Maker Pro
Maker Pro

TO3 & HEATSINK

E

exxos

Jan 1, 1970
0
Hi all,

Instead of mounting the TO3 transistor ont he flat part (bottom) onto the
heatsink, do you lot recon I could drill a large hole in a heatsink and
mount the can side into it, and put heatpaste around the can and screw it
down from the top ? does anyone have any thoughts if this would work as good
or worse than the bottom just being bolted to the heatsink ?

The transistor I have in mind is a 2N5038, a little higher wattage than the
3055, I plan to have around 50watts constant disipation, it should be ok,
right ??

Thanks
chris
 
N

Norm Dresner

Jan 1, 1970
0
exxos said:
Hi all,

Instead of mounting the TO3 transistor ont he flat part (bottom) onto the
heatsink, do you lot recon I could drill a large hole in a heatsink and
mount the can side into it, and put heatpaste around the can and screw it
down from the top ? does anyone have any thoughts if this would work as good
or worse than the bottom just being bolted to the heatsink ?

The transistor I have in mind is a 2N5038, a little higher wattage than the
3055, I plan to have around 50watts constant disipation, it should be ok,
right ??

Thanks
chris

The actual die is mounted to the base of the package. Your scheme increases
the thermal resistance from the junction to the case.

Norm
 
C

CWatters

Jan 1, 1970
0
It will be worse. You are increasing the thermal resistance between the die
and heatsink quite a lot - the die is on the center of the base plate
normally only 1 or 2 mm away from the heatsink. Try this.. build it or a
test rig as you describe. Power it up and measure the temperature in the
middle of the T03 base plate. Then calculate the die junction temperature
(Tj) using....

Tj = (WxR)+Tc

where

R= Thermal res of the case = 1.25 C/W for a 2N5038 in a T03
W= power in watts = 50
Tc= Temp of case measured in the center of the base.

The calculated value of Tj must be below 200 C according to the data and
150C would give it longer life.

It's impossible to calculate how big a heatsink you will need because the
case res R is specified from the die junction to center of the base not the
mounting holes.
 
N

Norm Dresner

Jan 1, 1970
0
Norm Dresner said:
The actual die is mounted to the base of the package. Your scheme increases
the thermal resistance from the junction to the case.

Norm

After rereading your post, I see an even worse situation than poor thermal
conduction (which can be alleviated with an appropriate H/S and that's the
fact that the transistor case isn't at ground potential. With traditional
mounting, you can get nice mica (or other material) insulators to isolate
the transistor's electric potential from the H/S but not (as far as I know)
with your scheme.

Exactly what are you trying to accomplish?

Norm
 
E

exxos

Jan 1, 1970
0
Norm Dresner said:
After rereading your post, I see an even worse situation than poor thermal
conduction (which can be alleviated with an appropriate H/S and that's the
fact that the transistor case isn't at ground potential. With traditional
mounting, you can get nice mica (or other material) insulators to isolate
the transistor's electric potential from the H/S but not (as far as I know)
with your scheme.

Exactly what are you trying to accomplish?

Norm


Its not a good idea then, I was just trying to see if I could avoid routing
the connection wires though the finn side of the heatsink. The device I
have found is the 2N5038, its for a audio amplifier, its about the best gain
and wattage I can find, The only other transistor is the TIP3055 which is no
problem to mount, but lower gain and wattage, perhaps someone where knows
of a good gain transistor ? 5amps 80V 100W or there about with a good gain,
more there better I guess.

Chris
 
N

Norm Dresner

Jan 1, 1970
0
exxos said:
Its not a good idea then, I was just trying to see if I could avoid routing
the connection wires though the finn side of the heatsink. The device I
have found is the 2N5038, its for a audio amplifier, its about the best gain
and wattage I can find, The only other transistor is the TIP3055 which is no
problem to mount, but lower gain and wattage, perhaps someone where knows
of a good gain transistor ? 5amps 80V 100W or there about with a good gain,
more there better I guess.

A little more expensive but good performance in the MJ150xx series of power
transformers. The complementary pair MJ15003/15004 are higher voltage than
you need but especially good for high power audio amplifiers. They're still
TO-3 but there's generally no real problem in routing wires through the heat
sink if you protect them against rubbing and overheating by installing a
gromet in the hole. The 3055 is an old workhorse, best for low frequency
designs like subwoofers.

For TO-220 packages, consider a MOSFETs instead of bipolar output stage --
there are some really good medium power units available in this package.
Norm
 
E

exxos

Jan 1, 1970
0
A little more expensive but good performance in the MJ150xx series of
power
transformers. The complementary pair MJ15003/15004 are higher voltage than
you need but especially good for high power audio amplifiers. They're still
TO-3 but there's generally no real problem in routing wires through the heat
sink if you protect them against rubbing and overheating by installing a
gromet in the hole. The 3055 is an old workhorse, best for low frequency
designs like subwoofers.

For TO-220 packages, consider a MOSFETs instead of bipolar output stage --
there are some really good medium power units available in this package.
Norm


The MJ21194 power NPN looks nice, I have already tried some mosfets, They
seem to sound the same, but when I look at set freqencys on the scope the
wave is very much distorted, I've been trying this out for the past 2 days,
ive tried everything I can think of, they just dont seem to work as good as
transistors do. I'm all for using mosfets, though I think for my next
project I'll design a class D amp and use some mosfets for that.

Chris
 
E

exxos

Jan 1, 1970
0
CWatters said:
Oh is that all... What you normally do is mount the transistor on the fin
side so the legs stick through into the box.

Found this example (long URL)

http://homepage.mac.com/alrossnin/....-29 22.49.28 -0700/Image-90A1C1E0925E11D7.jpg

id like to do that, though I cant find any "good" TO3 heatsinks, most
heatsinks are expensive in general so I was just seeing if there is any
other solutions. I can get heatinks where are fair in price, though id have
to drill holes and route the wire between the finns, looks like this is the
only solution.

Chris
 
N

Norm Dresner

Jan 1, 1970
0
exxos said:
The MJ21194 power NPN looks nice, I have already tried some mosfets, They
seem to sound the same, but when I look at set freqencys on the scope the
wave is very much distorted, I've been trying this out for the past 2 days,
ive tried everything I can think of, they just dont seem to work as good as
transistors do. I'm all for using mosfets, though I think for my next
project I'll design a class D amp and use some mosfets for that.

Chris
MOSFETs require different biasing and somewhat different drivers than
BJTs -- you can't just substitute one for the other. There are a few "good"
books on the subject -- Doug Self and Randy Slone are the latest ones I've
read.

Norm
 
E

exxos

Jan 1, 1970
0
Norm Dresner said:
MOSFETs require different biasing and somewhat different drivers than
BJTs -- you can't just substitute one for the other. There are a few "good"
books on the subject -- Doug Self and Randy Slone are the latest ones I've
read.

Norm

yeah I spent a while alterting the bias, the audio output sounded ok, though
somewhat lower volume, on sine wave test the sine on the falling edge seemed
to slow down as it came to the crossover point, I imagine the capacitance of
the fet isn't being discharged fast enough. I will look into those books,
should make interesting reading. whats the titles of the books ?


chris
 
N

Norm Dresner

Jan 1, 1970
0
exxos said:
yeah I spent a while alterting the bias, the audio output sounded ok, though
somewhat lower volume, on sine wave test the sine on the falling edge seemed
to slow down as it came to the crossover point, I imagine the capacitance of
the fet isn't being discharged fast enough. I will look into those books,
should make interesting reading. whats the titles of the books ?
I happen to have G. Randy Slone's "The Audiophile's Project Sourcebook"
sitting on the desk next to me. He's also got "High Power Audio Amplifier
Construction Manual" that I've seen. Doug Self's Audio Power Amplifier
Design Handbook is somewhat more academic but contains a lot of good stuff.
IIRC Amazon's prices aren't bad.

Norm
 
A

Airy R. Bean

Jan 1, 1970
0
Whereabouts in UK are you?

I have 4 heatsinks together with professionally-mounted 2N3055
for disposal. You can have them for free if you're close to me, but
I can't be arsed to post them!

In your suggestion, you need to be aware that the Collector is normally
the case connection. Do you really want your heatsinks to be at something
other than ground to external screwdrivers and the like?
 
E

exxos

Jan 1, 1970
0
Airy R. Bean said:
Whereabouts in UK are you?

I have 4 heatsinks together with professionally-mounted 2N3055
for disposal. You can have them for free if you're close to me, but
I can't be arsed to post them!

In your suggestion, you need to be aware that the Collector is normally
the case connection. Do you really want your heatsinks to be at something
other than ground to external screwdrivers and the like?

I live in Uttoxeter, Staffordshire, yeah I know the collector is the case,
didnt think about the electrical connection, the heatsink would be live, not
good!. I'm looking into using CPU heatsinks at the moment, very cheap plus
on most you get a fan, all for under a tenner! might give it a try.....

Chris
 
A

Airy R. Bean

Jan 1, 1970
0
Sorry but you're too far away to make the gift worthwhile.

Don't forget if you've the heat sink vertical not to make
the valleys vertical, otherwise only the bottom of the
heat sink works. - The air there gets hot and starts to rise, but
further up the valley there's no temperature difference and so
the heat sink doesn't give off any heat. Mount the heatsink
so that the valleys are horizontal - the turbulent flow that results
gives cooling from the whole area. (Little known factoid)
 
Top