The Bench: DMM, Supply, and Generator

Your circuit needs eyes and hands

You have come a long way up this ladder: voltage dividers, biased transistors, op-amp stages, filters, oscillators, regulated supplies, even a sketched PCB. But every one of those lived somewhere transparent — in your head, on a breadboard you could poke, or inside a simulator where each node's voltage is simply a number printed on screen. The real world is opaque. Solder a board, switch it on, and it is a sealed black box: nothing glows, nothing speaks, and your beautiful design either works or sits there dead with no clue why. The bench is how you pry that box open. Its instruments give a circuit eyes — to see what its voltages and signals are doing — and hands — to feed it power and stimulus on demand. Without them, debugging is just guessing.

The bench divides cleanly into two camps. Sources push something in: the bench power supply delivers a controllable DC voltage with a built-in safety limit, and the function generator delivers a controllable repeating waveform. Measurers read something out: the digital multimeter (DMM) hands you a single trustworthy number, while the oscilloscope draws the actual shape of a signal over time. A few specialists wait in the wings for later jobs — the logic analyzer for watching many digital lines at once, the spectrum analyzer for seeing a signal's frequency content, the LCR meter for measuring a component's actual inductance, capacitance, and resistance. This guide meets the three you reach for ninety percent of the time: supply, generator, and DMM.

The DMM: the careful surveyor

You first met the multimeter back in the foundations rung; now let us go deeper. A DMM is a careful surveyor: it samples slowly — a few readings per second — and averages over that window to hand you one steady, trustworthy number. That patience is both its gift and its hard limit. It gives you precise DC and slowly-changing values, but it is essentially blind to anything fast: a brief glitch, the shape of a waveform, a ringing edge. Seeing those is the oscilloscope's job, not the meter's. What a DMM does measure: DC and AC voltage, resistance, current, plus handy extras — a continuity beeper, a diode test that reads a junction's forward voltage, and often capacitance and frequency.

Two golden habits separate a measurement from a mistake. Voltage is measured in parallel — you lay the two probes across the thing of interest, like reading water pressure across a valve without ever cutting the pipe. For that touch not to disturb the circuit, the meter must barely sip any current, so a voltmeter is built with a huge input resistance, typically 10 Mohm. Current is measured in series — you must break the loop and let the entire flow pass through the meter, like splicing a flow gauge into the pipe; an ammeter is therefore built to be nearly zero ohms. This is why placing a low-resistance ammeter across a voltage is a dead short and a blown fuse, and why even a 10 Mohm voltmeter still slightly loads a very high-impedance node — the full loading story is guide 3.

  MEASURE VOLTAGE: meter in PARALLEL          MEASURE CURRENT: meter in SERIES
  (touch across -- break nothing)             (cut the loop -- all flow goes through)

     +9V                                          +9V
      |                                            |
    [R1] 10k                                     [R1] 10k
      |                                            |
      +------(V)  reads 3.00 V                     +--/ /--(A)  reads the loop current
      |       ~10 Mohm, sips ~no current         cut here    ~0 ohm, must be in-line
    [R2] 5k                                      [R2] 5k
      |                                            |
     GND                                          GND

   3.00 V = 9 x 5 / (10 + 5)        |   an ammeter (~0 ohm) ACROSS a voltage = dead short -> blown fuse

Read the meter's own honesty too. A '3.5-digit' DMM shows up to 1999 'counts', and its accuracy is quoted as a percentage of the reading plus a few counts of fixed error — say ±(0.5% of reading + 2 counts). Read 5.00 V on that meter and the true value lies within 0.5% of 5 V (25 mV) plus 2 counts (20 mV), so anywhere from about 4.955 V to 5.045 V; do not trust the last digit blindly. One more trap on AC: the RMS value is neither the average nor the peak. A cheap meter secretly assumes a clean sine and only scales the average, so it mis-reads anything else; a true-RMS meter computes the real heating value and reads any waveform honestly. Match the meter to the wave.

The bench supply: a source with a safety net

The bench power supply is the clean, adjustable DC source your whole circuit drinks from — essentially the regulated supply from the power rung, but with front-panel knobs and one feature that will save you more smoked boards than any other: a current limit. It has two controls, voltage and current, and two matching displays so you can always see both what it is trying to give and what the circuit is actually drawing.

The killer feature is constant-voltage / constant-current (CV/CC) crossover. Normally the supply holds the voltage you dialed and pours out whatever current the load asks for — that is CV mode. But you also set a current ceiling. The instant the load tries to pull more than that ceiling — a short, a backwards chip, a slipped probe — the supply gives up holding voltage and instead holds the current pinned at the limit, letting the output voltage sag as far down as it must. That is CC mode, and in it the supply behaves like a current source. It is really a feedback loop, exactly like a thermostat: negative feedback watches both rules at once, and whichever one is about to be broken takes command. Set the limit to just above expected draw before you connect, and a wiring blunder becomes a glowing 'CC' light instead of a puff of smoke.

1. With the output OFF, dial the voltage knob to your target (say 5.0 V) and read it back on the display.
2. Briefly short the output leads together with a clip; the supply jumps into CC and now the current display shows the limit you are about to set.
3. Turn the current knob down to just above your expected draw — for a small logic board pulling ~60 mA, set perhaps 100 mA.
4. Remove the short, connect your circuit, switch the output on, and watch the current meter as it powers up.
5. If the supply slams into CC and the voltage collapses, stop — something is shorted or miswired, and the limit just saved your board.

The function generator: a controllable signal

The function generator is the supply's restless cousin: instead of steady DC it produces a repeating waveform you fully control — most often a sine, a square, or a triangle. You dial its frequency (from a fraction of a hertz up to many MHz), its amplitude (usually quoted as a peak-to-peak voltage), and a DC offset that slides the whole wave up or down. It is how you ask a circuit a live question: 'what do you do with this signal?' Feed a filter a slow frequency sweep and watch what survives; ring an amplifier and see if it rings cleanly; clock a digital input at a known rate.

Now an honest trap that bites everyone once. A generator is not an ideal source; it has a built-in source resistance, almost always 50 ohm. Its front panel can be told to assume either a 50-ohm load (the convention for impedance matching with RF gear) or a high-impedance ('Hi-Z') load. Here is the bite: if the generator is set for '50 ohm' but you actually connect it to a high-impedance input — an op-amp, a scope, a breadboard node — you get double the amplitude you dialed. The instrument was quietly halving its internal voltage expecting a matched 50-ohm load to take the other half; your near-infinite input takes nothing, so the full, doubled swing appears. Fix it by setting the output mode to match reality (usually Hi-Z on the bench), or simply measure the real amplitude before trusting the dial.

Put it to work on a circuit you already know. Take the RC low-pass with R = 1.6 k and C = 100 nF, whose cutoff is f = 1 / (2 times pi times R times C) ≈ 1 kHz. Set the generator to a 1 Vpp sine and sweep slowly upward: below 1 kHz the output comes through nearly full size, at 1 kHz it has fallen to about 0.707 of the input (the −3 dB point), and above it the output keeps dropping by 20 dB for every tenfold rise in frequency. Point by point, you have traced the filter's Bode plot by hand — a measured curve to compare against the low-pass filter theory you learned earlier. (Watching that whole curve live on a screen, rather than one point at a time, is exactly what the oscilloscope in guide 2 makes effortless.)

Putting the bench to work — and sharing a ground

Now watch the three instruments cooperate on a real bring-up. Power the board from the current-limited supply and ease the output on, eyes on the current meter for any unexpected surge. With the board alive, walk the DMM around the key nodes and check each DC operating point against what your schematic and simulator predicted — is the regulator output where it should be, is the transistor's bias correct, is every supply pin actually at the rail. Only once the DC story checks out do you inject a signal: clip the generator to the input and follow the waveform through the circuit. That deliberate order — power first, then bias, then signal — is the backbone of the systematic debugging you will build in guide 4.

These three instruments are only ever as good as how you connect them — and connecting them is itself a circuit, because the act of measuring current breaks a loop, and the act of touching a probe to a node adds load. The next guide takes apart the most capable tool on the bench, the oscilloscope, and the surprisingly subtle x10 probe that tames it. Then guide 3 turns squarely to that uncomfortable truth about probe loading: that every measurement disturbs the thing it measures, and the craft is in disturbing it as little as you can. Learn this bench, and a dead board stops being a mystery and becomes a puzzle you can actually solve.