Constant-Force Springs

Constant ForceContinuousCascadeCOTS PartsCustom Parts

Coiled stainless-steel strip springs (constant-force springs) mounted on spools. Force stays within roughly 1–2% across extension, and stage-specific spring ratings can track a continuous lift's stepped gravity load.

Best for

Continuous lifts that should hold position anywhere unpowered, with per-stage springs matched to each stage's moving mass.

Poor fit for

Teams unable to mount spools and supported shafts, or lifts with no room for spring coils at the base of each stage.

ComplexityModerate

COTS springs plus printed spools, shafts, and mounts; one spring/spool assembly per compensated stage.

Force accuracyHigh

Spring force is nearly constant (~1–2% variation across wraps). Accuracy is limited mostly by available spring ratings versus actual stage masses.

PackagingModerate

Coils and spools need clearance at each stage; staggered spool placement avoids interference.

Tuning effortSelect, then verify

Tuning is spring selection, not knob turning: approximate from CAD mass, then verify with a luggage-scale measurement per stage.

In-season riskest.Moderate

Springs are reliable once installed, but replacing a mis-sized spring means new hardware and possibly reprinted spools mid-season.

Click a metric to see why it’s rated that way. Ratings marked est. are editorial estimates awaiting test data.

Quick recommendation

Use this when

You want the lift to hold position at any height with the motor unpowered
Your lift is continuous and you can fit a spool at each stage
You can measure or estimate per-stage moving mass from CAD and verify with a scale
You can 3D print spools and mount short supported shafts

Avoid this when

Available spring ratings don't divide well across your stage masses
You cannot package the spring coils and spools within your slide envelope
The lift design will keep changing mass late in the season (springs must be re-selected)
You need the counterspring force itself to vary within a stage

Mechanism overview

A constant-force spring is a strip of thin stainless steel (think 0.05–0.15 mm) that has been pre-stressed — put under tension and wrapped onto itself or a drum — so that it wants to stay coiled. Pulling the strip off the coil delivers a restoring force that barely changes with extension.

Annotated diagram — to be addedA constant-force strip spring: a coil of thin stainless steel with the free end extended, force arrows showing near-constant pull back toward the coil. From the source guide (image1).

Because the steel is so thin, the radius of each successive wrap is nearly the same, so the force stays nearly constant. The variation between wraps is usually only around 1–2% of the rated force — negligible for a lift.

To counterspring a lift, the coil sits on a spool near the base of a stage and the free end of the strip attaches to the moving stage above it. Gravity pulls the stage down; the spring pulls it back with a constant force. Match the spring rating to the weight the stage carries, and the stage floats.

Why is the force nearly constant? (deeper explanation)

An ordinary extension spring stores energy in stretch, so force grows with displacement (Hooke's law). A constant-force spring instead works by bending: as you pull the strip, material at the tangent point continuously unbends from the coil radius to flat. The energy cost per unit length of strip is set by the strip's thickness, width, and the coil radius — none of which change meaningfully as the coil pays out (the coil's radius shrinks only slightly since the strip is so thin). Constant energy per unit length means constant force. The small 1–2% variation comes from that slight change in coil radius between the outer and inner wraps.

Terminology used on this page

Rated force — the constant pull the spring delivers, from the manufacturer's catalog.
Spool — the cylinder the coil rides on. It idles; it is not a winch.
Stage weight — the total weight a given stage must move: everything mounted above it.
Full extension length — the strip length; must exceed the stage's travel with margin.

Continuous and cascade implementations

This is the architecture the method was practically made for, and the architecture covered by the source guide.

On a continuous lift, stages extend one after another, so the moving weight increases in steps (why). The fix is direct: give each stage its own spring, rated for the weight that stage carries.

What moves: each stage relative to the stage below it.
Where the spring attaches: the coil sits on a spool mounted at the base of stage n; the strip's free end attaches to stage n+1 (the stock mounting hole in the strip works for slide inserts; the last stage's spring screws into the carriage plate).
Force profile produced: constant per stage, stepping with the stage transitions — the same staircase shape as the gravity load.
Packaging: a coil + spool at the base of every compensated stage. Stagger the spools along the slide so they don't interfere.
Advantages: the only COTS approach that tracks the continuous staircase; the lift can float anywhere; nothing to re-tension.
Limitations: spring ratings come in discrete steps, so each stage is slightly under- or over-sprung; hardware multiplies by stage count (and by two if mounted symmetrically).

The heaviest spring goes on the outermost fixed rail (it carries everything) and is fixed to the side plate — on robots with static slides, attach it via an outside insert or the drivetrain plate. Each successive spring is lighter, because it carries one fewer stage.

Physics and force matching

The goal is for the springs to cancel gravity at every extension, so the motor only supplies acceleration and friction.

Gravity load on a stage

F_{g} = m_{\text{carried}}\, g \sin\theta

In plain terms: Each stage must counteract the weight of everything it carries — the carriage plus every stage above it — reduced by the lift's angle (a fully vertical lift has sin θ = 1).

Assumes: Rigid stages · Angle measured from horizontal

Per-stage spring requirement

F_{\text{spring},\,i} \;=\; \frac{m_{\text{carried},\,i}\; g \sin\theta}{n_{\text{sides}}}

In plain terms: Divide each stage's gravity load by the number of spring sides (two if you mount symmetrically) to get the rating each individual spring needs.

Matching error

e(x) = F_{\text{counterspring}}(x) - F_{\text{gravity}}(x)

In plain terms: The leftover force at each extension. Positive means over-sprung (the lift creeps upward); negative means under-sprung (it settles downward). With per-stage constant-force springs, the error within each stage is just the gap between the catalog rating and the true stage weight.

Variable definitions
Symbol	Quantity	Unit	Notes
$m_{\text{carried},\,i}$	Mass carried by stage i	kg	Carriage + all stages above stage i
$g$	Gravitational acceleration	m/s²	9.81
$\theta$	Lift angle from horizontal	deg	90° = vertical
$n_{\text{sides}}$	Spring sides	—	1 or 2 (symmetric mounting)
$e(x)$	Counterspring error	N	Target: small and slightly negative

Worked example: Springs for a three-stage vertical lift

Carriage + outtake ≈ 1.2 kg; each slide stage + inserts ≈ 0.45 kg; vertical lift (θ = 90°); springs on both sides.

Stage weights are cumulative — each stage carries everything above it:

Stage 1 (innermost spring, carries carriage + 2 stages): (1.2 + 2×0.45) × 9.81 ≈ 20.6 N → 10.3 N per side
Stage 2 (carries carriage + 1 stage): (1.2 + 0.45) × 9.81 ≈ 16.2 N → 8.1 N per side
Stage 3 (carries carriage only): 1.2 × 9.81 ≈ 11.8 N → 5.9 N per side

Rounding down to catalog ratings (McMaster lists strip springs in roughly half-pound ≈ 2.2 N steps): 8.9 N, 6.7 N, and 4.5 N per side. Each stage ends up a few tenths of a newton under-sprung — the motor supplies the difference, and the lift settles instead of creeping up.

Note on what these springs don't cancel: perfectly countersprung gravity still leaves inertial force — the force to accelerate the mass upward. There is no practical passive way to counterspring acceleration (a normal spring can't follow an acceleration profile; you'd need a specialized variable-rate element or an actuator).

Interactive force graph

Adjust the lift to match yours. With rating per stage, the spring force tracks the stepped gravity load; the residual is the quantization error from discrete catalog ratings. Switch to a single rating to see why one spring can't cover a continuous lift.

Force vs. extension — Constant force

Illustrative model, not measured data

RMS error 0.8 NNormalized RMS 5%Max |error| 0.8 N

Stringing architecture

Carriage / outtake mass1.20 kg

Mass per stage0.45 kg

Stages3

Travel per stage30 cm

Lift angle from horizontal90 °

Spring arrangement

Available rating step2.2 N

View chart data as a table

Extension (cm)	Gravity (N)	Counterspring (N)	Motor (N)
0.0	11.8	11.0	0.8
7.5	11.8	11.0	0.8
15.0	11.8	11.0	0.8
22.5	11.8	11.0	0.8
30.0	11.8	11.0	0.8
37.5	16.2	15.4	0.8
45.0	16.2	15.4	0.8
52.5	16.2	15.4	0.8
60.0	16.2	15.4	0.8
67.5	20.6	19.8	0.8
75.0	20.6	19.8	0.8
82.5	20.6	19.8	0.8
90.0	20.6	19.8	0.8

Design and sizing

Sizing happens in two passes: an approximation from CAD, then a confirmation measured on the real lift. You could probably get away with either alone, but the two together catch mistakes before you order springs or print spools.

Approximating from CAD

Using Onshape with the FTC Parts Library, COTS parts carry accurate weights; use the material library for plates (reasonably accurate carbon-fiber data is available from your plate vendor). For 3D-printed parts, a community materials library's PLA option tends to be close — but printed parts are the least accurate, so if you care, use your slicer's weight estimate or weigh them for real.

Add up, per stage:

Outtake / carriage — arm, claw, mounting plates, screws if you want precision
Slides — one set of rails per stage
Inserts — aluminum / carbon fiber / printed inserts; you can include the spring's own weight in final numbers

The first (heaviest) spring covers the full total; subtract one set of slides + inserts for each spring after that.

Constant-force spring sizing (per stage)

Moving stages

count

Carriage / outtake mass

Mass per slide stage

Spring sides

1 or 2

Lift angle

° from horizontal

Required spring force per stage
Stage	Mass carried	Force per side	≈ in lbf
1	2.10 kg	10.3 N	2.32 lbf
2	1.65 kg	8.1 N	1.82 lbf
3	1.20 kg	5.9 N	1.32 lbf

Round DOWN to the nearest available spring rating — a slightly under-sprung lift settles gently; an over-sprung one climbs on its own. Verify with a scale measurement on the assembled lift before ordering final springs.

Pretension and travel

Constant-force springs need little pretension in the classic sense, but the strip should never fully rewind onto the coil in operation: keep roughly 1–1.5 wraps engaged at minimum extension, and choose a strip long enough to cover the stage's full travel with margin at maximum extension.

Safety factor

Choose ratings against your measured stage weights (procedure below), not CAD alone, and bias down. If the design will gain mass late in the season — it will — recheck the measurement after major changes; a heavier outtake un-floats the lift.

CAD, manufacturing, and assembly

Open CAD: Countersprung Slides (Onshape)Interactive CAD embed planned; opens in Onshape for now.

Reference implementation from the source guide: staggered constant-force springs on Misumi slides with 3D-printed two-part spools.

Annotated diagram — to be addedAnnotated staged-spring layout: heaviest spring outermost, fixed to the side plate, each spring's strip reaching the next sliding stage. From the source guide (image2).

Recommended mounting architecture

The reference design staggers the spools along the slide so they don't interfere with each other. Each spool rides on a short, fully supported shaft: two single-block mounts with a 20 mm spacer between them create a rigid shaft that's supported on both ends.

Robot photo — to be addedClose-up of the spool assembly: GoBilda single block mounts, 20 mm metal spacer, two-part 3D-printed spool with flanges, constant-force spring coil. From the source guide (image3/image4).

Assembly notes from the reference build:

Double-threading: running one screw through two sets of threads is usually frowned upon, but it worked here. If you're worried about a screw seizing, drill out the first block mount's threads.
Metal spacers over printed: printed spacers can crush when tightened; metal costs negligible weight and keeps the stack rigid.
Two-part printed spools: printing the spool in two halves lets you slide the coil on sideways instead of unwinding the spring to thread it — assembly takes seconds.
Flanges: the spools have flanges to keep the strip from wandering. Possibly unnecessary if the strip extends without twist — untested either way, so they stay.
No bearings: a slightly oversized bore creates negligible friction, on par with a bearing, without the weight and part count. A bearing is still a viable option.
Symmetric mounting: springs on both sides distribute the counterspring force evenly. One-sided can work if the stages are mechanically linked with real rigidity — and note that catalogs offer more low-force options, so two light springs can often match the target more accurately than one heavy one.

Known design flaw and the improved alternative

In the reference design, spools mounted on the inside of the plates interfere with any plates mounted to the slides on the spring side. Mounting the spools on the outside face fixes the interference. The revised version replaces the block mounts + spacer with beams and a standoff — one beam needs countersinking.

Annotated diagram — to be addedImproved mounting: spools moved to the outside face of the plate, beams with a countersunk standoff replacing block mounts and spacer. From the source guide (image5).

Assembly sequence

Print spool halves; verify the coil slides over the assembled spool.
Install block mounts (or beams) and spacer; torque the shaft stack.
Slide each coil onto its spool, heaviest spring at the outermost fixed rail.
Route each strip flat — no twist — to its stage and fasten through the strip's stock hole (screw into the carriage plate for the last stage).
Cycle the lift by hand slowly, watching each strip wrap and unwrap cleanly.

Safety and spring retention

Testing and validation

Measuring the real stage weights

The most accurate way to choose final springs is to measure the assembled lift — strung, wired, on the robot — ideally before the motor and transmission are connected, since they add friction and tension that contaminate the reading.

Attach a force sensor or luggage scale to the lift (most have a hook — loop it over a shaft).
Pull the first stage up at any point before the slide reaches its endstop and read the force: that's everything.
Hold the first stage down (a hand over it works) and extend the second stage — now you're weighing everything minus stage one.
Continue stage by stage until all are measured.

Judging over- vs. under-countersprung

With springs installed and the motor still disconnected, move the lift by hand to several heights and let go:

Stays put or settles slowly downward → correctly or slightly under-sprung. Good.
Creeps upward → over-sprung; drop a rating on the offending stage.
Falls freely → under-sprung or a spring isn't engaged; recheck strip attachment.

Identify which stage misbehaves by holding the stages below it and testing one at a time — the same isolation trick as the measurement procedure.

Watch: Countersprung lift holding position (unpowered)Inline player planned; opens externally for now.

Reference build: the lift is moved to arbitrary positions and released — it holds. (The last stage has no spring only because the season ended first.)

Powered testing

Watch: Countersprung lift under powerInline player planned; opens externally for now.

Reference build running on the robot.

Once the motor is connected, validate the original goal:

Current — log lift motor current over a full practice cycle; compare against the pre-spring baseline if you have one.
Speed — time full extension and retraction; verify retraction hasn't become the new bottleneck (the motor now works against the springs going down… which is exactly the gravity assist you wanted).
Thermal — motor temperature after a full match's worth of cycles.
Cycle test — run continuous extend/retract cycles and re-inspect strips, spool retention, and fasteners afterward.

Real-world examples

Site author's reference build (FTC 7172 / 23381)

Constant Force

Season not specified in source · continuous stringing

Misumi-slide continuous lift with staggered constant-force springs on 3D-printed two-part spools, mounted symmetrically on both sides.

What worked: Lift held position unpowered at arbitrary heights; assembly was fast thanks to two-part spools; measurements matched CAD within ~0.2 lb.
What failed: Inside-mounted spools interfered with slide-mounted plates on the spring side; the last stage never received its spring before the season ended.
Would change: Mount spools on the outside face of the plate and use beams + countersunk standoff instead of block mounts and a spacer.

CAD Video

Your team here

Constant Force

— · architecture unrecorded

A verified competition implementation with CAD, video, and honest lessons learned belongs in this slot.

Placeholder card — a real, verified team example belongs here. Contributions welcome.

Common failure modes

Stage creeps upward on its ownMedium severity

Cause: Spring rating above the true stage weight (over-sprung), often after weight was removed from the outtake.
Symptoms: Lift extends slowly with the motor idle; motor draws current to hold the lift down.
Prevention: Round ratings down; re-measure after any mass change.
Fix: Swap the offending stage to the next rating down.

Strip disengages or rewinds off the spoolHigh severity

Cause: Too little strip left engaged at minimum extension, or the free-end fastener pulled out.
Symptoms: Sudden loss of counterspring on one stage; loose coil rattling on the spool.
Prevention: Keep 1–1.5 wraps engaged at minimum extension; use the strip's stock hole with a proper screw and washer.
Fix: Re-tension the strip onto the spool and re-fasten; replace the strip if kinked.

Spool interference with slide platesMedium severity

Cause: Spools mounted on the inside face of the plate (the documented design flaw).
Symptoms: Stages bind near specific extensions; visible rubbing on the spool flanges.
Prevention: Mount spools on the outside face; check swept clearances in CAD.
Fix: Re-mount per the improved alternative design.

Crushed printed spacer loosens the shaft stackLow severity

Cause: 3D-printed spacer compressed when the stack was torqued.
Symptoms: Spool wobbles; strip tracks off-axis and rubs the flanges.
Prevention: Use metal spacers in the shaft stack.
Fix: Replace with a metal spacer and re-torque.

Strip edge injury during assemblyHigh severity

Cause: Handling a loose coil without gloves; coil snapping closed.
Symptoms: —
Prevention: Gloves and eye protection; keep coils captive until installed.
Fix: First aid; then add retention so it can't recur.

How this compares

Bungee

Bungee is cheaper and faster to install, but its force grows with extension — it can't hold every position the way stage-matched constant-force springs can.

Guide

Constant Torque

Constant torque delivers one force level at the winch; constant force can step per stage. On a continuous lift that difference is decisive — on cascade, constant torque packages better.

Guide

Sprung Linkage

A linkage can theoretically match any profile, including this method's staircase — but it trades a catalog part for custom geometry and weeks of tuning.

Guide

For the full matrix, force-curve overlay, and decision summary, see Compare Methods.

References

Perfect Lift Gravity Counterbalance Explanation (Continuous) — source guideEngineering source
Primary source for this page: the site author's own implementation writeup.
McMaster-Carr constant-force spring catalogManufacturer data
The guide's recommended source — reliable and fast. Amazon carries similar springs but often at higher prices.
Countersprung Slides CAD (Onshape)CAD
Reference implementation on Misumi slides.
Unpowered holding demonstrationVideo
Powered lift demonstrationVideo