pinkmouse said:
let's say that you have 200mm long heatsink;
his C/W ratio is ,say, 0,5 ;
you put some 100W of dissipation on that heatsink
math is simple (this time without too much precision)
Tsum= 25C +100x0,5x1,1(for mica etc) = 80C
if you have same heatsink with 400mm length,his C/W ratio is only in our dreams just half of previous heatsink;anyway-not 0,5/2
hypotethically- his C/W is somewhere around 0,38
so we have :
25C + 200x0,38x1,1 = 108,6C
my conclusion is that we must put on same heatsink only those components which must have same temp and mutual temp tracking,in other words-only components of one channel.
sorry-I checked this math in flesh long time ago,but I don't have in that time nothing to measure temp ,except my hand;
Heatsinking Small Qs
I'm finally populating my Pinkmouse Boards.
Regarding mounting of the bias sensing transistor.....I found your prior post with reference to Q111 being mounted off of the Driver Board, and on the warmest part of the heatsink. Is there any point to mounting the entire driver board on the heatsink so that Q109 and Q110 are also on the heatsink, or is this a bad idea? If you mount Q111 off pcb and on the main sink, should Q109 and Q110 have their own small sinks?
I also found some prior posts discussing the desirability of heatsinking Q107 and Q108....some said yes, some said its not needed.....what is the current thinking on this? If some sinking is advised, how much?
Robert
lgreen said:
I'm finally populating my Pinkmouse Boards.
Regarding mounting of the bias sensing transistor.....I found your prior post with reference to Q111 being mounted off of the Driver Board, and on the warmest part of the heatsink. Is there any point to mounting the entire driver board on the heatsink so that Q109 and Q110 are also on the heatsink, or is this a bad idea? If you mount Q111 off pcb and on the main sink, should Q109 and Q110 have their own small sinks?
I also found some prior posts discussing the desirability of heatsinking Q107 and Q108....some said yes, some said its not needed.....what is the current thinking on this? If some sinking is advised, how much?
Robert
Q107/8 get warm, no more, so they should be ok with no sinks. I mounted the driver board on the main heatsink and that tracks the temperature well. I see no reason not to use the big main heatsink as Q109/10 get very warm in use, so you might as well mount them to the biggest lump of ally you have, and save the expense of another one.
Hi Choky,
I think your conversion from 200mm wide to 400mm wide is wrong.
When you increase the length (height) of a heatsink you apply the square root rule i.e. 200 to 400 high increases the heat dissipation by sqrt(400/200) =1.414 times.
When you increase the width, the heat dissipation is proportional to the increase in width.
There would be a slight problem with left channel feedback when right channel is developing the heat. But the high resistance in the horizontal direction along the angle will remove much of this interaction. This amp seems to need more coupling between bias controller and driver temp than output temp. Although personally I don't agree. But the alternatives are less attractive;- monitor both driver and output on a common sink raises the driver temps needlessly or monitor the outputs only and try hard to keep the driver at a constant low temp. I would tend towards the last option.
If you do adopt the last option then cross coupling may mislead the two feedback paths when gross power imbalance exists betwen channels. However the stages are running in ClassA and run at constant power dissipation until you drive them out of ClassA, so the problem goes away until you ask too much on a continuous basis, then it's no longer a domestic situation.
I think your conversion from 200mm wide to 400mm wide is wrong.
When you increase the length (height) of a heatsink you apply the square root rule i.e. 200 to 400 high increases the heat dissipation by sqrt(400/200) =1.414 times.
When you increase the width, the heat dissipation is proportional to the increase in width.
There would be a slight problem with left channel feedback when right channel is developing the heat. But the high resistance in the horizontal direction along the angle will remove much of this interaction. This amp seems to need more coupling between bias controller and driver temp than output temp. Although personally I don't agree. But the alternatives are less attractive;- monitor both driver and output on a common sink raises the driver temps needlessly or monitor the outputs only and try hard to keep the driver at a constant low temp. I would tend towards the last option.
If you do adopt the last option then cross coupling may mislead the two feedback paths when gross power imbalance exists betwen channels. However the stages are running in ClassA and run at constant power dissipation until you drive them out of ClassA, so the problem goes away until you ask too much on a continuous basis, then it's no longer a domestic situation.
anyway-it's almost unimportant in Mark's case,just because he will use forced cooling;
I said what I said from one reason only- for me looks that Mark loves to keep things tidy and proper ...
anyway- from my xperience-it's always better to have two independent smaller heatsinks than mutual bigger one (no matter in what dimension is division made-L or W) ...
but-mebbe I'm wrong
Why are you guys screwing around with Plitron? They raised their prices, nothing is ever in stock.
Avel.......OTOH......has stock, is cheaper, and despite obviously being made in China now, are good. Besides, how do you know that Plitron does not wind the primaries overseas, and finishes them in Canada? Hmmmmm?
Jocko
Avel.......OTOH......has stock, is cheaper, and despite obviously being made in China now, are good. Besides, how do you know that Plitron does not wind the primaries overseas, and finishes them in Canada? Hmmmmm?
Jocko
On those Avel's at Parts Express... I picked 2 special order 800VA 30-0-30 (not in catalog) from them vs/ their lower VA instock's
About a 2 week turnaround at 76$ each. China whatever, They match perfectly to each other.
You might mention they need to pack them better as mine were put in a box too big with too little foam packing.
They survived unscratched but could of been broken easily.
David
About a 2 week turnaround at 76$ each. China whatever, They match perfectly to each other.
You might mention they need to pack them better as mine were put in a box too big with too little foam packing.
They survived unscratched but could of been broken easily.
David
Mark A. Gulbrandsen said:
Just got a response from Anatek (1 week later), they are getting 25.5-0-25.5 400VA in stock in about 5 weeks.
25.5 x sqrt(2) = 36V
Using the same math (if it's right) 40V rails would be a ~28V transformer?
I'm not sure I want to wait for their arrival given PartsExpress has AVEL's in stock in the same price ballpark.
~Brad