Lab 4.1 — MCP6002 Voltage Follower
← Course 2 syllabus · Module 4 · Prev: « Lab 3.5 · Next: Lab 4.2 »
Goal
Build the simplest and most useful op-amp circuit — the unity-gain buffer (voltage follower) — and use it to understand why buffering matters. You will bring up the MCP6002 on a single supply (correct pinout, decoupling, mid-rail sense of “ground”), then prove the buffer’s value empirically: drive a load from a high-impedance resistive divider directly and watch the voltage sag, then insert the follower and watch the sag disappear. The buffer’s job is to convert a fragile, high-source-impedance node into a stiff, low-output-impedance one — exactly what you must do before an ADC input, a filter stage, or any load that would otherwise disturb the node you’re measuring. That impedance-decoupling instinct is foundational for every mixed-signal front end you’ll build in Modules 5–6.
Recommended reading
- O-S&S Ch. 1 — signals and systems, and the idea of an ideal system that passes its input unchanged (unity gain, the follower’s mathematical identity). Focus on system properties: linearity, memorylessness, and the notion of loading. → Oppenheim S&S
- The MCP6002 datasheet — read the pinout, the absolute-maximum ratings, and the input common-mode / output-swing (rail-to-rail) sections.
- PEI Ch. 8 — Operational Amplifiers: the ideal op-amp, the voltage follower/buffer, and input/output impedance.
- Course 1: Week 1–3 (linear algebra / least-squares) for the Thévenin-equivalent view of a source impedance; the follower is the physical realization of “make the source impedance ≈ 0.”
Equipment & parts
- WANPTEK 30 V / 10 A supply (used at 5.0 V, current-limited).
- MCP6002 dual op-amp, DIP-8.
- Fluke 117 DMM.
- Siglent SDS1104X-E oscilloscope + 10× probe (optional, to see AC follow-through).
- Resistors from the kit: two 100 kΩ (the high-impedance divider), one 1 kΩ and one 10 kΩ (loads), plus one 100 Ω.
- One 0.1 µF ceramic decoupling capacitor.
- Breadboard + jumper wires.
- Signal source (for the AC check): the MCP4725 DAC from Lab 3.3, or the scope’s ~1 kHz probe-compensation square wave.
Safety & don’t-break-it
- ESD is the real hazard here. The MCP6002 is a bare CMOS DIP and is not ESD-rugged. Before touching it: power off, touch a grounded metal object (or wear a wrist strap), and handle the chip by the plastic body, never by the pins. Insert and remove it only with the supply off. A static zap you can’t feel can silently degrade an input.
- Seat the DIP correctly. Straddle the breadboard center channel, confirm the pin-1 notch/dot orientation before power. Pin 1 = OUTA is at the notch end; getting this backward swaps VDD (pin 8) and VSS (pin 4) and can destroy the part.
- Respect the input common-mode and supply limits. Inputs must stay within roughly (VSS − 0.3 V) to (VDD + 0.3 V). On a 0–5 V single supply, never drive an input negative or above 5 V.
- Decouple before you power. Place the 0.1 µF cap from VDD (pin 8) to VSS (pin 4) right at the chip before applying the rail; without it the op-amp can oscillate.
- Current limit ~100 mA on the WANPTEK, set before connecting (Lab 0.1 habit).
Background
An ideal op-amp has infinite open-loop gain \(A\), infinite input impedance, and zero output impedance. Wire the output straight back to the inverting input (100% negative feedback) and the closed-loop relationship is
\[ V_\text{out} = A\,(V_+ - V_-) = A\,(V_\text{in} - V_\text{out}) \quad\Longrightarrow\quad V_\text{out} = \frac{A}{1+A}\,V_\text{in} \xrightarrow{A \to \infty} V_\text{in}. \]
So the follower’s ideal transfer function is unity: \(V_\text{out} = V_\text{in}\). It has no voltage gain — its entire value is impedance transformation. Feedback divides the op-amp’s already-huge input impedance up and its output impedance down by the loop gain \((1+A)\), so the input looks nearly open-circuit and the output looks nearly like a stiff voltage source.
Why that matters — the loading problem. A source with open-circuit voltage \(V_s\) and internal (Thévenin) resistance \(R_s\) driving a load \(R_L\) delivers only
\[ V_L = V_s\,\frac{R_L}{R_s + R_L}. \]
When \(R_s\) is comparable to (or larger than) \(R_L\), the node sags badly. Take two 100 kΩ resistors dividing 5 V: the open-circuit midpoint is 2.5 V, but the Thévenin source resistance seen at that node is \(R_s = 100\text{k} \parallel 100\text{k} = 50\text{ k}\Omega\). Hang a 10 kΩ load on it directly and
\[ V_L = 2.5\text{ V}\cdot\frac{10\text{k}}{50\text{k}+10\text{k}} = 2.5\cdot\tfrac{10}{60} \approx 0.42\text{ V}. \]
That is a collapse to about 0.42 V, an 83% error. A follower inserted between the divider and the load draws essentially no current from the 50 kΩ node (so the divider stays at 2.5 V) and sources the load current from its low-impedance output, holding \(V_L \approx 2.5\) V. That single demonstration is the whole point of the lab.
Procedure
Part A — Bring up the MCP6002 on a single supply.
- Supply off, current limit ~100 mA. Place the MCP6002 across the breadboard center channel, pin-1 notch confirmed.
- Wire VDD (pin 8) → +5 V rail and VSS (pin 4) → ground rail. This is a single-supply op-amp: “ground” is 0 V and the output can only swing between 0 and ~5 V.
- Place the 0.1 µF cap directly between pin 8 and pin 4.
- We will use amplifier A: OUTA = pin 1, IN A− = pin 2, IN A+ = pin 3.
Part B — Baseline the loading problem (no buffer).
- Build the divider: +5 V → 100 kΩ → node M → 100 kΩ → ground. With the Fluke on DC volts, measure M unloaded: expect ≈ 2.50 V.
- Now hang a 10 kΩ load from M to ground. Re-measure M. Expect it to collapse to ≈ 0.42 V (per the Background). Record this — it is the “before.”
- Remove the 10 kΩ load; confirm M springs back to 2.5 V.
Part C — Insert the follower.
- Wire the follower: node M → IN A+ (pin 3). Connect OUTA (pin 1) directly to IN A− (pin 2) — that wire is the feedback.
- Power on. Measure OUTA (pin 1) with the Fluke: it should read ≈ 2.50 V, equal to the unloaded divider — the follower copies M.
- Now hang the 10 kΩ load from OUTA to ground. Re-measure both M and OUTA. Critically: M stays at 2.5 V (the follower isn’t loading it) and OUTA stays ≈ 2.5 V (the follower sources the load current). Record this “after.”
Part D — AC follow-through (optional, with a source).
- Replace the divider with an AC source: either program the MCP4725 (Lab 3.3) to output a low-frequency sine centered near mid-rail, or feed the scope’s ~1 kHz probe-comp square wave into IN A+ through a series 100 kΩ (making it deliberately high-impedance).
- Put scope CH1 on IN A+ and CH2 on OUTA. With no load the two traces overlap. Add the 10 kΩ load to OUTA: CH1 (the high-Z input) barely moves while CH2 stays with it — visually, the buffer holds the loaded output up.
Deliverable & expected results
A bench note (docs/lab-4-1.md) with the before/after divider measurements and a one-line statement of what the follower did to the source impedance.
| Quantity | Predicted | Measured |
|---|---|---|
| Divider midpoint M, unloaded | 2.50 V | … |
| M with 10 kΩ load, no buffer | 0.42 V | … |
| OUTA with 10 kΩ load, buffer inserted | ≈ 2.50 V | … |
| M with 10 kΩ load, buffer inserted | ≈ 2.50 V | … |
| Follower voltage gain | 1.00 (0 dB) | … |
Thévenin source resistance of the divider: \(R_s = 100\text{k}\parallel 100\text{k} = 50\text{ k}\Omega\).
Analysis & reconciliation
Compute the loaded divider voltage \(V_L = 2.5\cdot R_L/(R_s+R_L)\) by hand and compare to the “no buffer” measurement — they should agree within resistor tolerance (±5%). Then confirm the buffered case: the divider node should be within a few millivolts of its unloaded value, proving the follower drew negligible current. Any residual droop on OUTA under load reflects the op-amp’s finite output impedance divided by loop gain plus its output-stage current limit — the MCP6002 can source only a few mA before its rail-to-rail output starts to fall short of the rail. If OUTA can’t reach exactly 2.5 V, check whether you’re near the output current limit (try a 100 kΩ load instead of 10 kΩ and the residual error should shrink). Note also that “rail-to-rail” is near-rail: the output typically gets to within tens of millivolts of VDD/VSS, not exactly to them.
Going further
- Measure the output impedance directly: apply two different loads (e.g. 10 kΩ then 1 kΩ) and compute \(R_\text{out} = \Delta V / \Delta I\) from the small change in OUTA. You should get a fraction of an ohm.
- Use the second amplifier (B: pins 5/6/7) to buffer a second divider — practice with both halves of the dual package.
- Push the input toward the rails (divider set near 0.1 V and near 4.9 V) and find where the output stops tracking — this maps the input common-mode range and previews the swing limits you’ll exploit deliberately in Lab 4.3.
- This buffer is the standard front end for the ADS1115 and STM32 ADC inputs in Module 5 — a high-impedance sensor should always see a follower before the converter.