Thermal Relief Pad

[[email protected]]

I'm having a PSU board laid out at the moment that is conducting some
reasonable current (8A). The ground connections have been made to the
ground plane with what i believe is called thermal relief - there
isn't a full ring of copper around the hole to connect it to the
ground plane, instead there are 4 thin tracks - to help when soldering
the thing together. I have a couple of questions about this.

1.) Does anyone have experience of soldering devices in without this
thermal relief? If its not much worse then this is surely the way to
go to ensure enough current handling / low resistance to ground.

2.) If its a royal pain to solder is there a better alternative - for
example thickening one of the spurs to the thickness of a track.

Thanks

Andrew

 

[Eeyore]

I'm having a PSU board laid out at the moment that is conducting some
reasonable current (8A). The ground connections have been made to the
ground plane with what i believe is called thermal relief - there
isn't a full ring of copper around the hole to connect it to the
ground plane, instead there are 4 thin tracks - to help when soldering
the thing together. I have a couple of questions about this.

1.) Does anyone have experience of soldering devices in without this
thermal relief?
Yes.

If its not much worse then this is surely the way to
go to ensure enough current handling / low resistance to ground.

It is MUCH more difficult. There's a higher chance in particular of a cold/dry
joint.

2.) If its a royal pain to solder is there a better alternative - for
example thickening one of the spurs to the thickness of a track.

If you're worried, alter the pattern of it. 8 Amps is nothing though.

Graham

 

[John Larkin]

I'm having a PSU board laid out at the moment that is conducting some
reasonable current (8A). The ground connections have been made to the
ground plane with what i believe is called thermal relief - there
isn't a full ring of copper around the hole to connect it to the
ground plane, instead there are 4 thin tracks - to help when soldering
the thing together. I have a couple of questions about this.

1.) Does anyone have experience of soldering devices in without this
thermal relief? If its not much worse then this is surely the way to
go to ensure enough current handling / low resistance to ground.

2.) If its a royal pain to solder is there a better alternative - for
example thickening one of the spurs to the thickness of a track.

Thanks

Andrew

We "flood over" pads and vias all the time. It's no big deal. If the
board is surface mount, or wave soldered, it's fine. Our boards are
mostly surface mount, with a few thru-hole parts hand soldered after
reflow. With a good assembler and a good Metcal iron, soldering the
non-thermal pads is still no big deal.

Thermal resistance and electrical resistance are inseparable. For most
metals, including copper, the ratio is about 140,000 K/w thermal
resistance per ohm electrical. So a typical thermal pad, with four
spokes (1 net square of 1 oz copper) and maybe 0.5 milliohm of
resistance, has a thermal conductivity of about 70 K/w to the plane.

John

 

[Mike Harrison]

We "flood over" pads and vias all the time. It's no big deal. If the
board is surface mount, or wave soldered, it's fine. Our boards are
mostly surface mount, with a few thru-hole parts hand soldered after
reflow. With a good assembler and a good Metcal iron, soldering the
non-thermal pads is still no big deal.

Yes - This is the sort of job where a Metcal really shows how much better it is than most other
irons.

 

[DJ Delorie]

Yes - This is the sort of job where a Metcal really shows how much
better it is than most other irons.

Yup, best money I ever spent on a tool.

Wouldn't oven reflow soldering be more tolerant of thermal-less pads?
I would think that they'd heat up the solid copper as much as
everything else, which avoids the cold joint problem. Same with
hotplate reflow like I use; the ground planes are plenty hot by the
time the paste melts.

 

[John Larkin]

Yes - This is the sort of job where a Metcal really shows how much better it is than most other
irons.

But heating those flooded pads still slows things down. If a milliohm
of electrical resistance is OK, the 140 k/w thermal resistance will
make soldering a lot easier. You've just got to run the numbers.

John

 

[[email protected]]

But heating those flooded pads still slows things down. If a milliohm
of electrical resistance is OK, the 140 k/w thermal resistance will
make soldering a lot easier. You've just got to run the numbers.

John

A milliohm shouldn't be a problem (0.064W / 0.008V drop), my concern
was the current handling capacity of the spokes. After doing the
calculations for the track width the spokes didn't seem to add up.

Infact doing my own calculations on the spokes give each one to
0.5milliohms (10mil x 10mil spokes @ 1oz), with 4 this then further
reduces the resistance to ground. I guess the current carrying
potential isn't really the issue as they are so short and therefore
wont heat up.

Thanks for all your replies.

Andrew

 

[Mike Harrison]

A milliohm shouldn't be a problem (0.064W / 0.008V drop), my concern
was the current handling capacity of the spokes. After doing the
calculations for the track width the spokes didn't seem to add up.

Infact doing my own calculations on the spokes give each one to
0.5milliohms (10mil x 10mil spokes @ 1oz), with 4 this then further
reduces the resistance to ground. I guess the current carrying
potential isn't really the issue as they are so short and therefore
wont heat up.

Thanks for all your replies.

Andrew

Re. current-carrying ability, remember that each end will be heavily heatsinked, so the temp rise of
the spoke will be a lot lower than a piece of track of the same width.

 

[John Larkin]

Re. current-carrying ability, remember that each end will be heavily heatsinked, so the temp rise of
the spoke will be a lot lower than a piece of track of the same width.

Again, approximate math is simple. a 10x10 mil spoke has 500 uohms
resistance. Run one amp through it and it dissipates half a milliwatt.
The thermal resistance from the center to the pad/plane is about 35
k/w, and the average is about half that. So the center hot-spot temp
rise is ballpark 10 millikelvins.

John