and R2s, and each has a separately established maximum permitted value, the smallest reachable ratio uses the smallest numerator and largest denominator. The largest ratio uses the largest numerator and smallest denominator. A target outside this interval cannot be reached by the available upward adjustments, regardless of measurement precision.

For example, assume starting values of 9.8 kilohms and 10.1 kilohms and a target ratio of one. Increasing the 9.8 kilohm denominator to 10.1 kilohms can reach the target, provided that value is within its qualified trim window. Trimming the 10.1 kilohm numerator first moves the ratio farther away and forces still more adjustment in the denominator. The final ratio might remain reachable, but the sequence has wasted useful margin.

Preserve the absolute impedance window

A ratio of one can be formed by many absolute values. Choosing 10 kilohms or 20 kilohms changes divider current, output impedance, noise and power even though the ratio is unchanged. Before trimming, intersect the ratio-reachable interval with the absolute limits of both elements. A valid ratio solution outside either absolute window is not an acceptable network.

Where one element must meet a tight absolute target, it can be useful to establish that value first and then adjust the paired element for the ratio. However, the second element must begin on the correct side of the required endpoint. If starting distributions overlap unfavorably, choosing a fixed sequence for every unit may reject otherwise reachable networks. Evaluate whether the controller should select the direction from the measured starting pair.

Resolve shared-element dependencies before cutting

Consider three resistors with requirements Q21 equal to R2 divided by R1 and Q31 equal to R3 divided by R1. Changing the shared denominator moves both ratios in the same direction. If R2 is trimmed to satisfy Q21 and R1 is subsequently increased to help Q31, the first ratio is disturbed. This is a dependency problem, not evidence that the trim endpoint was inaccurate.

One useful approach is to position the common element inside a permitted absolute range before fine-adjusting the two numerators. Another is to solve all target values jointly and verify that every required change is nonnegative and physically reachable. The right approach depends on starting values and available access. Record the dependency structure explicitly so a later schematic revision does not retain an incompatible historical trim order.

Choose a sequence from the actual starting condition

The sequence should respond to measured values and circuit constraints rather than a universal rule to trim the largest or smallest resistor first. Use the table as a decision structure and calculate the full effect before each irreversible operation. When no valid solution remains, stop instead of continuing cuts that cannot restore the required relationships.

Decide between passive ratio measurement and active function trim

Passive ratio trimming uses electrical measurements of resistor relationships. Active trimming monitors the powered circuit function, such as an output voltage or gain. Active measurement can include the real receiver and other circuit effects, but it can also hide errors that depend on supply, temperature or operating state. A single calibrated output does not prove that each underlying resistor is within its permitted range.

Use a powered schematic and identify all conditions held constant during an active trim. Check whether amplifier offset, reference error or receiver loading is being absorbed into the resistor setting. After adjustment, test the other relevant states. Manufacturer matched-network documentation is useful for understanding circuit-level matching requirements, but a particular component's performance cannot establish the behavior of a custom printed array.

Reserve a controllable final approach

Coarse and fine stages serve different purposes. The coarse stage moves efficiently toward the feasible target region; the fine stage must respond to a small remaining error without overshoot. As a cut advances, the same additional length can produce a different resistance increment because the current path is changing. Characterize that sensitivity near the intended endpoint rather than assuming the first cut increment predicts the last.

Establish a stopping band that accounts for measurement uncertainty, response delay and the required post-trim condition. If the ratio lies inside the acceptance band, another cut made merely to center a display may worsen the outcome. Repeatedly approaching the target from alternating directions is impossible with an irreversible single-element cut; any apparent bidirectional controller behavior must actually come from changing different resistors or from unstable measurements.

Retain the network history needed to explain the result

Record starting values, selected adjustment order, final values and the affected circuit outputs. Include the sensing arrangement and the substrate condition at each reading. Inspect the remaining resistor geometry and repeat the electrical verification after the specified stabilization and subsequent processing. This connects the final network result to the physical path used to achieve it.

For a revised design, compare not only final ratio distributions but also the amount of adjustment and the frequency of unreachable starting combinations. A design requiring less extreme correction may offer more useful robustness even when both designs initially meet the same ratio tolerance. The RFQ should therefore communicate absolute windows, ratio functions and accessible trim regions together, allowing the adjustment problem to be evaluated before artwork is committed.