Another heatsink question!
 User Name Stay logged in? Password
 Home Forums Rules Articles Store Gallery Blogs Register Donations FAQ Calendar Search Today's Posts Mark Forums Read Search

 Solid State Talk all about solid state amplification.

 Please consider donating to help us continue to serve you. Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
 24th July 2004, 06:46 AM #1 diyAudio Member   Join Date: Apr 2001 Another heatsink question! Hi! I built Holtons lateral amp with +-65V voltage rails (so something like 160W/module) and got a 0,4K/W heatsink. Will it be ok to mount both modules on this heatsink or is it too small? my calculations...don't know if they are right... ok... assuming 320W power output and 30% additional power that goes just into heat, about 100W (96W) of heat is produced. with my heatsink (0.4 K/W) i have a teperature raise of about 40°. Ambienttemperature is 20°C-25°C (on a hot day where I live). So the amp temperature will be something from 60°C to 65°C at full power? The outputdevices are rated at 55W/pcs at 65°C. Four devices in the output produce 160W, every device has to provide 40W. So I think the heatsink will do it...but not sure if my calculations are right... What temperature is ok for a poweramp? Please correct me! Konstantin
 24th July 2004, 09:13 AM #2 diyAudio Member     Join Date: Feb 2004 Location: Recife - Brasil Northeast Just as reference, i do not know how to calculate But made some thousand amplifiers, this way, life experience and i use one rule.... each 10 RMS watts need 3 by 3 inches side aluminium with 2 milimeters thick... this way 9 squared inches each 10 watts. My environment temperature normally is 30 degrees Celsius, the centigrade scale. they are working. Your amplifier is 160 watts power, so, 16 aluminium fins, 3 X 3 inches, if square shape thin, and 16 thins if outside the equipment, without cooler and correct position for convection air movement. Sorry not to use calculations, since this simple idea works, i started to use simple mathematicas.... 9 square inches corresponds to 10 watts X square inches corresponds to 160 watts. 10X=9X160 Carlos __________________ Sons visiting uncle charlie; http://www.youtube.com/watch?feature...&v=5T76qsvZrdw; http://www.youtube.com/watch?v=Hb_wolQGU5Y
 24th July 2004, 09:42 AM #3 diyAudio Member     Join Date: Jan 2002 Location: Earth I'm curious as to how you got 160W per module. To get 160W average power into an 8-ohm load requires an amplitude of 51V. An amplitude of, say, 63V would give 250W. If your figures are correct then the psu is +/-65V and the output amplitude is +/-51V. **Assuming** the output transistors connect between the speaker and the 65V power rails, Worst case, each transistor will dissipate about 50W average and 130W peak. And this assumes a resistive 8-ohm load. Do you have the transistor part numbers? If they are rated at 55W/65C they will not be up to the job.
 24th July 2004, 11:24 AM #4 diyAudio Member   Join Date: Apr 2001 Output devices are 2SK1058/2SJ162. Anthony writes that the amp will deliver: +-28VDC -> 50W/module +-50VDC -> 100W/module +-70VDC -> 200W/module (approxmately, all into 8 Ohms ) If you draw a little grph with this values you get (about, +-30W or something) 160W per modules with +-65VDC supply rails. The schematic (and some pics) can be found on Anthonys site: http://www.aussieamplifiers.com/lateral.htm I guess I need a bigger heatsink... Konstantin
diyAudio Member

Join Date: Jan 2002
Location: Earth
Ok, I see now that each module has 4 output transistors. This helps.
With +/-65V rails and an 8-ohm resistive load, each module will dissipate 100W average and each transistor will dissipate 25W average and 66W peak. The 2SK1058 is rated at 100W peak power at 25C. So if your heatsink is at 25C and your speaker is 8-ohms resisitve you'll definitely be ok as regards safe operating area of the transistors.

As each module could dissipate 100W average, your shared heatsink must dissipate 200W. At 0.4K/W and air temperature of 25C, the heatsink temperature will reach 105C. At this case temperature, your transistors are only rated at about 40W - thus they will be damaged because they will be dissipating 66W peak and this infringes on the diagonal part of the SOA curve. To allow 66W the case temperature must be kept below about 67C thus requiring a heatsink of 0.2K/W.

So your heatsink must be at least twice as big/efficient as the one you are contemplating.

Now, from the other thread about device failure it emerges that speakers are not pure resistors and their impedance drops at certain frequencies and so on. It seems to be a sensible rule of thumb to design for half the nominal speaker load as a pure resistance. For an 8-ohm speaker, design for a 4-ohm resistor.

I've attached the load lines per transistor for your amp and power rails. I've shown curves for both 8 and 4 ohm resistive loads. You can see that the SOA is just about ok for 4 ohms, so that's fine too at 25C. A problem occurs when we look at the temperature derating. The transistors can only dissipate 280W peak at 25C and, perhaps 140W peak at 67C. In other words, if you are running your amp continuously at 160W per channel, your heatsink will warm up to 67C and your transistors will then only just work for 4-ohms resistive.

The design needs transistors that are rated at 130W at the required heatsink temperature where the SOA curve starts to be pulse length limited. It appears that the 2SK1058 is just about ok at 67C. It is hard to judge what the SOA curves do as case temperature rises - the only data is for continuous power. So I've assumed that at 67C the power handling is about half that at 25C on all parts of the SOA curve.

I must say, however, that I wouldn't be happy with the reliability of this amp with +/-70V rails and 70C heatsink temperature into real speaker loads.
Attached Images
 soa holden amp.jpg (72.5 KB, 65 views)

 24th July 2004, 02:39 PM #6 diyAudio Member     Join Date: Jan 2002 Location: Earth This is quite a complex thing to work out. On reflection for a moment, it all depends how you interpret the temperature derrating of the 2SK1058. Another way to look at is it to identify the maximum channel power dissipation of the device, at time periods likely with a music signal. From the SOA curves I look at the point where the curve starts to slope downwards. Here the power is between 210W and 280W depending upon duration. Let's use 280W. The device can dissipate 280W at 25C. Now to derrate this is uncertain. One way is to look at the continuous power derrating curve and use that. So at 67C the continuous power is reduced by about 40%. If we apply that to the peak we get 280*0.6 = 170W. This is ample for 66W peak dissipation. At 105C (your 0.4K/W heatsink) the continuous power is derated by about 65%. Peak power may now be only 100W. This is fine for 66W peak but not enough for 4-ohms which needs 130W - a real possibility with a speaker. To be safe for a 4-ohm resistive load you need 130W peak which is about 54% derating. This implies the heatsink needs to keep the case teperature below 92C. By this reckoning, you need a heatsink dissipation of better than 0.3K/W at 25C free air temperature, not 0.2K/W as I stated earlier. I think 0.3K/W will be fine if your room is at normal temperatures.
 24th July 2004, 03:22 PM #7 diyAudio Member   Join Date: Apr 2001 Thank you guys! You helped a lot! I think I will try to get another heatsink of the same typ... Thank you! Konstantin
 24th July 2004, 04:31 PM #8 diyAudio Member   Join Date: Oct 2003 Location: L.A., CA Get as much heatsink as you can afford. The bigger the better, the safer the amp and will last longer. __________________ If it sounds good... it is good!

 Posting Rules You may not post new threads You may not post replies You may not post attachments You may not edit your posts BB code is On Smilies are On [IMG] code is On HTML code is OffTrackbacks are Off Pingbacks are Off Refbacks are Off Forum Rules

 Similar Threads Thread Thread Starter Forum Replies Last Post Apogee Pass Labs 195 23rd July 2008 08:07 PM xdissent Solid State 13 18th June 2004 02:44 AM kheldar Chip Amps 8 30th January 2004 02:50 PM tpenguin Solid State 3 23rd December 2003 09:47 PM BrianGT Pass Labs 3 19th September 2002 07:55 PM

 New To Site? Need Help?

All times are GMT. The time now is 07:42 AM.