SOT23-6 PCB deadbug epoxy?

[Peacefrog]

I need to deadbug an SO23-6 part to a PCB to add ESD protection to a USB port. Can anyone recommend an suitable epoxy to affix the part to the board - preferably with a fine tip?

[ataradov]

Why do you need epoxy? Once the pins are soldered, it should be held fine just from that.

If you really need to fix it, then any epoxy would work, all you need is a tiny dot. But I usually try to identify a couple pins that would align with traces on the board and place the part to those pins first (at least ground should be easy enough to match). Then run wires to the rest. Use exiting solder mask as insulator.

[Peacefrog]

Thanks. There aren't really any pads that align - I might be able to get away with ground, but I'd sooner secure the package to avoid ripping off the pad if disturbed.

[ataradov]

It is hard to tell without looking at the board and existing traces, but I really don't understand how that is possible. If you have those USB traces running parallel for 1/2", you should be able to match SOT23-6 pins to those traces in some orientation. Can you post pictures of that part of the board and device datasheet?

In any case, after 6 pins are soldered down, it would be very hard to disturb it.

[jpanhalt]

Any "consumer" epoxy will work.  You may want to add "fill" such as fumed silica to thicken it.  Myself?  I would use hot melt.

[Peacefrog]

- I'm implementing a fix for a design flaw in a Nord G2 synthesizer whereby the USB chip is easily blown due to lack of ESD protection (see the service manual here: https://bespin.federalproductions.com/mycloud/Organ%20Documents/Owners%20and%20Service%20Manuals/Nord-Clavia/nord%20modular%20G2%20service%20manual%20v1.2.pdf)

- Rather than use the 1N4004 diodes suggested, I'd like to wire in an IC that's designed for this purpose (https://www.digikey.com/en/products/detail/USBLC6-4SC6/497-4492-2-ND/725216)

- Given the disparate points at which VBUS, GND, D+ and D- are available, and the small size of the package, I'd like to epoxy the IC down and wire to it - I'm also only going to be using four of the six pins)

Maybe there's a better option, but it'd be great to know if there's a suitable epoxy I could use (say, versus superglue). Thanks!

[KE5FX]

Try UV cure adhesive (check your local fly-fishing outlet, or buy the suspiciously-strong Chinese stuff from Amazon.)  You won't go back to epoxy or cyanoacrylate.

I use a blue laser pointer to cure it instantly, but that is best done only if you have access to good eye protection at both ends of the spectrum.

[ataradov]

Any epoxy or other  glue would work.

In the document the advice is tailored for discreet components. With a single chip, I personally would put it before EMI filters, close to the actual USB connector.

[T3sl4co1l]

You want to use a high-Tg epoxy, so that it doesn't turn to goo (or worse) at soldering temperature.  Hardware store 1-minute-cure stuff seems to be rather poor at this.  Probably filler helps, but also the slower cure types seem to do better (30 minute, "2 ton", etc.).

Also just bog standard JB Weld, it's rated for quite high temperatures and is already filled.  Downside, the filler is metal; it's not conductive over any sample I've seen, but it's maybe not a good look on electronics.  But any of the comparable ceramic-filled types will do, for sure.

And yeah, UV cure epoxy can do quite well.  Soldermask itself is the same stuff.  You'll need a UV LED pen or other light, or enough sunlight exposure, to cure it.  Not sure how well it'll cure underneath a black component, maybe it just takes longer?  Can test it a few times -- if it snaps off from reasonable forces, or like bubbles up under soldering heat, just clean it off and try again.

As for soldering, I don't have a problem with that, tacking the part to a ground pad that is.  If the pad pulls off that easily, it was probably too high electrical impedance as well, to do the job!  Preferably it will be tied directly to ground plane, which the USB traces will also be routed over (for immunity as well as controlled impedance).

Tim