Coin Cell Spacer and Spring Selection: Stack-Height Guide

Most CR2032 build sheets say "one 1.0 mm spacer, one spring" and stop there. Run the numbers on that default and you will often find the spring compressed by less than a tenth of a millimetre — barely a spring at all — which is why nominally identical cells scatter across a plate. This guide replaces the default with a stack-height budget you can run in a spreadsheet before you order parts, plus the bench checks that catch a bad build in thirty seconds instead of three weeks.

The short answer: match spacer thickness to spring travel

Coin cell spacer and spring selection comes down to one number: how far the spring is compressed once the case is crimped. Target 30–60% of the spring’s usable travel — enough to hold the stack under load, with headroom left for electrode swelling. For a standard CR2032 running a 1.20 mm free-height wave spring, that usually means 1.5 mm of total spacer, not the 1.0 mm most kits ship with. Diameter and grade matter less: Ø15.8 mm in 304 stainless covers the majority of lithium-ion work.

The 30–60% window is a working heuristic from bench practice, not a published standard. Treat it as a starting point, then validate it against your own hardware and chemistry. The arithmetic that gets you there is the part worth learning.

What the spacer and the spring actually do

The spacer is a thickness dial, not a filler

A spacer is a solid stainless disc that does not compress in any useful way, so its real job is setting how far the spring gets squeezed. Add 0.5 mm of spacer and you add 0.5 mm of spring deflection. No other component in the stack gives you that much leverage over stack pressure.

It carries two secondary jobs: conducting current between the electrode and the case, and spreading the spring’s contact points so they do not press into the separator.

The spring is a pressure source with a hard ceiling

Every wave or conical spring has a free height (uncompressed) and a solid height (flattened, roughly equal to its material thickness). The difference between them is usable travel. A Ø15.4 mm wave spring with a 1.20 mm free height stamped from 0.20 mm stock has about 1.00 mm of travel.

Below roughly 20% of that travel you are relying on crimp deformation rather than the spring, and tolerance differences of a few microns swing your results. Past 100% the spring is a rigid washer: crimp load transfers straight into the separator and coating, which is how you get soft shorts and crushed electrodes.

How to run a coin cell spacer and spring selection budget

Build the budget in five steps before you order anything. It takes about ten minutes the first time and thirty seconds after that.

1. Measure the internal cavity height. A CR2032 is 20 mm across and 3.2 mm tall under IEC 60086 nomenclature. Subtract both case walls — typically 0.20–0.30 mm each, depending on supplier [VERIFY against your supplier’s drawing]. The examples below use 2.70 mm.
2. Micrometer every layer. Use measured thicknesses, not datasheet nominals. Coating weight routinely lands 5–10 µm off spec, and that is the same order as the tolerance you are trying to control.
3. Sum the uncompressed stack, counting the spring at its free height.
4. Calculate deflection: stack sum minus cavity height. A negative result means the spring never engages at all.
5. Divide by spring travel (free height minus material thickness) and check the percentage against the 30–60% window.

Worked example: a lithium-metal half cell with an NMC cathode, run three ways.

Layer (measured)	Config A: default kit	Config B: +0.5 mm spacer	Config C: 2 × 1.0 mm
Cathode: 15 µm Al + 75 µm coating	0.090 mm	0.090 mm	0.090 mm
Separator (25 µm PP/PE/PP)	0.025 mm	0.025 mm	0.025 mm
Lithium chip, Ø15.6 mm	0.450 mm	0.450 mm	0.450 mm
Spacer(s), Ø15.8 mm 304SS	1.000 mm	1.500 mm	2.000 mm
Wave spring at free height	1.200 mm	1.200 mm	1.200 mm
Uncompressed stack total	2.765 mm	3.265 mm	3.765 mm
Internal cavity	2.700 mm	2.700 mm	2.700 mm
Spring deflection	0.065 mm	0.565 mm	1.065 mm
% of 1.00 mm usable travel	6.5%  — too loose	57%  — on target	107%  — bottomed out

 

Config A is the default kit, and it is the one that burns people. At 6.5% of travel the spring is effectively a washer, contact depends on how hard you happened to crimp, and replicate scatter follows.

Config B is the fix: one 1.0 mm plus one 0.5 mm spacer, landing at 57%. Cells behave the same way build to build because the spring — not the crimp — is setting the load.

Config C is what happens when "just add a spacer" gets treated as universal advice. The stack demands 1.065 mm of compression from a spring with 1.00 mm to give. The extra 65 µm comes out of your separator and coating, and you will see it as early soft shorts.

Reality check from the literature. Zhou and colleagues documented exactly this lever in an anode-free lithium-metal study: their standard 2032 build measured 3.10 mm as-crimped, and adding one spacer on the cathode side took it to 3.33 mm, with the extra spring deflection improving cycling performance. That 0.23 mm difference is easy to resolve with a hand micrometer — which is precisely why crimped height makes a good process control.

Picking spacer thickness, diameter and grade

Thickness

Stock two thicknesses, not one. A structural spacer carries the load; a thin spacer trims the stack into the travel window.

Thickness	What it is for	Watch out for
0.2–0.3 mm	Fine trim to land inside the travel window	Dishes noticeably under crimp; never use as the only spacer
0.5 mm	The standard tuning increment; pair with a 1.0 mm	Two of these are not equivalent to one 1.0 mm — you add an interface
0.8–1.0 mm	Main structural spacer; resists dishing, carries load	Leaves little height for a second spacer in thick stacks
1.5 mm	Very thin electrodes, or 2430/2450 formats	Rarely fits a 2032 alongside a full-height spring

 

Diameter

Ø15.8 mm is the safe default for CR2032, CR2025 and CR2016. Two rules govern the choice.

• The spacer should exceed your electrode diameter by at least 2 mm, so load is carried past the electrode edge rather than concentrating on it.
• It must clear the gasket’s inner ring without binding. One supplier publishes roughly 16.39 mm for the 2032 gasket inner-ring ID, which is why 16.1 and 16.2 mm spacers exist and why 16.2 mm is a genuinely tight fit [VERIFY on your own gasket].

If a spacer catches on the gasket it cocks by a degree or two, and you get a wedge-shaped pressure profile that no amount of crimp force will correct.

Grade

Material	Use it for	Limitations
304 stainless	Standard Li-ion half and full cells up to about 4.3 V. Cheapest and most widely stocked.	Becomes slightly magnetic after cold working — relevant if you use magnetic tooling or magnetometry
316L stainless	Sulfur chemistries, higher voltages, long-duration cells. Molybdenum improves pitting resistance.	Carries a modest price premium; not a substitute for the right case material at high voltage
Aluminium-clad positive case + Al foil disc	Cathodes cycled above ~4.3 V, to limit oxidative corrosion of steel on the cathode side.	Adds cost and one more layer to your height budget — recalculate before switching

 

Burr orientation: the ten-second check almost nobody does

Stamped spacers have two different faces. One side has a rounded rollover edge; the other has a sharp burr. Tilt the disc under a lamp and the burr side shows a bright ring around the rim.

Put the burr side toward the spring and the rollover side toward the separator. It costs ten seconds per cell and it removes a whole class of soft short that looks exactly like an electrolyte problem on the tester.

Spring types compared

The spring always sits against the negative cap, with a spacer between it and the electrode stack. Placing a spring directly against a separator is a reliable way to puncture it.

Type	Typical free height	Pressure profile	Best for
Wave spring (3-peak)	1.1–1.5 mm, 0.2–0.3 mm stock	Three contact bands; high at the peaks, low at the centre	General lithium-ion work — the default in most kits
Conical (cone) spring	1.1–1.5 mm, 0.25 mm stock	Ring contact at the rim, unloaded centre	Thicker stacks that need more deflection per millimetre
Belleville / disc washer	0.5–1.0 mm	Stiff, short travel, high force	Very thin stacks where almost no height budget remains

 

A twenty-cell QC routine that costs thirty seconds per cell

A stack-height budget only helps if the cell you built matches the cell you designed. Four checks catch nearly every assembly error before a cycler channel is committed.

1. Weigh every finished cell on a 1 mg balance. A Ø15.8 × 0.5 mm 304 spacer weighs about 0.78 g and a 1.0 mm about 1.55 g. A missing or doubled spacer cannot hide from that measurement.
2. Micrometer the crimped height and log it. Set a control limit of ±0.03 mm from the batch median and quarantine outliers rather than cycling them.
3. Record OCV at 30 minutes and again at 12 hours. A cell that drifts down by more than a few millivolts overnight usually has a mechanical problem, not a chemical one.
4. Run one EIS sweep before formation. Series-resistance outliers almost always trace back to a burr, a cocked spacer, or a spring that never engaged.

The payoff is finding the three bad cells out of twenty before you tie up channels for three weeks. In practice the mass check alone catches more assembly errors than every other step combined — it is the cheapest instrument in the lab doing the most useful work.

Where this approach stops working

A stack-height budget improves reproducibility. It does not turn a coin cell into a pouch cell, and the ceiling is worth stating plainly.

Soulen and colleagues at UC San Diego mapped pressure inside 2032 cells using pressure-sensitive film and finite element modelling. During a 50 MPa crimp, pressure at the edge of the stack exceeded the film’s 49 MPa measurement ceiling while the centre registered roughly 0.5 MPa — the wave spring has no material at its centre, so it cannot load one. Lithium plated in that low-pressure centre came out over 600% thicker than at the edge in one electrolyte. Replacing the spring and one spacer with a silicone rubber disc of equivalent thickness flattened the distribution and raised 100-cycle capacity retention from 58.6% to 65.6%.

So the advice above does not apply cleanly in these cases:

• Solid-state cells, which need sustained pressure in the tens of MPa. No 2032 spring holds that; use a dedicated pressure die.
• Supercapacitor and high-porosity electrodes, which need more compression than a single spring provides — one study found a single-spring coin cell produced a compression ratio below 1.05.
• Any result you need to report at a known absolute stack pressure. A crimped coin cell cannot give you one; use a pouch cell with a load cell.
• Commercial CR2032 primaries, which are built differently. Lab hardware is a test fixture, not a route to a sellable battery.

What to actually order

For a lab running lithium-ion half cells, three line items cover almost everything:

• Ø15.8 × 1.0 mm 304SS spacers — the structural spacer.
• Ø15.8 × 0.5 mm 304SS spacers — the tuning increment.
• Ø15.4 × 1.2 mm wave springs in 0.2 mm stock — the workhorse spring.

Spacers are cheap enough that stocking two thicknesses is not really a decision. Canrd lists 304SS spacers at $66.45 per 1,000 pieces for the 16.1 × 0.8 mm size (canrud.com, checked August 2026; pricing and stocked sizes change). Buying a single thickness to save a few dollars costs you the ability to tune the stack, which is the entire point.

Order springs and spacers as separate line items rather than pre-boxed case sets. Kits fix the ratio at 1:1, and a properly budgeted stack often needs two spacers per spring.

Frequently asked questions

How thick should a CR2032 spacer be?

Most CR2032 lithium-ion builds need 1.5 mm of total spacer, usually a 1.0 mm plus a 0.5 mm disc. The correct figure depends on your electrode and lithium thicknesses: pick whatever lands spring deflection between 30% and 60% of the spring’s usable travel. A single 1.0 mm spacer frequently leaves the spring almost uncompressed.

Do I need one spacer or two?

Two, in most lithium-metal half cells. One spacer typically leaves under 10% of spring travel engaged, so stack pressure comes from crimp deformation instead of the spring. Adding a second spacer of 0.5 mm brings deflection into a controlled range. Stop before the total exceeds the spring’s travel, or the spring bottoms out.

Should I use 304 or 316L stainless spacers?

Use 304 for standard lithium-ion work up to about 4.3 V — it is cheaper and universally stocked. Choose 316L for sulfur chemistries, higher voltages and long-duration cells, where its molybdenum content improves pitting resistance. Above roughly 4.3 V, also consider an aluminium-clad positive case rather than relying on grade alone.

Which side does the coin cell spring go on?

The spring goes against the negative cap, with a spacer between the spring and the electrode stack. Never let a spring touch a separator directly — its contact points will puncture it. If you need extra height on the cathode side, add a second spacer there instead of relocating the spring.

Why do identical coin cells still give scattered results?

The most common cause is a spring sitting near the bottom of its travel, where crimp force rather than spring force sets stack pressure. Micron-level tolerance differences then produce visible performance spread. Check your stack-height budget first, then burr orientation, then crimping pressure consistency.

Can I reuse coin cell spacers and springs?

Reuse is possible but rarely worth it. Spacers dish permanently under crimp and springs take a set, so both come back with altered thickness and spring rate — exactly the variables you are trying to control. At roughly seven cents per spacer, new parts are cheaper than the data you risk.

Conclusion

Spacer and spring choice is not a parts-list decision — it is a thickness calculation with one output: how far your spring is compressed. Measure the cavity, micrometer the layers, sum the stack, and land the deflection between 30% and 60% of spring travel. Then verify with a balance and a micrometer before anything reaches a cycler.

Next step: build the budget for the chemistry you are running right now. Open a spreadsheet, enter your six measured thicknesses, and see where your current default lands. If it comes out below 20%, add 0.5 mm of spacer and rebuild five cells — in most labs the scatter collapses on the first attempt.