ZENCIRCUITLAB

How accurate is the lab?

Every circuit below was run in Zen Circuit Lab and in ngspice-42, the open-source SPICE simulator engineers use, with the same part models. 42 of 42 give the same answer; the worst difference anywhere is 0.330 %.

Re-run the check in your browser
42 / 42reference circuits agree with SPICE
0.330 %worst difference, any terminal, any time
31 + 11DC answers + waveforms compared
4 / 4fast converter models match their switching twins

What was compared

Each circuit is built from ordinary parts and wires. The lab writes it as a SPICE netlist (the same File → Export SPICE netlist anyone can use), ngspice solves it, and every terminal voltage is compared with the lab's own answer: at the operating point for DC circuits, and sample by sample for waveforms. DC answers must agree within 2 %, waveforms within 5 %.

What this shows: the lab's solver gets the same answer SPICE does for the same circuit, with the lab's diode, LED, Zener, transistor and MOSFET parameters and every wire and contact as the resistor the lab uses for it. It doesn't show that a part's model matches a real part's datasheet; that is checked part by part. Not in the suite yet: op-amps, the 555 and transformers.

CircuitKindWorst differenceLimit
Voltage divider, 12 V, 10 k / 4.7 kDC✓ < 0.01 %2.00 %
Series-parallel network, 5 resistorsDC✓ < 0.01 %2.00 %
Unbalanced Wheatstone bridgeDC✓ < 0.01 %2.00 %
Two sources, three resistors (superposition)DC✓ < 0.01 %2.00 %
9 V battery with its internal resistance, 100 Ω loadDC✓ < 0.01 %2.00 %
Potentiometer at 30 %, loaded by 10 kDC✓ < 0.01 %2.00 %
1N4148 with 1 k from 5 VDC✓ < 0.01 %2.00 %
1N4007 with 10 Ω from 12 V (1 A)DC✓ < 0.01 %2.00 %
1N5819 Schottky with 100 Ω from 5 VDC✓ < 0.01 %2.00 %
Red LED with 330 Ω from 5 VDC✓ < 0.01 %2.00 %
Blue LED with 150 Ω from 5 VDC✓ < 0.01 %2.00 %
5.1 V Zener with 470 Ω from 12 VDC✓ < 0.01 %2.00 %
5.1 V Zener regulator, 1 k loadDC✓ < 0.01 %2.00 %
NPN common emitter, divider biasDC✓ < 0.01 %2.00 %
NPN switch driving an LED (saturated)DC✓ < 0.01 %2.00 %
PNP high-side switch, 100 Ω loadDC✓ < 0.01 %2.00 %
Darlington emitter followerDC✓ < 0.01 %2.00 %
Two-transistor current mirrorDC✓ < 0.01 %2.00 %
N-MOSFET switch, 5 V gate, 10 Ω loadDC✓ < 0.01 %2.00 %
N-MOSFET in saturation, 2.5 V gate, 1 k drainDC✓ < 0.01 %2.00 %
P-MOSFET high side (AO3401), gate pulled lowDC✓ 0.080 %2.00 %
Diode OR from 12 V and 9 VDC✓ < 0.01 %2.00 %
LED constant-current sink (two NPN)DC✓ < 0.01 %2.00 %
Bridge rectifier on DC (two diode drops), 100 ΩDC✓ < 0.01 %2.00 %
RC charge: 10 k, 100 µF electrolytic, 5 V (τ = 1 s)Waveform✓ 0.181 %5.00 %
RC charge: 1 k, 1 µF ceramic, 5 V (τ = 1 ms)Waveform✓ 0.120 %5.00 %
RL rise: 1 H (8 Ω inside) + 100 Ω, 10 V (τ ≈ 9.3 ms)Waveform✓ 0.020 %5.00 %
Series RLC ring: 10 mH, 1 µF film, 10 Ω (1.6 kHz)Waveform✓ 0.330 %5.00 %
Half-wave rectifier: 12 V 60 Hz, 1N4007, 470 µF, 1 kWaveform✓ 0.301 %5.00 %
Bridge rectifier: 12 V 60 Hz through 2.2 Ω (the transformer's winding), 1000 µF, 220 ΩWaveform✓ 0.266 %5.00 %
RC low-pass on 60 Hz AC (fc = 60 Hz)Waveform✓ 0.051 %5.00 %
Buck from parts: 12 V, 50 kHz, duty 45 %, 100 µH, 220 µF, 10 ΩWaveform✓ 0.234 %5.00 %
Boost from parts: 5 V, 50 kHz, duty 50 %, 100 µH, 220 µF, 47 ΩWaveform✓ 0.029 %5.00 %
Non-inverting amplifier: op-amp as VCVS (gain 100 000), 9 k / 1 k, 1 V inDC✓ < 0.01 %2.00 %
VCCS 2 mS driven by 1.5 V, into 1 k ∥ 2.2 kDC✓ < 0.01 %2.00 %
Transimpedance (CCVS 2.2 kΩ): 5 V through 4.7 k into the sense pair, 1 k loadDC✓ < 0.01 %2.00 %
Current gain (CCCS β = 50): 3 V through 100 k sets Ic, output into 2.2 kDC✓ 0.021 %2.00 %
Formula source V = 0.5·a² − b + 1, a from a divider, b = 2 VDC✓ < 0.01 %2.00 %
Formula current source I = 2 mA · tanh(a / 1 V), a = 0.8 V, into 1.5 kDC✓ < 0.01 %2.00 %
Formula source 5·sin(2π·50·t) into an RC low-pass (1 k, 10 µF)Waveform✓ 0.060 %5.00 %
Subcircuit: transistor stage (bias divider, Rc, Re, NPN) packed into a block, driving an LEDDC✓ < 0.01 %2.00 %
Subcircuit: RC (1 k, 1 µF) packed into a block, 5 V step, 10 k load outsideWaveform✓ 0.127 %5.00 %

Fast converter models against switching twins

Buck and boost modules run as averaged models, so a big solar or battery bench stays fast. Each one is checked against a twin built from parts (MOSFET, diode, choke, capacitor and a PWM controller) that switches at 150 kHz, cycle by cycle, under the same load. Output within 3 % and input power within 15 % pass: the twin uses generic parts, so its losses are close to, not the same as, the module's.

ModuleOutput: fast / switchingInput power: fast / switchingResult
LM2596 buck module, 12 V → 5 V at 0.5 A4.973 / 5.010 V2.94 / 2.75 W✓ 0.7 % · 7.0 %
LM2596 buck module, 12 V → 5 V at 1 A4.947 / 5.015 V5.78 / 5.54 W✓ 1.3 % · 4.4 %
LM2596 buck module, 24 V → 12 V at 2 A11.894 / 11.934 V27.60 / 25.10 W✓ 0.3 % · 9.9 %
MT3608 boost module, 5 V → 12 V at 0.2 A11.989 / 11.791 V2.76 / 2.45 W✓ 1.7 % · 12.3 %

Check it yourself

Open the lab, then Menu → Simulate → Accuracy against SPICE, and press Run here: your browser repeats the lab side of every circuit against the stored SPICE answers. Any bench, including your own, exports as a netlist (Menu → File → Export SPICE netlist) for ngspice or LTspice.

Open the check in the lab About the lab

Checked 2026-10-22 with ngspice-42.