Mr Fryettes opinion on pcb vs hw

[Ufoscorpion]

 

[sdn25]

This is beautiful.

 

[PelliX]

Hmmm, heatsink in an RoHS board. Not sure what the LM is powering, but if that heatsink is truly necessary, then I'd have kept it off the board. Is CR4 is a diode, by the way?

Serious points for silkscreening the component values on the board though! Appreciated by techs the world over!

 

[Chris-in-LA]

Is that real?

 

[Pete Farrington]

Hmmm, heatsink in an RoHS board

What’s the issue?

if that heatsink is truly necessary, then I'd have kept it off the board

There’s a couple of potential issues -
1/ if the heatsink has no mechanical support other than the component it’s cooling, then over time, the component legs will get stressed and may fatigue fracture.
In this case, it looks like the heatsink may be soldered to the board somehow.
Kinda like the support brackets on board mounted pots.
Another option is to lay it flat and screw it down.

2/ if the thermal circuit and environment is such that the heatsink gets too hot, the heat from it can damage the board.
Hopefully it’s been designed to all but eliminate the likelihood of that occurring (ie to run pretty cool).

 

[PelliX]

1/ if the heatsink has no mechanical support other than the component it’s cooling, then over time, the component legs will get stressed and may fatigue fracture.

Agree, though those LM's should be fine supporting a lightweight heatsink like that. Equally, heating the board will cause stress on the solderjoints and the board itself eventually. Maybe I was too harsh and the LM is near idling - no idea what it's driving. Probably just some switching electronics.

In this case, it looks like the heatsink may be soldered to the board somehow.
Kinda like the support brackets on board mounted pots.

Yup, a real pain to remove, those, often.

Another option is to lay it flat and screw it down.

Or angle the component parallel with the board and have the heatstink sticking up in the air, indeed.

Still, I appreciate anyone who has the common courtesy to silkcreen the values on the board.

 

[Ufoscorpion]

Is that real?

It’s real Stevie Freyette is not only a guitar amplifier guru he has a great sense of humour, a legend .

 

[Ufoscorpion]

It’s real Stevie Freyette is not only a guitar amplifier guru he has a great sense of humour, a legend .

He’s pretty tasty on the guitar aswell .

 

[Pete Farrington]

though those LM's should be fine supporting a lightweight heatsink like that

You'd think that, but it's not been my experience. Another guy on my team assumed it'd be fine, and had the embarressment and hassle of dozens of them failing in the field, and our field support techs had to jet around the world replacing boards with a reworked version where the heatsink was screwed down.
Maybe there's something about that type of component leg that makes them prone to fatiguing.

 

[PelliX]

You'd think that, but it's not been my experience. Another guy on my team assumed it'd be fine, and had the embarressment and hassle of dozens of them failing in the field, and our field support techs had to jet around the world replacing boards with a reworked version where the heatsink was screwed down.
Maybe there's something about that type of component leg that makes them prone to fatiguing.

Just curious; what kind of application are we talking about here? Any extreme conditions like heat or shock?

 

[Pete Farrington]

Just curious; what kind of application are we talking about here? Any extreme conditions like heat or shock?

No, it was lab equipment. As I recall, mostly it was fauly as soon as the customer powered it up, so the damage (vibration fatigue to the regulator legs I suppose) must have occured during shipping.

 

[PelliX]

I'll bear that in mind as I've got a project coming up for which I'm going to have to some board design and a couple of LM's will be on BOM, I'm pretty sure.

Fair deal of heat and a bit of vibration. Noted.