Maximum allowable voltage between PCB traces

Question

I have some confusion regarding the guidelines for clearance between surface uncoated PCB traces (specifically component pads). IPC-2221 guidelines and other guidelines and standards indicate spacings that don't seem to make much sense.

I was intending to use a connector with 48V DC until I realized that it was rated for 30V maximum, on trying to figure out why this was I realized that the pads were at a 1.25mm pitch, which left less than 0.6mm between pads, thus violating the indications of IPC-2221. The connector pins themselves are spaced at 2.5mm leaving plenty of clearance otherwise.

These indications jump from 0.1mm for V<=30V to a minimum of 0.6mm for 30V < V < 150V (other standards indicate 0.4mm for V<50V, I have no access to EN 60664-1). Such jumps seem rather excessive.

I understand that air can withstand a field of 3kV/mm (i.e, 300V @ 0.1mm), so clearly this must be a surface contamination and creepage requirement more than a clearance requirement. Or perhaps it is also meant to take into account field increases due to 90° corners. But I even found a military test document that recommends as little as 0.13mm for 100V for altitudes of <3000ft.

I am pretty sure that many manufacturers go below the IPC recommendations (e.g., I am using a DC-DC converter rated for 60V input in a BGA package with a 1mm ball pitch and a maximum ball dimension of 0.7mm which leads to <0.3mm pcb spacing), the manufacturer does not even suggest using a sealer for the underside of the IC.

Besides subjecting the assembled board to (expensive) testing, is there any way to responsibly specify a smaller distance between traces?

---

Some related questions:

• Low Voltage PCB Creepage Requirements (No answer)
• Creepage distance over pcb
• PCB track clearance (Conformal coating)

Answer 1

I've done a lot of research on this and unfortunately it's not as straight-forward as you would think. Everyone refers to IPC2221A because that's the defacto standard but the tests that UL performed were limited and subsequent testing has found that the outter/inner layer differences are basically negligible despite the standard showing a large difference.

The pollution degree and type of insulation of the board plays a huge role and can vary the clearance numbers by a factor of ~6x when combined. According to the first table below, 0.5mm will handle 150V. Other tables say different things (see second table). Also, keep in mind that when you design a board to work at 150V, UL tests at a much higher voltage to ensure compliance. There's a huge safety margin built-in to these standards. The last test I was involved in ran at 1000V plus twice the operating voltage of 138V peak which gives around 1280V. The main reason they do these tests is in the event of a large voltage surge on the line from an exploding transformer or similar. They don't want to kill someone who is holding the device. If your device won't put anyone in harm's way, you have much less to worry about.

It is my opinion that you're fine with that connector if you don't need to pass any major standards.

At the end of the day, you need to test your unit in the environment you plan on using it in to determine if it will work.

Answer 2

The machines that manufacture the traces, pads, etc. are not 100% accurate. They have tolerances likely specified as +/-d, d being some distance. These tolerances are likely accounted for in the conservative figures for clearances. I don't think this is a complete answer to the question, but it is one of the factors for consideration.