Reference guide
How Much Squeeze Should an O-Ring Have?
Published squeeze bands by seal type and AS568 cross-section, the 0.007 in floor, and what temperature, swell and stretch do to the number.
An O-ring seals by being squeezed, and the published manufacturer guidelines put the squeeze in bands by seal type: roughly 19 to 32 percent for a static face seal on the smallest AS568 section, 15 to 20 percent for a static radial seal on the largest, and 8 to 16 percent for dynamic reciprocating seals on the mid and large sections. Parker's handbook sets the floor for every case: never design below 0.007 in of absolute squeeze, because below about 0.005 in the ring takes close to 100 percent compression set and stops sealing. This guide tabulates the bands by seal type and family, explains why the percentage falls as the section grows, and covers what temperature, swell and stretch do to the number.
Working squeeze by seal type and cross-section
Percentages are of the ring cross-section; the inch figures are the same squeeze in absolute terms. AS568 industrial convention (Parker and MIL-G-5514 lineage), cross-checked against ERIKS, Trelleborg and Precision Associates.
| AS568 family (cross-section) | Static face, percent | Static face, inches | Static radial, percent | Static radial, inches | Dynamic reciprocating, percent | Dynamic, inches |
|---|---|---|---|---|---|---|
| -0xx (0.070 in) | 19 to 32 | 0.013 to 0.023 | 22 to 32 | 0.015 to 0.023 | 15 to 25 | 0.010 to 0.018 |
| -1xx (0.103 in) | 20 to 30 | 0.020 to 0.032 | 17 to 24 | 0.017 to 0.025 | 10 to 17 | 0.010 to 0.018 |
| -2xx (0.139 in) | 20 to 30 | 0.028 to 0.042 | 16 to 23 | 0.022 to 0.032 | 9 to 16 | 0.012 to 0.022 |
| -3xx (0.210 in) | 21 to 30 | 0.043 to 0.063 | 15 to 21 | 0.032 to 0.045 | 8 to 14 | 0.017 to 0.030 |
| -4xx (0.275 in) | 21 to 29 | 0.058 to 0.080 | 15 to 20 | 0.040 to 0.055 | 11 to 16 | 0.029 to 0.044 |
| Rotary (all) | 0 to 10 percent, with the ring sized 2 to 5 percent LARGER than the shaft, not smaller (Trelleborg); a separate design case | |||||
Two patterns are deliberate. The percentage falls as the section grows because the absolute squeeze in inches is what bridges tolerances, and a big ring does not need proportionally more of it (ERIKS). And a face seal takes the most squeeze because nothing moves and the flanges carry the load; the Parker handbook calls face seals the most tolerant gland type with the best low-temperature margin.
Why the squeeze exists at all
Trelleborg lists five jobs for initial compression: achieve the initial seal, bridge production tolerances, set a defined friction force, compensate for compression set, and compensate for wear. The fourth is the one that decides where in the band to start: a compound that takes a large set has to begin higher so it still has squeeze left after the set is taken, which is why the compression set guide and this one are read together. ERIKS adds the ceiling: initial compression above 25 percent is undesirable because it becomes over-compression at higher temperature, and in dynamic service it means excessive friction.
What changes the number
| Condition | Effect on squeeze | Published guidance |
|---|---|---|
| Low temperature or vacuum | Raise initial squeeze; the ring shrinks away from the gland when cold | ERIKS (FFKM illustration: static 18 percent rising to 27 percent on the 1.78 mm section); the direction generalizes, the values do not |
| High temperature | Reduce initial squeeze; the ring expands into the gland | ERIKS (FFKM: 18 percent falling to 16 percent above 200 C); volumetric expansion of an FFKM runs to about 20 percent at 316 C |
| Fluid swell 15 to 25 percent | Raise minimum squeeze AND open the groove: on a 0.139 in axial gland Precision Associates' high-swell chart moves minimum squeeze from 0.025 to 0.030 in and widens the groove from 0.185 to 0.215 in | Precision Associates; sealing a fluid that swells the ring more than 25 percent is not practical |
| Swell in dynamic service | Keep swell near 10 percent (Parker) or under 15 to 20 percent (ERIKS); above that, friction problems | Parker ORD 5700; ERIKS |
| Shrinkage in dynamic service | At most 4 percent before leakage | ERIKS |
| Installed stretch | Stretch thins the ring; cross-section reduction is about half the ID stretch percentage, so recompute squeeze on the reduced section | Marco Rubber; Parker: over 2 to 3 percent stretch needs a gland-depth correction, over 5 percent is not recommended |
| Compression set | Start higher in the band for a compound with poor set | Trelleborg |
| Hardness above 70 Shore A | More friction at the same squeeze, not less | ERIKS |
Squeeze and the other two gland rules
Squeeze works with gland fill and stretch. Fill the groove to 60 to 85 percent of its volume (Parker's optimum is 75 percent), checked at maximum ring against minimum groove, and keep assembled stretch inside the published limits: 0 to 2 percent on a face seal under internal pressure, 1 to 3 percent under external pressure, 2 to 8 percent on a static piston and 2 to 5 percent on a dynamic piston on Trelleborg's figures, zero on a rod. Parker does not recommend more than 5 percent; above 2 to 3 percent, correct the gland depth. The groove dimensions that produce the squeeze bands above are tabulated in the groove design guide. Friction in dynamic service has one more published number worth knowing: ERIKS puts breakout friction after ten days standing at 2 to 5 times the running friction of a lightly loaded seal.
The ranges in this guide are the published design guidelines of the manufacturers named beside them (Parker, ERIKS, Trelleborg, Precision Associates, Marco Rubber, Apple Rubber), re-arranged by seal type and AS568 cross-section family. They are not Conquest Seal engineering recommendations: Conquest distributes seals and cuts gaskets, and does not design parts. If a drawing calls out SAE AS4716, SAE AS5857, ISO 3601-2 or MIL-G-5514, the standard governs and these ranges are context.
Sources
- Parker O-Ring Handbook ORD 5700A/US, sections 3.5 to 3.7 (stretch, squeeze and gland fill), read 2026-08-12
- AS568 industrial gland tables (Parker and MIL-G-5514 lineage), cross-checked against the manufacturer guides below, read 2026-09-06
- ERIKS nv, O-ring Technical Handbook, chapter 12, O-ring Gland Design, read 2026-09-06
- Trelleborg Sealing Solutions, O-Rings and Back-up Rings catalog (June 2024 edition), read 2026-09-06
- Precision Associates, Application Data: Seal Gland Design (2011), read 2026-09-06
- Marco Rubber (Marco Sealing Solutions), Understanding O-Ring Squeeze, Stretch and Compression, read 2026-09-06
- Apple Rubber Products, Seal Design Guide (Rules of Thumb; O-Ring Installation), read 2026-09-06
- SAE AS4716 and AS5857, ISO 3601-2 and MIL-G-5514: titles and scope only, from the SAE, ISO and ASSIST records read 2026-09-06
Reviewed by Grant Midstokke, Conquest Seal Corporation. Last updated September 2026. Spec values in this guide are drawn from the sources listed above; the compound datasheet and your application conditions govern the final choice.
Frequently asked questions
Is more squeeze always safer?
No. Above about 25 percent initial compression the ring over-compresses when it warms, takes set faster and, in dynamic service, drags. Stay inside the published band for the seal type.
What is the absolute minimum squeeze?
0.007 in, per Parker's handbook; below about 0.005 in everything takes 100 percent compression set. That is an absolute figure, not a percentage, which is why small sections carry high percentages.
Does squeeze change with material?
The bands are the same for all elastomers; what changes is where in the band to start (higher for poor-set compounds and cold service) and, for silicone and fluorosilicone, the clearance allowed beside the gland, which ERIKS cuts by 50 percent and Parker by 60 percent.
How do I check squeeze on a metric ring?
The same way; use the ISO 3601-2 housing tables if the hardware is metric, and note that ISO and the AS568 industrial convention differ slightly on static depth for the large sections. The metric O-ring sizes guide covers the size systems.
Which durometer for a given squeeze?
Durometer is chosen for pressure and clearance, not squeeze; see the durometer guide. Above 70 Shore A the same squeeze produces more closure force and more friction.
What does the squeeze come to in inches on the common sections?
For a static face seal the published range is 0.013 to 0.023 in on the 0xx section, 0.020 to 0.032 in on 1xx, 0.028 to 0.042 in on 2xx, 0.043 to 0.063 in on 3xx and 0.058 to 0.080 in on 4xx. Reciprocating glands run much lighter: 0.012 to 0.022 in on the 2xx section.
How does swell change the squeeze?
Precision Associates publishes two complete chart sets, one for normal swell of 0 to 15 percent and one for high swell of 15 to 25 percent. On a 2xx face-seal gland the high-swell chart raises the minimum squeeze from 0.025 to 0.030 in, makes the gland shallower and the groove wider, and no published source considers it practical to seal a liquid that swells the ring more than 25 percent.
How much stretch is allowed on the inside diameter?
Trelleborg allows a static piston gland 2 to 8 percent stretch, a dynamic piston 2 to 5 percent, and a face seal under external pressure 1 to 3 percent; rod glands are sized with no stretch. Parker does not recommend more than 5 percent; above 2 to 3 percent, correct the gland depth. Stretch thins the cross-section and eats into the squeeze, which is why Parker, Marco Rubber and Apple Rubber converge on about 5 percent as the practical ceiling.
Related guides
Send the part number and we quote it
Part number or drawing, compound if you know it, and the quantity. A Conquest representative replies, not an auto-responder.
Request a quote