Electronics > Projects, Designs, and Technical Stuff

Thermal relief vs direct connection for heat dissipation

<< < (3/3)

T3sl4co1l:
Vias are plated through holes, that's how they work. :)

If you mean as a THT pad, the only difference is whether you stick something in it before soldering.  Which is going to be challenging under 0.5mm i.d., and I think the smallest you usually see is 0.8.  Hm, there are 1.27 and 1mm pitch headers available in THT, I forget offhand what hole size they use -- that's probably the practical limit.

And yeah, most any kind of soldering on multilayer gets difficult without extreme heat (iron >400C?) or preheating.  Even if you don't have a proper hot air machine or board preheater, a hot air gun can be used to help out -- keep it at enough distance, or keep it moving, so it doesn't burn anything: typical one will be MUCH hotter than anything on a PCB should get, so don't use it up close and still.  Be patient, it takes minutes of heating for the board and components to come up to temperature.  In this way, you can even do SMT (reflow) soldering with a generic hot air gun, but it's a heck of a lot easier to screw up is the thing (sit still for just a little too long).

Tim

Geoff-AU:

--- Quote from: T3sl4co1l on May 16, 2022, 09:14:02 am ---Vias are plated through holes, that's how they work. :)

--- End quote ---
Of course.  I was tongue in cheek, you could technically call every PTH a via but I wouldn't.  For me, PTH is something you plan to stick a component in and apply solder.  A via's primary job is swapping layers for a trace.  Mechanically, they are interchangeable... it's purely semantics.  Test points are for probing or soldering things to temporarily, of course a via can be used in a pinch but test points come with silkscreen labels which makes finding it easier.  Anyway, it becomes a religious argument after a while.

I've had boards that resisted even preheating with hot air or a radiant PCB heater from underneath.  Lead-free solder doesn't help.  Some boards just want you to curse them enthusiastically before they let go.  Sometimes getting the header out is OK but you'll never wick the solder out.  Need to find a PCB drill and a hand chuck.... more cursing.

ajb:

--- Quote from: T3sl4co1l on May 05, 2022, 08:00:24 pm ---So what good are they in soldering?  Soldering is done at much higher heat density.  If the spokes drop say 1 K/W, that's inconsequential for the say 5W you might get out of a D2PAK; but if you're putting 50W into it by soldering iron, or who knows what by hot air -- now you're talking the difference between the joint reaching MP (180 to 230°C depending on alloy) or not at all (versus a cold board / no preheater, and a hot end at 200-300°C say).
--- End quote ---

To add to this, in terms of helping to intuitively understand the thermal dynamics, it's useful to remember that thermal behavior is analogous to electrical behavior. 

Heat flux ~= current
Temperature ~= voltage
Thermal resistance ~= electrical resistance
Thermal mass ~= capacitance

And with those equivalencies, Ohm's law holds.

So if you consider that the thermal relief spokes add a fixed amount of thermal resistance to the path between the package and the ambient via the PCB, then it's easy to see that the temperature rise across the thermal relief will be proportionally low with the relatively low heat flux induced by the package.  On the other hand, the much higher heat flux available in the soldering process results in a much higher temperature difference.  Of course reality is more complicated than this, because you have multiple heat paths, and there is some thermal resistance in the soldering equipment as well as time component via the thermal mass of the various components, but you can use the same basic principles to expand the model to whatever level of complexity you need. 

T3sl4co1l:
Yup.  Thermal flow has the advantage that there's no wave effects*, just diffusion, so it's very easy to calculate.  Even in bulk, it's just applying Ohm's law to volumes; in general you might need to integrate over many small volumes, but a lot of practical problems can easily be broken down to flat or linear geometry that can simply be added up.  And I mean adding up by hand, few pieces -- like spokes and pours.

So like with spokes, the pad can be considered ideal (unipotential), the spokes are just rectangular strips, and the spoke-pour transition is complex (think of it as a transition between 1-D trace and 2-D pour) but we can approximate the transition region as a circular section with some fixed resistance and skip having to compute it every time.

The actual value of the transition region will go something like inverse or logarithmic with radius, in turn is given by the scale factor of the spoke width and spacing.  So it only needs to be computed once, then scaled appropriately.

*Not until such energies as explosive shockwaves; or in free space, relativistic energy densities.  Indeed, it turns out nothing is linear, there are always limits; but obviously this particular limit is very safe to ignore for our purposes. ;D

Tim

Navigation

[0] Message Index

[*] Previous page