Is it allowed to skip a pin to increase creepage distance?
From IEC-62368 for signals below 30kHz the creepage distance for pollution degree 2 and material group IIIa,IIIb (FR-4 boards are IIIa) at 40V is specified to be 1.1mm. We have a connector that has many BGA pads, and the copper-to-copper distance is 0.635mm:
Putting a slit between is not an option, the only way I could see is to add coating after assembly that would allow to use values for pollution degree 1, which would be 0.16mm, quite a drastic reduction.
One of the engineers mentioned that leaving a pin floating in between also counts towards increasing creepage, but I am not sure about that. The definition of creepage that is used in standard:
Creepage is the shortest path between two conductive parts measured along the surface of a solid insulating material
It would imply that it does not matter if the pin in between is floating or not, but on the other hand, how would the voltage be defined if one of the pins is floating? The voltage could be anything between those two pins! In my mind the pin would have to be physically removed from the connector body for this to be valid, and the pad removed from PCB.
Can anyone comment on the validity of such interpretations of creepage? If you have experience with other standards please share it, IEC is just where I got the numbers from.
3 answers
The following users marked this post as Works for me:
| User | Comment | Date |
|---|---|---|
| Elleanor Lopez | (no comment) | Jun 18, 2026 at 07:43 |
Olin's answer lead me to look for illustrations and I found a reference around creepage definition in IEC 62368 to another IEC norm, namely IEC 60664-1: "Insulation coordination for equipment within low-voltage supply systems". Within this norm, chapter 5.3.3.6 "Creepage distances split by floating conductive parts" specifies: A creepage distance may be split into several parts, made with the same insulation material, including or separated by floating conductors as long as the sum of the distances across each individual part is equal or greater than the creepage distance required if the floating part did not exist.
And the accompanying illustration:
This is exactly what I was looking for, now I now for sure that this approach is recognized by the standard and I can apply it.
0 comment threads
The following users marked this post as Works for me:
| User | Comment | Date |
|---|---|---|
| Elleanor Lopez | (no comment) | Jun 18, 2026 at 07:44 |
I don't have experience with that particular standard, but do with other standards that also specify clearance and creepage distances, like Intrinsic Safety.
I would be very surprised if just leaving the pin unconnected would allow adding the creepage distances on either side. As you say, you can't guarantee what the voltage on the pin will be. It would be floating, and could pick up stray signals by capacitive coupling, static, etc.
Do you really have a 40 V signal? What voltage does the 0.635 mm distance allow? Can you arrange all signals adjacent to the sensitive one to have a sufficiently low maximum guaranteed voltage difference?
I don't see how you can add a coating after assembly to fix this. The material would have to somehow get between a BGA part and the board it's mounted to, while guaranteeing that all voids are filled. That doesn't sound likely.
I have heard several times from engineers that skipping pin is allowed, but none of them have been able to give a reference or explain how it is compliant. I will see if there are any ways to contact someone from IEC to ask.
It might be easier to talk to someone at whatever place you will have certifying this product. If your company builds these kinds of products regularly, they probably already have an established relationship with a testing house. You are their customer, so they are more likely to talk to you when they have a reasonable expectation of getting the testing/certifying work. They are also the people that will be making the judgement of whether skipping a pin is acceptable according to the standard or not.
Added 17 June 2026
I looked in the Intrinsic Safety standard I have used before to see how it defines clearance and creapage. Here is one of the diagrams and accompanying key:
In this case it does say you can add the creepage on both sides of an unused pin, as long as that pin is "not electrically connected". Maybe your standard has a similar diagram.
[ If this is a certification and compliance matter, then you should run your question and your findings by your test lab, certifying body, specialist safety reviwer. ]
Gaps on either side of a floating pin count towards creepage. I've seen it done when Molex KK 396 connector carries mains AC voltage. The distance between adjacent didn't satisfy creepage requirement. We changed the connector from 2-pin to 3-pin with the middle pin floating.
[...] the only way I could see is to add coating after assembly that would allow to use values for pollution degree 1, which would be 0.16mm, quite a drastic reduction.
Underfill under BGAs is an established process. Even if you have voids, the sealed space under the connector could give you pollution degree 1.
I've also seen underfill for high voltage insulation too. It was an analog mux for ultrasound. ±100V excitation pulses went through it.
1 comment thread
1 comment thread