Quarter Car Suspension: Ride, Tyre Grip, Spring and Damper

Simulate a two-mass quarter car over a bump, pothole or ISO 8608 road. Compare ride acceleration, tyre loading, travel, natural frequencies and spring/damper choices.

#adaptive integration #damping-ratio #ISO 8608 #quarter-car #ride-comfort #spring-rate #suspension #tyre-load

This quarter car model compares how a spring and damper affect ride comfort, tyre loading and suspension travel at one wheel. It is intended for engineering students learning vehicle dynamics and for enthusiasts comparing measured coilover settings. Both the sprung corner mass and the unsprung wheel assembly move, so body bounce and wheel hop remain separate parts of the calculation.

Enter the sprung mass supported by one suspension corner, then the unsprung mass of that wheel, hub and associated suspension parts. Enter the coil spring rate and the spring motion ratio. Here the motion ratio means spring travel divided by wheel travel and already includes the installation angle. The wheel rate therefore contains the square of this ratio. The damper has its own motion ratio and separate bump and rebound coefficients at the shaft; use slopes from a force–velocity dyno curve over the velocity range of interest.

Choose a rounded bump, a rounded pothole or a random road described by an ISO 8608 class, then enter vehicle speed. Bumps and potholes have a raised-cosine profile with the specified length and height or depth. The random road uses a band-limited displacement power spectral density, rather than arbitrary noise. Keep its seed, frequency band and harmonic count unchanged when comparing settings. Increase record duration and try several seeds before drawing conclusions about a road class.

Press play to see the road scroll while the body, wheel, spring and damper follow the simulated response. The printed vertical magnification stays fixed throughout playback and bump-height dragging; its envelope includes the largest allowed obstacle. The contact patch colour and upward arrow show the predicted tyre load relative to static load. The charts show displacement transmissibility, body acceleration, tyre loading and wheel travel. The body acceleration RMS is unweighted: it supports comparisons but is not an ISO 2631 comfort rating. For an obstacle, RMS covers the whole record, including the flat approach and settling tail; use the same duration for comparisons. Random-road RMS excludes the first two seconds.

Use the suspension natural frequency calculator results to distinguish the coupled body mode from the simpler wheel-fixed estimate in the calculation steps. The spring rate calculator also gives the coil rate for a target coupled body frequency, with an equivalent damper coefficient for a target projected modal damping ratio. These are starting values for comparison, not a unique optimal tune.

This coilover spring rate calculator does not include pitch, roll, anti-roll bars, tyre enveloping, tyre damping, bump stops, top-out impacts or damper digression. Tyre load variation is a grip indicator; lateral grip and braking distance require a tyre model and whole-vehicle analysis. Once the predicted load reaches zero or entered travel is exceeded, the subsequent response falls outside the model. Check candidate settings against measured travel and dyno data, packaging and whole-vehicle behaviour.

How it is calculated

The equations are written about static equilibrium, positive upward:

msz¨s=−kw(zs−zu)−cw(z˙s−z˙u),m_s\ddot z_s=-k_w(z_s-z_u)-c_w(\dot z_s-\dot z_u),
muz¨u=kw(zs−zu)+cw(z˙s−z˙u)+kt(zr−zu).m_u\ddot z_u=k_w(z_s-z_u)+c_w(\dot z_s-\dot z_u)+k_t(z_r-z_u).

Wheel rates are kw=ksrs2k_w=k_s r_s^2, cbw=cbrd2c_{bw}=c_b r_d^2 and crw=crrd2c_{rw}=c_r r_d^2. Bump applies when z˙s−z˙u<0\dot z_s-\dot z_u<0 (the suspension is shortening); otherwise rebound applies. These balances agree with Soong et al. (2017), Eqs. 1–2; the direction-dependent damper is solved without replacing it by its average in time.

Natural frequencies solve det⁡(K−ω2M)=0\det(K-\omega^2M)=0 for both masses. With mode shape [1,q]T[1,q]^T, the reported projected damping is ζ=cw(1−q)2/[2ω(ms+muq2)]\zeta=c_w(1-q)^2/[2\omega(m_s+m_uq^2)]. The exact linear frequency response solves the complex two-mass dynamic stiffness at 801 logarithmic frequencies, using ceq=(cbw+crw)/2c_{eq}=(c_{bw}+c_{rw})/2. The separate-coefficient chart shows two additional linear comparisons; these are not the nonlinear system's exact sinusoidal response.

Time integration uses the second-order implicit midpoint rule. The relative-velocity damping rate is λc=max⁡(cbw,crw)(1/ms+1/mu)\lambda_c=\max(c_{bw},c_{rw})(1/m_s+1/m_u). If the requested step times this rate exceeds 0.25, adaptive step doubling compares one midpoint step with two half steps, using position, velocity and both mass accelerations to reject or refine a step. The accepted solution is the two-half-step result. A direct 2 × 2 midpoint solve determines the damping branch from the sign of its relative-velocity numerator; this sign is independent of the nonnegative damping coefficient.

The output grid is no coarser than the entered maximum, 1/25 of the undamped wheel-hop period normally (1/12.5 with adaptive stiff-damping control), 1/50 of obstacle crossing time, or 0.08 of the shortest road harmonic period, as applicable. Adaptive internal steps may be much smaller. Peak acceleration, minimum tyre load and maximum jounce/droop are accumulated over every accepted internal endpoint, rather than only the output grid. Endpoint road samples and midpoint work preserve the discrete energy identity at every accepted substep. Reported RMS quantities use trapezoidal integration on the output grid.

The two-slope viscous model is positively homogeneous in road amplitude: multiplying a bump or pothole by a nonnegative factor multiplies all positions, velocities and accelerations by that factor, and work/energy by its square. A reference-amplitude response is integrated once and scaled exactly for the selected height. The scene uses that same reference response and the fixed 150 mm input limit to choose a constant magnification that contains the body, wheel and road excursions. Random-road magnification is chosen from its complete record. Neither playback time nor the draggable obstacle height changes the scene scale.

For random roads, ISO 8608:2016 displacement PSD is implemented as Gd(n)=G0(n/0.1)−2G_d(n)=G_0(n/0.1)^{-2}, with nn in cycles/m and G0=16×10−64jG_0=16\times10^{-6}4^j m³ for class index j=0j=0 (A) through 7 (H). The profile is zr(x)=∑j2Gd(nj)Δnjcos⁡(2πnjx+ϕj)z_r(x)=\sum_j\sqrt{2G_d(n_j)\Delta n_j}\cos(2\pi n_jx+\phi_j). The bins are logarithmic, evaluated at geometric centres; independent repeatable pseudo-random phases come from the entered seed. See Rao (2016), printed pp.21–24 (attached pp.29–32), for the PSD/phase synthesis. Temporal frequency is f=vnf=vn.

Target spring rate is obtained by inverting the exact lower-mode characteristic equation, not the wheel-fixed approximation. The target mean damper matches the projected lower-mode damping of that target spring. It does not determine a unique bump/rebound split.

Assumptions

One vertical wheel station, rigid sprung and unsprung masses, constant motion ratios, linear coil and tyre stiffness, and a two-slope passive viscous damper. Tyre damping is zero. No aerodynamic load or longitudinal/lateral load transfer. The tyre remains in contact and the suspension remains within its available travel.

Displacements are relative to the static ride height; static suspension deflection is msg/kwm_sg/k_w, while static tyre compression is (ms+mu)g/kt(m_s+m_u)g/k_t. Random-road integration starts with both masses at the initial road height and zero vertical velocity; the first 2 s are excluded from RMS summaries. Peak acceleration, minimum tyre load and travel demands cover the full record, including startup. All result times and road details are retained in the traces.

Where it applies

Use for ride and setup comparison, not whole-vehicle handling prediction. Body acceleration is unweighted and excludes seat and occupant dynamics; no ISO 2631 exposure or comfort classification is claimed. Dynamic tyre-load RMS is measured about static load, not about the record mean. Negative predicted tyre load signals loss of validity: the bilateral tyre spring continues mathematically, but free flight and recontact are not simulated. Bump stops, top-out and spring bind are also outside the model.

Natural frequencies are exact undamped two-mass values. Damping ratios are projections on those undamped mode shapes; strong nonproportional or overdamped behaviour needs a damped-pole analysis. Frequency curves use the mean viscous coefficient and omit harmonics and mean shifts caused by bump/rebound asymmetry. The target damper is a modal starting point, not a dyno recommendation.

Random roads are synthetic realizations, not measured profiles. The default band is 0.02–3 cycles/m with 384 harmonics; long waves, short-wave texture and tyre enveloping outside that band are omitted. A short record and a single seed have sampling uncertainty; compare several seeds and longer records. The 2 s warm-up may be insufficient for very lightly damped settings. Refine the time step to verify peaks and RMS. Default peak/RMS convergence and strong asymmetric-damping peak convergence are checked explicitly. Strongly damped settings can take several seconds: their adaptive internal steps resolve the fast dynamics while the stored history uses up to 12.5 samples per undamped wheel-hop period. This coarser stored grid can smooth very short features in the charts and RMS integration; accepted-substep extrema remain available in the results. Playback linearly interpolates that history. Vertical displacements and road heights use the constant magnification printed on the scene, chosen for the complete excursion envelope and kept fixed during playback and bump-height dragging. The load arrow is schematic; its colour follows the load/static scale. The 0.33 m drawing tyre radius is illustrative and does not enter the dynamic model.

ISO 8608:2016, edition 2, is used for road spectral description, not a road safety or design criterion; check against the standard in force. Formulae and drawings are original; no standard tables or figures are reproduced.

References

  • Gillespie, T. D., Fundamentals of Vehicle Dynamics, SAE, 1992 edition. Bibliographic background on ride and quarter-car models. https://doi.org/10.4271/R-114
  • Milliken, W. F. and Milliken, D. L., Race Car Vehicle Dynamics, SAE, 1995 edition. Suspension and damper setup background. https://www.millikenresearch.com/rcvd.html
  • ISO 8608:2016, edition 2, Mechanical vibration — Road surface profiles — Reporting of measured data. https://www.iso.org/standard/71202.html
  • Soong, Ramli and Saifizul (2017), Between simplicity and accuracy: Effect of adding modeling details on quarter vehicle model accuracy, PLOS ONE 12(6). https://doi.org/10.1371/journal.pone.0179485
  • Thite (2012), Development of a Refined Quarter Car Model for the Analysis of Discomfort due to Vibration. The present tool implements the ordinary two-mass model, without the paper's extra series compliance. https://doi.org/10.1155/2012/863061
  • Rao, A. M. (2016), A Structured Approach to Defining Active Suspension Requirements, Virginia Tech thesis, pp.21–24. https://vtechworks.lib.vt.edu/bitstream/10919/72247/1/Rao_AM_T_2016.pdf
  • Malgaca, Uyar and Lök (2018), Vibration Control of a 2-DOF Quarter Car with an Actuator by State Space and System Identification Methods, 2nd International Students Science Congress, pp.171–176. https://acikerisim.ikcu.edu.tr/dosyalar/int_62d7dd425e885.pdf
  • Swift / YCW Suspension, Suspension 101, worked motion-ratio, wheel-rate and critical-damping example. https://www.swiftspringsasia.com/suspension101
  • Independent numerical verification: Python 3.10.12 / SciPy 1.15.3, Radau stiff integrator with dense-output peak search. Source s7 archives settings, source code and results. Solver documentation: https://docs.scipy.org/doc/scipy/reference/generated/scipy.integrate.solve_ivp.html

Inputs

InputSymbolUnitDefault
Sprung corner massm_skg300
Unsprung massm_ukg40
Coil spring ratek_sN/mm30
Spring motion ratio
Spring travel divided by wheel travel, including installation angle.
r_s0.8
Damper bump coefficientc_bN*s/m1500
Damper rebound coefficientc_rN*s/m2500
Damper motion ratio
Damper shaft travel divided by wheel travel.
r_d0.8
Tyre vertical stiffnessk_tN/mm200
Available wheel jounce
Wheel travel from static ride height to the bump stop.
j_avmm80
Available wheel droopd_avmm70
Road excitationroad_kindRaised-cosine bump
Vehicle speedspeedkm/h30
Bump height / pothole depthHmm50
Obstacle lengthL_bm3
ISO 8608 classroad_classClass B
Road realization seed
Keep the seed fixed when comparing suspension settings.
seed37
Record durationT_ends6
Animation timets0.98
Target body natural frequencyf_targetHz1.3
Target modal damping ratioz_target0.3
Maximum time step
Halve this to check RMS and peak results for time convergence.
dts0.002
Lowest road spatial frequencyn_lo1/m0.02
Highest road spatial frequencyn_hi1/m3
Road harmonicsN_h384

Results

With the default inputs above, the calc gives:

ResultValue
Body acceleration RMS (unweighted)0.704 m/s²
Tyre load RMS / static (%)6.35
Body bounce natural frequency1.216 Hz
Wheel-hop natural frequency11.79 Hz
Body modal damping, mean0.232
Wheel-hop modal damping, mean0.221
Body modal damping, bump0.174
Body modal damping, rebound0.29
Wheel spring rate19.2 N/mm
Required wheel jounce32.04 mm
Required wheel droop38.56 mm
Minimum predicted tyre load2,143 N
Peak body acceleration4.04 m/s²
Peak damper displacement30.85 mm
Coil rate for target body frequency34.8 N/mm
Mean shaft damping for target mode2,849 N·s/m
RMS evaluation duration6 s
Tyre contactTyre remains loaded
Suspension travelWithin entered travel

Checked against sources

Each case below was calculated by this calc and agrees with the value from its source (within the tolerance where one is shown).

Manufacturer motion-ratio worked example

Source: Swift / YCW Suspension, Suspension 101, Natural Ride Frequencies and Damping sections. Published example: 10 kgf/mm, spring/wheel travel ratio 0.97, sprung mass 400 kg; wheel rate 92,270.77 N/m and shaft critical coefficient 12,913.64 N·s/m. Their unit spellings N-m and N-m/sec denote rate and damping coefficient, not torque. https://www.swiftspringsasia.com/suspension101

Inputs: Sprung corner mass = 400 kg; Coil spring rate = 98.067 N/mm; Spring motion ratio = 0.97; Damper motion ratio = 0.97; the others as above.

ResultSourceThis calc
Wheel spring rate92,271 N/m92,271 N/m
Wheel-fixed critical damper at shaft12,914 N·s/m12,914 N·s/m

Published ANSYS two-mass natural frequencies

Source: Malgaca, Uyar and Lök, Vibration Control of a 2-DOF Quarter Car with an Actuator by State Space and System Identification Methods, 2nd International Students Science Congress, 2018, printed p.172 (attached p.2). Independent MASS21/COMBIN14 ANSYS modal analysis. Only the undamped frequencies are checked; the paper's stated Δt=0.407 is inconsistent with resolving an 11 Hz mode and is not used. https://acikerisim.ikcu.edu.tr/dosyalar/int_62d7dd425e885.pdf

Inputs: Sprung corner mass = 240 kg; Unsprung mass = 36 kg; Coil spring rate = 16 N/mm; Spring motion ratio = 1; Damper motion ratio = 1; Damper bump coefficient = 980 N·s/m; Damper rebound coefficient = 980 N·s/m; Tyre vertical stiffness = 160 N/mm; the others as above.

ResultSourceThis calc
Body bounce natural frequency1.238 Hz1.238 Hz
Wheel-hop natural frequency11.135 Hz11.135 Hz

Independent stiff ODE peak, strong bump-only damping

Source: Independent reviewer supplied approximately 2.652 m/s² for this case. Independently reproduced with scipy.integrate.solve_ivp(method='Radau', rtol=1e-10, atol=1e-12, max_step=0.002 s), dense output and a scalar search for peak |body acceleration|, in stiff-reference.py in the task workspace. Expected 2.6520640954 m/s² at 0.748445852 s. This is an independent numerical implementation of the two-mass equations, not a published measurement. The 2 s regression record contains the peak; the independent Radau search covered 6 s.

Inputs: Sprung corner mass = 50 kg; Unsprung mass = 5 kg; Damper bump coefficient = 20,000 N·s/m; Damper rebound coefficient = 0 N·s/m; Damper motion ratio = 2; Bump height / pothole depth = 10 mm; Record duration = 2 s; the others as above.

ResultSourceThis calc
Peak body acceleration2.652 ± 0.003 m/s²2.65 m/s²

Independent stiff ODE peak, strong rebound-only damping

Source: Independent numerical implementation using SciPy Radau at rtol=1e-10, atol=1e-12, max_step=0.002 s, followed by dense-output peak search. Expected 3.4707009797 m/s² at 0.868777369 s; full settings are in stiff-reference.py in the task workspace. This checks the opposite damping branch; it is not a published reference case. The 2 s regression record contains the peak; the independent Radau search covered 6 s.

Inputs: Sprung corner mass = 50 kg; Unsprung mass = 5 kg; Damper bump coefficient = 0 N·s/m; Damper rebound coefficient = 20,000 N·s/m; Damper motion ratio = 2; Bump height / pothole depth = 10 mm; Record duration = 2 s; the others as above.

ResultSourceThis calc
Peak body acceleration3.471 ± 0.004 m/s²3.473 m/s²

Related Resources

Responses

Your email address will not be published. Required fields are marked *

L ading...