What does "on top" mean?
If you do some reading you will find a range of solutions, each of which needs to form part of your design.
First, a few questions:
- What device?
- Insulated package or with exposed metal thermal pad?
- How much dissipation? And how large a heatsink, i.e. what tolerance do you have for a thermal gradient across the case - heatsink junction?
Most commonly you will, these days, find:
- to-264 / to-3p plastic package with
- uninsulated package, i.e. a copper thermal pad, connected to the collector if the device is a BJT
- Usually on a decent heatsink.
For hobbyists silicone pad insulators are popular. They also generally have a fairly high thermal resistance.
That said, the next most common option, Mica insulators along with silicone grease, is better but more fiddly to install. I must also say that over the last couple of decades mica has become a lot less common. Probably because:
- mica will take a lot longer to install. In a product this is BAD
- mica and grease is fiddly, and much easier to get wrong in terms of inadequate or excess silicone grease. I have seen this cause failure.
I have seen Kapton and some exotics like alumina in the past. I doubt they have much to offer in a general / audio application over silpad / mica options.
I note you sat class A. How much dissipation are you looking for from each device?
- you can work out the thermal drop across the insulator from its specs.
- unless you are really pushing the power, it would be surprising to hear of there being a significant issue with using silicone pads vs mica vs something exotic.
Thermal pads are popular because they work. Start there.
On a copper shim on the "top"... if this means the top of the plastic package of a T0-264, don't waste your effort. Again assuming TO-264 and BJT:
- the collector f the transistor is bonded to the copper heat spreader, as the collector is where power is dissipated.
- The thermal resistance collector - case is in your spec sheet.
- In a TO-264, the "top" is epoxy. It has crap thermal conductivity.
- Don't worry about trying to pull heat through epoxy, it will add essentially nothing to the thermal design and make your life hard!
If you do some reading you will find a range of solutions, each of which needs to form part of your design.
First, a few questions:
- What device?
- Insulated package or with exposed metal thermal pad?
- How much dissipation? And how large a heatsink, i.e. what tolerance do you have for a thermal gradient across the case - heatsink junction?
Most commonly you will, these days, find:
- to-264 / to-3p plastic package with
- uninsulated package, i.e. a copper thermal pad, connected to the collector if the device is a BJT
- Usually on a decent heatsink.
For hobbyists silicone pad insulators are popular. They also generally have a fairly high thermal resistance.
That said, the next most common option, Mica insulators along with silicone grease, is better but more fiddly to install. I must also say that over the last couple of decades mica has become a lot less common. Probably because:
- mica will take a lot longer to install. In a product this is BAD
- mica and grease is fiddly, and much easier to get wrong in terms of inadequate or excess silicone grease. I have seen this cause failure.
I have seen Kapton and some exotics like alumina in the past. I doubt they have much to offer in a general / audio application over silpad / mica options.
I note you sat class A. How much dissipation are you looking for from each device?
- you can work out the thermal drop across the insulator from its specs.
- unless you are really pushing the power, it would be surprising to hear of there being a significant issue with using silicone pads vs mica vs something exotic.
Thermal pads are popular because they work. Start there.
On a copper shim on the "top"... if this means the top of the plastic package of a T0-264, don't waste your effort. Again assuming TO-264 and BJT:
- the collector f the transistor is bonded to the copper heat spreader, as the collector is where power is dissipated.
- The thermal resistance collector - case is in your spec sheet.
- In a TO-264, the "top" is epoxy. It has crap thermal conductivity.
- Don't worry about trying to pull heat through epoxy, it will add essentially nothing to the thermal design and make your life hard!
What does "on top" mean?
If you do some reading you will find a range of solutions, each of which needs to form part of your design.
First, a few questions:
- What device?
"Jean Hiraga 8W le monstre"
- Insulated package or with exposed metal thermal pad?
Not sure?
- How much dissipation? And how large a heatsink, i.e. what tolerance do you have for a thermal gradient across the case - heatsink junction?
Making my own chassis and heatsink structure
Most commonly you will, these days, find:
- to-264 / to-3p plastic package with
- uninsulated package, i.e. a copper thermal pad, connected to the collector if the device is a BJT
- Usually on a decent heatsink.
NJW0302G+NJW0281G are the transistors being used
For hobbyists silicone pad insulators are popular. They also generally have a fairly high thermal resistance.
I bought silicone pad insulators and a good quality Silicone based, with 20% metal oxide conductive compound paste
That said, the next most common option, Mica insulators along with silicone grease, is better but more fiddly to install. I must also say that over the last couple of decades mica has become a lot less common. Probably because:
- mica will take a lot longer to install. In a product this is BAD
- mica and grease is fiddly, and much easier to get wrong in terms of inadequate or excess silicone grease. I have seen this cause failure.
I have seen Kapton and some exotics like alumina in the past. I doubt they have much to offer in a general / audio application over silpad / mica options.
I note you sat class A. How much dissipation are you looking for from each device?
Good question and one i'm not sure of! I know they run hot, bias will be set to accordingly but generally 0.8 -
1amp/COLOR]
- you can work out the thermal drop across the insulator from its specs.
- unless you are really pushing the power, it would be surprising to hear of there being a significant issue with using silicone pads vs mica vs something exotic.
Thermal pads are popular because they work. Start there.
Sounds like a good start!
On a copper shim on the "top"... if this means the top of the plastic package of a T0-264, don't waste your effort. Again assuming TO-264 and BJT:
- the collector f the transistor is bonded to the copper heat spreader, as the collector is where power is dissipated.
- The thermal resistance collector - case is in your spec sheet.
- In a TO-264, the "top" is epoxy. It has crap thermal conductivity.
- Don't worry about trying to pull heat through epoxy, it will add essentially nothing to the thermal design and make your life hard!
On a copper shim on the "top"... if this means the top of the plastic package of a T0-264, don't waste your effort. Again assuming TO-264 and BJT:
I was going to use the copper shim as a washer on top of the transistor to bolt it to the heatsink.
- the collector f the transistor is bonded to the copper heat spreader, as the collector is where power is dissipated.
- The thermal resistance collector - case is in your spec sheet.
- In a TO-264, the "top" is epoxy. It has crap thermal conductivity.
- Don't worry about trying to pull heat through epoxy, it will add essentially nothing to the thermal design and make your life hard![/QUOTE]
I was going to use the copper shim as a washer on top of the transistor to bolt it to the heatsink.
- the collector f the transistor is bonded to the copper heat spreader, as the collector is where power is dissipated.
- The thermal resistance collector - case is in your spec sheet.
- In a TO-264, the "top" is epoxy. It has crap thermal conductivity.
- Don't worry about trying to pull heat through epoxy, it will add essentially nothing to the thermal design and make your life hard![/QUOTE]
Hello again. Yes there is a lot of good info there. Don't use thermal paste with silicone pads. Belleville washers are very good for mountingHey Scott! Thanks for this link, good reading and education for me! Nice
I think i was just going with the silicone pads! Yes i just ordered some belleville washers! I ordered a range of parts to see what will work for me and what you guys suggest.
hi,
Those heatsinks are chunky. They will be in the region of a few tenths of a degree per watt.
The thermal resistance junction to case for your transistors is 0.83Degrees / watt. This is straight from the datasheet.
SIL pads vary, say between 0.6 and 1.5 degrees C per Watt.
NOTE: you do NOT, repeat, not, use grease with SIL pads. Seriously, don't - it wont help and will only stuff you up.
NOTE: The transistors are not insulated. this is why you need the SILPAD. The metal on the "back" is physically connected to the collector of the device.
I do not have the exact details of the Le Monstre you are building. I saw several variants, one a 15 Watt version. Lets assume it is that.
For a 15Watt Class A amplifier into 8 ohms, the current will be about 1.4A RMS, or 2 amps peak.
This means your bias current will be 2 Amps or so, and the rail voltage will be a touch more than 15 volts.
This means each output device will be dissipating 30 watts or so each. This is actually a fair bit.
Thermal resistance sums are a bit like resistance with current.
If you put in 30 watts (at the transistor junction), you can add up the "drops" across each junction.
- 0.83 deg / W transistor junction to case
- 0.6 to 1.5 deg /W for the SILPAD The top end of this is pretty pessimistic.
- 0.5 (guess) deg / W for the heatsink - this will be a LOT less if that heatsink is as big as I think it is.
This means the temperature drops are worst case:
30*(0.83 + 1.5 + 0.5) = 30 *2.83 = less than 90 degrees C.
The last tricky fact is that the heatsink thermal resistance is "to ambient" or the room temp. This is 21 degrees, or say 35 if it is HOT.
This means that worst case the transistor junction is 125 degrees C.
- Which is well below the rated 150C maximum junctio temperature, and
- Based on some wildly pessimistic assumptions of
- A poor thermal conductivity of 1.5C/W for the SILPAD, and
- 0.5C/W for your heat sink. It looks more like 0.1 ot 0.2C/W to me.
Build it up and have fun.....
Those heatsinks are chunky. They will be in the region of a few tenths of a degree per watt.
The thermal resistance junction to case for your transistors is 0.83Degrees / watt. This is straight from the datasheet.
SIL pads vary, say between 0.6 and 1.5 degrees C per Watt.
NOTE: you do NOT, repeat, not, use grease with SIL pads. Seriously, don't - it wont help and will only stuff you up.
NOTE: The transistors are not insulated. this is why you need the SILPAD. The metal on the "back" is physically connected to the collector of the device.
I do not have the exact details of the Le Monstre you are building. I saw several variants, one a 15 Watt version. Lets assume it is that.
For a 15Watt Class A amplifier into 8 ohms, the current will be about 1.4A RMS, or 2 amps peak.
This means your bias current will be 2 Amps or so, and the rail voltage will be a touch more than 15 volts.
This means each output device will be dissipating 30 watts or so each. This is actually a fair bit.
Thermal resistance sums are a bit like resistance with current.
If you put in 30 watts (at the transistor junction), you can add up the "drops" across each junction.
- 0.83 deg / W transistor junction to case
- 0.6 to 1.5 deg /W for the SILPAD The top end of this is pretty pessimistic.
- 0.5 (guess) deg / W for the heatsink - this will be a LOT less if that heatsink is as big as I think it is.
This means the temperature drops are worst case:
30*(0.83 + 1.5 + 0.5) = 30 *2.83 = less than 90 degrees C.
The last tricky fact is that the heatsink thermal resistance is "to ambient" or the room temp. This is 21 degrees, or say 35 if it is HOT.
This means that worst case the transistor junction is 125 degrees C.
- Which is well below the rated 150C maximum junctio temperature, and
- Based on some wildly pessimistic assumptions of
- A poor thermal conductivity of 1.5C/W for the SILPAD, and
- 0.5C/W for your heat sink. It looks more like 0.1 ot 0.2C/W to me.
Build it up and have fun.....
Another word on the heatsinks.
These could be very efficient as cooling devices using fan forced air for which they were almost certainly designed. With natural convection only, they need to be used vertically, as your pic shows but with their closely spaced fins, the air velocity and hence volume of air moving upwards naturally, will be quite restricted so there won't be anything like the performance these extrusions are capable of.
A tall, narrow amplifier is unusual but there's no problem if the cooling is efficient. Unfortunately, tall, narrow and shallow heatsinks are about the worst combination of conditions for convection cooling and they will be inefficient without some help from fans or just more of them than you think might be necessary.
Class A is a right PITA. I love the smooth and spacious sound quality but with any decent sized amplifier, the heat is constant and pervasive. In time, everything gets annoyingly hot, from the connectors to the electronics, case and controls - even the surrounding equipment so you do need to take a fun attitude, use more than enough heatsinking or go small....
These could be very efficient as cooling devices using fan forced air for which they were almost certainly designed. With natural convection only, they need to be used vertically, as your pic shows but with their closely spaced fins, the air velocity and hence volume of air moving upwards naturally, will be quite restricted so there won't be anything like the performance these extrusions are capable of.
A tall, narrow amplifier is unusual but there's no problem if the cooling is efficient. Unfortunately, tall, narrow and shallow heatsinks are about the worst combination of conditions for convection cooling and they will be inefficient without some help from fans or just more of them than you think might be necessary.
Class A is a right PITA. I love the smooth and spacious sound quality but with any decent sized amplifier, the heat is constant and pervasive. In time, everything gets annoyingly hot, from the connectors to the electronics, case and controls - even the surrounding equipment so you do need to take a fun attitude, use more than enough heatsinking or go small....
Another word on the heatsinks.
These could be very efficient as cooling devices using fan forced air for which they were almost certainly designed. With natural convection only, they need to be used vertically, as your pic shows but with their closely spaced fins, the air velocity and hence volume of air moving upwards naturally, will be quite restricted so there won't be anything like the performance these extrusions are capable of.
Ian you got it spot on mate, avionic heatsinks which indeed did have a fan blowing through it.
A tall, narrow amplifier is unusual but there's no problem if the cooling is efficient. Unfortunately, tall, narrow and shallow heatsinks are about the worst combination of conditions for convection cooling and they will be inefficient without some help from fans or just more of them than you think might be necessary.
As you said a tall narrow amplifier is unusual hence why im trying to build one based on the heatsinks. Yes i agree i am looking at ultra slim fans, quiet PC fans with a temperature sensor to control them
Class A is a right PITA. I love the smooth and spacious sound quality but with any decent sized amplifier, the heat is constant and pervasive. In time, everything gets annoyingly hot, from the connectors to the electronics, case and controls - even the surrounding equipment so you do need to take a fun attitude, use more than enough heatsinking or go small....
They make a really nice hand warmer!
hopefully these heatsinks will suffice. It's part of the learning experience, i just like to give thngs a go and see where it takes me.- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.