I'm with Jan on this.
However, we need to take into account the artists' demands. For example, on the LP of The Animal's Greatest Hits, the piece "Sky Pilot" has a clealry heard drum pedal squeak in it, but on the CD version, that has been edited out (I have both).
The most obvious general difference between say 60-ies and those from the 70-ies is a lack of tape hiss everywhere near the "audible" mark. I expect noise reduction systems like Dolby A and dBX came down in price and were much more widely spread, this improving the whole scene in general.
However, we need to take into account the artists' demands. For example, on the LP of The Animal's Greatest Hits, the piece "Sky Pilot" has a clealry heard drum pedal squeak in it, but on the CD version, that has been edited out (I have both).
The most obvious general difference between say 60-ies and those from the 70-ies is a lack of tape hiss everywhere near the "audible" mark. I expect noise reduction systems like Dolby A and dBX came down in price and were much more widely spread, this improving the whole scene in general.
Jan: +1. I like the performances and am often suprised at the quality. But spin a few old Deccas and then play the TELARC 1812 and tell me there hasn't been progress. And since then things have improved even more, sadly few leverage that.
Another different topic if I may:
Suppose I have a PCB area used as a heatsink, for instance an extended copper area around a board-mounted TO220. Does it make a (big) difference for the heatsinking if I leave that area covered with soldermask? Or should I keep it uncovered for max heatsinking effect through the ambient air? I guess conduction through the PCB would be unaffected.
And please, if you donot have an answer but have many good ideas how to do better heatsinking; thanks I have those as well.
i really would like an answer, if possible, to this question.
Jan
Suppose I have a PCB area used as a heatsink, for instance an extended copper area around a board-mounted TO220. Does it make a (big) difference for the heatsinking if I leave that area covered with soldermask? Or should I keep it uncovered for max heatsinking effect through the ambient air? I guess conduction through the PCB would be unaffected.
And please, if you donot have an answer but have many good ideas how to do better heatsinking; thanks I have those as well.
i really would like an answer, if possible, to this question.
Jan
Last edited:
soldermask can be different thickness, and as a low thermal conductivity plastic it can't help thermal conductivity re copper
but it may not make a big difference in air conduction since air is even poorer thermal conductivity, especially in natural convection - but if you blow hard the film will be more noticeable
you certainly don't want the film over an area you might attach a heatsink to
even without you still have the question of what finish/plating over the bare copper
radiation cooling does't usually become important until heatsink to ambient delta approaches 100 C so the surface emissivity/color doesn't matter that much either
but it may not make a big difference in air conduction since air is even poorer thermal conductivity, especially in natural convection - but if you blow hard the film will be more noticeable
you certainly don't want the film over an area you might attach a heatsink to
even without you still have the question of what finish/plating over the bare copper
radiation cooling does't usually become important until heatsink to ambient delta approaches 100 C so the surface emissivity/color doesn't matter that much either
Last edited:
Affect of Solder Mask on Heat Dissipation
Jan - does this help:
Thermal conductivity [W/m-�C] * Silver: 418 * Aluminum: 403 * Copper [rolled annealed]: 392 * Copper [electrodeposited]: 392 * Gold: 297 * Nickel: 90.7 * Tin: 73 * SAC: 73 * Castin: 57 * Pb37Sn63: 51 * Lead: 35 * Epoxy, phenolic: 25-75 * Conformal coating [AR, ER, UR]: 17-21 * Epoxy, silicone: 13-26 * Epoxy, conductive: 5 * Molding compound: 0.63 * FR4: 0.35 * Solder mask: 0.245 * BT resin: 0.17
Phil
Jan - does this help:
Thermal conductivity [W/m-�C] * Silver: 418 * Aluminum: 403 * Copper [rolled annealed]: 392 * Copper [electrodeposited]: 392 * Gold: 297 * Nickel: 90.7 * Tin: 73 * SAC: 73 * Castin: 57 * Pb37Sn63: 51 * Lead: 35 * Epoxy, phenolic: 25-75 * Conformal coating [AR, ER, UR]: 17-21 * Epoxy, silicone: 13-26 * Epoxy, conductive: 5 * Molding compound: 0.63 * FR4: 0.35 * Solder mask: 0.245 * BT resin: 0.17
Phil
Yes, but with a thin conformal coating, the much greater resistance is from mask -> air or copper -> air.
We're talking a hundred microns or so of coating right?
Edit: seems my points were made.
We're talking a hundred microns or so of coating right?
Edit: seems my points were made.
No brainer, of course leaving the area free of soldermask is better.
How much exactly, that I am not sure of. Not sure if there is any material about the differences on the subject. Soldermask would acts a thermal insulator of course. I do not think it is a BIG difference. That would depend on the area used to begin with. Free air vs air flow also makes a difference in thermal efficiency of the radiator, I am sure you are well aware of.
Using a TO-263 package would be the proper part to use over a TO-220 imo, assuming the part that you are using is offered in a TO-263. I looked at the TI LM117 DS they state that they both have the same typical Theta jc 50 degrees C/W.
TO-263 is physically smaller thus you can have a bit more pcb area for heat dissipation. It is designed for this purpose. Also it allows a better thermal transfer if you bond it to the copper with a reflow of solder paste verses soldering it only around the edge of the device or a standard method of a TO-220 flange with screw attachment. With the solder mask area open it is easier to mount one of these HS unless you leave a SM opening to accommodate one in advance.
http://ca.mouser.com/Thermal-Manage...g0?P=1yztmo1Z1yzvkzl&Keyword=heatsink&FS=True
How much exactly, that I am not sure of. Not sure if there is any material about the differences on the subject. Soldermask would acts a thermal insulator of course. I do not think it is a BIG difference. That would depend on the area used to begin with. Free air vs air flow also makes a difference in thermal efficiency of the radiator, I am sure you are well aware of.
Using a TO-263 package would be the proper part to use over a TO-220 imo, assuming the part that you are using is offered in a TO-263. I looked at the TI LM117 DS they state that they both have the same typical Theta jc 50 degrees C/W.
TO-263 is physically smaller thus you can have a bit more pcb area for heat dissipation. It is designed for this purpose. Also it allows a better thermal transfer if you bond it to the copper with a reflow of solder paste verses soldering it only around the edge of the device or a standard method of a TO-220 flange with screw attachment. With the solder mask area open it is easier to mount one of these HS unless you leave a SM opening to accommodate one in advance.
http://ca.mouser.com/Thermal-Manage...g0?P=1yztmo1Z1yzvkzl&Keyword=heatsink&FS=True
My app is for a BUF634. I also was planning to place an enlarged pad on the back side of the PCB and stitch the two pads together. Or am I now violating my own request ;-)
Jan
Of course metallization matters, it depends on your soldering process. If you do not plan on reflow then go for the plating that has the best coplanarity, such as immersion gold. HASL is the worst for coplanarity.
I would use the BUF634 in the DDPAK(TO-263-5). I have heard gmarsh saying that he has done reflow in his skillet 🙂
Sure stitching the top and bottom together is great idea if you have 2-sided or multi-layer pcb, use lots of full contact thermal vias to transfer the heat.
There is figure 3 in the DS to give you the info you require. I am sure the assumption is soldermask free pcb area of course.
Another advantage of the DDPAK(TO-263-5) is that it is already lead formed for you.
I would use the BUF634 in the DDPAK(TO-263-5). I have heard gmarsh saying that he has done reflow in his skillet 🙂
Sure stitching the top and bottom together is great idea if you have 2-sided or multi-layer pcb, use lots of full contact thermal vias to transfer the heat.
There is figure 3 in the DS to give you the info you require. I am sure the assumption is soldermask free pcb area of course.
Another advantage of the DDPAK(TO-263-5) is that it is already lead formed for you.
Last edited:
NanoFoil® made by Indium Corporation looks interesting for PowerPad and other smt area bonding applications
further in they claim to have a eval kit with tin plated nanofoil for ~$200
Standard NanoFoil® sheets made by Indium Corporation
Buy NanoFoil® Activation Kit - Tin Plated / Indium Corporation's Online Store
further in they claim to have a eval kit with tin plated nanofoil for ~$200
Standard NanoFoil® sheets made by Indium Corporation
Buy NanoFoil® Activation Kit - Tin Plated / Indium Corporation's Online Store
> doesn't really make a hell of a lot difference when there's no forced air cooling.
BUT !!!!! ......... Since with S.S. device cooling you are dealing with the
control of possible 'runaway' or 'fractal' type events, it is always wise to
have the best cooling possible within your budget ! (even small increments
of safety can lead to 'make or break' under certain circumstances .........
BUT !!!!! ......... Since with S.S. device cooling you are dealing with the
control of possible 'runaway' or 'fractal' type events, it is always wise to
have the best cooling possible within your budget ! (even small increments
of safety can lead to 'make or break' under certain circumstances .........
PCB copper thickness, number of layers all come into play. 1oz. copper not very good, 2oz. much better, 14 layer 2oz. every layer with thermal vias quite good.
1oz copper anything over about 1.5 inch square a waste of time 2 oz 2-2.5 inch square is max.
Need to determine how much heat you need to get rid of... Do quite a few of these thermally dense designs, can be fun... but I have the advantage of many layers, lots of copper and thermal vias, also if required extra heatsinking. A single layer design (or even 2) can only handle a tiny amount of heat. Because of the copper thickness it does not travel far through the copper hense the sizes above.
1oz copper anything over about 1.5 inch square a waste of time 2 oz 2-2.5 inch square is max.
Need to determine how much heat you need to get rid of... Do quite a few of these thermally dense designs, can be fun... but I have the advantage of many layers, lots of copper and thermal vias, also if required extra heatsinking. A single layer design (or even 2) can only handle a tiny amount of heat. Because of the copper thickness it does not travel far through the copper hense the sizes above.
- Status
- Not open for further replies.