cviller said:
Thanks for your encouraging comments and for this hint.
I know that the efficiency of the heatsinks is worse, when used in this orientation, but I simply like the looks of horizontal fins more
As I wrote, the enclosure will be made of quite solid aluminum chunks, too. They will be in direct contact to the heatsinks and I hope they will contribute to heat-dissipation a bit.
I will watch this for a while and if the amps really get too hot, I will change the direction of the heatsinks. But it doesn´t look bad now - I think temperature is still in the "green range".
Cheers
I was at my local metal surplus/supply place looking for heatsinks for this project. They usually don't have anything close to massive enough.
Then I found the attached! Tragically they only had one, but I bought it anyway. After all- it HAD to be big enough for an F4. The base is 18mm thick! so no problem spreading the heat! The fins 60mm long! weighs 7 Kg.
To satisfy the educators here, I did my homework:
schuhbu's report included pictures so I counted the number of fins per heatsink. Using his size information I calculated his sq cm of fin surface as about 4,000. Mine came out to 6,700. So, not quite enough for 2 channels since he says his runs pretty hot. BUT my fins are spaced further apart and I could orient the fins in the "correct" direction vertically so I'll give it a try.
Then I found the attached! Tragically they only had one, but I bought it anyway. After all- it HAD to be big enough for an F4. The base is 18mm thick! so no problem spreading the heat! The fins 60mm long! weighs 7 Kg.
To satisfy the educators here, I did my homework:
schuhbu's report included pictures so I counted the number of fins per heatsink. Using his size information I calculated his sq cm of fin surface as about 4,000. Mine came out to 6,700. So, not quite enough for 2 channels since he says his runs pretty hot. BUT my fins are spaced further apart and I could orient the fins in the "correct" direction vertically so I'll give it a try.
Attachments
If it runs too hot still, I have a backup plan:
The "twin carburetor" version.
I put the 120mm fans in series and they are quite quiet-at half speed.
Finally a question:
Would it be overly optimistic to think that I can double the capacity of the heatsinks with the fans? I can feel air blowing out the slots between the fins.
The "twin carburetor" version.
I put the 120mm fans in series and they are quite quiet-at half speed.
Finally a question:
Would it be overly optimistic to think that I can double the capacity of the heatsinks with the fans? I can feel air blowing out the slots between the fins.
Attachments
You would more than double the capacity 
You are planning to use the fan's in series.
Are they run by 12VDC or 24VDC ?
If 12VDC or similar for each, then you can easily use zeners for lovering the voltage, by putting a zener in series with the supply.
You could use a switch with a couple of different zeners and you have a nice fan speed control.
All you have to be sure about, is the wattage of the zeners and not going to low on the voltage, or the fans will not start.
You are planning to use the fan's in series.
Are they run by 12VDC or 24VDC ?
If 12VDC or similar for each, then you can easily use zeners for lovering the voltage, by putting a zener in series with the supply.
You could use a switch with a couple of different zeners and you have a nice fan speed control.
All you have to be sure about, is the wattage of the zeners and not going to low on the voltage, or the fans will not start.
The most critical part of the whole thermal resistance equation is actually the transistor itself.
If the thermal resistance of the active devices can be reduced, the heatsink requirements are lessened, resulting in a pleasing increase in reliability, and an equally pleasing reduction in the cost of the necessary heatsink.
The thermal resistance from transistor case to heatsink is also very important, and proper mounting technique can result in a significant improvement.
If these two thermal resistances can be reduced enough, the heatsink itself can be quite a bit smaller than would otherwise be the case.
And then with a fan you increase your heatsinks capabillity significant.
But make sure your circuit shuts down when you have a fan failure.
Which is one of the reasons i never use fans. The other is i do not want to hear the fan(s) at any time.
If the thermal resistance of the active devices can be reduced, the heatsink requirements are lessened, resulting in a pleasing increase in reliability, and an equally pleasing reduction in the cost of the necessary heatsink.
The thermal resistance from transistor case to heatsink is also very important, and proper mounting technique can result in a significant improvement.
If these two thermal resistances can be reduced enough, the heatsink itself can be quite a bit smaller than would otherwise be the case.
And then with a fan you increase your heatsinks capabillity significant.
But make sure your circuit shuts down when you have a fan failure.
Which is one of the reasons i never use fans. The other is i do not want to hear the fan(s) at any time.
The fans are 12v and 2 in series start fine. I put 3 in series and no-go.
At correct voltage they are very noisy, at half, they are hard to hear.
Mounted correctly they will be even quieter. I noticed you said WHEN I have fan failure I will put one of those AC thermal switches on the sink as a simple protection.
At correct voltage they are very noisy, at half, they are hard to hear.
Mounted correctly they will be even quieter. I noticed you said WHEN I have fan failure
I will put one of those AC thermal switches on the sink as a simple protection.
OK , I had another idea. I confess that I'm intrigued with the balanced configuration. I plan to build the Russ White balanced DAC when it matures. I have the X'ed Balanced Pass Linesource. A balanced power amp is something I've always wanted.
Even without a fan it appears that my heatsink is more than big enough for a single channel with its 6 mosfets. So if a made four boards, each with only two Mosfets, and combined them in pairs in balanced configuration, it seems that I would get about 35 watts per channel. If I decide to run them unbalanced at some other time, I guess they'd put out about 8 watts times four.
My current speakers are about 96 dB efficient, so those numbers sound like plenty to me.
That means eight mosfets total, which the heatsink should handle fanless..
Ok, I guess the question is whether to go with 2 unbalanced channels with 4 MOSFETS each or 2 balanced channels with 4 MOSFETS each(2 per board)
BALANCED (4 circuits with 2 MOSFETS each)
So the bad for balanced is:
only 20 damping factor?
Less distortion per watt ?
What else?
The good is:
Common mode rejection
Conceptually matches my other components
Power for 2 ch: 33w/ch?
What else?
UNBALANCED (2 circuits with 4 MOSFETS each)
So the bad for unbalanced is:
Slightly noisier
Doesn't match my other balanced components
Half Power for 2 ch: 16 w/ch?
What else?
The good is:
40 damping factor
More distortion per watt output?
What else?
THE MOST IMPORTANT QUESTION:
So either way, to have only 2 or four output devices per board (circuit) , how do I proceed? I suppose I just leave off the excess mosfets, and each one's two associated resistors.
Then I guess I have to lower the bias? I use P1 to do that? Is a different value for the pot and maybe other resistors required?
Other things to do? I just slowly bias it until it reaches the correct temp? That can't be right: if I had a bigger heatsink, I would end up biasing it more.
moving on...
Ok, now on to the question everyone asks Nelson in addition to "How big should my heatsinks be? " That is " I have a tube of IRFP 140's-that will be OK , right? or am I stressing them too much? I think he'll say it's OK.
Even without a fan it appears that my heatsink is more than big enough for a single channel with its 6 mosfets. So if a made four boards, each with only two Mosfets, and combined them in pairs in balanced configuration, it seems that I would get about 35 watts per channel. If I decide to run them unbalanced at some other time, I guess they'd put out about 8 watts times four.
My current speakers are about 96 dB efficient, so those numbers sound like plenty to me.
That means eight mosfets total, which the heatsink should handle fanless..
Ok, I guess the question is whether to go with 2 unbalanced channels with 4 MOSFETS each or 2 balanced channels with 4 MOSFETS each(2 per board)
BALANCED (4 circuits with 2 MOSFETS each)
So the bad for balanced is:
only 20 damping factor?
Less distortion per watt ?
What else?
The good is:
Common mode rejection
Conceptually matches my other components
Power for 2 ch: 33w/ch?
What else?
UNBALANCED (2 circuits with 4 MOSFETS each)
So the bad for unbalanced is:
Slightly noisier
Doesn't match my other balanced components
Half Power for 2 ch: 16 w/ch?
What else?
The good is:
40 damping factor
More distortion per watt output?
What else?
THE MOST IMPORTANT QUESTION:
So either way, to have only 2 or four output devices per board (circuit) , how do I proceed? I suppose I just leave off the excess mosfets, and each one's two associated resistors.
Then I guess I have to lower the bias? I use P1 to do that? Is a different value for the pot and maybe other resistors required?
Other things to do? I just slowly bias it until it reaches the correct temp? That can't be right: if I had a bigger heatsink, I would end up biasing it more.
moving on...
Ok, now on to the question everyone asks Nelson in addition to "How big should my heatsinks be? " That is " I have a tube of IRFP 140's-that will be OK , right? or am I stressing them too much? I think he'll say it's OK.
I'm thinking about integrating preamplifier into this project. Few years ago I made such amplifier (it's basing on Dariusz Kulesza design):
Idea is very similar to F4 but NFB includes the output pair to. Schematic can be found here:
http://www.unisonus.com/pdf/dracula.pdf
An externally hosted image should be here but it was not working when we last tested it.
Idea is very similar to F4 but NFB includes the output pair to. Schematic can be found here:
http://www.unisonus.com/pdf/dracula.pdf
Variac said:A balanced power amp is something I've always wanted.
So the bad is:
only 20 damping factor
not as "manly" as the stock amp
What else?
The good is:
Common mode rejection
Adequate power
Conceptually matches my other components
No fans
Less energy used
Adaquate power.
What Else?
Then I guess I have to lower the bias? I use P1 to do that?
Ok, now on to the question everyone asks Nelson in addition to "How big should my heatsinks be?
First off, I have thought that you are a quiet old man. From today, you are not.
The balanced amp invites 3rd harmonics. Are you familiar with 3rd harmonics? I have thought that Papa developed F4 mainly for whom like 2nd harmonics.
Damping factor of 20 . . . what? Mine (not tube) is lower than that. I'm fine.
The good is no fan? Why fan with such big fins?
I think if P1 is adjusted lower, the bias would be more. By thw way, your speakers have 96dB SPL (I suppose 8-ohms). Why more bias?
Papa always likes to have heatsink temp about 50-55. This must be good temp with respect to all. My fingers are still safe, and I need some temp for a short heat-up time and for the fast reach to the steady-state heat after the turn-on.
We have a saying: "Better to say nothing and be possibly thought a fool than to open your mouth and remove all doubt"
But as I posted here before, this amp is just too great to ignore!OK I have rewritten my previous post that you are commenting on. I hope it makes more sense! I meant to say "Do I adjust P1 to lower the bias?" I think I need to lower the bias due to fewer output devices.
Because the people that have built this say you need a LOT of fins to keep it cool. I calculated my heatsink compared to theirs, and it is possibly not quite big enough for 2 channels.. OF course I can try it, and if the temp is reasonable with the fewer MOSFETS, I can add more
Thak you for your Babo-post. Quite an honor..but it wasn't dumb
Variac said:
look:
http://www.diyaudio.com/forums/showthread.php?postid=1151521#post1151521
you don't need to touch anything;
comparing to Papa's full F4 -with 2 pairs of mosfets you'll have 2/3 of bias;
with 1 pair just 1/3 of bias..........off course-presuming that you somewhere in the past put 3 pairs in it and adjusted bias as Papa commandeddddddddddddddddd........
conclusion- if you have them-put all 3 pairs in amp-spread them nice on heatsink and put that silly pot where you can- to not burn your little fingers
Thanks Zen-Choky . I hadn't seen your post.
The problem is I wanted two channels on this heatsink, If I built it "stock" I would have 12 mosfets on it. I will try to study the Pass Commandment of Biasing.
EDIT: I found the Wisdom of Pass: " .25 volts across the .47 ohm source resistors" So there it is I think. I just build it with as many output device pairs as I think my heat sink will stand, then bias per
The Wisdom of Pass using P1 ?
Also, does the balanced have less distortion for a given output than the unbalanced? That seems to be what I'm seein in the specs...
. I hadn't seen your post. The problem is I wanted two channels on this heatsink, If I built it "stock" I would have 12 mosfets on it. I will try to study the Pass Commandment of Biasing.
EDIT: I found the Wisdom of Pass: " .25 volts across the .47 ohm source resistors" So there it is I think. I just build it with as many output device pairs as I think my heat sink will stand, then bias per
The Wisdom of Pass using P1 ?
Also, does the balanced have less distortion for a given output than the unbalanced? That seems to be what I'm seein in the specs...
Variac said:
you again didn't read my post.......see 0,5A mark somewhere up?
that's Papa's .25 across .47
anyway-put 4 +4 (two pairs/channel) on sink and then crank it as much you can ; even with 1A per vertical pair that's just 22W per mosfet
and yes-balanced always mean less distortion.....at least on paper
cancellation........
