A customer in Shenzhen called us last month with a strange one. Their firmware team had logged intermittent key errors on a 24-panel industrial HMI: about 0.3% of presses on the F3 key registered as a double-click, but only in winter mornings, and only on units that had sat in a cold warehouse for more than 48 hours. Their hardware team had already swapped the microcontroller, re-flashed firmware, and reworked the FPC - all with no change. The actual root cause turned out to be a fiber of packaging dust trapped under one metal dome, sitting on the PCB contact pad. Cold ambient temperature made the dome stiffer, the fiber compressed less, and contact resistance bounced between 50 ohms and 2k ohms depending on how the operator pressed. Once we peeled the dome sheet, blew the pocket with ionized air, and re-seated it, the failure rate went to zero.
Intermittent contact is the single most frustrating failure mode in snap-dome switch design because it never reproduces on the bench at 100% rate. In our application-engineering queue, it is also the most self-inflicted: about 70% of cases we debug trace back to five physical root causes that were visible at the design or assembly stage, not to the dome itself being defective.
Opened metal dome panel - the kind of teardown that usually reveals the root cause within minutes.
The five physical causes of intermittent contact
Before you touch firmware, isolate which of these five mechanisms you are dealing with. Each has a different signature.
1. Dust or fiber contamination under the dome
The most common cause by a wide margin. A single 50 micron fiber on the contact pad changes the dome-to-pad contact from a clean metal-to-metal click to a bouncy partial contact. Symptoms: the key works when you press hard, fails when you press lightly; works after you tap the panel; works better in a warm office than on a cold factory floor.
How to confirm: Peel the dome sheet array off the PCB under a magnifier or stereo microscope. Look for dark speckles on the center contact pad. If you see them, the fix is ionized-air blowdown and re-lamination in a cleanroom environment. If your assembly line is not cleanroom-rated, this will keep happening.
2. Cracked or fatigued dome
A snap dome is a stainless-steel disc that snaps through under force. Once it accumulates micro-cracks - typically after 60-80% of rated life, or from a single over-torque press - the click becomes mushy and contact becomes intermittent. Symptoms: the key feels different from the others; actuation force has dropped visibly; the failure gets worse the more you press.
How to confirm: Peel off the dome and look at the outer rim under 10x magnification. Radial cracks in the SUS301 rim are the smoking gun. This is a lifecycle problem, not a dust problem - the fix is either a stronger dome variant or a derated actuation force.
3. PCB pad misalignment or via too close to the pad
If the via that drains the contact pad sits inside the center pad itself, solder wicking creates a bump under the dome that prevents full snap-through. Symptoms: the key feels mushy instead of crisp; intermittent contact appears only on one specific key out of a matrix.
How to confirm: Compare the failing key pad layout to a working key. Per our own design rules, the center pad diameter should be 2.5 to 3.0 mm and any via should sit at least 1.0 mm outside the pad edge. If the via is inside the pad, the PCB Gerber is the problem, not the dome.
4. Worn-out or wrong-force dome
We see this in consumer remotes that ship with a 300g dome but the end-user membrane overlay needs a 400g dome to actuate reliably through the overlay thickness. The dome works on a bench test but feels weak in the finished device. Symptoms: contact registers only on the sharpest part of the click, not the return; customers report they have to press really hard.
How to confirm: Measure actuation force on a force-curve tester. If you specified 300g but the application needs 400g through a 1.5 mm silicone overlay, re-spec the dome.
5. Adhesive transfer or PSA peeling off
The pressure-sensitive adhesive that holds the dome sheet to the PCB can, over time and under heat, transfer onto the pad and insulate the contact. Symptoms: the key works when new, starts failing after 3-6 months in a hot environment; the dome sheet itself is loose when you peel it.
How to confirm: Peel the dome sheet. If the contact pad has a hazy film of glue on it, the PSA has outgassed or sheared. This is a material-spec problem - use a higher-temperature acrylic PSA for above-60C applications.
Four-legged snap domes - the leg geometry matters more than engineers think for contact stability.
The 10-minute debug procedure
When a customer sends us a failing unit, we run this exact sequence. It isolates the cause without needing an oscilloscope on day one.
- Listen and feel. Actuate the suspect key ten times. Is the click crisp like the neighboring keys, or mushy? Mushy points to cracked dome or wrong force. Crispy-but-intermittent points to dust or pad alignment.
- Swap the dome sheet. Put a known-good, brand-new dome sheet on the same PCB. If the failure follows the sheet, it is the dome. If the failure stays on the same key position, it is the PCB pad.
- Peel and inspect under 10x. Look for dust fibers, glue residue, or cracked rim.
- Measure force-curve. On a calibrated snap-dome tester, record actuation force, travel, and contact resistance. Compare to the datasheet. A force that has dropped more than 20% from spec means fatigue.
- Re-flow the pad. If the contact pad has oxidation or flux residue, a light re-flow or IPA scrub usually restores contact.
How to prevent it in the first place
Ninety percent of intermittent-contact failures are designed in, not assembled in. The cheapest fixes happen at the Gerber and BOM stage:
- Specify a 1.0 mm minimum via-to-pad clearance, and never place a via inside the center contact pad.
- Use a 2.5 to 3.0 mm center pad for 8 to 12 mm domes, with a 0.3 to 0.5 mm clearance ring around the outer ring contact.
- Account for overlay stack-up when choosing actuation force. A 1.0 mm silicone overlay adds roughly 80g of felt force - over-specify accordingly.
- Ask your dome supplier for a die-cut module with integrated PSA that matches your PCB surface finish (ENIG vs HASL). We see more intermittent-contact cases from mismatched PSA surface energy than from dome failures.
- For outdoor or high-humidity products, specify gold-plated contacts. Tin oxidizes and produces exactly this symptom after 6-12 months.
When it is actually firmware, not hardware
Worth saying because we have been on the wrong end of this call: about 10% of the time, intermittent contact is a debouncing problem in the microcontroller firmware. If the debounce window is shorter than the dome typical 5 to 15 ms bounce time, a single press can register as two. The way to tell apart: probe the contact pins with an oscilloscope. If you see clean single contacts on the scope but double events in software, the dome is innocent and the debounce routine needs lengthening. If you see bounce or intermittent opens on the scope itself, it is one of the five hardware causes above.
Need a second set of eyes?
If you are debugging an intermittent-contact problem on a consumer electronics dome, automotive dome, or silicone button assembly, send us the failing unit and we will run the full tear-down and force-curve test at our Qingdao lab. We return a measured report - not a guess. Typical turnaround is 5 working days from receipt.
Get in touch with Shanyo engineering team to start a failure analysis. We will tell you honestly if the root cause is on our side or on your PCB design side.
- John Smith, Senior Application Engineer, Shanyo
A customer in Shenzhen called us last month with a strange one. Their firmware team had logged intermittent key errors on a 24-panel industrial HMI: about 0.3% of presses on the F3 key registered as a double-click, but only in winter mornings, and only on units that had sat in a cold warehouse for more than 48 hours. Their hardware team had already swapped the microcontroller, re-flashed firmware, and reworked the FPC - all with no change. The actual root cause turned out to be a fiber of packaging dust trapped under one metal dome, sitting on the PCB contact pad. Cold ambient temperature made the dome stiffer, the fiber compressed less, and contact resistance bounced between 50 ohms and 2k ohms depending on how the operator pressed. Once we peeled the dome sheet, blew the pocket with ionized air, and re-seated it, the failure rate went to zero.
Intermittent contact is the single most frustrating failure mode in snap-dome switch design because it never reproduces on the bench at 100% rate. In our application-engineering queue, it is also the most self-inflicted: about 70% of cases we debug trace back to five physical root causes that were visible at the design or assembly stage, not to the dome itself being defective.
Opened metal dome panel - the kind of teardown that usually reveals the root cause within minutes.
The five physical causes of intermittent contact
Before you touch firmware, isolate which of these five mechanisms you are dealing with. Each has a different signature.
1. Dust or fiber contamination under the dome
The most common cause by a wide margin. A single 50 micron fiber on the contact pad changes the dome-to-pad contact from a clean metal-to-metal click to a bouncy partial contact. Symptoms: the key works when you press hard, fails when you press lightly; works after you tap the panel; works better in a warm office than on a cold factory floor.
How to confirm: Peel the dome sheet array off the PCB under a magnifier or stereo microscope. Look for dark speckles on the center contact pad. If you see them, the fix is ionized-air blowdown and re-lamination in a cleanroom environment. If your assembly line is not cleanroom-rated, this will keep happening.
2. Cracked or fatigued dome
A snap dome is a stainless-steel disc that snaps through under force. Once it accumulates micro-cracks - typically after 60-80% of rated life, or from a single over-torque press - the click becomes mushy and contact becomes intermittent. Symptoms: the key feels different from the others; actuation force has dropped visibly; the failure gets worse the more you press.
How to confirm: Peel off the dome and look at the outer rim under 10x magnification. Radial cracks in the SUS301 rim are the smoking gun. This is a lifecycle problem, not a dust problem - the fix is either a stronger dome variant or a derated actuation force.
3. PCB pad misalignment or via too close to the pad
If the via that drains the contact pad sits inside the center pad itself, solder wicking creates a bump under the dome that prevents full snap-through. Symptoms: the key feels mushy instead of crisp; intermittent contact appears only on one specific key out of a matrix.
How to confirm: Compare the failing key pad layout to a working key. Per our own design rules, the center pad diameter should be 2.5 to 3.0 mm and any via should sit at least 1.0 mm outside the pad edge. If the via is inside the pad, the PCB Gerber is the problem, not the dome.
4. Worn-out or wrong-force dome
We see this in consumer remotes that ship with a 300g dome but the end-user membrane overlay needs a 400g dome to actuate reliably through the overlay thickness. The dome works on a bench test but feels weak in the finished device. Symptoms: contact registers only on the sharpest part of the click, not the return; customers report they have to press really hard.
How to confirm: Measure actuation force on a force-curve tester. If you specified 300g but the application needs 400g through a 1.5 mm silicone overlay, re-spec the dome.
5. Adhesive transfer or PSA peeling off
The pressure-sensitive adhesive that holds the dome sheet to the PCB can, over time and under heat, transfer onto the pad and insulate the contact. Symptoms: the key works when new, starts failing after 3-6 months in a hot environment; the dome sheet itself is loose when you peel it.
How to confirm: Peel the dome sheet. If the contact pad has a hazy film of glue on it, the PSA has outgassed or sheared. This is a material-spec problem - use a higher-temperature acrylic PSA for above-60C applications.
Four-legged snap domes - the leg geometry matters more than engineers think for contact stability.
The 10-minute debug procedure
When a customer sends us a failing unit, we run this exact sequence. It isolates the cause without needing an oscilloscope on day one.
- Listen and feel. Actuate the suspect key ten times. Is the click crisp like the neighboring keys, or mushy? Mushy points to cracked dome or wrong force. Crispy-but-intermittent points to dust or pad alignment.
- Swap the dome sheet. Put a known-good, brand-new dome sheet on the same PCB. If the failure follows the sheet, it is the dome. If the failure stays on the same key position, it is the PCB pad.
- Peel and inspect under 10x. Look for dust fibers, glue residue, or cracked rim.
- Measure force-curve. On a calibrated snap-dome tester, record actuation force, travel, and contact resistance. Compare to the datasheet. A force that has dropped more than 20% from spec means fatigue.
- Re-flow the pad. If the contact pad has oxidation or flux residue, a light re-flow or IPA scrub usually restores contact.
How to prevent it in the first place
Ninety percent of intermittent-contact failures are designed in, not assembled in. The cheapest fixes happen at the Gerber and BOM stage:
- Specify a 1.0 mm minimum via-to-pad clearance, and never place a via inside the center contact pad.
- Use a 2.5 to 3.0 mm center pad for 8 to 12 mm domes, with a 0.3 to 0.5 mm clearance ring around the outer ring contact.
- Account for overlay stack-up when choosing actuation force. A 1.0 mm silicone overlay adds roughly 80g of felt force - over-specify accordingly.
- Ask your dome supplier for a die-cut module with integrated PSA that matches your PCB surface finish (ENIG vs HASL). We see more intermittent-contact cases from mismatched PSA surface energy than from dome failures.
- For outdoor or high-humidity products, specify gold-plated contacts. Tin oxidizes and produces exactly this symptom after 6-12 months.
When it is actually firmware, not hardware
Worth saying because we have been on the wrong end of this call: about 10% of the time, intermittent contact is a debouncing problem in the microcontroller firmware. If the debounce window is shorter than the dome typical 5 to 15 ms bounce time, a single press can register as two. The way to tell apart: probe the contact pins with an oscilloscope. If you see clean single contacts on the scope but double events in software, the dome is innocent and the debounce routine needs lengthening. If you see bounce or intermittent opens on the scope itself, it is one of the five hardware causes above.
Need a second set of eyes?
If you are debugging an intermittent-contact problem on a consumer electronics dome, automotive dome, or silicone button assembly, send us the failing unit and we will run the full tear-down and force-curve test at our Qingdao lab. We return a measured report - not a guess. Typical turnaround is 5 working days from receipt.
Get in touch with Shanyo engineering team to start a failure analysis. We will tell you honestly if the root cause is on our side or on your PCB design side.
- John Smith, Senior Application Engineer, Shanyo
