Thermal or solid pad connection to ground plane, does it matter?

[Nocaster Cat]

Title says it all, thanks.

[mth5044]

Thermal pads are easier to solder since they keep the heat from spreading around the ground planes as quickly as it would if it were just a hole.

That means less time soldering and less time with your tip on a component.

[armdnrdy]

Everything that Matt said. 100%

Let me add...a solid ground pad on a ground plane isn't really a pad at all is it?

It's a hole in the ground plane.

You will spend a lot of time (and solder) trying to make the solder connection on a hole in a ground plane.

The results don't look too tidy either! 

[R.G.]

Let me add...a solid ground pad on a ground plane isn't really a pad at all is it?

It's a hole in the ground plane.

You will spend a lot of time (and solder) trying to make the solder connection on a hole in a ground plane.

The results don't look too tidy either! 

Well, one could opine that it's really an opening in the solder mask around a hole in the ground plane. That actually helps.

But as you guys noted, results without thermal breaks are *ugly* and in fact would be rejected for high volume manufacture because the wave-solder percent yield would be in the single digits.

[Nocaster Cat]

Awesome, thanks guys!

[chaseblissaudio]

A couple years ago I was designing a PCB for a company that was making agricultural drones.  It was a pretty complex project, with lots of RF considerations.  A section of the PCB had some really sensitive GPS related parts on it, and the RF engineer was absolutely insistent that the shield pin holes for that section *not* have thermal relief.    I didn't understand all the specifics, but apparently it was so that there would be a ground shield around that entire section with no gaps of any kind for RF to leak through. 

That said, it was a 6 layer board, with lots of ground planes and it was damn near impossible to solder that shield down. 

[R.G.]

... and the RF engineer was absolutely insistent that the shield pin holes for that section *not* have thermal relief.    I didn't understand all the specifics, but apparently it was so that there would be a ground shield around that entire section with no gaps of any kind for RF to leak through. 

One hopes the RF engineer knew his stuff. RF will leak through a hole which is equal to or larger than a quarter wavelength at the frequency.  A quarter wavelength exactly will behave as a "slot antenna". Of course, the main physical dimension of the gaps around a pad for a thermal relief are so small that it's hard for me to credit that even GPS stuff would see it as an opening.

I would worry more about the dimension of the cracks between the legs of the ground shield. Those would be much longer and hence leakier at much lower frequencies. Effective EMI shielding needs to bet tagged together every quarter inch or so, or better yet, forced together with some kind of compliant and conductive gasket to eliminate these leaks.

That said, it was a 6 layer board, with lots of ground planes and it was damn near impossible to solder that shield down.

Yep. Damn near impossible is what I'd guess. It's not all that easy to get good solder joints on the ground legs of decoupling caps on four-and six-layer motherboards even *with* thermal reliefs.

So for audio pedal work... use thermal reliefs. Unless your guitar puts out sub-millimeter-wave audio.     

... quick! How short does an 0.010" E string have to be and what tension does it need to have an output in millimeter-band RF?

[Transmogrifox]

... quick! How short does an 0.010" E string have to be and what tension does it need to have an output in millimeter-band RF?

How thick do the leather bands joining plates of armor need to be in order to prevent mortal wounding when a standard unicorn thrusts it's emerald horn between the plates at max velocity?

[armdnrdy]

How thick do the leather bands joining plates of armor need to be in order to prevent mortal wounding when a standard unicorn thrusts it's emerald horn between the plates at max velocity?

As a former student of Unicornology...It is common knowledge that a Unicorn's horn is not made of Emerald.

[PRR]

> How short does an 0.010" E string have to be and what tension does it need to have an output in millimeter-band RF?

Speed of sound in a steel string is a fishy thing, but clearly MUCH slower than speed of electromagnetic waves in space.

Therefore the guitar string must be MUCH shorter than these "millimeter waves".

If you literally mean MILLIMETER wave, 300GHz, then the string is MUCH shorter than the width of a fret (~~1mm).

[chaseblissaudio]

One hopes the RF engineer knew his stuff. RF will leak through a hole which is equal to or larger than a quarter wavelength at the frequency.  A quarter wavelength exactly will behave as a "slot antenna". Of course, the main physical dimension of the gaps around a pad for a thermal relief are so small that it's hard for me to credit that even GPS stuff would see it as an opening.

I would worry more about the dimension of the cracks between the legs of the ground shield. Those would be much longer and hence leakier at much lower frequencies. Effective EMI shielding needs to bet tagged together every quarter inch or so, or better yet, forced together with some kind of compliant and conductive gasket to eliminate these leaks.

It was very likely overkill and over-engineered.   I wasn't in a position to question anything.  The ground shield chassis thing actually extended all the way down to the PCB where it met with solder paste lines (in between the ground shield pins) - so, it was completely enclosed by ground except for a few vias on the inside of the shield that traveled to traces on inner planes.  I think it was just the top layer copper that didn't have thermal reliefs on the ground plane (that probably helped make it from impossible to *near* impossible to solder it down).  There were 3 or 4 traces that entered this section of the PCB, but they traveled on an inner plane.  That section was crazy shielded.

BTW, I've learned more from you about electronics than anyone in my life (as a forum lurker) - so it's fun to respond to one of your posts.  If I wanted to learn about relays, for example, I'd typically just google "R.G. Keen relays" and read your posts.  I honestly can't thank you enough for all of that you've given to everyone.