As a general rule I allow for transistor dissipation to be the same as the required output. In a class AB amp the theoretical efficiency is never achieved, and 50% is a reasonable estimate if square waves are considered. You have to add up the thermal resistances and check that the transistor temperature is not exceeded and that the power at that temperature is still OK for the amp/required output.
Plastic transistors have about 0.625 C/W thermal resistance. YOu will need a good insulating washer so as not to degrade this too much, like the Berquist sil-pad which some types have 0.2C/W resistance, but are quite expensive.
Four transistors in parallel with these insulators on a 0.4 C/W heatsink can dissipate about 160W at ambient temp of 50 C. HOwever, the heatsink will reach 114C- so I don't think these styles of amp are really intended to run at full power for long term. Even at a modest room temp of 25C the heatsink could reach 90C. The amplifier chassis should be designed with the heatsink enclosed in a safety guard for such powers!
I'd say with these calculations you could run a 150W output load, or 200W if the output is not a square wave. You could in principle run a 400W amp provided the duty cycle was low, and for music it generally tends to be in the order of 10-20%,and as long as the quiescent dissipation does not cause too much heating either.
John
Plastic transistors have about 0.625 C/W thermal resistance. YOu will need a good insulating washer so as not to degrade this too much, like the Berquist sil-pad which some types have 0.2C/W resistance, but are quite expensive.
Four transistors in parallel with these insulators on a 0.4 C/W heatsink can dissipate about 160W at ambient temp of 50 C. HOwever, the heatsink will reach 114C- so I don't think these styles of amp are really intended to run at full power for long term. Even at a modest room temp of 25C the heatsink could reach 90C. The amplifier chassis should be designed with the heatsink enclosed in a safety guard for such powers!
I'd say with these calculations you could run a 150W output load, or 200W if the output is not a square wave. You could in principle run a 400W amp provided the duty cycle was low, and for music it generally tends to be in the order of 10-20%,and as long as the quiescent dissipation does not cause too much heating either.
John
Hi John,
You are so right. I am yet to find anyone who listens to an average power of 400 watt. That is the reason most commercial stuff, although it can measure the power output for short durations, few if any are made to be utilized at 100% duty cycle.
They came to realize few people will listen to continuous sine or worse, square waves for any length of time at the amp's full-power rating, else all equipment will be made extreme high-end and cost several thousand dollars due to the heat sink and power supply requirements.
Nico
You are so right. I am yet to find anyone who listens to an average power of 400 watt. That is the reason most commercial stuff, although it can measure the power output for short durations, few if any are made to be utilized at 100% duty cycle.
They came to realize few people will listen to continuous sine or worse, square waves for any length of time at the amp's full-power rating, else all equipment will be made extreme high-end and cost several thousand dollars due to the heat sink and power supply requirements.
Nico
That chassis looks very nice but all the photos show Krell branding. I doubt they're actually selling Krell chassis, so what you see may not be what you get.
My class-AB amps have 0.4c/W sinks. The bigger ones are 60 and 80W RMS per channel respectively (P101 and SymaSym). I tested them at full power into 8ohm resistive and the sinks get quite warm, can still keep a hand on them though. What's interesting, and probably unsurprising, is that in any actual use I've given them, they rise barely above ambient.
testing with continuous sinewave and a reduced output power (~half to 2/3 of maximum output voltage) will make the amplifier output stage run hotter. This is a severe test and in my view should only be applied to PA/Disco equipment.
Long term testing like this should never be required for domestic use amplifiers. It is an inappropriate test for a Music reproduction device used in the home.
Short term testing to maximum output voltage into worst case impedances is far more appropriate. That more closely replicates what our music does when replayed loud with minimal clipping.
Yes, this is the point. I have 0.35C/W sinks on my 100W/channel amp, and maximum dissipation in sinewave mode is where the output peak voltage is 2/pi times the supply. But after hours in idle they are barely warm, and after running hours of music at moderate to loud levels still aren't too hot to touch.
There is another point about these larger sinks- they have a high thermal mass which also takes a long time to reach equilibrium, so are well able to handle short test periods at full power and deliver music with peaks at up to 200W while providing much less average power.
John
There is another point about these larger sinks- they have a high thermal mass which also takes a long time to reach equilibrium, so are well able to handle short test periods at full power and deliver music with peaks at up to 200W while providing much less average power.
John
Seems the power transformer requirements should closly follow HS requirements for Avg Music Power in the home environment VS Heavy Duty PA use. The PS capacitor storage requirements should allow dynamic capability for peak power times X number of seconds. This a little more difficult to spec so that the avg user could make a choice given his speakers, max levels, and music. Granted that is done already for us in newer multichannel HT gear. ie Yes it does NOT make sense to have a transformer sized for continuous maximum power. But for DIY everything bigger is better of coarse.
You can have fans if you run them at a low speed. I used fast 120mm thin-fans as exhaust and intake in a 400WRMS amp I built 5 yrs ago, it also has 4 80mm fans directly on the heatsinks, and run them all at low speed with a variable linear voltage controller - to tune the fan noise to near inaudible levels, but still cool well.
I also like to play my music loud, and many others do, and just like a fast car, if you have a big amp, you will crank it up all the way once in a while, and listen to a few songs, so that tells me that continuous RMS heat design is important. I can crank it up and the heatsinks are only warm, never hot. Only time they get real warm is when it's hot in the room and no AC, and it's being played hard. You should always have enough cooling.
I used dual heatsinks 8 inches long, 3.5 inches tall, 1.5 inch long fins, and two 80mm fans slow speed on each. Since the OP is building 400W this could help.
I also like to play my music loud, and many others do, and just like a fast car, if you have a big amp, you will crank it up all the way once in a while, and listen to a few songs, so that tells me that continuous RMS heat design is important. I can crank it up and the heatsinks are only warm, never hot. Only time they get real warm is when it's hot in the room and no AC, and it's being played hard. You should always have enough cooling.
I used dual heatsinks 8 inches long, 3.5 inches tall, 1.5 inch long fins, and two 80mm fans slow speed on each. Since the OP is building 400W this could help.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.