Arcing between drain/source/gate pads of TO220 FET in Offline Flyback SMPS?

Hello,
What are the chances of getting arcing between in-line TO220 pads on a 60W Offline Flyback SMPS FET. (FR4 PCB coated with solder resist)
Our contractor has designed a 60W offline (240VAC) Flyback LED driver (PFC’d) for us. It is for outdoor use.
The footprint for the TO220 primary FET is just three PTH pads in a line, ie, no bending out of the middle leg of the TO220 to facilitate greater clearance between the Drain pad and the other two pads. The PCB is not conformal coated, but just normal solder resist coated.
There is just 0.7mm of clearance between drain pad copper and source/gate pad copper.
The PCB Enclosure is plastic, and “snap-fit”, …in other words, it has “seam” lines through which moisture laden air could enter the PCB enclosure.
The PCB enclosure will sit inside an aluminium enclosure, which does have a rubber gasket around its closure-contact surfaces….so I suppose this is some surety that moisture laden air will not enter into the aluminium outer enclosure…however, we believe that moisture can actually “breathe” through such rubber seals?
Anyway, do you think that the 0.7mm clearance between the drain pad and other pads will cause problems due to moisture ingress? We are going to do a test whereby we spray a film of water over the TO220 footprint, and then power up the LED driver. However, we wonder if tap water is going to give a realistic indication of whatever the atmospheric moisture might be like?
We don’t want a footprint where the drain pad is bent out, because its an extra assembly cost.
What will happen in our “sprayed tap water” test, and will it be realistic to what real atmospheric moisture will be like?


Thread #19 of the following says that arcing is a possibility between the pads, but how likely is this in reality?...
https://www.edaboard.com/showthread...FET-and-legs-not-bent-Too-close-to-each-other
…according to this, the TO220 leg spacing on the body of the TO220 is insufficient to meet standards, and so flashover may occur here?...in this case, this means that it would be a waste of time to bend out the middle (drain) leg of the TO220.
Post #10 of the following shows a good way to maximise the clearance between in-line TO220 pads….
https://www.edaboard.com/showthread...learance-between-drain-and-source-pads-on-PCB

according to this, the TO220 leg spacing on the body of the TO220 is insufficient to meet standards, and so flashover may occur here?...in this case, this means that it would be a waste of time to bend out the middle (drain) leg of the TO220.

Click to expand...

The tab clearance of 1.27 mm is surely sufficient for a 230V switcher (Vpk < 300V), but the inline pad clearance isn't. If flashover occurs, it happens at the pads, in case of failure, you'll tell the customer "I know the pad clearance is too small, but the package clearance isn't much better"?

I experienced cheap Chinese switchers failing with drain pad flashover and notice that quality SMPS are equipped with milled slits between the pads.

Thansk, we looked into a TO220 "leg beding tool"

https://www.olamefusa.com/e_spec_handtools.htm#PN505015

..and we came up with the attached alternative footprint....it has 0.3mm cutout channels between drain and source and drain and gate pads.....these channels are only 0.3mm wide....do you think the PCB making machines will be able to produce such a narrow slot?

As said, milled slits are standard for HV circuits and even used for standard (non-bended) TO220 footprint. 0.3 mm isn't feasible and unreasonable small for the shown footprint. Ask your PCB house what they can provide.

Thanks, do you believe a milled slit would need to be at least 1mm wide?
Also, is it unreasonable to have it "bent" like i have?....i did try to make it up out of straight sections.

Would it kill you to use one of the wider-spaced TO-2xx package
styles that are meant for exactly this problem?

Thanks, yes, i now think the cost of the TO247 would easily be made acceptable by the cost of bending the legs of a TO220.

I just wonder if we will still be ok attaching our TO220 heatsink to a TO247 package?.....(fischer SK 469 35 STS)
https://www.fischerelektronik.de/we...ded heatsinks for PCB mounting/VA/SK46935STS/

Smaller slits than 1 mm are possible, the tools are also used for throughplated slits. But milling speed is low and tool wear higher.

Thanks, I have replaced the “TO220 with bent legs” with a TO247 FET with straight legs……but now I find I need to add slits behind the gate and source pads because they are too near to the heatsink as it rests on the PCB behind the TO247. (it’s the TO247’s own heatsink, to which it is screwed). The slits are 1mm wide (as in the attached)…do you think these slits are manufacturable?...are they too narrow?