# Home brew 0-32VDC Power Supply

Discussion in 'General Electronics Chat' started by assuc, Apr 6, 2012.

1. ### assuc Thread Starter Member

Mar 23, 2012
77
1
Assalam o alaikum,
I have been digging up many threads to find myself information regarding Building up a variable power supply 0-32Vdc and 3-4.5Amp(5amp). And found this thread to be more easy rather others.
So here are my following questions with reference to post # 5 @sgtwookie:
1.You said
So by any means we cannot regulate the voltages by using 5k-10k variable resistor(potentiometer)?
2.Regarding the power dissipation at low voltages we have to use the tapped voltages from the transformer so can you please guide me how can we do it...by simple means of switching or using the 0-12v input into another circuit with it.
Also i m posting the circuit constructed by me on proteus oif u can help me with this circuit,it will be highly appreciable.

Others are also welcome to give suggestions
Regards,

2. ### studiot AAC Fanatic!

Nov 9, 2007
5,005
515
Hello assuc.

1) You cannot regulate with a pot. You can use one to provide a variable division of the voltage, but you cannot draw significant current from it.

2) Use a multitap transformer and a switch say 6, 12, 18, 24, volts AC.

This will give you about 9, 18, 27, 36 volts DC. You need at least 2.5 volts greater than you max psu output. Note the max input to a 338 is 32 volts.

3. ### Wendy Moderator

Mar 24, 2008
20,772
2,540
Welcome to AAC!

A thread belongs to the OP (original poster). Trying to take over someone elses thread is called hijacking, which is not allowed at All About Circuits. I have therefore given you a thread of your very own.

This was split from LM338 PSU 0-30v (hunting for good circuit).

Also note the thread was over 2 years old.

This may be of some help...

Basic Bench Top Power Supplies

4. ### bountyhunter Well-Known Member

Sep 7, 2009
2,498
507
If you think you can build a 160W linear supply (32V @ 5A) using an LM317 or any other linear reg you need to do some research. They can't handle anything near that.

In reality, you would need to use several discrete power transistors in parallel because of something called safe operating area.

160W would also take a massive heatsink and probably a fan.

I would recommend a switcher approach.

5. ### assuc Thread Starter Member

Mar 23, 2012
77
1

I completely get the idea of tapping the transformer to get various voltages but what i can think of is to use a multi way switch to get various voltages for example; i want to have 18v so in order to get this i have to design a schematic which cutoff all other voltages and provides me 18vdc.
I have blown two 5k pots by extracting 28vdc at high currents and it will be really helpful if u explain it briefly, as i am new to it.

Last edited: Apr 7, 2012
6. ### assuc Thread Starter Member

Mar 23, 2012
77
1
Basically, my intentions weren't to takeover someone threads but what i thought that creating multiple threads about similarly related psu is not right. Anyways thanks for my seperate thread now it would be more helpful.
on the topic, the article u provided of Basic bench top power supplies is the basic idea of my psu, so a question arises that won't that circuit provide me with 4.5A at 30Vdc by using Lm338K Metal can package along with good heatsink?

7. ### studiot AAC Fanatic!

Nov 9, 2007
5,005
515
I assume you understand that power = volts x amps?

In this case the 'power' is the power dissipated in the regulator,
the 'volts' are the difference between the output voltage and the input voltage
and the 'amps' is the current in the load and therefore also in the regulator.

So if, as you originally asked, you want a variable supply from 0 - 32 volts at 4.5 amps you will have to arrange for the regulator to dissipate 35 x 4.5 = 157.5 watts.

This is out of the reach of an LM338T and extremely difficult with an LM338K.

I have just had this discussion with someone I supplied with a bunch of LM338s for 4.5 amp supplies.

I wonder why they have suddenly become so popular.

Further I hope you realise that the 338 requires a negative rail if you are going to make zero your lowest voltage.

8. ### panic mode Senior Member

Oct 10, 2011
1,328
305
linear regulators drop voltage. this voltage drop multiplied by regulator current (same as output current) is power that turns to heat in the regulator. this is why linear regulators need cooling and usually are considered for applications with small voltage drop. using switching mode power supply eliminates this.
in SMPS transistor is not used as analog device but as switch (either on or off).
when it's on, current can be large but voltage drop is small (and so is the power). when it is off, voltage drop is large but there is no current (no power).
for example LTC3891, LTC3878 etc. (with external transistors) can do same or similar job much more efficiently.
as for getting down to 0V, i see no need for it. psu is there to provide power and i can't recall many circuits that needed less than 1.2V to run.

Mar 23, 2012
77
1
10. ### studiot AAC Fanatic!

Nov 9, 2007
5,005
515
You don't answer the questions we asked you so why do you expect us to answer yours?

Building a 5 amp power supply requires some serious considerations about cooling.
How do you plan to accomplish this?

Yes the proposed circuit will just about work. Each 3055 will dissipate 85 watts, worst case.
I have doubts about the ability of the BD140 to supply enough base drive however to achieve good regulation.

11. ### bountyhunter Well-Known Member

Sep 7, 2009
2,498
507
Which will give a junction temperature rise of about 100C above the heatsink temp....... design not feasible with only two transistors. I would put four 2N3055 devices in parallel.

And you will need a heatsink whose effective thermal resistance is less than about 0.4 C/W, and that is very difficult to do. A fan will be required.

12. ### assuc Thread Starter Member

Mar 23, 2012
77
1
Regarding your last post I didn't understand the input and output differential correctly,means one has to consider the differential volt input in order to get the power rating??
And i read the lm338(TO-3 metal can package) datasheet more closely and found out that it can handle max 50W so cascading 2 lm338 is one option to optimized the circuit and for cooling i will be using metallic heatsinks or a fan may be

Apr 2, 2009
8,785
771

14. ### bountyhunter Well-Known Member

Sep 7, 2009
2,498
507
You need to do some research before you build this. If you don't understand how to calculate the power dissipation in a linear regulator, start with that.

And as for:

I think you meant "parallel" them and that is not advisable. They are not intended to work that way.

You need to understand the basics of how to calculate a heatsinks rating so you can design this. Here are two references. APPB explains about power dissipation and heatsinking.

File size:
40.6 KB
Views:
30
File size:
27.1 KB
Views:
41
15. ### assuc Thread Starter Member

Mar 23, 2012
77
1
i know how to calculate the power but the thing that buzzing me is the differential voltage thing i got through few posts but not found a valid explanation and thanks for the APPB , i'll surely go through it.

16. ### studiot AAC Fanatic!

Nov 9, 2007
5,005
515
Both types of linear regulator that you showed (the 2N3055 and LM338 circuits) are known as series regulators.

In a series regulator the regulating element is in series with the load and acts as a variable resistor to drop the all the input voltage that is not adropped across the load.

Because this is in series all the current that flows through the load also flows through the regulating element.

Power in load = 1 volt x 5 amps = 5 watts

but the power in the regulating element is

(35 -1) volts x 5 amps = 170 watts.

Total power = 175 watts

If your output is set to 25 volts at 5 amps then similar calculations lead to

Regulator power = 50 watts

Total power = 175 watts

So in fact the load and the regulator share the 175 watts that are always drawn from the input 35 volt power supply at 5 amps current.
The total power drawn will be increase with increasing current.

Does this help?

If you have understood this there are other types of regulator we could discuss.

17. ### assuc Thread Starter Member

Mar 23, 2012
77
1
^^ yeah that really helped me and also i read the book by Boylested "Electronic devices and circuit theory" so it was helpful for me to grab your explanation.
Now the point comes that if the current is supposed to vary according to the voltage drop on the type of load present and hence i m using lm338 i.e., is to work at 50watt max as given in datasheet.
Therefore, if i put a load let say a device that works on 10V and draws 5amp so the total power dissipated would be 50watt and if the device that works on 20v so max current it can draw is 2.5amperes as I=50w/20v=2.5amp... current exceeding will make the lm338 shutdown. so in order to obtain 5 amp i need to cascade or parallel it with another lm338 regulator ic.

18. ### studiot AAC Fanatic!

Nov 9, 2007
5,005
515
Now that you have begun to appreciate the difficulties of a high power supply would be a good time to review your requirements.

What voltage and current do you actually want?

Why do you need this? What sort of project is it?

Members seeking power supply designs post frequently here and if you do a forum search you will be able to read their experiences.

If you really need a supply capable of supplying 0 - 32 volts, without range switching , at 5 amps you must expext a design that looks like a small heater.

I would not recommend running voltage regulators in parallel. A bypass transistor is a much better option.

19. ### panic mode Senior Member

Oct 10, 2011
1,328
305
I didn't look too closely but that design seem to be over optimistic. According to datasheet 723 cannot regulate below 2V, also it cannot regulate in one continuous range from 2-30V either, there is switch-over at 7V (2-7, 7-30).