my heatsink calcs for Zen v4

[faustian bargin]

the quest is for:

in general, understanding of heatsink selection; and

in specific, minimum heatsink limitations for the Zen v4 amp.

here are my calcs.

given:
the zen amp 'idles' at 2A. being totally new to this, i'm not certain what 'idle' means. anyway i'm assuming the transistors each have 2A running through them.

voltage across Q1 = 22v.
voltage across Q2 = 19.34v.
voltage across Q5 = 8v.

the wattage at each Q is therefore:

Q1 = 44W.
Q2 = 38.68W.
Q5 = 16W.

at a temperature rise of 30°C (recommended by Mr. Pass in his project writeup), each Q will need to have heatsinking at the following ratings, not to exceed:

Q1 = 0.68°C/W
Q2 = 0.78°C/W
Q5 = 1.88°C/W

-----------
as an aside, using a more conservative 25°C rise gives the following max ratings:

Q1 = 0.57°C/W
Q2 = 0.64°C/W
Q5 = 1.65°C/W
-----------

because this is a project guided by economics i want the most heatsink/dollar. to date the best i've found is this $30 extrusion at MECI, which has the properties shown here at aavid thermalloy.

the total Rth for that extrusion at 24.75" is 0.28°C/W. i plugged in a few alternate lengths for that profile on aavid's worksheet:

1" = 1.39°C/W
2" = 0.96°C/W
3" = 0.80°C/W
4" = 0.69°C/W
5" = 0.62°C/W
6" = 0.57°C/W

so, the following lengths of extrusion could work for the components, individually:

Q1 = 4.125"
Q2 = 3.25"
Q5 = 1"

total length = 8.375" for one channel. assuming i can use the heatsink as a continuous piece, with the components simply spaced apart according to their required lengths, there ought to be plenty of extrusion to fit 2 channels on a single heatsink.

-------
(even using the 24° rise figures, the required lengths still work: Q1 = 6", Q2 = 5", Q5 = 1"; total = 12".)
-------

so, the big question is, does it work well enough to have components spaced apart?

i think i'm willing to give it a try. if i'm totally barking up the wrong tree, perhaps someone will point out the error of my ways. otherwise, well maybe i'll burn through a few FET's.

plus, i think there's great 'cool factor' potential (pun, haha) - an amp design with a single long extrusion will be interesting...

at the very least, i hope this will help other fellow newbies who may be trying to get a handle on the economics of heatsinking.

/andrew - now tries to think of a way to explain a 25" hunk of metal in the living room to the wife...

[till]

hmm. one heatsink 85*150*250 mm each channel works well at 2 Ampere Bias.

[faustian bargin]

Originally posted by till
hmm. one heatsink 85*150*250 mm each channel works well at 2 Ampere Bias.

good to hear...the extrusion i'm looking at is 80*178*629...pretty comparable.

[ultrachrome]

Using the 24" MECI heatsink vertically is not supposed to be very efficient according to Rod Elliot's heat sink article. There's an Zen on passdiy that has that configuration and the builder notes that even with two 24" heatsinks, it is still very hot.

You might be better off slicing them up into 8" tall pieces or around there and setting them side by side. I plan to do this. They are actually 24.75" long so you could get three 8" or four 6" pieces out of it very easily.

If you don't have a saw big enough to cut it, you could rent a 14" chop saw from Home Depot. Using an abrasive blade will take a little while as I found with my 12" DeWalt but it will do it. I've since moved to a Freud LU89M012 per PDaniel's recommendation.

Also, if you don't have these heatsinks already, you should know that the back is not completely flat. There are four machined pads 2.25" wide and I estimate between 1/32" and 1/64" deep.
You could tie these together using copper bar from http://www.onlinemetals.com or simply angle aluminum.

Here's a picture of an 8" piece I cut and the uncut piece next to it.

[UrSv]

Originally posted by faustian bargin
the quest is for:

in general, understanding of heatsink selection; and

in specific, minimum heatsink limitations for the Zen v4 amp.

here are my calcs.

given:
the zen amp 'idles' at 2A. being totally new to this, i'm not certain what 'idle' means. anyway i'm assuming the transistors each have 2A running through them.

voltage across Q1 = 22v.
voltage across Q2 = 19.34v.
voltage across Q5 = 8v.

the wattage at each Q is therefore:

Q1 = 44W.
Q2 = 38.68W.
Q5 = 16W.

-- snip --
/andrew - now tries to think of a way to explain a 25" hunk of metal in the living room to the wife...

Only Q1 and Q2 carry that high current. You have noticed that they will dissipate about the same power and the total power is the supply voltage times the current which is 42 times a little more than 2 A which is close to 90 W and the article states 100 W so we use that number. For the others transistors you don't have to worry. For those 100 W you want the 30 degree increase which gives you 0.3 K/W heatsink for one channel. If you wish you may split that into 2 (or more providing you join them well thermally) pieces at 0.6 K/W. You cannot take a 0.6 K/W heatsink and double the length and use that as suggested. The pieces need to be separate and side by side, not on top of eachother. As you have stated the total is 0.28 K/W for the WHOLE extrusion and that is about what you need for one channel. But as you also stated it is 0.57 for a 6 inch piece so cutting the 24 inches into 4 equal pieces and using them side by side gets you 0.14 K/W instead of the whole length 0.28. That is how you should use it - 4 pieces side by side (e.g. 2 mono blocks or a big stereo amp with two pieces each side).

[FrankDIY]

I am planning to use the same trick (4 pieces of 9'' section for each channel). I will put one transistor on each piece. The fourth piece will be unused (no transistor on it).

[faustian bargin]

Originally posted by ultrachrome
Using the 24" MECI heatsink vertically is not supposed to be very efficient according to Rod Elliot's heat sink article. There's an Zen on passdiy that has that configuration and the builder notes that even with two 24" heatsinks, it is still very hot.

....

If you don't have a saw big enough to cut it, you could rent a 14" chop saw from Home Depot.
....

ah. well, there's my answer. if i can't get what i need from an uncut 24.75" extrusion, then it will probably be more cost-effective for me to use 4 of the 5.25" pieces (which perform at 0.60°C/W) MECI is selling for $9.95 each. if i wanted to cut the longer one, i'd have to rent a saw, and the cost of that would be more than the difference.

[ultrachrome]

That's a good point. I spent about $80 on a blade to cut those heatsinks but since I plan to use it for other things and it gives me the flexibility to use the heatsinks the way I want to, I decided it was worth it.

Sometimes I wonder if I am buying so much equipment to build my amps that I could have just saved myself the effort and bought a nice production amp. But then I remind myself that I would then miss out on half the fun.

Even still, I have to justify every tool purchase. I don't want to end up with a garage full of tools that I don't use.

[planet10]

Originally posted by François
I am planning to use the same trick (4 pieces of 9'' section for each channel). I will put one transistor on each piece. The fourth piece will be unused (no transistor on it).

Did you get that batch of heatsinks you were talking about on the trading post?

dave

[FrankDIY]

No I lost the bid and I forgot to update the previous thread. I have now to use the same heatsink they are talking here. I have a 6 feet bar that I will cut into 8 pieces of 9''. I am now wondering if I can use longer wire to mount all the transistor on a separate section of the heatsink. It may be overkill but safe.