TO-3 heatsinks

[Johan Wagener]

I have a power supply with 4 2N3055 (to-3 package) transistors. Any ideas
how I should mount them to a heatsink? How big heatsink do I need? The
supply will deliver about 3A at 20V

 

[default]

On Fri, 5 Sep 2003 21:54:29 +0200, "Johan Wagener"
<pil(nospam)@webmail.co.za> wrote:

I have a power supply with 4 2N3055 (to-3 package) transistors. Any ideas
how I should mount them to a heatsink? How big heatsink do I need? The
supply will deliver about 3A at 20V

A single 2N3055 will handle all 3 amps and then some. Four for 3 amps

is overkill.

The size of the heat sink will depend on how much power is dissipated
in the pass element (output transistor) ambient temperature, how hot
you want the transistor to get, thermal conductivity of the heat sink
metal, what color the heat sink material is, if it is painted - how
much insulation value the paint adds, and how fast heat is removed
from the heat sink.

The amount of power dissipated in the pass element will depend on how
much voltage is across the transistor, Collector to Emitter, and what
current it is.

With a typical design linear regulator you would need a minimum of 24
volts into the regulator and keep 4 across the transistor, at three
amps that's 12 watts.

I would allow about 20 square inches MINIMUM for 12 watts, but that
will depend on air flow, how much safety factor you want and what
input voltage range you want it to run.

Say you want 10% input range, that would be two more volts at 3 amps
or 18 watts or 25 square inches of 1/8" aluminum.

The thickness is desirable under the transistor to carry heat away -
further from the transistor the metal can be thinner. Poor air
circulation, or nearby heat sources, and you'd want it make it larger.

The transistor itself shouldn't get over 105 centigrade under any
circumstance. My rule of thumb is design so it so I can keep my hand
on it without getting burned. (much less than 105 - and about twice
as large as I think it should be)

Someone else may have a formula or two for you -


-----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
-----== Over 100,000 Newsgroups - 19 Different Servers! =-----

 

[Wouter van Ooijen (www.vo]

A single 2N3055 will handle all 3 amps and then some. Four for 3 amps
is overkill.

That might be true when focussing on current alone, but paralleling
power transistors has the benefit that it also parallels the
junction-to-heatsink thermal resistance.



Wouter van Ooijen

-- ------------------------------------
http://www.voti.nl
PICmicro chips, programmers, consulting

 

[Stepan Novotill]

On Fri, 5 Sep 2003 21:54:29 +0200, "Johan Wagener"
<pil(nospam)@webmail.co.za> wrote:

I have a power supply with 4 2N3055 (to-3 package) transistors. Any ideas
how I should mount them to a heatsink? How big heatsink do I need? The
supply will deliver about 3A at 20V

To calculate the size of the heatsing, you need to know how much

voltage is between emitter and collector, as well as the current, so
you can calculate power V*I

....But I would just mount them on the aluminum box of the power
supply. Why TO-3 though? those are a lot of drilling and screwing.

 

[Stepan Novotill]

On Sat, 06 Sep 2003 11:24:50 GMT, wouter@voti.nl (Wouter van Ooijen
(www.voti.nl)) wrote:

A single 2N3055 will handle all 3 amps and then some. Four for 3 amps
is overkill.

That might be true when focussing on current alone, but paralleling
power transistors has the benefit that it also parallels the
junction-to-heatsink thermal resistance.

I beleive you can run them quite hot.

 

[default]

On Sat, 06 Sep 2003 11:24:50 GMT, wouter@voti.nl (Wouter van Ooijen
(www.voti.nl)) wrote:

A single 2N3055 will handle all 3 amps and then some. Four for 3 amps
is overkill.

That might be true when focussing on current alone, but paralleling
power transistors has the benefit that it also parallels the
junction-to-heatsink thermal resistance.



Wouter van Ooijen
Quite true.

But he's using a transistor capable of a maximum of 15 amps for 3.
He doesn't need four - unless that's a typo and he meant to write 30
amps . . . Using four would increase the base drive requirement and
necessitate emitter bias resistors. A lot of extra work, for gross
overkill in my opinion.


-----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
-----== Over 100,000 Newsgroups - 19 Different Servers! =-----

 

[Johan Wagener]

I can't seem to get the circuit posted on here. I would like you to have a
look at it. The reason why I have 4 2n3055's in there is so I can upgrade
the PSU later on. I tried to attach a gif file to my messages but it won't
show up on the newsgroup

 

[default]

On Sat, 06 Sep 2003 11:51:51 GMT, Stepan Novotill
<snovotill@hotmail.com> wrote:

On Sat, 06 Sep 2003 11:24:50 GMT, wouter@voti.nl (Wouter van Ooijen
(www.voti.nl)) wrote:

A single 2N3055 will handle all 3 amps and then some. Four for 3 amps
is overkill.

That might be true when focussing on current alone, but paralleling
power transistors has the benefit that it also parallels the
junction-to-heatsink thermal resistance.

I beleive you can run them quite hot.
Operating case temp range is 200 Celsius. Military design might

require operational testing of a supply to 125 (one place I worked at
did).

Realistically, in most design, 100 is plenty. Avoids problems with
insulation of the wires softening, circuit board turning brown,
accidental burns etc.


-----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
-----== Over 100,000 Newsgroups - 19 Different Servers! =-----

 

[default]

On Sat, 6 Sep 2003 15:15:04 +0200, "Johan Wagener"
<pil(nospam)@webmail.co.za> wrote:

I can't seem to get the circuit posted on here. I would like you to have a
look at it. The reason why I have 4 2n3055's in there is so I can upgrade
the PSU later on. I tried to attach a gif file to my messages but it won't
show up on the newsgroup

Alt electronics won't take a binary, can you get

alt.binaries.schematics.electronic? They do take binaries.

The other alternative is to use ascii to draw the schematic in text
form.

If you actually plan to use more current, you'd have to size the heat
sink with the anticipated current in mind.

Consider 2N3771; a 30 amp transistor with similar specs to the 3055?

If you really intend to use a lot of current, consider improving the
efficiency of the supply by driving the NPN pass transistor from a
higher bias source. For instance, your high current source is 20-24
volts (whatever it is) at the rated current. The NPN is not "self
biased" (that's where the 4 volt drop requirement comes in) but biased
from a low current supply that supplies an additional 4 volts (at the
base drive current) to the pass transistor to turn it on. Then the
drop between collector and emitter is .6 volts. The small 4 volt
supply is referenced to the + of the high current supply, and adds 4
volts for base bias alone.

And if you really want serious power and good efficiency consider
using an Thyrister (SCR) to provide rectification and control in one
element (gets rid of the rectifier heat sink and diode drop of the
rectifiers - with a center tapped transformer).


-----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
-----== Over 100,000 Newsgroups - 19 Different Servers! =-----

 

[default]

On Sat, 6 Sep 2003 15:15:04 +0200, "Johan Wagener"
<pil(nospam)@webmail.co.za> wrote:

I can't seem to get the circuit posted on here. I would like you to have a
look at it. The reason why I have 4 2n3055's in there is so I can upgrade
the PSU later on. I tried to attach a gif file to my messages but it won't
show up on the newsgroup

One little trick I used with my last hi current PS design was to use

an angle aluminum extrusion and bolt a CPU heat sink between the two
SCR's to increase surface area.

I noticed in your other post that you were thinking of
thermostatically controlling a fan. Lots of fans built for computers
already have thermistors and control circuits built in - and at least
in the US are cheap and readily available on the surplus market.

I added one to my computer case, it idles along, to carry air across
the sensor, until the temp gets around 35C, then kicks into high
speed.

Your country code indicates you are from South Africa?


-----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
-----== Over 100,000 Newsgroups - 19 Different Servers! =-----

 

[Johan Wagener]

Yes, I am from South Africa. Buying a thermal fan could work, but it would
be best if I built it myself because this is actually for a subject called
projects. So the more I do myself the more marks I will receive. How would I
go about getting the circuit to ascii? I have it drawn on circuitmaker.

 

[default]

On Sat, 6 Sep 2003 18:23:51 +0200, "Johan Wagener"
<pil(nospam)@webmail.co.za> wrote:

Yes, I am from South Africa. Buying a thermal fan could work, but it would
be best if I built it myself because this is actually for a subject called
projects. So the more I do myself the more marks I will receive. How would I
go about getting the circuit to ascii? I have it drawn on circuitmaker.

Your picture in alt.binaries------ works fine. No need for ascii

unless you want to post in two newsgroups. The binaries group doesn't
care if you post text there it is loosely run or un moderated. I
think most of the ascii posters just do it by looking at it (using a
fixed, not variable, width font) Personally it is too limiting and I
don't like reading it that way - but whatever works for you is fine.

I use Design Works and just do a widows copy from Design Works then
paste it into Irfanview (graphics program) then do a "save as" from
there to put it in a GIF attachment file. That may work with Circuit
Maker if it doesn't allow you to export or save as a different format.

I don't understand how your adjustable current works. You're taking
advantage of the current limit of the 7815? Give me a theory of
operation on the current limiting?

Does this also allow you to heatsink the pass transistors to the same
heat sink as the 3055's and buy you thermal limiting as well?

Provisions are there for remote voltage sensing from the looks of it.

My efficiency idea of the additional small PS would work. Totally
overkill for a mere three amps.

OK what do you have in mind for the fan? Most of the brushless DC
fans, as well as brush types, have a low voltage that the motor won't
turn at. With some brushless designs you have to hit it with voltage,
if you creep up on the voltage the fan won't turn even when its
maximum operating voltage is reached. I built two voltage controlled
fans one brushless the other brush type - neither uses a thermistor,
but I have designed a few thermostats as well.

You planning on trying for proportional speed control based on
temperature or just turn it on when the heatsink gets hot? Where will
you sense? ambient or heat sink?

You'll doubtlessly draw comments from the schematics group - they
can't resist critiquing a design.


-----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
-----== Over 100,000 Newsgroups - 19 Different Servers! =-----

 

[default]

On Sat, 6 Sep 2003 18:23:51 +0200, "Johan Wagener"
<pil(nospam)@webmail.co.za> wrote:

Yes, I am from South Africa. Buying a thermal fan could work, but it would
be best if I built it myself because this is actually for a subject called
projects. So the more I do myself the more marks I will receive. How would I
go about getting the circuit to ascii? I have it drawn on circuitmaker.

Correction:

Scratch:

Does this also allow you to heatsink the pass transistors to the same
heat sink as the 3055's and buy you thermal limiting as well?

Instead:
Do you plan to use the internal thermal limiting of the 7815 to
provide thermal limiting for the pass transistors?


-----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
-----== Over 100,000 Newsgroups - 19 Different Servers! =-----

 

[Rev. 11D Meow!]

use a circuitmaker to ascii converter program . . .



"Johan Wagener" <pil(nospam)@webmail.co.za> wrote in message
news:3f5a0a17$0$64720@hades.is.co.za...

Yes, I am from South Africa. Buying a thermal fan could work, but it would
be best if I built it myself because this is actually for a subject called
projects. So the more I do myself the more marks I will receive. How would
I
go about getting the circuit to ascii? I have it drawn on circuitmaker.

 

[Bob W.]

I didn't think they still made 2N3055 transistors anymore. Got that
idea from "Troubleshooting Analog Circuits" by Pease <a National Semi
engineer>. They were used as audio output devices and power supply
pass transisitors back in the day.

 

[default]

On 9 Sep 2003 15:47:02 -0700, dobertw@yahoo.com (Bob W.) wrote:

I didn't think they still made 2N3055 transistors anymore. Got that
idea from "Troubleshooting Analog Circuits" by Pease <a National Semi
engineer>. They were used as audio output devices and power supply
pass transisitors back in the day.
The venerable 3055? That's like the '57 Chevy. Wash your mouth out

with soap.

I too bow to Pease. Bob Pease is right up there . . .(but he is a
National Semi. spokesperson, and newer ain't always better, or cost
effective).

Yeah, there are better transistors, but Johann is in South Africa, and
perhaps they are behind, or maybe he's using what he has or was told
to use - this is some engineering course. Schools are always way
behind the state of the art (forgetting Cal Tech, MIT, Stanford, et
al)

This design has to be scalable to a higher current. Until Johann
steps in again (if ever) we won't know. We don't know what all the
constraints are . . . There could be good reasons for what he's doing.

The object is to stroke the professor, not to design the dynamite,
best ever, power supply.

BTW: I worked under, and shared a brew or two with, Fred Baker.


-----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
-----== Over 100,000 Newsgroups - 19 Different Servers! =-----