A 4k7 resistor (4.7kΩ) sits at the sweet spot for pull-ups, voltage dividers, and signal conditioning - strong enough to hold logic levels stable without bleeding power, but not so high that it slows down switching or starves analog inputs. Its ubiquity in the E-series makes it the go-to value when you need predictable impedance without overthinking tolerances or sourcing headaches. Below, I'll walk through its color code, real-world applications, power considerations, and how to replace it when the exact value isn't in your bin.
Key Takeaways:
- 4k7 (4.7kΩ) balances signal speed and power draw better than 10k in pull-ups, especially for fast I²C or GPIO switching.
- Standard tolerance is ±5% (carbon film), but ±1% metal film costs only pennies more and eliminates drift in precision voltage dividers.
- Power rating matters more than you'd expect: a ¼W resistor dissipates only 21 mW at 10V, but pack four in a tight cluster on a PCB and you've quadrupled local heating.
- Series combinations (2.2k + 2.5k) work in a pinch, but parallel substitutions (two 9.1k resistors) require verification that your circuit tolerates the 3% deviation.
- The 4-band color code (Yellow-Violet-Red-Gold) is universal, but 5-band precision versions swap the gold for brown (±1%) and add a third digit band for tighter manufacturing control.
What "4k7" Actually Means
4k7 is shorthand for 4.7 kilo-ohms, or 4,700Ω. The "k" replaces the decimal point - a convention borrowed from European component marking to avoid confusion when decimal dots fade on schematics or get lost in photocopies. You'll see the same notation for other values: 1k2 (1.2kΩ), 10k (10kΩ), 100R (100Ω). It's not marketing; it's anti-ambiguity.
The 4.7kΩ value belongs to the E12 and E24 resistor series, which means every supplier stocks it, every automated pick-and-place machine has a reel loaded, and every hobbyist kit includes a handful. That ubiquity translates to lower cost, faster sourcing, and fewer PCB redesigns when you need to swap a part.
Reading the Color Code: Yellow-Violet-Red-Gold
A standard through-hole 4k7 resistor uses four color bands:
- Yellow (4) - first digit
- Violet (7) - second digit
- Red (×100) - multiplier
- Gold (±5%) - tolerance
Formula: 47 × 100 = 4,700Ω.
The gold band tells you the resistor can legally measure anywhere from 4,465Ω to 4,935Ω and still meet spec. In practice, modern carbon-film parts cluster near the nominal value; I've measured hundreds of ±5% resistors and rarely seen one drift beyond ±3% fresh out of the tape. But heat, humidity, and soldering stress will nudge that number over time, so don't assume gold means "good enough forever."
Precision versions swap gold for brown (±1%) and add a fifth band to encode three significant digits instead of two. For a 4.7kΩ ±1% resistor, you'd see Yellow-Violet-Black-Brown-Brown: 470 × 10 = 4,700Ω, ±1%. That tighter window (4,653Ω to 4,747Ω) matters in op-amp feedback networks, reference dividers, and anywhere you're counting on exact ratios.
Tolerance Classes and When They Matter
Tolerance isn't just a datasheet footnote - it defines whether your circuit drifts, oscillates, or fails calibration after a year in the field.
| Tolerance | Range (4.7kΩ nominal) | Typical Material | When to Use |
|---|---|---|---|
| ±1% | 4.653-4.747kΩ | Metal film | Precision voltage dividers, sensor conditioning, audio preamps |
| ±5% | 4.465-4.935kΩ | Carbon film | Pull-ups, pull-downs, general digital logic, LED current limiting |
| ±10% | 4.23-5.17kΩ | Older carbon comp | Legacy repairs, non-critical hobby projects |
A 5% resistor in a 1:1 voltage divider can shift your midpoint by 10% if both parts sit at opposite tolerance extremes. A 1% pair keeps that error under 2%. The price delta is negligible (often $0.01 vs. $0.02 per unit), so I default to 1% metal film unless I'm buying in five-figure volumes and the circuit truly doesn't care.
One thing datasheets don't shout about: tolerance tightens or loosens with temperature. A ±5% carbon film resistor might hold ±5% at 25°C but wander to ±8% at 85°C. Metal film parts spec a temperature coefficient (TCR) of ±50 to ±100 ppm/°C, meaning a 4.7kΩ resistor shifts roughly 2-5Ω per degree Celsius. If your board sees 40°C swings and you need 0.1% accuracy, you're looking at thin-film SMD parts with ±25 ppm/°C TCR, not standard through-hole.
Where You'll Actually Use a 4k7 Resistor
Pull-Up and Pull-Down Resistors
A 4.7kΩ pull-up holds a GPIO pin at logic HIGH when nothing else is driving it, preventing the floating-input lottery that leads to random resets, phantom interrupts, or power spikes. Why 4.7k instead of 10k? Speed. A 10kΩ pull-up into 10 pF of pin capacitance gives you an RC time constant of 100 ns; 4.7kΩ cuts that to 47 ns. For I²C running at 400 kHz, that difference shows up as sluggish rising edges and failed ACKs. For a button with 50 ms debounce, it doesn't matter.
Downside: Lower resistance burns more idle current. A 5V pull-up through 4.7kΩ wastes 1.06 mA continuously. Multiply by eight GPIOs and you've added 8.5 mA to your sleep budget. Battery-powered designs sometimes compromise at 47kΩ or use internal weak pull-ups (typically 30-50kΩ) to keep quiescent draw under 10 µA.
Voltage Dividers
A 4.7kΩ + 4.7kΩ divider splits an input voltage exactly in half - assuming the load impedance is at least 100× higher. Connect that divider to a microcontroller ADC with 10 MΩ input impedance and you're fine. Feed it into an op-amp with 1 MΩ input and you've loaded the divider enough to shift the midpoint by 1%. Feed it into a 10kΩ load and your math is off by 30%.
I've seen engineers reach for 4.7k resistors in dividers because "they're on the BOM already," then wonder why their 3.3V rail drifts under load. The fix is to recalculate for actual load current or drop to 1kΩ resistors if you're driving something real. Conversely, if you're dividing a high-impedance sensor output for an ADC, using 47kΩ or 100kΩ resistors instead saves power and reduces thermal noise contribution.
Analog Signal Conditioning
In op-amp feedback networks, 4.7kΩ shows up as a gain-setting resistor because it sits in the middle of the "don't fight the output stage, don't pick up hum" zone. Too low (e.g., 100Ω) and the op-amp output has to source tens of milliamps, heating up and distorting. Too high (e.g., 1 MΩ) and parasitic capacitance turns your clean amplifier into an accidental integrator, and every PCB trace becomes an antenna.
Paired with a 100 nF capacitor, a 4.7kΩ resistor makes a single-pole RC low-pass filter with a cutoff around 340 Hz - useful for cleaning up PWM ripple before it hits an ADC, or attenuating 50/60 Hz mains hum in sensor lines.
Current Limiting (Low-Power LEDs and Pin Protection)
A 4.7kΩ resistor limits current to roughly 1 mA when dropping 5V. That's enough to light a modern high-efficiency LED (think 3-5 mcd at 1 mA) as a dim indicator, but nowhere near the 10-20 mA you'd need for a bright panel LED. I use 4.7k resistors in status LEDs that only need to be visible in low ambient light, or as series protection on ADC inputs where I want to absorb voltage spikes without clamping diodes turning on.
Reality check: If you're driving a standard 20 mA LED from 5V, you need more like 150-220Ω, not 4.7kΩ. But for logic-level translation buffers, UART RX lines, or anywhere a 1-2 mA current limit prevents damage without killing functionality, 4.7k is exactly right.
Timing Circuits (555, RC Oscillators)
In a 555 timer astable circuit, the charge/discharge resistor sets frequency. With a 4.7kΩ resistor and a 100 nF capacitor, you get roughly 1 kHz. Swap in 47kΩ and you drop to 100 Hz; go down to 470Ω and you're at 10 kHz. The value is predictable, the part is cheap, and the tolerance error (±5%) translates directly to ±5% frequency error - acceptable for PWM dimming, debounce timers, or clock generation where crystal precision isn't required.
Types of 4k7 Resistors and When Each Matters
| Type | Construction | Tolerance | Noise | TCR (ppm/°C) | Power Typical | Best For |
|---|---|---|---|---|---|---|
| Carbon film | Carbon layer on ceramic | ±5% | Moderate | ±200 to ±500 | ¼W, ½W | General logic, pull-ups, hobbyist builds |
| Metal film | Thin metal alloy | ±1% | Low | ±50 to ±100 | ¼W, ½W | Precision dividers, sensor interfaces, audio |
| Wire wound | Resistive wire coil | ±1% or better | Very low (but inductive) | ±20 | 1W-10W+ | Power supplies, load testing (avoid RF) |
| Thick film SMD | Screen-printed paste | ±1% to ±5% | Moderate | ±100 to ±200 | 1/16W-¼W | Consumer electronics, automated assembly |
| Thin film SMD | Sputtered metal | ±0.1% to ±1% | Very low | ±25 to ±50 | 1/16W-⅛W | Precision instrumentation, RF matching |
Practical note: Wire-wound resistors have inductance - typically a few tens of nanohenries - which makes them useless above a few hundred kilohertz. I once debugged an oscillator that wouldn't start above 500 kHz because someone swapped a metal-film timing resistor for a wire-wound "precision" part. The inductance formed an LC tank that killed the feedback. Lesson: precision doesn't mean universally better.
Power Rating: The Forgotten Design Constraint
Power dissipation for a 4.7kΩ resistor is calculated as P = V² / R or P = I² × R.
Example: 10V across 4.7kΩ dissipates (10² / 4700) = 0.021W, well under a ¼W (0.25W) resistor's rating.
But: Four 4.7kΩ resistors in a tight row, each dissipating 20 mW, collectively dump 80 mW into a 1 cm² PCB area. If there's no copper pour for heat spreading, that cluster can hit 50-60°C above ambient, enough to shift resistance by 1% and cook nearby electrolytic caps.
Derating rule: Don't run resistors above 70% of rated power in production. A ¼W resistor should stay under 0.175W continuous, and if the ambient is 50°C (inside a sealed enclosure), derate further or jump to ½W.
I've seen designs where a 4.7kΩ resistor failed not because it exceeded ¼W on paper, but because it sat between two MOSFETs that each radiated 0.5W. The resistor cooked to 120°C, drifted to 5.8kΩ, and the voltage divider it fed mis-calibrated the control loop. Moving the resistor 5 mm away solved it.
Finding a Replacement: Exact Match vs. Close Enough
| Scenario | Solution | Notes |
|---|---|---|
| Exact 4.7kΩ available | Use it | Match or exceed original power rating and tolerance |
| No 4.7kΩ, have 2.2k + 2.7k | Series = 4.9kΩ | ~4% high; fine for pull-ups, marginal for precision dividers |
| No 4.7kΩ, have two 9.1kΩ | Parallel = 4.55kΩ | ~3% low; acceptable if load impedance is 100× higher |
| No 4.7kΩ, have 10kΩ | Check circuit type | Works for slow pull-ups, fails for I²C or fast switching |
| SMD replacement | Match footprint (0603, 0805, 1206) | Wrong size won't solder reliably |
Critical mistake to avoid: Never replace a ½W resistor with a ¼W part, even if the nominal resistance matches. The lower-power part will overheat, drift, or crack the solder joint, and you'll spend hours chasing an intermittent fault.
4-Band vs. 5-Band: What You're Really Paying For
- 4-band (Yellow-Violet-Red-Gold): Two significant figures, ±5%, carbon film. Costs $0.01-0.02. Good enough for 95% of hobbyist and general-purpose industrial circuits.
- 5-band (Yellow-Violet-Black-Brown-Brown): Three significant figures, ±1%, metal film. Costs $0.02-0.03. Necessary when tolerance stacking or temperature drift will break your spec.
The third digit doesn't make the resistor "more accurate" in absolute terms - it means the manufacturer measured and binned it to a tighter distribution. You're paying for QC, not magic.
Common Pitfalls I've Seen (and Made)
Using 4.7k as a one-size-fits-all pull-up: Fast I²C needs it. Slow button inputs would be fine with 47k and waste less power. Match the resistor to the rise-time and current budget, not to what's already on the BOM.
Ignoring ESR in timing circuits: Carbon film resistors have 1-5% voltage coefficient - resistance changes slightly under high voltage. In a precision RC timer running near a resistor's voltage rating, that can shift frequency by 2-3%. Metal film stays flat.
Assuming ±5% means ±5% forever: Resistors age. Pulse loading, thermal cycling, and humidity all nudge the value. A ±5% part can drift to ±8% after a year in a hot, vibrating environment. If you need long-term stability, spec ±1% and test at end-of-life temperature.
Substituting 10k "because it's close": I watched a junior engineer swap 4.7k pull-ups for 10k to reduce power draw, then spend two days debugging why I²C worked on the bench but failed in the field. The longer RC time constant made the bus marginal at 400 kHz, and a little extra capacitance from a longer cable pushed it over the edge.
What to Watch Going Forward
As supply chains tighten and manufacturers rationalize SKU counts, oddball E24 values (like 3.3kΩ, 5.6kΩ) sometimes drop out of stock while E12 staples (4.7kΩ, 10kΩ, 22kΩ) stay available. Designing around the most common values reduces your BOM risk, but it also means everyone else is doing the same - so lead times for 4.7kΩ resistors can spike during shortages. Keep a list of acceptable substitutions (series/parallel combinations, next E24 neighbor) in your design notes so your CM can build without waiting for your approval.
The other trend is package miniaturization. 0402 and 0201 resistors are now standard in phones and wearables, but their power density is brutal - 0402 resistors are rated ≤1/16W, so the same 4.7kΩ that handled 10V in 0805 will burn out in 0402. If you're shrinking a design, re-run your power budget.
Conclusion
The 4k7 resistor earns its ubiquity. It's cheap, stocked everywhere, and sits at the inflection point where current draw, signal speed, and noise all converge into "good enough" for most digital and analog circuits. But "good enough" isn't the same as "optimal." Choosing between 4.7k and 10k for a pull-up is a tradeoff between speed and power. Picking ±5% carbon film over ±1% metal film is a bet that drift and tolerance stacking won't bite you. And assuming a ¼W rating is always sufficient ignores the thermal reality of densely packed PCBs.
The best component choice isn't the one that works on the bench - it's the one that still works after 10,000 power cycles, 100,000 units, and three years in a 60°C enclosure. That's where understanding the color code, tolerance, and power rating stops being academic and starts being the difference between a product that ships and one that gets a running change six months into production.
