i have searched the forum for similar threads but didn't find a lot,
how ever, i searched on google, and i learnt it's not good engineering to use 2 of them in parallel...

i found in many places a design that uses many transistors... or other deisgns that use ballas resistor...


now i am proposing this design, i'll try it tomorow, but before i damage my componenets, or waste my time... am i on the good track?

i expect the output to be more or less 5V..

also, what is a good value for the resistor for delivering say... 3A ?

thx guys

Attached Thumbnails

 

__________________
Ibrahim Kamal
check my electronics and robotics page: http://www.ikalogic.com/

 

Try consulting the datasheet for the 7805, which shows how to do it - the method shown above is really poor!.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

 

Hi, Kamal,

you get equivalent devices with 3 amps-- Google for 78S05

__________________
Regards,
Sarma.

 

Recently I needed to repair a power supply - a RS 12 vdc/3 amp regulated supply. The repair was required because the ripple was incredible.

The supply included a 2N3055 mounted to a heat sink that appeared to function properly. I had an LM317 on hand so I wired it so that the 2N3055 was wrapped around it (sorry, don't have time to post schematic). I knew the LM317 had substantial ripple reduction capability but wondered how this arrangement would perform.

Tests were done with power resistors at 1 amp up to 3.2 amps. Ripple at the power supply terminals was less than 5 mv peak to peak. I measured the ripple just off the filter capacitor prior to the regulator and determined that this arrangement reduced ripple by a factor of 90 (39 db). Not as good as I'd like but acceptable. Waveform of the post-regulator ripple confused me for a moment until I realized that it was the result of the regulator's reactions. Only thing is that the output will sag 1/2 volt or so because of the voltage drop of the 2N3055 but that was ok in my situation.

This configuration might be described on a 3 terminal regulator datasheet. No power resistor is required.

__________________
stevez

 

"Wrapped around". Very descriptive.

Was the 2N3055 used as an emitter-follower of the output of the LM317?

__________________
Uncle $crooge

 

Wrap around is used by the authors of the handbook I used (ARRL Handbook) though I can see where it doesn't tell the whole story. I'll have a look tonight to see if I can describe the arrangement better. While not the ideal approach it was quite the time saver. I only needed to remove the control board and replace it with perf board and the LM317. There was room on the front panel for a pot so I added the adjustment feature.

My first choice would have been an LM338 in a TO-3 in place of the 3055 but I would have to lift boxes to get to the collection I have - prohibited activity as I recover from back surgery. Next in line would have been an LM723 but the "wrap-around" a 317 was easy - and I needed something easy to do. I knew about the problems with it but they aren't a problem as a supply to my handheld transceiver. I could have used a 7812 but the adjustment feature is nice.

__________________
stevez

 

Perhaps you should consider a switching regulator. You can make a cheap one from an LM317 and a few common components.

See the datasheet for more detail.

__________________

I do not answer private messages asking for help because no one else can: benefit from advice I may give or correct me if I'm wrong.

Please ask on the open forum if you have a question and I'll be happy to help, if I know the answer.

 

The switching supply does look simple enough. I checked cost of inductor at Jameco - less than $3 US. Might need to substitute transistor and diode because of cost/availability. I wonder what the switching frequency is - and what the output ripple might be. Probably a good choice for some applications.

The "wrap-around" that I described yesteday is such that the output of the LM317 goes to the base of the 2N3055. The input (filtered DC +) of the LM317 is tied to the collector of the 2N3055 and the emitter becomes the new output. It's not pretty or efficient but it does work.

__________________
stevez

 

There must be better, less crude ways?.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

 

While the regulation isn't great it is regulated, in a relative sense. In my own case the regulation plays out in terms of very low ripple - if the average voltage dips somewhat it's not a problem.

Sure, there are less crude ways. This happens to be one way that is sufficient for some needs. I tried it and was suprized that it worked as well as it did. I could make the argument that it should not be more complicated than necessary to accomplish the task.

I do have concerns that someone who is not aware of the shortcomings might run off thinking that this is the ideal high current approach so I explain it as best I can. Thankfully posts like this result in numerous suggestions so that the OP might realize that there are many ways to accomplish a task - some being better than others. Better could be simpler - better could also be improved regulation/ripple rejection, or efficiency, or cost. For me, at the time, this was best because I had the parts and could accomplish the modifications with limited physical effort.

For what it's worth the authors also show a PNP pass transistor with a single power resistor that maintains the regulation of the 317.

A note of caution to the less experienced people - take a close look at the inside of a quality lab power supply and you'll see that it's not quite as simple as some of the simple power supply diagrams you've seen. Look thru the datasheet on an LM317 and you'll see the simple diagrams - followed by others that still do the same thing, in a basic sense. The extra stuff is there for many reasons - some nicely explained in the datasheets - which might help to explain why good equipment isn't quite as simple as it seems it might be.

TGIF, my second day back to work. Got to get some work done.

__________________
stevez

 

I haven't built the circuit from the previoud post.

I imagine that the switching frequency will vary depending on the load.

I'm not sure what the efficiency will be, it's obviously better than a linear regulator but not as good as most modern ICs, I'll guess around 80%.

I imagine that the inductor isn't too critical, most inductors of a similar inductance will probably work; If I built it myself I would probab;y wind my own inductor.

__________________

I do not answer private messages asking for help because no one else can: benefit from advice I may give or correct me if I'm wrong.

Please ask on the open forum if you have a question and I'll be happy to help, if I know the answer.

 

Check to see if National Semiconductor simple switcher IC's are available where you live. They have several types such as the LM2576T-5.0 which is good for 5V@ 3A. They are much more efficiant than linear regulators, and only require a few parts to make them work, a coil a diode and a few caps. They also have online design software on their website to help you select the component values.

 

I agree heartily with Brevor - you *really* want to use a simple switcher from Nat Semi... one regulator IC, 1 coil, 1 low-esr cap, and a common-as-dirt schottky diode means it's *very* simple to build.

Not only that, but in a design I just did the 7805 needed a decent heatsink to dissipate the input voltage even when supplying low current (100-200mA), and now the LM2575 I'm using doesn't need a heatsink at all :-)

Never mind that 2N3055's are in nasty TO-3 cases that need bigger than a matchbox heatsinks... the LM2575, coil, diode and cap take about 1 square inch of board space on a single-sided board :-)

/rant off

P.

 

Often, especially for larger power supplies, switching regulators often work out cheaper when you factor in the cost of the heatsinks and fans.

__________________

I do not answer private messages asking for help because no one else can: benefit from advice I may give or correct me if I'm wrong.

Please ask on the open forum if you have a question and I'll be happy to help, if I know the answer.

 

Hero999 - precisely! Also, there may be "collateral damage" from the heat being disipated by linear regulators, especially in enclosed boxes where it can shorten the life of *all* the other components...

P.