- up to 60deg in hottest summer .
I would say no, because there are other parts in the case. With higher heatsink temperature, case interior will dramatically rise its temperature, which will affect driver board, caps... I think we should try to stay with lower temps, unless output devices are in the separate enclosure alone. 
Very often the "0 Temperature coiffecint point" is at a higher Id than we would like. Lateral FETs seam to be more ideal at this trait. At that point we would get the least amount of threshold shift going on maintaining a consistant bias current regardless of temp, in the output
Chances are these devices are similar or even higher Id, I would guess, in terms of the "0" tempco point
Chances are these devices are similar or even higher Id, I would guess, in terms of the "0" tempco point
with high die temperature the transconductance will be smaller for the same bias current, this is one that I can think of.
But this is about die temperature.
With heat sink at 50+degC (about there for 5 second rule), there is 120degC of difference between the maximum junction temp and the heatsink temp. With the RTH between the two at about only 2degC/W (RTHjc + RTH of insulation), the junction seems to be running a lot cooler than 175degC
But this is about die temperature.
With heat sink at 50+degC (about there for 5 second rule), there is 120degC of difference between the maximum junction temp and the heatsink temp. With the RTH between the two at about only 2degC/W (RTHjc + RTH of insulation), the junction seems to be running a lot cooler than 175degC
I would say no, because there are other parts in the case. With higher heatsink temperature, case interior will dramatically rise its temperature, which will affect driver board, caps... I think we should try to stay with lower temps, unless output devices are in the separate enclosure alone.
This sounds possible.
If we rise the heatsink temperature to 75degC, it will at least double the heat dissipation of a given heatsink (for ambient of 25degC, compared to normal heatsink temp of 55degC, and considering both radiation of convection rise faster than linear with increasing temperature delta).
The interior front-end of the amp is most likely to be 10deg cooler than the heatsink, if there is ventilation holes on the bottom and top cover, due to air circulation. That will bring the interior temperature to 65degC.
But then Capacitors due to their ESR, will dissipate heat on themself too. assuming reasonable good caps, it rise the Cap temp to 75degC again.
So if we are using 105degC caps, as most DIYers would anyway, the standard 2000hour part will last for for 16,000 hours, which looks reasonable.
There is a lot of estimates here, so to be safe, a 5 second fit heatsink will still be better.
AR2 is right. I have been playing with high temp on the heatsinks in the past, which turns the box into an oven, If the box is 75C, many of the devices inside the box, will be at a much higher temp.
I though think Zhou is on to something, especially regarding the Semisouth devices, as they sure can deal with the higher temp. So a separate box for the PSU, and make a cap-less amp (think F5), and you're going to be able to put a lot of power into a quite small box.
Magura
I though think Zhou is on to something, especially regarding the Semisouth devices, as they sure can deal with the higher temp. So a separate box for the PSU, and make a cap-less amp (think F5), and you're going to be able to put a lot of power into a quite small box.
Magura
I have another idea.
Have traditional Class-A amp design with anodized heatsink on both side will definitely make the whole amp enclosure almost as hot as the heatsink.
So to achieve higher heatsink temperature and save on heatsink weight, we can use heatsinks that are inside the case, relying solely on natural convection.
My ideal heatsink will be like Thermalright HR-01 Plus.
have something like this but lie horizontally, letting air get hot and rise. Each of these can get rid of 70W easily without rise the die temp too much. With SiC kind of die temp, each can do 100W or more.
And with passive convection, Stack effect effect is what decides the overall efficiency.
So I think we should redesign the traditional amp case format, and have something that is smaller footprint but taller, with a chimney build in.
Have traditional Class-A amp design with anodized heatsink on both side will definitely make the whole amp enclosure almost as hot as the heatsink.
So to achieve higher heatsink temperature and save on heatsink weight, we can use heatsinks that are inside the case, relying solely on natural convection.
My ideal heatsink will be like Thermalright HR-01 Plus.
An externally hosted image should be here but it was not working when we last tested it.
have something like this but lie horizontally, letting air get hot and rise. Each of these can get rid of 70W easily without rise the die temp too much. With SiC kind of die temp, each can do 100W or more.
And with passive convection, Stack effect effect is what decides the overall efficiency.
So I think we should redesign the traditional amp case format, and have something that is smaller footprint but taller, with a chimney build in.
cpu tower heatsinks are not usually very good for passive cooling. If you are going to put fans in your amplifier, then, they are probably the best and have excellent Q/C ratio.
You can sometimes find reviews of those heatsink without fans on silentpcreview, and you can ask there which would be the best passive cooler. I dont remember now if it is the HR-01 or another model, since i use highspeed fans in my PC and i dont care a biip about noise
But I use a scythe ninja mini passive in my dedicated pc for audio. It is compact size and I would say about 20-30% less efficient as the best bigger towers.
This thing is promising anyway, keep going
You can sometimes find reviews of those heatsink without fans on silentpcreview, and you can ask there which would be the best passive cooler. I dont remember now if it is the HR-01 or another model, since i use highspeed fans in my PC and i dont care a biip about noise
But I use a scythe ninja mini passive in my dedicated pc for audio. It is compact size and I would say about 20-30% less efficient as the best bigger towers.
This thing is promising anyway, keep going
I forget to mention that this figure is for passive cooling.
HR-01 plus is the best passive cooler reviewed on silentpcreview, at least at the time it was reviewed.
Whether CPU heatsink are good for passive cooling is rather easy to tell. You need larger spacing between fins to allow air to move without much resistance, which contribute to good passive cooling. Whereas the best active cooler typically have narrow spacing between fins to accommodate more fins (thus larger surface area and better cooling) in a given physical dimension, since the fan can produce sufficient pressure to move the air.
Even with the larger fin separation, the cooling capability of these modern heatpipe+fin cooler are in a total different league from traditional dummy heatsink.
In size of 11x12x5cm like the HR-01 Plus, it can passively dissipate close to 100W or so.
With such heatsink technology and the high temperature capability of SiC, it does not take a very large case to accommodate to build a 500W (power consumed) class-A amp while still keep all the capacitors and front end relatively cool. Such an amp can actually be quite small, maybe 30x40cm foot print. You might need to add a removable chimney on top of it for it to work more efficiently, unless you want to build a tall and vertical amp in the first place.
In size of 11x12x5cm like the HR-01 Plus, it can passively dissipate close to 100W or so.
With such heatsink technology and the high temperature capability of SiC, it does not take a very large case to accommodate to build a 500W (power consumed) class-A amp while still keep all the capacitors and front end relatively cool. Such an amp can actually be quite small, maybe 30x40cm foot print. You might need to add a removable chimney on top of it for it to work more efficiently, unless you want to build a tall and vertical amp in the first place.
See, this is actually one of those rare cases, where water cooling becomes interesting for an amp.
Nice small box, lots of heat.....somewhere else. With the Semisouth devices, it would even be possible to make a rather nifty solution. Phase change cooling, using plain demineralized water.
FLG has done something along those lines in another context, just using oil instead.
Magura
Nice small box, lots of heat.....somewhere else. With the Semisouth devices, it would even be possible to make a rather nifty solution. Phase change cooling, using plain demineralized water.
FLG has done something along those lines in another context, just using oil instead.
Magura
All high power class-A amp seems well suited for water-cooling.
I even think that silicon device with 125degC limit need water-cooling more.
And talking about small case, a good SMPS will be useful too.
A near silent (<30dB) active cooled heatsink (heatpipe + fins), with SMPS, we could get a self-contained high power class-A amp in a real small package. This will be cool.
Last edited:
Now we are talking. I agree that water cooling is perfect for what we are about. Now question is what water cooling. The first choice would be the one uses in computers, but my feeling is that those are not going to be enough? What else than? Something bigger, than we have a separate set of problems...
I believe Grey Rollins uses water cooling in his amp. It would be good to hear the experience of existing users.
I believe Grey Rollins uses water cooling in his amp. It would be good to hear the experience of existing users.