0-30 Vdc Stabilized Power Supply

[Tedy1]

Hi!
Me again
I got 3 pieces of OPA445AP from TI as free samples 
And I wonder on what I should be careful, because I dont want them burn or damaged.
According to datasheet of OPA445 I should use 100k trimpot, is that correct?

 

[audioguru2]

Hi all, I'm new to the EL forums, but not new to the website, nice stuff here

 

[audioguru2]

Hi!
Me again
I got 3 pieces of OPA445AP from TI as free samples

 

[Tedy1]

Thanks for your fast response
Yes, free samples are nice and I also get them very quick 

one more question, could be anything wrong, because I am using 2n2219(w/ heatsink) instead of TIP31 with modified parts written in a text file few pages back?

 

[audioguru2]

Hi Tedy,
A little old 2N2219 in a heatsink can dissipate only a couple of watts and will be at its absolute max temp. A TIP31 can dissipate 20W to 40W. If your project is set for a low output voltage and a high current, you are going to melt the little 2N2219! ;D

 

[Tedy1]

But is it possible to use 2N2219 with modified parts?
I was just asking this, because I dont have TIP31 yet. As soon as I will get out to city Ill buy it and another 2N3055 (have heatsink for 2 of them).

 

[audioguru2]

But is it possible to use 2N2219 with modified parts?

Only if you use a heatsink with an infinite size, with a high velocity fan blowing on it.
Also, only if both your 2N3055 transistors have more gain than their guaranteed minimum.
Also, Only if you don't operate the power supply project at a high output current at a low voltage output.
Otherwise the little 2N2219 will melt, probably short and destroy your load. ;D

Your 2N3055 transistors might also melt sharing a heatsink with up to 120W of dissipation. Even if you don't use transistor insulators they will be twice as hot as they should be. 60W for each transistor on its own large heatsink is a lot of heat.
What is the thermal resistance rating of your heatsink? We will calculate the temperature and see if a fan is required.

 

[maksar]

Hi!

Why not to say simply, that we need 4 transistors for 3A project and 5 for 5A project? But still a large heatsink is needed (I probably will buy this one: http://www.semikron.com/internet/ds.jsp?file=648.html) with fan.
Or it is posible to use more powerfull transistors with thermal resistence ~0.5 C/W (maybe there are even better transistors) and not 1.5 C/W as 2n3055 has. It will be enough to use 2-3 of better transistors).
Also the best solution may be to do teramal control based on termal sensor (sample from Maxim semi, for example) and opamp. So the fan won't work all the time, but only when it is needed (rarely).

Maksar

 

[Tedy1]

HI
I went to city and bought 2N3055, but they didnt have TIP31 so I bought TIP41

http://pubpages.unh.edu/~aperkins/pdf/TIP-devices/TIP31.pdf TIP31 datasheet
http://pubpages.unh.edu/~aperkins/pdf/TIP-devices/TIP41.pdf TIP41 datasheet

TIP41 is similiar to TIP31 but is made for higher power

Heatsink is almost same as the one at original page but it is slightly longer, to fit 2x 2N3055.
I dont know thermal resistance of heatsink, because friend gave it to me, but it was used in 100W amplifier.

 

[audioguru2]

Hi Maksar,
That's a nice heatsink, but I don't think it is made to fit TO3 cased transistors like 2N3055's.
It has a space of only 27mm, and with tolerance might be slightly less. The max width of a 2N3055 is 26.67mm so it would make an extremely tight fit. I don't know if the machined surface inside the heatsink is completely flat for the entire 27mm, it might have small shoulders beside the fins.

It would make an interesting mechanical problem if the transistor didn't quite fit the space in the heatsink:
If you freeze the transistor so it shrinks, will the space in the heatsink expand when it is heated, or shrink as all parts of it around the space expand towards themselves? ???

View attachment 37408

 

[audioguru2]

I went to city, but they didnt have TIP31 so I bought TIP41

Hi Tedy,
A TIP41A would be fine. Most would have less gain than most TIP31A's but their minimum gains are about the same at the fairly low current for it in this project. An "ordinary" TIP31 or TIP41 is rated at only 40V, while the "A" is rated at 60V. With the project powering a low voltage load with a low current, then the output voltage would rise if the 40V transistor breaksdown.

Heatsink is almost same as the one at original page but it is slightly longer, to fit 2x 2N3055.
I dont know thermal resistance of heatsink, but it was used in 100W amplifier.

You shouldn't use transistor insulators and will need a fan with your "little" heatsink. The transistors in a 100W amp have a dissipation of only about 40W only for the short duration periods when the amp is clipping. The transistors in this project must dissipate about 120W continuously when it drives a 3A load at a low output voltage. Figure how to insulate the heatsink from the chassis for its mounting. ;D

 

[maksar]

Hi audioguru!

Interesting point, I didn't think about this!    But I don't think that this is a big problem. You can make the "shoulders" the shape you want them to be with a dremel. Also you can enlarge the space near the bottom surface of the heatsink, so that 2n3055 will match. Or you can use dremel to match the transistors to the heatsink 
By the way there is other heatsink with mach more space: http://www.semikron.com/internet/ds.jsp?file=649.html
But due to it's dimensions, I think, it should be the top wall of the psu box.

Maksar

P.S.  Paint rulez  ;D

View attachment 37409

 

[audioguru2]

Hi Maksar,
He, he. The way you show cutouts in the heatsink to fit the shoulder of the transistor, you would need to cut the heatsink down the middle for them to be assembled. He, he. ;D ;D

Let's calculate the temp of a transistor's chip on that heatsink:
1) The heatsink is 0.33 degrees C/W with only one transistor.
2) The transistor is 1.52 degrees C/W.
3) A tight fit with good thermal grease is 0.1 degrees C/W.
4) The ambient is 30 degrees C.
5) With the total of 1.95 degrees C/W and a dissipation of 60W in the transistor, its chip temp is 147 degrees C. Not bad since its max rated temp is 200 degrees C. ;D

I think heatsinks are rated with max convection airflow. So their fins must be vertical and open at top and bottom. Therefore the heatsinks should be the sides or back of the case. ;D

 

[maksar]

Hi audioguru!

He, he. The way you show cutouts in the heatsink to fit the shoulder of the transistor, you would need to cut the heatsink down the middle for them to be assembled. He, he. ;D ;D

I think you didn't understand me.  If the transistors won't fit the heatsink in ambient temperature, I'll cut a little one of it's sides. The cutouts near the "shoulders" of the heatsink are only for providing the room for the transistor and heatsink to enlarge on high temperatures.

5) With the total of 1.95 degrees C/W and a dissipation of 60W in the transistor, its chip temp is 147 degrees C. Not bad since its max rated temp is 200 degrees C. ;D

First of all, I don't understand why we have to refer to 200 degrees temperature? This is the max temp. for 0W dissipation. According to the datasheet of 2n3055 there is the max temp. for each value up to 115W. So for ~60W we have to keep transistor only at ~100 degrees, and not at 147C. Am I right?
If so, 3 transistors for 5A project, for example, is not enough.

For 4 transistors:
Dissipation for each 2n3055 is 190W/4=47.5W.
If total resistance is 2C/W (I am complitely agree with your calculations) the temperature is 125C, while the max for 47.5W according to the datasheet is a little above 125C.

For 5 transistors we will get much better situation. So I think that the best option for 5A project is 5 2n3055 transistors. 

I think heatsinks are rated with max convection airflow. So their fins must be vertical and open at top and bottom. Therefore the heatsinks should be the sides or back of the case. ;D

:-\ I want to have only one but long piece of this heatsink on back wall, so that it's lengh will be the weidth of the case.
What is it "natural convection"? And what do you mean by "max convection airflow"?
Do you think it is critical if the fins won't be vertical (in the case of back wall of the case)?
I think the hot air will go up by itself...  ???

Maksar

 

[audioguru2]

If the transistors won't fit the heatsink in ambient temperature, I'll cut a little one of it's sides.

OK.

First of all, I don't understand why we have to refer to 200 degrees temperature? This is the max temp. for 0W dissipation. According to the datasheet of 2n3055 there is the max temp. for each value up to 115W.

The max temp for the inside of the transistor is 200 degrees C. The heatsink can't cool its inside, only its outside (its case). The thermal resistance from the inside to the outside is 1.52 degrees C/W, so if you could keep its case at 25 degrees C somehow, its inside temp with 115W dissipation is 115 x 1.52 = 174.8 degrees, plus the 25 degree ambient = 199.8 degrees C.

So for ~60W we have to keep transistor only at ~100 degrees, and not at 147C. Am I right?
If so, 3 transistors for 5A project, for example, is not enough.

The transistors in the 5A project must dissipate a total of 195W max. Therefore each of the 3 transistors dissipates 65W. Your heatsink is still very good for each of them.

For 4 transistors:
Dissipation for each 2n3055 is 190W/4=47.5W.
If total resistance is 2C/W (I am complitely agree with your calculations) the temperature is 125C, while the max for 47.5W according to the datasheet is a little above 125C.

Correct. The internal temp of each transistor is 125 degrees C if the ambient is 30 degrees C.
Since their max temp is 200 degrees C then your heatsink could be smaller for each transistor.

I want to have only one but long piece of this heatsink on back wall, so that it's length will be the width of the case.

I didn't calculate it but it would be a very tall heatsink and would need a high velocity fan.

What is it "natural convection"? And what do you mean by "max convection airflow"?

Natural convection is cooling caused by hot air rising away from the fins and replaced by cool ambient air from the bottom.
Max convection airflow is when the air isn't blocked from moving easily by an obstruction such as the fins.

Do you think it is critical if the fins won't be vertical (in the case of back wall of the case)?
I think the hot air will go up by itself...

 

[maksar]

Hi!

Thanks audioguru. I think this point is clear now 

Here are the main components I have for the project (5A).
By the way, I made PCB by myself. If some one is interested in it, I can post the Ultiboard files.

Maksar

View attachment 37440

 

[audioguru2]

Hi Maksar,
Your parts look good and your pcb is very nicely done. ;D
I guess the enormous heatsinks and high velocity fan wouldn't fit into your picture!

 

[mongseng]

Hi all,
I plan to build this project cos the voltage can be reduced to 0v, variable current limitting, over current operation led, etc but my question is that I only need variable voltage ranged from 0v-12v and current from 1mA - 1A. Can I use a wall wart with 15vdc, 1.5A output as the input supply instead of using stepdown transformer?

Thanks for reply.

 

[audioguru2]

Hi Mongseng,
Welcome to our forum. ;D
To have a 12V max output voltage at up to 1A this circuit needs at least 16V to 17V DC input because of the output and driver transistors Vbe voltage drops, opamp saturation voltage and current regulator sense resistor (R7) voltage drop.
A few of its resistors will need their values reduced to supply enough current with the reduced input voltage.
An additional negative supply is also needed, but could be made from an oscillator charge-pump circuit since it requires a low current.

It would probably work with a 15V input if the output transistor is a PNP and connected to the driver transistor as a Suzlaki (spelling?) pair, a low saturation voltage opamp such as the MC34071 was used and the current regulator circuit was re-designed for more gain.
The negative supply circuit could be made for it. ;D 

 

[mongseng]

Hi audioguru,
Thanks for your super fast reply.
I'm totally new in electronic. I don't think I can build this circuit without the re-designed schematic. If possible, could you please come out with the re-designed schematic or detail in text what and where to change?

Thanks