top of page

ISO 815-1
Compression Set of Vulcanized and Thermoplastic Rubber

Universal Grip ISO 815-1 compression set clamping device for rubber

Downloads: ISO 815-1 Compression Set Clamping Device for Rubber – Universal Grip Data Sheet (PDF)

Written by Chris Thomas, Universal Grip Company · Reviewed by Frank Bertini · Updated October 24, 2026

ISO 815-1 is the international standard for compression set of vulcanized and thermoplastic rubber at ambient or elevated temperature. A cylindrical disc is squeezed to a fixed strain between parallel plates, held for 24, 72 or 168 hours at a set temperature, released and allowed to recover for 30 minutes, and the thickness it fails to recover is reported as a percentage of the compression that was applied. It is the test every seal, O-ring, gasket and weatherstrip specification leans on, because a rubber that takes a permanent set under load is a rubber that stops sealing.

This guide covers the current edition, ISO 815-1:2019, and how it differs from earlier ones; the hardness-based strain levels and the two test pieces; why the standard specifies a strain tolerance rather than a spacer height and what that means for the spacers you machine; the durations, temperatures and the three release methods A, B and C, which can move the result by a third on the same compound; the plates, thickness gauge, lubricant and conditioning rules; the calculation with a worked example; the common errors; a numeric comparison with ASTM D395; and the clamping device, plates, cutters and ovens Universal Grip supplies for the test.

In this guide: What ISO 815-1 Measures · Editions: 1991, 2008, 2014, 2019 · Compression Strain by Hardness · Test Pieces: Type A and Type B · Clamping Device, Plates, Spacers and Gauge · Durations, Temperatures and Conditioning · Methods A, B and C: Release and Recovery · Step-by-Step Procedure · Calculation and Worked Example · Common Errors and How Much They Cost · ISO 815-1 vs ASTM D395 · Equipment for ISO 815-1 · FAQ

What ISO 815-1 Measures

Compression set is the permanent deformation left in a rubber after it has been held compressed and then released. ISO 815-1 expresses it as a percentage of the compression applied, not of the original thickness: a result of 20 % means the disc recovered 80 % of the deflection it was given. Low values mean the elastomer network returned to shape; high values mean physical relaxation, chemical cross-link scission or new cross-links formed under strain have locked in the deformed shape. For a seal that is the difference between holding pressure after five years and leaking after one.

The standard applies to vulcanized and thermoplastic rubbers of 10 to 95 IRHD hardness. It gives one method at ambient temperature and three at elevated temperature, called Methods A, B and C, which differ only in how the test piece is released and recovered. The scope note is worth repeating: a short 24 hour test at elevated temperature is a cure-state and quality-control check, while a test of around 1,000 hours is the one that predicts service life of seals. The low-temperature version is a separate standard, ISO 815-2, where the whole measurement is made at the test temperature. For the rubber tensile and tear properties that are usually measured alongside, see ISO 37 and ISO 34-1; the elastomer testing page gives the lab overview.

Editions: 1991, 2008, 2014, 2019

ISO 815 of 1991 was a single standard covering ambient, elevated and low temperatures. In 2008 it was split into Part 1 (ambient and elevated) and Part 2 (low temperature), with Part 2 replacing the old ISO 1653, which several lab pages still cite. The 2014 second edition of Part 1 tidied the methods, and the 2019 third edition is the current one: it updated the normative references to ISO 188 (ovens), ISO 18899 (calibration) and ISO 23529 (conditioning and test-piece preparation), and added a precision statement as Annex A, with a normative calibration schedule as Annex B. The standard was reviewed and confirmed in 2025, so no new edition is due. ISO 815-2 was reissued in 2019 as well and confirmed in 2024.

Compression Strain by Hardness

ISO 815-1 ties the amount of compression to the hardness of the rubber, which is the main difference from the fixed 25 % of ASTM D395 Method B. The tolerances are tight and they are on the strain, not on a spacer dimension:

  • Below 80 IRHD: 25 ± 2 % compression. This is the everyday case for seals, gaskets and most moulded goods.

  • 80 to 89 IRHD: 15 ± 2 %. A hard compound compressed 25 % stores so much force that the plates bend and the piece shears at the edges.

  • 90 IRHD and above: 10 ± 1 %. Very hard rubbers and some thermoplastic elastomers.

Because the tolerance is on strain, the spacer height follows from the actual thickness of the test piece. For a nominal 12.5 mm Type A disc, 25 % compression means a spacer of 0.75 × 12.5 = 9.375 mm; for a nominal 6.3 mm Type B disc it means 4.725 mm. Measure the disc, not the drawing: a Type A piece at the low end of its tolerance, 12.0 mm, compressed onto a 9.375 mm spacer sees 21.9 % strain, which is outside the ± 2 % window. Good practice is to sort discs and select spacers, or to machine the spacer set to the lot, so that every piece falls inside 23 to 27 %.

Test Pieces: Type A and Type B

Two cylindrical test pieces are defined, and results from them are not comparable with each other, so the type always goes in the report.

  • Type A: diameter 29.0 ± 0.5 mm, thickness 12.5 ± 0.5 mm. The preferred piece for low-set rubbers because the larger recovery is measured with less relative error. It is the same disc as ASTM D395 Type 1.

  • Type B: diameter 13.0 ± 0.5 mm, thickness 6.3 ± 0.3 mm. Preferred when pieces must be cut from finished products; take them from near the centre of the product with the disc axis parallel to the direction of service compression where possible. It is not the ASTM Type 2 disc, which is 6.0 ± 0.2 mm thick.

Moulding is preferred. Cut pieces follow ISO 23529: cut oversize and trim, or use a sharp rotating cutter with lubricant, so that the faces stay flat and the edges are not cupped. Where the product is too thin, up to three plies may be stacked without adhesive, lightly pre-compressed for a minute and then measured as one piece; ASTM allows more plies, ISO stops at three. A minimum of three test pieces is tested, separately or as a set in one clamp. After the test each piece is cut in half across a diameter and inspected; gas bubbles or internal defects mean a repeat. Universal Grip supplies Ø29 and Ø13 mm cutting dies and die presses for cutting Type A and B pieces from sheet.

Clamping Device, Plates, Spacers and Gauge

The apparatus is a stack of parallel plates clamped together with spacers between them, which the standard says a simple screw device is adequate for. The plates are the critical part:

  • Plates: polished chromium-plated steel, stainless steel or another corrosion-resistant material; they must not bend more than 0.01 mm under the load of the compressed pieces; surface finish about Ra 0.4 µm; large enough that the piece can expand laterally without touching anything.

  • Spacers: set the compression to the strain tolerance above and must never touch the test piece. Rings or bars placed outside the disc footprint are both used.

  • Clamp: bolted tie rods and nuts, tightened evenly until the plates sit on the spacers. Uneven torque tilts the plate and gives an uneven strain across a set of pieces.

  • Thickness gauge: accurate to ± 0.01 mm, with a flat circular foot of 4.0 ± 0.5 mm diameter exerting 22 ± 5 kPa (10 ± 2 kPa for rubber under 35 IRHD). A 0.001 mm resolution digital gauge is typical. Thickness is read at the centre of the disc.

  • Oven: an ISO 188 method A or method B oven; the standard warns that the two oven types can give different results and that the interior of the clamp should reach steady state within 3 hours of a 24 hour test, which argues for thin plates and a well-circulated oven rather than a massive stack.

  • Lubricant: a thin film on the plate faces of a fluid that does not affect the rubber, typically silicone or fluorosilicone fluid of about 100 mm²/s viscosity; silicone fluid is not used on silicone rubber. The lubricant, or its omission, is reported, because friction at the faces changes the stress state and the set.

The Universal Grip ISO 815-1 compression set clamp described at the end of this guide is this device: ground parallel plates, spacer stops, tie rods and a two-tier option for oven capacity.

Durations, Temperatures and Conditioning

Test duration is timed from the moment the clamp goes into the oven and is one of 24 h (+0, −2 h), 72 h (+0, −2 h), 168 ± 2 h, or a multiple of 168 h. The elevated temperatures are the ISO 188 series: 40, 55, 70, 85 and 100 °C held to ± 1 °C, and 125, 150, 175, 200, 225 and 250 °C held to ± 2 °C. Ambient tests run at the standard laboratory temperature, 23 ± 2 °C or 27 ± 2 °C. The standard notes that results at the higher temperatures bring thermal stability and surface oxidation into play and do not correlate in any simple way with set at room temperature; that is why a seal specification quotes a temperature and time with every limit.

Timing before the test matters as much as during it. At least 16 hours must pass between vulcanization and testing; for non-product testing the maximum is 4 weeks, and product testing should ideally be within 3 months of manufacture. Pieces are protected from light and heat in the meantime and conditioned at least 3 hours at the standard laboratory temperature. Thermoplastic rubbers are annealed first, 70 °C for 30 minutes is the suggested treatment, then conditioned, because their set depends strongly on the thermal history of the hard phase.

Methods A, B and C: Release and Recovery

All three elevated-temperature methods compress the piece the same way; they differ in what happens when the clock stops, and that difference is large. Rubber recovers faster when it is hot, so a disc released at the oven temperature springs back further in its 30 minutes than one released cold.

  • Method A (the default when nothing else is specified): remove the clamp from the oven, release the test piece immediately while hot, move it quickly to a wooden bench and let it recover 30 ± 3 min at the standard laboratory temperature, then measure.

  • Method B: remove the clamp from the oven and let it cool for 30 to 120 min to the standard laboratory temperature with the piece still compressed, then release, recover 30 ± 3 min and measure. The cold release gives a higher set.

  • Method C: release the piece inside the oven, let it recover 30 ± 3 min at the test temperature, then a further 30 ± 3 min at the standard laboratory temperature, then measure. The hot recovery gives the lowest set.

  • Ambient method: release, recover 30 ± 3 min at the standard laboratory temperature on a wooden bench, measure.

The bench matters: a poor thermal conductor lets the hot piece cool slowly and evenly. The size of the method effect is not academic. On one nitrile compound tested at 70 °C for 24 hours the ASTM D395 guide shows about 14 % by hot release against about 21 % by cold release; a purchase specification that does not name the method cannot be met or failed with confidence. Method B corresponds most closely to how a seal is actually relieved in service, cold, and is what many automotive specifications ask for; Method A is the ISO default and the one that matches ASTM D395 Method B.

Step-by-Step Procedure

  • Confirm the cure date and the conditioning: 16 h minimum after vulcanization, at least 3 h at 23 ± 2 °C, thermoplastic rubbers annealed first.

  • Clean the plates, apply the lubricant film if one is used, and record it.

  • Measure each disc thickness at the centre to 0.01 mm with the 4 mm, 22 kPa foot; this is h0. Measure the spacers; this is hs.

  • Place the discs on the lower plate clear of the spacers and bolts, fit the spacers, add the upper plate and tighten the nuts evenly until the plates seat on the spacers. Check that no piece touches a spacer.

  • Put the clamp in the centre of the preheated oven without delay and start the clock. Note the oven type.

  • At the end of the period, release and recover by Method A, B or C as specified, 30 ± 3 min each stage, on a wooden bench.

  • Measure the recovered thickness h1 at the centre with the same gauge and foot.

  • Calculate the set for each piece, report the individual values and the median or mean, then cut each piece in half and inspect for bubbles and defects.

Calculation and Worked Example

The compression set is Cs = (h0 − h1) / (h0 − hs) × 100, where h0 is the initial thickness, h1 the thickness after recovery and hs the spacer height. The denominator is the compression that was applied, which is why the spacer must be measured rather than assumed.

Worked example, Type A disc, 25 % compression, 72 h at 100 °C, Method A. Measured h0 = 12.52 mm; spacer hs = 9.38 mm, so the applied compression is 12.52 − 9.38 = 3.14 mm and the strain is 3.14 / 12.52 = 25.1 %, inside 25 ± 2 %. After the 30 min recovery h1 = 11.90 mm. Unrecovered deflection is 12.52 − 11.90 = 0.62 mm, so Cs = 0.62 / 3.14 × 100 = 19.7 %. Three pieces at 19.7, 20.4 and 18.9 % give a median of 19.7 % and a mean of 19.7 %, which is what a seal specification of 25 % maximum at 100 °C / 72 h would be judged against.

Note how the arithmetic punishes a sloppy spacer. Had an ASTM-style 9.50 mm spacer been used without re-measuring it, the denominator would be 3.02 mm and the same recovered thickness would report as 20.5 %; on a Type B disc with a 4.50 mm spacer the strain is 28.6 % and the result is simply outside the standard.

Common Errors and How Much They Cost

  • ASTM spacers on ISO Type B discs. A 4.50 mm spacer compresses a 6.3 mm disc by 28.6 %, outside 25 ± 2 %. Use 4.725 mm nominal and select to the measured lot.

  • Unstated release method. Method A and Method B differ by roughly a third of the result on the same compound; a certificate without the method is not comparable with anything.

  • Pieces touching spacers or bolts. The piece is partly restrained and reads low; the standard's instruction to keep spacers clear of the disc is a validity rule, not a convenience.

  • Massive clamps in a slow oven. If the stack takes four hours to reach 100 °C the 24 h test is really a 20 h test at temperature; keep plates thin and ovens circulating, and record the oven method.

  • Testing too soon after cure. Within the first 16 hours post-cure continues and set reads high; equally, non-product pieces older than 4 weeks are outside the standard.

  • Measuring the hot piece. Reading h1 before the 30 min recovery, or on a steel bench that quenches it, moves the result by several points. Use the wooden bench and the timer.

  • Wrong gauge pressure. A 22 kPa foot on a 25 IRHD rubber sinks into it; the standard drops to 10 ± 2 kPa below 35 IRHD.

  • Mixing Type A and Type B data. The two pieces do not give the same number for the same compound; Type A reads lower set on soft rubbers. Report the type and compare like with like.

  • No lubricant record. Lubricated and dry faces give different edge stresses; the standard requires the lubricant, or its omission, in the report.

ISO 815-1 vs ASTM D395

The two standards are close relatives and a lab can usually run a single clamp to satisfy both, but they are not interchangeable and a report must name one. The numbers that differ:

  • Test pieces: ISO Type A = ASTM Type 1, both 29.0 ± 0.5 × 12.5 ± 0.5 mm. ISO Type B is 13.0 ± 0.5 × 6.3 ± 0.3 mm; ASTM Type 2 is 13.0 ± 0.2 × 6.0 ± 0.2 mm, 0.3 mm thinner with a tighter diameter tolerance.

  • Compression: ISO specifies strain, 25 ± 2 / 15 ± 2 / 10 ± 1 % by hardness. ASTM Method B specifies spacers of 9.5 ± 0.02 mm and 4.50 ± 0.01 mm (nominally 25 %); ASTM Method C mirrors the ISO hardness logic.

  • Constant force: ASTM Method A applies 1.8 kN constant force; ISO 815-1 has no constant-force method.

  • Durations: ISO 24 / 72 / 168 h (+0, −2 h); ASTM 22 / 70 / 166 h minimum with windows of 22 to 24, 70 to 72 and 166 to 168 h, so a run finished between 22 and 24 h satisfies both.

  • Temperatures: the same ISO 188 / ASTM D1349 series, 40 to 250 °C.

  • Cure-to-test: ISO 16 h minimum, 4 weeks maximum for non-product pieces; ASTM 24 h minimum, 60 days maximum.

  • Release: ASTM D395 Method B releases hot and recovers 30 min on a poor conductor, which is ISO Method A; ISO adds the cold-release Method B and the hot-recovery Method C.

  • Plates: ISO about Ra 0.4 µm, bend ≤ 0.01 mm; ASTM Ra 0.25 µm.

  • Plies and replicates: ISO up to 3 plies and a minimum of 3 pieces; ASTM up to 7 plies (Type 1) or 4 (Type 2), average of 2 or median of 3.

  • Precision (ASTM): at a 13.7 % mean, repeatability 1.7 and reproducibility 4.4 percentage points; at 52.8 %, 1.6 and 16.8 points. ISO 815-1:2019 gives its own interlaboratory data in Annex A.

ISO 815-2 covers the low-temperature case, −10 to −70 °C in practice, with Method 1 derived from Part 1 and Method 2 using a recording device to log thickness during recovery, all measured at the test temperature. For the constant-strain relaxation that complements set, see ISO 3384; for the broader rubber programme, the elastomers page.

Equipment for ISO 815-1

Universal Grip compression set clamping device

The Universal Grip ISO 815-1 device (part UGC-ISO815-1-CP) is a bolted multi-plate clamp built for bench and oven use with no load frame required. The data sheet describes ground, parallel stainless or plated steel plates with ground, polished faces for uniform compression; spacer stops that set the compression for Type A and Type B test pieces; room for multiple test pieces per level so a set of three or more runs in one clamp; a two-tier stack that doubles capacity in one oven footprint; corrosion-resistant construction for elevated-temperature ageing; and bolted tie rods with nuts for clamping. When ordering, specify the test piece type (A or B), the number of test pieces, the compression ratio and the maximum ageing temperature.

One point to get right at order time: the sheet lists approximately 9.5 mm and 4.5 mm spacers, which are the ASTM D395 Type 1 and Type 2 heights. For ISO 815-1 the spacers are made to 0.75 × the nominal thickness, about 9.38 mm for Type A and 4.72 mm for Type B, and for 80 IRHD and harder compounds to 0.85 or 0.90 × thickness (10.6 and 11.25 mm on Type A). Ask for the ISO set, the ASTM set, or both, and for spacers to be selected to your measured disc thickness when the tolerance window is tight. Compression platens in ground hardened steel are available as plate stock for larger custom stacks.

Cutters, ovens and the rest of the rubber lab

  • Cutting dies: Ø29 mm and Ø13 mm rotary or punch cutters and a die press for cutting Type A and B pieces from sheet and from products, per ISO 23529.

  • Ageing ovens and chambers: an ISO 188 method A or B oven with forced circulation and ± 1 °C control to 100 °C, ± 2 °C to 250 °C; for combined ageing and mechanical testing, environmental chambers on the testing machine.

  • Thickness gauge: dial or digital gauge with a 4 mm foot and 22 kPa (or 10 kPa) pressure, 0.01 mm accuracy.

  • Companion tests: ISO 37 tensile dumb-bells, ISO 34-1 tear pieces and the ASTM D412 and ASTM D624 equivalents are cut on the same press; the ASTM D395 guide covers the American method in the same depth as this page.

ISO 815-1 FAQ

What is ISO 815-1?

ISO 815-1 is the international standard for compression set of vulcanized and thermoplastic rubber at ambient or elevated temperature. A disc is compressed to a fixed strain, usually 25 %, held for 24, 72 or 168 h at a set temperature, released, recovered for 30 min and re-measured; the unrecovered thickness is reported as a percentage of the compression applied. The current edition is ISO 815-1:2019, confirmed in 2025.

What are the ISO 815-1 test piece sizes?

Type A is a disc 29.0 ± 0.5 mm in diameter and 12.5 ± 0.5 mm thick, preferred for low-set rubbers. Type B is 13.0 ± 0.5 mm in diameter and 6.3 ± 0.3 mm thick, preferred for pieces cut from products. Results from the two types are not comparable, and Type B is not the same as ASTM D395 Type 2, which is 6.0 mm thick.

What spacer height gives 25 % compression?

The standard specifies strain, 25 ± 2 %, not a spacer height, so the spacer is 0.75 × the measured thickness: nominally 9.375 mm for a 12.5 mm Type A disc and 4.725 mm for a 6.3 mm Type B disc. An ASTM 4.50 mm spacer on a 6.3 mm disc gives 28.6 %, outside the ISO tolerance.

What compression is used for hard rubbers?

25 ± 2 % below 80 IRHD, 15 ± 2 % for 80 to 89 IRHD and 10 ± 1 % for 90 IRHD and above. On a Type A disc that means spacers of about 9.38, 10.6 and 11.25 mm respectively.

What are the ISO 815-1 test durations and temperatures?

24 h (+0, −2 h), 72 h (+0, −2 h), 168 ± 2 h or multiples of 168 h, at ambient (23 ± 2 °C or 27 ± 2 °C) or at 40, 55, 70, 85, 100 °C (± 1 °C) or 125, 150, 175, 200, 225, 250 °C (± 2 °C). Time runs from the moment the clamp enters the oven.

What is the difference between Methods A, B and C?

Method A releases the piece hot, straight from the oven, then recovers it 30 ± 3 min at room temperature; Method B cools the clamp 30 to 120 min with the piece still compressed, then releases and recovers 30 min; Method C releases inside the oven, recovers 30 min hot and 30 min at room temperature. Method A is the default; Method B reads highest, Method C lowest, and the spread can be a third of the result.

How is compression set calculated?

Cs = (h0 − h1) / (h0 − hs) × 100, where h0 is the initial thickness, h1 the thickness after recovery and hs the spacer height. A 12.52 mm disc on a 9.38 mm spacer that recovers to 11.90 mm has a set of 0.62 / 3.14 × 100 = 19.7 %.

How many test pieces does ISO 815-1 require?

At least three, tested separately or as a set in one clamp; individual values and the median or mean are reported. Up to three plies may be stacked without adhesive when the product is too thin for a single disc.

What are the plate and gauge requirements?

Polished chromium-plated or stainless steel plates about Ra 0.4 µm that bend no more than 0.01 mm under load, with spacers that never touch the piece; a thickness gauge accurate to 0.01 mm with a 4.0 ± 0.5 mm flat foot at 22 ± 5 kPa, or 10 ± 2 kPa below 35 IRHD.

Is ISO 815-1 the same as ASTM D395?

Close but not identical. Type A equals ASTM Type 1 but Type B is 0.3 mm thicker than Type 2; ISO specifies strain by hardness where ASTM Method B specifies 9.5 and 4.50 mm spacers; ISO durations are 24/72/168 h against ASTM 22/70/166 h minimums; ASTM Method A constant force (1.8 kN) has no ISO equivalent; and ASTM D395 Method B release matches ISO Method A only.

How long after vulcanization can I test?

Not before 16 h and, for non-product pieces, not after 4 weeks; product testing should be within 3 months of manufacture. Condition at least 3 h at 23 ± 2 °C before measuring, and anneal thermoplastic rubbers (70 °C for 30 min is suggested) first.

Do I need a testing machine for ISO 815-1?

No. The test uses a bolted clamp with spacers in an oven and a thickness gauge; no load frame is involved. The Universal Grip ISO 815-1 clamping device holds multiple Type A or B pieces per level and stacks two tiers in one oven footprint.

Need a compression set clamp, spacers, cutting dies or an ageing oven setup for ISO 815-1? Contact Universal Grip with your test piece type, compression ratio and maximum ageing temperature and we will quote the right device.

bottom of page