Home |
Search |
Today's Posts |
#1
Posted to rec.audio.tech
|
|||
|
|||
Thermal pads for transistor mounts
Am I right in my belief that putting heatsink compound on silicon pads gives
less thermal conductivity than pads alone? Cheers, Gareth. |
#2
Posted to rec.audio.tech
|
|||
|
|||
Thermal pads for transistor mounts
"Gareth Magennis" wrote in message
news Am I right in my belief that putting heatsink compound on silicon pads gives less thermal conductivity than pads alone? Cheers, Gareth. My understanding is that thermal compound couples the bottom of the device to the heatsink by filling-in the tiny interstices that would otherwise exist between the slightly rough device and the slightly rough heatsink. Consequently, heat is conducted better than it would be if the device were clamped directly to the heatsink. S. -- http://audiopages.googlepages.com |
#3
Posted to rec.audio.tech
|
|||
|
|||
Thermal pads for transistor mounts
"Serge Auckland" wrote in message ... "Gareth Magennis" wrote in message news Am I right in my belief that putting heatsink compound on silicon pads gives less thermal conductivity than pads alone? Cheers, Gareth. My understanding is that thermal compound couples the bottom of the device to the heatsink by filling-in the tiny interstices that would otherwise exist between the slightly rough device and the slightly rough heatsink. Consequently, heat is conducted better than it would be if the device were clamped directly to the heatsink. S. -- http://audiopages.googlepages.com A few years back Yamaha was having higher than normal failures of some amp channels using the conductive rubber. As part of the repair the techs were to add silicone grease to the outputs and rubber sheets. Mark Z. |
#4
Posted to rec.audio.tech
|
|||
|
|||
Thermal pads for transistor mounts
"Serge Auckland" wrote in message ... "Gareth Magennis" wrote in message news Am I right in my belief that putting heatsink compound on silicon pads gives less thermal conductivity than pads alone? Cheers, Gareth. My understanding is that thermal compound couples the bottom of the device to the heatsink by filling-in the tiny interstices that would otherwise exist between the slightly rough device and the slightly rough heatsink. Consequently, heat is conducted better than it would be if the device were clamped directly to the heatsink. My understanding was that silicon pads already have the optimum contact with both surfaces due to their squidginess. Putting 2 layers of thermal compound with less than 100% thermal conductivity between them is only going to make things worse. I just seem to remember reading something about this somewhere, but that does not of course make it true. Gareth. |
#5
Posted to rec.audio.tech
|
|||
|
|||
Thermal pads for transistor mounts
"Gareth Magennis" wrote in message
... "Serge Auckland" wrote in message ... "Gareth Magennis" wrote in message news Am I right in my belief that putting heatsink compound on silicon pads gives less thermal conductivity than pads alone? Cheers, Gareth. My understanding is that thermal compound couples the bottom of the device to the heatsink by filling-in the tiny interstices that would otherwise exist between the slightly rough device and the slightly rough heatsink. Consequently, heat is conducted better than it would be if the device were clamped directly to the heatsink. My understanding was that silicon pads already have the optimum contact with both surfaces due to their squidginess. Putting 2 layers of thermal compound with less than 100% thermal conductivity between them is only going to make things worse. I just seem to remember reading something about this somewhere, but that does not of course make it true. Gareth. Do you mean Silicon, or Silicone? Silicon is a pretty hard material and would behave as I have described. I have no idea about silicone, I thought it was used in car wax and breast implants. S. -- http://audiopages.googlepages.com |
#6
Posted to rec.audio.tech
|
|||
|
|||
Thermal pads for transistor mounts
Serge Auckland wrote:
"Gareth Magennis" wrote in message ... "Serge Auckland" wrote in message ... "Gareth Magennis" wrote in message news Am I right in my belief that putting heatsink compound on silicon pads gives less thermal conductivity than pads alone? Cheers, Gareth. My understanding is that thermal compound couples the bottom of the device to the heatsink by filling-in the tiny interstices that would otherwise exist between the slightly rough device and the slightly rough heatsink. Consequently, heat is conducted better than it would be if the device were clamped directly to the heatsink. My understanding was that silicon pads already have the optimum contact with both surfaces due to their squidginess. Putting 2 layers of thermal compound with less than 100% thermal conductivity between them is only going to make things worse. I just seem to remember reading something about this somewhere, but that does not of course make it true. Gareth. Do you mean Silicon, or Silicone? Silicon is a pretty hard material and would behave as I have described. I have no idea about silicone, I thought it was used in car wax and breast implants. The usual white heat transfer compound is a silicone compound. At least some of those products exibit rather crappy heat transfer characteristics, plus it hardens and deteriorates with age. While I haven't use this for audio, serious computer builders use products like Titan Silver Grease or Artic Silver, which contains very fine grained silver flakes in a non-silicone compound. Note that some similar products are silicone based, to be avoided. Of cause, for some applications electric isolation is needed; a silver based compound may be less applicable for this Maybe a hard thin mica or ceramic isolation shimmer with said compound centered (to avoid contact round edges) can be used. YMMV.. And as stated by others, the point with any compound is filling an uneven surface, i.e. as little as possible should be applied. If two surfaces could have a perfect meet, any compund would only result in lesser contact. A Bit the same as with lipstick.. -- Kind regards, Mogens V. |
#7
Posted to rec.audio.tech
|
|||
|
|||
Thermal pads for transistor mounts
Mogens V. wrote:
While I haven't use this for audio, serious computer builders use products like Titan Silver Grease or Artic Silver, which contains very fine grained silver flakes in a non-silicone compound. Note that some similar products are silicone based, to be avoided. I think it is aluminium dust rather that silver dust. Kind regards Peter Larsen |
#8
Posted to rec.audio.tech
|
|||
|
|||
Thermal pads for transistor mounts
Peter Larsen wrote:
Mogens V. wrote: While I haven't use this for audio, serious computer builders use products like Titan Silver Grease or Artic Silver, which contains very fine grained silver flakes in a non-silicone compound. Note that some similar products are silicone based, to be avoided. I think it is aluminium dust rather that silver dust. You may very well be right. Still is a good thermo coupler -- Kind regards, Mogens V. |
#9
Posted to rec.audio.tech
|
|||
|
|||
Thermal pads for transistor mounts
Gareth Magennis wrote: Am I right in my belief that putting heatsink compound on silicon pads gives less thermal conductivity than pads alone? Do you mean silicone pads ? There's simply no purpose in using heatsink compound with them. In as much as it adds something to the thermal path it must increase the thermal resistance at least by a tiny bit. Graham |
#10
Posted to rec.audio.tech
|
|||
|
|||
Thermal pads for transistor mounts
Serge Auckland wrote: "Gareth Magennis" wrote Am I right in my belief that putting heatsink compound on silicon pads gives less thermal conductivity than pads alone? My understanding is that thermal compound couples the bottom of the device to the heatsink by filling-in the tiny interstices that would otherwise exist between the slightly rough device and the slightly rough heatsink. Consequently, heat is conducted better than it would be if the device were clamped directly to the heatsink. Irrelevant to silicone pads as they confrom to the mating surfaces themselves. Graham |
#11
Posted to rec.audio.tech
|
|||
|
|||
Thermal pads for transistor mounts
"Mark D. Zacharias" wrote: A few years back Yamaha was having higher than normal failures of some amp channels using the conductive rubber. As part of the repair the techs were to add silicone grease to the outputs and rubber sheets. Bizarre. Using a better grade of pad would have been far more intelligent. Graham |
#12
Posted to rec.audio.tech
|
|||
|
|||
Thermal pads for transistor mounts
Gareth Magennis wrote: My understanding was that silicon pads already have the optimum contact with both surfaces due to their squidginess. Putting 2 layers of thermal compound with less than 100% thermal conductivity between them is only going to make things worse. I just seem to remember reading something about this somewhere, but that does not of course make it true. Your understanding is perfectly correct. Graham |
#13
Posted to rec.audio.tech
|
|||
|
|||
Thermal pads for transistor mounts
"Eeyore" wrote in message
... Serge Auckland wrote: "Gareth Magennis" wrote Am I right in my belief that putting heatsink compound on silicon pads gives less thermal conductivity than pads alone? My understanding is that thermal compound couples the bottom of the device to the heatsink by filling-in the tiny interstices that would otherwise exist between the slightly rough device and the slightly rough heatsink. Consequently, heat is conducted better than it would be if the device were clamped directly to the heatsink. Irrelevant to silicone pads as they confrom to the mating surfaces themselves. Graham Ah, so it's silicone, not silicon...quite a difference. S. -- http://audiopages.googlepages.com |
#14
Posted to rec.audio.tech
|
|||
|
|||
Thermal pads for transistor mounts
Eeyore wrote:
Gareth Magennis wrote: Am I right in my belief that putting heatsink compound on silicon pads gives less thermal conductivity than pads alone? Do you mean silicone pads ? There's simply no purpose in using heatsink compound with them. In as much as it adds something to the thermal path it must increase the thermal resistance at least by a tiny bit. Bring back mica and berlyium oxide. geoff |
#15
Posted to rec.audio.tech
|
|||
|
|||
Thermal pads for transistor mounts
On Sat, 24 Nov 2007 05:45:38 +0000, Eeyore
wrote: Am I right in my belief that putting heatsink compound on silicon pads gives less thermal conductivity than pads alone? My understanding is that thermal compound couples the bottom of the device to the heatsink by filling-in the tiny interstices that would otherwise exist between the slightly rough device and the slightly rough heatsink. Consequently, heat is conducted better than it would be if the device were clamped directly to the heatsink. Irrelevant to silicone pads as they confrom to the mating surfaces themselves. My experience is that the pads are worse than just a thin layer of thermal "grease". The layer should be very thin, and if so, it has little influence on the overall thermal conductivity. I find that some of the high thermal conductivity compounds are too thick in consistency, and don't allow a thin layer. The thermal compound MUST a thin film! Beware if your grease settles out (like organic peanut butter) and you get lumpy hard-to-spread grease. It's worse than useless in that state. However..... you sometimes need to ELECTRICALLY insulate the device from the heatsink. If you have this situation YOU CANNOT USE THERMAL GREASE ALONE. Then you need an (electrically) insulated pad. You shouldn't need to use both methods at the same time, since the silicone should conform under pressure. If it doesn't conform, you need a new pad. We used to use Mica films, but these weren't flexible, and needed the thermal compound to fill in the gaps with something thermally conductive. Just to be controversial, you don't need something all that thermally conductive.... a thinner film (possibly flatter surface) is more effective than the high thermal conductivity greases. We used to use plain silicone grease without the aluminum/zinc oxide. It worked OK.... Back in the old days they used beryllium oxide too, but it was deadly poisonous! |
#16
Posted to rec.audio.tech
|
|||
|
|||
Thermal pads for transistor mounts
In article ,
Paul wrote: Back in the old days they used beryllium oxide too, but it was deadly poisonous! Only if you pulverized it and snorted it. Isaac |
#17
Posted to rec.audio.tech
|
|||
|
|||
Thermal pads for transistor mounts
geoff wrote: Eeyore wrote: Gareth Magennis wrote: Am I right in my belief that putting heatsink compound on silicon pads gives less thermal conductivity than pads alone? Do you mean silicone pads ? There's simply no purpose in using heatsink compound with them. In as much as it adds something to the thermal path it must increase the thermal resistance at least by a tiny bit. Bring back mica and berlyium oxide. I prefer live heatsinks myself. That helps keep the device die that much cooler. BeO2 is potentially dangerous (carcinogenic). Graham |
#18
Posted to rec.audio.tech
|
|||
|
|||
Thermal pads for transistor mounts
Paul wrote: Eeyore wrote: Am I right in my belief that putting heatsink compound on silicon pads gives less thermal conductivity than pads alone? My understanding is that thermal compound couples the bottom of the device to the heatsink by filling-in the tiny interstices that would otherwise exist between the slightly rough device and the slightly rough heatsink. Consequently, heat is conducted better than it would be if the device were clamped directly to the heatsink. Irrelevant to silicone pads as they confrom to the mating surfaces themselves. My experience is that the pads are worse than just a thin layer of thermal "grease". Yabbut .... A thin layer of grease doesn't provide any insulation. And I don't use 'grease' either. I use the 'compound' that's filled with aluminium oxide. It has better thermal conductivity. Graham |
#19
Posted to rec.audio.tech
|
|||
|
|||
Thermal pads for transistor mounts
"Eeyore" wrote in message ... Gareth Magennis wrote: Am I right in my belief that putting heatsink compound on silicon pads gives less thermal conductivity than pads alone? Do you mean silicone pads ? There's simply no purpose in using heatsink compound with them. In as much as it adds something to the thermal path it must increase the thermal resistance at least by a tiny bit. Graham That was exactly my thinking. Wharfedale are putting compound on some of their pads, and I clean it off when I find it. Gareth. |
#21
Posted to rec.audio.tech
|
|||
|
|||
Thermal pads for transistor mounts
In article , Paul wrote:
On Sat, 24 Nov 2007 05:45:38 +0000, Eeyore wrote: Am I right in my belief that putting heatsink compound on silicon pads gives less thermal conductivity than pads alone? My understanding is that thermal compound couples the bottom of the device to the heatsink by filling-in the tiny interstices that would otherwise exist between the slightly rough device and the slightly rough heatsink. Consequently, heat is conducted better than it would be if the device were clamped directly to the heatsink. Irrelevant to silicone pads as they confrom to the mating surfaces themselves. My experience is that the pads are worse than just a thin layer of thermal "grease". The layer should be very thin, and if so, it has little influence on the overall thermal conductivity. I find that some of the high thermal conductivity compounds are too thick in consistency, and don't allow a thin layer. The thermal compound MUST a thin film! Beware if your grease settles out (like organic peanut butter) and you get lumpy hard-to-spread grease. It's worse than useless in that state. Most of the time, too much grease is used. The grease should be rubbed around, and you should see visable metal. I have found too many parts that have dried out over the years, but too much grease in the first place will create a worsening situation. greg However..... you sometimes need to ELECTRICALLY insulate the device from the heatsink. If you have this situation YOU CANNOT USE THERMAL GREASE ALONE. Then you need an (electrically) insulated pad. You shouldn't need to use both methods at the same time, since the silicone should conform under pressure. If it doesn't conform, you need a new pad. We used to use Mica films, but these weren't flexible, and needed the thermal compound to fill in the gaps with something thermally conductive. Just to be controversial, you don't need something all that thermally conductive.... a thinner film (possibly flatter surface) is more effective than the high thermal conductivity greases. We used to use plain silicone grease without the aluminum/zinc oxide. It worked OK.... Back in the old days they used beryllium oxide too, but it was deadly poisonous! |
#22
Posted to rec.audio.tech
|
|||
|
|||
Thermal pads for transistor mounts
Eeyore wrote:
Bring back mica and berlyium oxide. I prefer live heatsinks myself. That helps keep the device die that much cooler. BeO2 is potentially dangerous (carcinogenic). So are live heatsinks with a PD of 180VDC a centimetre apart ! At least chances are that the DC fuses have long ago popped when the PCB components burned and tracks vapourised, eh Graham ;-) geoff |
#23
Posted to rec.audio.tech
|
|||
|
|||
Thermal pads for transistor mounts
GregS wrote:
In article , wrote: Peter Larsen wrote: Mogens V. wrote: While I haven't use this for audio, serious computer builders use products like Titan Silver Grease or Artic Silver, which contains very fine grained silver flakes in a non-silicone compound. Note that some similar products are silicone based, to be avoided. I think it is aluminium dust rather that silver dust. You may very well be right. Still is a good thermo coupler Aluminum is 40 times worse than diamond. greg Difficult cutting them tiny flakes off a diamond for the heat compound.. -- Kind regards, Mogens V. |
#24
Posted to rec.audio.tech
|
|||
|
|||
Thermal pads for transistor mounts
Mogens V. wrote:
I think it is aluminium dust rather that silver dust. You may very well be right. Still is a good thermo coupler Aluminum is 40 times worse than diamond. Difficult cutting them tiny flakes off a diamond for the heat compound.. You will never make it in high end audio with that attitude, do you want flakey sound? - if you want clean, transparent sound, then you cut them in thin slices and use them like mica was used. Single ended designs may be attractive for budget conscious .... Kind regards Peter Larsen |
#25
Posted to rec.audio.tech
|
|||
|
|||
Thermal pads for transistor mounts
Peter Larsen wrote:
Mogens V. wrote: I think it is aluminium dust rather that silver dust. You may very well be right. Still is a good thermo coupler Aluminum is 40 times worse than diamond. Difficult cutting them tiny flakes off a diamond for the heat compound.. You will never make it in high end audio with that attitude, do you want flakey sound? - if you want clean, transparent sound, then you cut them in thin slices and use them like mica was used. Single ended designs may be attractive for budget conscious .... But audiences are all flakes too. And musicians. geoff |
#26
Posted to rec.audio.tech
|
|||
|
|||
Thermal pads for transistor mounts
geoff wrote: Eeyore wrote: Bring back mica and berlyium oxide. I prefer live heatsinks myself. That helps keep the device die that much cooler. BeO2 is potentially dangerous (carcinogenic). So are live heatsinks with a PD of 180VDC a centimetre apart ! Only if you open up the box. No different to most electronics really. And far less dangerous than a CRT. At least chances are that the DC fuses have long ago popped when the PCB components burned and tracks vapourised, eh Graham ;-) Blame that ****wit Paul Belcher whose bright idea that was. I'm sick of being criticised for his stupid ideas. Graham |
#27
Posted to rec.audio.tech
|
|||
|
|||
Thermal pads for transistor mounts
geoff wrote: Eeyore wrote: I prefer live heatsinks myself. That helps keep the device die that much cooler. BeO2 is potentially dangerous (carcinogenic). So are live heatsinks NO. They're not carcinogenic. Graham |
#28
Posted to rec.audio.tech
|
|||
|
|||
Thermal pads for transistor mounts
"geoff" wrote in message ... Eeyore wrote: Bring back mica and berlyium oxide. I prefer live heatsinks myself. That helps keep the device die that much cooler. BeO2 is potentially dangerous (carcinogenic). So are live heatsinks with a PD of 180VDC a centimetre apart ! At least chances are that the DC fuses have long ago popped when the PCB components burned and tracks vapourised, eh Graham ;-) geoff The EV P3000/2000 etc have live heatsinks less than 2mm apart, they seem to do just fine. Gareth. |
#29
Posted to rec.audio.tech
|
|||
|
|||
Thermal pads for transistor mounts
In article ,
Eeyore wrote: geoff wrote: Eeyore wrote: Gareth Magennis wrote: Am I right in my belief that putting heatsink compound on silicon pads gives less thermal conductivity than pads alone? Do you mean silicone pads ? There's simply no purpose in using heatsink compound with them. In as much as it adds something to the thermal path it must increase the thermal resistance at least by a tiny bit. Bring back mica and berlyium oxide. I prefer live heatsinks myself. That helps keep the device die that much cooler. And adds greatly to the capacitance of whichever electrode is connected to the case. Usually a bad thing, design-wise. Isaac |
#30
Posted to rec.audio.tech
|
|||
|
|||
Thermal pads for transistor mounts
isw wrote: Eeyore wrote: geoff wrote: Eeyore wrote: Gareth Magennis wrote: Am I right in my belief that putting heatsink compound on silicon pads gives less thermal conductivity than pads alone? Do you mean silicone pads ? There's simply no purpose in using heatsink compound with them. In as much as it adds something to the thermal path it must increase the thermal resistance at least by a tiny bit. Bring back mica and berlyium oxide. I prefer live heatsinks myself. That helps keep the device die that much cooler. And adds greatly to the capacitance of whichever electrode is connected to the case. Usually a bad thing, design-wise. Completely irrelevant at audio frequencies. Since the device case is almost invariably the collector with bipolars and this is usually the power supply, it is indeed totally irrelevant in such cases. Have you actually ever designed an amplifier ? I have designed a number of them and several hundred thousand of them in total have been made. Graham |
#32
Posted to rec.audio.tech
|
|||
|
|||
Thermal pads for transistor mounts
In article ,
Eeyore wrote: isw wrote: Eeyore wrote: geoff wrote: Eeyore wrote: Gareth Magennis wrote: Am I right in my belief that putting heatsink compound on silicon pads gives less thermal conductivity than pads alone? Do you mean silicone pads ? There's simply no purpose in using heatsink compound with them. In as much as it adds something to the thermal path it must increase the thermal resistance at least by a tiny bit. Bring back mica and berlyium oxide. I prefer live heatsinks myself. That helps keep the device die that much cooler. And adds greatly to the capacitance of whichever electrode is connected to the case. Usually a bad thing, design-wise. Completely irrelevant at audio frequencies. Well, probably, but the active devices need to have a bit more bandwidth than that. Since the device case is almost invariably the collector with bipolars and this is usually the power supply, it is indeed totally irrelevant in such cases. Have you actually ever designed an amplifier ? Ummm, yup. Stabilized them to prevent oscillation from too much capacitive coupling, too. Not all amplifiers use the topology you mentioned. And the difference between theta-j-a for a "hot" sink and a properly insulated one is really quite small. If it makes the difference between an amp that works and one that doesn't, I'd suggest that the heat sink is too small for reliable operation. There's a lot of power in the filter caps of a large amplifier; having it present on large chunks of aluminum is not a good idea. Isaac |
#33
Posted to rec.audio.tech
|
|||
|
|||
Thermal pads for transistor mounts
isw wrote: Eeyore wrote: isw wrote: Eeyore wrote: geoff wrote: Eeyore wrote: Gareth Magennis wrote: Am I right in my belief that putting heatsink compound on silicon pads gives less thermal conductivity than pads alone? Do you mean silicone pads ? There's simply no purpose in using heatsink compound with them. In as much as it adds something to the thermal path it must increase the thermal resistance at least by a tiny bit. Bring back mica and berlyium oxide. I prefer live heatsinks myself. That helps keep the device die that much cooler. And adds greatly to the capacitance of whichever electrode is connected to the case. Usually a bad thing, design-wise. Completely irrelevant at audio frequencies. Well, probably, but the active devices need to have a bit more bandwidth than that. Anywhere between 10 and 40 MHz for modernish designs but capacitance on the power supply nodes is GOOD anyway ! Since the device case is almost invariably the collector with bipolars and this is usually the power supply, it is indeed totally irrelevant in such cases. Have you actually ever designed an amplifier ? Ummm, yup. Stabilized them to prevent oscillation from too much capacitive coupling, too. Stability results from closing the feedback loop with adequate gain/phase margin. It has nothing to do with 'capacitive coupling' or SHOULDN'T do ! Not all amplifiers use the topology you mentioned. All amplifiers using bipolar devices have the collector as case which means that the power supply is the node that will typically be connected tot he heatsink. In the instances of QSC style designs (grounded collector), the heatsink is earthed anyway so there's no problem there either. And the difference between theta-j-a for a "hot" sink and a properly insulated one is really quite small. Nonsense. Take an MJ150xx type device dissipating an average of say 50W (which is quite typical). Adding a thermal washer will typically increase the interface resistance by about 0.4C/W .. https://www.onsemi.com/pub/Collateral/AN1040-D.PDF That'll take the junction temperature 20C higher. If it makes the difference between an amp that works and one that doesn't, I'd suggest that the heat sink is too small for reliable operation. Reducing device temp is ALWAYS beneficial. There's a lot of power in the filter caps of a large amplifier; having it present on large chunks of aluminum is not a good idea. It's a perfectly fine idea when it's inside the case. There's LOTS of live bits inside an amplifier including potentially lethal AC mains voltages., there's nothing special about a heatsink that means it can't be live. You're not meant to be pokoing about in there. Graham |
Reply |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Forum | |||
Thermal relays Info needed | Vacuum Tubes | |||
Mac OSX thermal CD printer for under $300.00? | Pro Audio | |||
Thermal Grease/Adhesive (?) | Tech | |||
Thermal rings | Pro Audio | |||
FS.Automated CD Duplicator / Thermal Printer | Pro Audio |