So I ordered a cheap SSR from Aliexpress and it finally arrived. I immediately disassembled it to look what's inside. It wasn't so bad -- BTA40-800, MOC3063.. assuming the components are not counterfeit of course. Pcb has some flux left and triac terminals were bent at some questionable angles and too close to the package.. but I can live with it. What surprised me though is how the triac was mounted.

SSR has an aluminum plate that you attach to the heatsink. I expected to see a screw, but there was none. The package (TOP3) was just soldered to the plate. Triac is 40A-rated, so maybe it's the way to minimize the thermal resistance. But I have questions.

1 Does the TOP3 package even allow that? I realize it's tab must be manufacturer-soldered for a reason, but the amount of the heat applied must be massive. I wonder if the triac is half-cooked already.

2 Will it last? Soldering aluminum is a tricky business, since the material is quite active chemically. So with the frequent heat-idle cycles, is there a danger it will "unstick"? Not unsolder due to the high temperature but just get off?enter image description here

  • 3
    \$\begingroup\$ That just looks wrong. The different coefficients of thermal expansion, the difficulty of soldering aluminum, potential for voids under the package.... \$\endgroup\$
    – John D
    Oct 20 '20 at 20:41
  • 1
    \$\begingroup\$ Does the TOP3 package even allow that? I'm not so much worried about the package but more about the silicon die that's inside it which cannot withstand high temperatures for that long (I'm assuming it takes some time to solder this and/or that there will be a temperature "shock": rapid heating/cooldown). There is no guarantee that soldering like that didn't exceed the manufacturer's recommendations. The fact that reputable manufacturers do not do this (soldering directly on aluminium) should tell you enough. \$\endgroup\$ Oct 20 '20 at 20:44
  • \$\begingroup\$ @Bimpelrekkie Well, it's the package (and the device itself of course) that defines what could be heated to which temp. and for how long. There are plenty of power elements in surface-mount packages like TO252. But yes, it's the die that concerns me, not the epoxy. \$\endgroup\$
    – rapucha
    Oct 20 '20 at 20:58
  • \$\begingroup\$ > The different coefficients of thermal expansion Actually not that much, typical brass and aluminum are about 10% different. \$\endgroup\$
    – rapucha
    Oct 20 '20 at 21:20
  • 1
    \$\begingroup\$ If an Active Solder were used, it could have been "welded" at a lower temperature with ultrasonics. \$\endgroup\$
    – rdtsc
    Oct 20 '20 at 21:40

It's possible that thermal cycling will break the bond, especially if the switching dissipation is high and the frequency is at some optimal spot for fatigue. The triac then overheats and fails 'on' typically.

I don't have a reference for you, but this was a serious failure mode of early SSRs and damaged the reputation of a couple US manufacturers (the SSR supplier and the OEM using them).

That said, this looks fairly decent. SSRs of any manufacture need careful care and feeding to last, especially heat sinking and protection against surges (voltage and current).

Edit: For what it's worth both triac makers ST and WeEn (an NXP China JV) address soldering the tab to the heatsink. WeEn says not to do it. ST says it can be done but they don't qualify their THT parts for that service (so don't do it).

enter image description here

That note is from 2015. It's possible that if they got an order for 100,000 pieces they might bother to qualify their parts for that service.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.