SIM & Memory Card Connectors

SIM Socket Contact Resistance After Cycling: What to Specify and How to Test

Tags:

“Contact resistance 100 mΩ maximum” is not a specification

Nearly every SIM socket datasheet lists a contact resistance figure. Almost none of them say when it is measured, at what current, or with what method. Those three omissions decide whether the number means anything.

A socket that measures 40 mΩ on a virgin part and 280 mΩ after 500 insertion cycles is compliant with a datasheet that says “100 mΩ maximum” as long as the reading was taken at zero cycles. The product will fail in the field, and the datasheet will have been accurate.

This article covers what the parameter actually means, how the standard test methods differ, and how to write a requirement that survives a qualification review.

Why contact resistance matters more in a SIM socket than in most connectors

The SIM interface carries a current profile that is not obvious from the pin descriptions.

State Typical current on VCC Duration Consequence of resistance
Idle Very low, milliamps or less Continuous Negligible drop
Transaction A few milliamps Bursts Small drop
2G transmit slot Can reach the low hundreds of milliamps Millisecond-scale bursts Voltage sag at the card

The 2G transmit burst is the case that breaks designs. The card draws a current pulse while the radio transmits, and the socket resistance sits directly in the supply path. A 20 mV sag is invisible. A 100 mV sag at a 1.8 V supply is a 5.5% deviation, and the interface tolerance is not much larger than that.

This is why a socket that behaves perfectly in a bench test with a 3 V supply and no radio activity can fail in the field on a 1.8 V class card during attach.

The voltage budget arithmetic

ISO 7816 defines three supply classes. The tightest budget is Class C.

Class Nominal VCC Typical tolerance Allowable total drop
A 5 V ±10% 500 mV
B 3 V ±10% 300 mV
C 1.8 V ±10% 180 mV

That 180 mV has to cover the regulator tolerance, PCB trace resistance, any series protection device, the socket contacts, and the return path. The socket is one line item, and it is not the largest one — but it is the one that degrades with use.

Socket resistance Drop at 200 mA burst Share of a 180 mV budget Assessment
40 mΩ (typical new) 8 mV 4% Comfortable
100 mΩ (datasheet max) 20 mV 11% Acceptable
200 mΩ 40 mV 22% Marginal in a tight design
300 mΩ 60 mV 33% Likely to cause resets

The transition from 100 to 300 mΩ is not hypothetical. It is what happens when a contact with thin plating oxidises, or when normal force falls below the level needed to break through the film.

How the measurement is made, and why it changes the answer

Contact resistance measured with a multimeter is not the parameter that matters. The standard method uses a low-level or dry-circuit condition, precisely because a higher test current can break down the film being measured.

Method Test current Open-circuit voltage What it measures
Dry circuit / low level Typically ≤100 mA, often much lower Typically ≤20 mV The film as it actually behaves at signal level
Rated current Up to the contact rating Higher Includes film breakdown; reads lower
Four-wire Kelvin Either Either Excludes lead and probe resistance

Two consequences follow. First, a two-wire measurement includes the probe and lead resistance, which can easily be tens of milliohms — comparable to the value being measured. Four-wire Kelvin is not optional for this parameter.

Second, a dry-circuit measurement will read higher than a rated-current measurement on the same contact, because the measurement current is too low to break the film down. The dry-circuit value is the honest one for predicting whether a SIM will work, since the card’s own current draw starts low and only rises after communication is established.

If a supplier quotes a single number without stating the method, ask which standard was used. IEC 60512-2-1 and EIA-364-23 are the usual references, and neither produces a number that can be compared with the other without knowing the test conditions.

Why the value after cycling is the only one worth specifying

Mechanism of degradation Effect on resistance Timescale
Plating wear at the contact point Gradual rise as base metal is exposed Hundreds to thousands of cycles
Fretting corrosion Sharp rise once oxide builds in a micro-motion zone Thermal cycling, vibration
Normal force relaxation Rise as the beam loses force and contact area shrinks Thermal exposure over years
Contamination film Rise depending on humidity and sulphur exposure Months in the field

All four mechanisms move the number in the same direction, and none of them appear on a virgin-part measurement. A specification that only covers initial value describes a part that has never been used.

The practical requirement is a maximum value stated at the end of the rated cycle life, measured by the dry-circuit method, on every contact, at the specified operating temperature.

Writing the requirement

Element Weak wording Wording that holds up
Value “Contact resistance 100 mΩ max” “Contact resistance shall not exceed 100 mΩ at any contact after the rated cycle life”
Method Not stated “Measured by the four-wire method per IEC 60512-2-1, dry circuit conditions”
Condition Not stated “After 500 mating cycles and 100 thermal cycles between the rated extremes”
Coverage Single value “All contacts individually; the maximum of the population governs”
Temperature Room ambient only “At 25 °C and at the upper rated temperature”

The “all contacts individually” clause matters because a socket can pass on average while one contact is at 400 mΩ. The SIM will fail, because every contact has to work.

Detecting a marginal socket in production

A full four-wire measurement on every contact of every unit is impractical at volume. The practical alternatives are a sampled qualification plus a functional test.

Approach What it catches Limitation
Sampled four-wire on first article and periodic lots Process drift, plating variation Does not catch individual defects
Functional activation with a test SIM Open contacts, wrong pinout, no connection Passes at 300 mΩ if the card tolerates it
Voltage-drop test under load High resistance contacts specifically Requires a load fixture and a threshold
Current-limited activation with margin Marginal contacts that fail under burst Needs a supply set at the tolerance limit

The last one is the most useful in production: run the activation with VCC set at the minimum allowed by the standard, not at nominal. A socket with a 300 mΩ contact will fail there and pass at nominal. That single change converts an invisible margin problem into a testable one.

FAQ

What contact resistance should I require? A maximum stated after cycling, not at zero cycles. A common working figure is 100 mΩ maximum at end of life, measured dry-circuit on every contact.

Why does my measurement differ from the datasheet? Almost always the method. Two-wire versus four-wire, and dry circuit versus rated current, produce very different numbers on the same contact.

Does contact resistance matter if the SIM works? It matters at the tolerance limit. A card supplied at the low end of the voltage range, plus a 200 mA burst, plus a 300 mΩ contact, is a reset that only appears on some units.

How often should resistance be re-measured? At first article, at each significant lot, and after any change in plating supplier or contact material. It should also be part of the qualification after thermal cycling.

Is gold plating enough on its own? No. Gold prevents oxide growth, but it does not prevent wear-through if the plating is thin or the normal force is too low. Both have to be specified.

What causes a sudden jump rather than gradual drift? Fretting corrosion. Micro-motion from thermal cycling or vibration wears through plating locally, and the resistance rises sharply once oxide forms in that spot.