Power Supply problem

[sherbo]

Hey guys, I'm trying to build a power supply for a car-based MP3 player (a stripped down computer).

I've tried this link for the +5v rail, it gives very good current (upto 10A now) but there's one CRITICAL problem: when not loaded, the output voltage is 12V !!

Just to make it clear, I'm using the 7805 regulator not the 12v ! (I'm stressing this because the description doesn't match the schematic)

The link:
http://www.mitedu.freeserve.co.uk/Circuits/Power/boosti.htm

What do you think needs to be done ?? I know that linear supplies would waste alot of energy, but I can't build a switching PSU - they're too complex & the ICs aren't easily obtained from where I live.

Thanks in advance for your help..

 

[audioguru2]

Hi Sherbo,
Welcome to our forum.
If you had 10A through your PNP transistor and 7V or more across it, you melted the poor thing. No wonder it leaks when not loaded.

 

[sherbo]

Audioguru: So what you're saying is the transistor is already damaged ?? Is this a common transistor fault ?

btw, the 10A came from 2 transistors in parallel - I added a similar one with the base resistor and all.

audioguru: So theoretically, should this circuit run fine under all loading conditions ?? If you give me a yes here I'll go and buy some replacement transistors (those things aren't cheap...)

Another question (sorry I'm kinda new)
I was trying to analyze the circuit, and wondered - is there a model for the 78XX regulators ??

Thanks for your time

 

[audioguru2]

Hi Sherbo,
I didn't know that you used two paralleled transistors.
You melted at least one of them, the one with the most gain.

Transistors aren't all the same. Some have high gain and others lower gain. When paralleling transistors either you must measure and sort them so that their gain is exactly the same, or use 0.1 ohm emitter resistors with each one to equalize their gain.

The transistors are rated to dissipate 115W when their case is at 25 degrees C. They might have an insulator between them and a heatsink and they won't make perfect thermal contact with the heatsink anyway. So even with an absolutely huge heatsink, a transistor will overheat when it dissipates about only 90W. I doubt that you used such a big heatsink.

Your transistor with the most gain passed the full 10A and tried to dissipate 70W. As it overheated, its gain increased even more (thermal runaway) and it melted inside. The transistor with the lower gain hardly conducted.

Replace them with three transistors on a fairly big heatsink. Each must have its own base and emitter resistor. You want them to each pass 3.2A, so its emitter resistor will drop 0.32V, therefore the value of the 1R current sensing resistor must be increased to about 1.32 ohms or a little more and its power rating increased by the same amount.
From no load to 10A the output voltage will be 5V.
In Canada, those transistors are very common and cheap.

You don't need a model, just look at the regulator's datasheet. It shows that the regulator will limit its 1.5A output current if it gets too hot or if its input voltage is higher than its output voltage by 15V or more. A model probably doesn't know that and would just average a maximum of only 1A output.

 

[sherbo]

Thanks audioguru for your great explanation

It'll take me a few days to try it (exams), but I'll keep you informed.

 

[Kevin Weddle]

I would guess you have the regulator hooked up wrong. It is easy to do. Once I could not find the correct pinout for my TO-3 regulator and may have burnt a few up before getting the circuit correct.

 

[audioguru2]

Kevin,
If the regulator was hooked up wrong then how did the circuit provide 10A at its designed 5V?