Hi all, I am building a KSA50 clone, the heatsink I have is the type that is in 4 seperate sections and sits on top of a 120mm fan.( looks to be the same as a genuine ksa50 only taller )The heatsink comes with insulating material that links the 4 sections together mechanically and keeps each section isolated from one another.
The question I want to ask is: from a heat dissapation point of view it would be better to mount the transistors ( TO3 ) without any mica washers, directly onto the heatsink. Would this cause any problems with respect to audio radiation that may be picked up by the amplifier/driver boards, or any other problems.
I have read with some amplifiers the output heatsink needs to be grounded.
Has anyone any experiance or views.
Thanks
Alan
The question I want to ask is: from a heat dissapation point of view it would be better to mount the transistors ( TO3 ) without any mica washers, directly onto the heatsink. Would this cause any problems with respect to audio radiation that may be picked up by the amplifier/driver boards, or any other problems.
I have read with some amplifiers the output heatsink needs to be grounded.
Has anyone any experiance or views.
Thanks
Alan
With respect to dissipation an advantage : a TO-3 mica washer adds 0.3C/W thermal resistance, raises the die temperature by about 10C in the KSA50.
Not much of an appeal to do if you're using a heatsink tunnel/fan setup, each heatsink section would have to be completely insulated from the others, the fan, and the bottom of the case.
A lot of additional labour, and two heatsinks intimately close, with the risk of carrying a different potential, is unhealthy engineering.
Smarter choice is picking other insulators, Kerafol comes in TO-3 shape, easy to mount and drops the die temperature rise to a level that makes all the hard work of insulating heatsinks not worthwhile.
KISS !
Serial manufacture example of live (no mica) heatsinks =>
Not much of an appeal to do if you're using a heatsink tunnel/fan setup, each heatsink section would have to be completely insulated from the others, the fan, and the bottom of the case.
A lot of additional labour, and two heatsinks intimately close, with the risk of carrying a different potential, is unhealthy engineering.
Smarter choice is picking other insulators, Kerafol comes in TO-3 shape, easy to mount and drops the die temperature rise to a level that makes all the hard work of insulating heatsinks not worthwhile.
KISS !
Serial manufacture example of live (no mica) heatsinks =>
Attachments
Last edited:
Thanks Jacco, I have read lots of your posts and I know you are one of the guys who always talks sense.
The heatsinks have a vertical rebate in the outside fin of all 4 and are linked together with insulation material that keeps them seperated by about5mm. The top and bottom surfaces also have an insulating disc and a mounting plate that the fan attaches to.This dosen't appear to be a "made up " design in appears to have been manufactured like it. I have removed the heatsink assembly from a piece of American industrial equipment.
I asked the original question as the heatsink looked too good an opertunity to miss, but if you think it is better to use TO3 insulators I will take your advice.
thanks
Alan
The heatsinks have a vertical rebate in the outside fin of all 4 and are linked together with insulation material that keeps them seperated by about5mm. The top and bottom surfaces also have an insulating disc and a mounting plate that the fan attaches to.This dosen't appear to be a "made up " design in appears to have been manufactured like it. I have removed the heatsink assembly from a piece of American industrial equipment.
I asked the original question as the heatsink looked too good an opertunity to miss, but if you think it is better to use TO3 insulators I will take your advice.
thanks
Alan
I have a couple of those (in factory oiled paper wrapping), does make it quite a bit easier.
Placing the heatsinks and fan on rubber suspension also helps reduce the mechanical noise level somewhat.
Bottom line is whether you need the extra reduction in thermal resistance, for your heatsink assembly and the bias level you have planned.
The example picture is of a 100W Class A power amp, both channels of 6 power devices each mounted on one blown heatsink duct, think ~450W continuous heat removal.
In my experience, running the dies of power devices at 87.5C max (midpoint of 25C ambient and 150C departure number) keeps them alive a very long time.
Without mica insulators, the same setup for 50W can handle enough bias for +70W Class A power in 8 Ohm (at identical die temperature).
Leaving the insulators out, at the 50W bias level, merely doubles the lifespan of the power devices to an average life expectancy few power amp owners will outlive (imo).
(make sense ? surely you are not suggesting i've spent years building a reputation of a village idiot for nothing ?)
Placing the heatsinks and fan on rubber suspension also helps reduce the mechanical noise level somewhat.
Bottom line is whether you need the extra reduction in thermal resistance, for your heatsink assembly and the bias level you have planned.
The example picture is of a 100W Class A power amp, both channels of 6 power devices each mounted on one blown heatsink duct, think ~450W continuous heat removal.
In my experience, running the dies of power devices at 87.5C max (midpoint of 25C ambient and 150C departure number) keeps them alive a very long time.
Without mica insulators, the same setup for 50W can handle enough bias for +70W Class A power in 8 Ohm (at identical die temperature).
Leaving the insulators out, at the 50W bias level, merely doubles the lifespan of the power devices to an average life expectancy few power amp owners will outlive (imo).
(make sense ? surely you are not suggesting i've spent years building a reputation of a village idiot for nothing ?)
If the heatsinks can be electrically isolated from possible shorts or human contact then you are better mounting the TO3s directly on the heatsinks as more heat will be transferred from the junction to the air.
HOWEVER. You will have to be VERY CAREFUL that the heatsinks are insulated from anything that could touch them.
I doubt very much that the heatsinks will radiate EMF radiation as you should still have the Collector connection close to the TO3 case - If not using the same bolt as the TO3 case is using. There will be virtually NO current flowing through the heatsink material.
This process has been used in HIGH POWER disco equipment where the heatsinks are internal to the chassis and fans are employed to get rid of the heat.
Most audiophiles would try to use FREE AIR cooling which means that the heatsinks are on the outside of the chassis and can be touched. In such instances the TO3s need to be isolated from the heatsinks for safety using mica, keratim, cryptonite or simple fibre insulators. OK Cryptonite is only available from B&Q(Krypton), the others get varied reviews depending on how much heat you are trying to dissipate. My favourite is Mica as its easy to get hold of.
HOWEVER. You will have to be VERY CAREFUL that the heatsinks are insulated from anything that could touch them.
I doubt very much that the heatsinks will radiate EMF radiation as you should still have the Collector connection close to the TO3 case - If not using the same bolt as the TO3 case is using. There will be virtually NO current flowing through the heatsink material.
This process has been used in HIGH POWER disco equipment where the heatsinks are internal to the chassis and fans are employed to get rid of the heat.
Most audiophiles would try to use FREE AIR cooling which means that the heatsinks are on the outside of the chassis and can be touched. In such instances the TO3s need to be isolated from the heatsinks for safety using mica, keratim, cryptonite or simple fibre insulators. OK Cryptonite is only available from B&Q(Krypton), the others get varied reviews depending on how much heat you are trying to dissipate. My favourite is Mica as its easy to get hold of.
Last edited:
He can be abusive sometimes. Once in a while he does come up with a useful comment.
Hi KatieandDad, thanks for the useful comments. I did construct this amplifier using passive cooling ( 400x150x40 on each side ) and although the heatsinks would run about 60C which was OK I decided, rightly or wrongly, that the components internal to the chassis,which was all alluminium, eg. transformers, filter caps etc were feeling too warm for my likeing.This may have been due to poor layout on my behalf.
But being a DIYer I thought I would try the forced air route in a different chassis and see what the results were. I am trying to make sure the warm air exits the case without getting the surrounding metalwork warm. Also putting a lot of effort into fan mounting and speed control for the minimum noise.
I'm happy to play and it's a joy to get good (or bad ) advice from this forum.
thanks
Alan
But being a DIYer I thought I would try the forced air route in a different chassis and see what the results were. I am trying to make sure the warm air exits the case without getting the surrounding metalwork warm. Also putting a lot of effort into fan mounting and speed control for the minimum noise.
I'm happy to play and it's a joy to get good (or bad ) advice from this forum.
thanks
Alan
Put the fan on so that it is blowing through the sink, not sucking.
Put the fan at the bottom so that natural convection assists the fan in getting heat out of the top.
To3 direct to sink gives about 0.2C/W Can this be improved upon with an insulator?
If it can't be bettered, then Rth c-s of 0.3C/W is very good in my book.
Put the fan at the bottom so that natural convection assists the fan in getting heat out of the top.
To3 direct to sink gives about 0.2C/W Can this be improved upon with an insulator?
If it can't be bettered, then Rth c-s of 0.3C/W is very good in my book.
many thanks for all the assistance, much appreciated.
Andrew T, yes the 120mm fan will be fitted at the bottom of the heatsink.
Slightly off subject, using K Amps Class A Calculator I get the following results:
23V rails
3Pairs output devices
0.5 Ohm emmiter resistors
1.6Amp bias per rail
= 41W RMS and 74watts per channel heat dissapation.
Does this sound correct?
thanks
Alan
Andrew T, yes the 120mm fan will be fitted at the bottom of the heatsink.
Slightly off subject, using K Amps Class A Calculator I get the following results:
23V rails
3Pairs output devices
0.5 Ohm emmiter resistors
1.6Amp bias per rail
= 41W RMS and 74watts per channel heat dissapation.
Does this sound correct?
thanks
Alan
You'll be very surprised as to how little air stirring you actually need in order to improve the heatsink transfer function of the heatsinks.
In my Pass Aleph 4 I'm ditching over 400W of heat into 8 heatsinks. Using two 50mm fans running at half voltage (ie the 2 x 12V fans in series across 12V). I've reduced the heatsink temperature by nearly 10 degrees and the fan noise is almost inaudible.
In my Pass Aleph 4 I'm ditching over 400W of heat into 8 heatsinks. Using two 50mm fans running at half voltage (ie the 2 x 12V fans in series across 12V). I've reduced the heatsink temperature by nearly 10 degrees and the fan noise is almost inaudible.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.