el cheapo power supply, ghetto style

[dannyf]

With all this talk about sophisticated power supplies, some of which can brew your coffee and make your laundry, I thought I would take a different approach - make the simplest power supply.

My starting point, obviously, is a 3-terminal regulator. It doesn't take lots of current so I would drop in a discrete device there.

Power source is to be a 20v laptop power supply. Potentially going higher, depending on the regulators used.

Display will be an el cheapo led volt/ammeter sourced from ebay a while ago.

Throw in a binding posts and we are ready to go.

[void_error]

A few graphs with load regulation and transient response would be nice .

[dannyf]

Trying out:

breadboard the circuit and it worked well, powering a 10ohm resistor.

The heatsink is a heat spreader (from a cpu cooler) + the project box (aluminium).

[void_error]

Hmm... what's the minimum required input voltage with respect to the output voltage? My guess is V_DROP[LM317]+V_GS[th]... a PNP would yield a lower minimum input-output voltage difference... but then there's the base current not present for the MOSFET.

[dannyf]

Got out my drill and here is the project box, with the volt/ammeter + holes for the binding posts.

The box is covered in painters tape.

[dannyf]

Yeah - my first try was actually on a MJE15030.

The maximum voltage drop would be 4v (for the irf9540) + 1.5v (lm317). So my output would max out at 19.5v (power supply) - 4 - 1.5 = 14v?

My pot is a 1k pot, with a 100ohm resistor between Vout and Vadj, so maximum output would be 1.25 * 11x = 14v.

The maximum input voltage for lm317 is 16v. That means you can use regulators like 78xx or 1117, etc.

[void_error]

Any plans for ghetto style adjustable current limiting?

[dannyf]

No. I don't think it would be ghetto enough,

[pickle9000]

I remember an old one from Radio Electronics Magazine (I think). The Junk Box Special, a 7805 with a ground ref varied through a pot. It gave about 5-15 volts.

[dannyf]

Powering a dummy load.

The pot to set output voltage is on the lower left. The 3 sets of binding posts on the right output a series of fixed voltages, from 20v (unregulated, 6amp max, top), to 12v (regulated, 1amp, middle), to 5v (regulated, 5v, 1amp, bottom).

[dannyf]

Here is the inside of the box - all genuine all 3D construction,

The IRF9540/LM317 are still not on a heatsink -> next on my to-do list.

[Skimask]

You said you used a laptop power supply.  Assuming it's one of the more-or-less standard 19V-21V "standard" power supplies with the recessed center pin that seem to be relatively common these days.
Got a Digikey type part number for the laptop side of those "standard" connectors?

[dannyf]

Put the IRF9540 on a heatsink and boxed it up.

[dannyf]

With a power resistor.

[dannyf]

No bad for a few hours of work.

And it is fairly easy as well.

I am sure one can improve it a gazillion different ways but the point I wanted to make is that a power supply doesn't have to be super-duper to be useful.

[dannyf]

One downside to the build:

the volt/ammeter I used is an el cheapo one. Its refresh rate is slow, more like 1hz or even worse. Because of that, the output voltage can seem "jumpy" when you turn the pot quickly.

A real coil meter here would have been super, but it would increase the cost significantly.

[dannyf]

Total parts cost:

I used existing parts so incremental cost for me is nothing.

If you have to buy parts to build it, it may look like this:

1) semiconductor: $1 for U1, $1 for M1; another $2 if you need other fixed output voltage;
2) volt/ammeter: $5;
3) binding posts: $5 total;
4) pot: $1;
5) project box: $5;
6) power supply: $10;
7) tools used: drill, bits, dremel, discs;
hours needed: 2 - 5 hours;
9) fun to have: priceless.

[dannyf]

One thing I didn't mention is heatsink. I used a recycled heatspreader from a cpu heatsink + the project box as a heatsink. It can comfortably dissipate 10 - 15 watts -> the box gets fairly warm, but not toasty.

This approach allows me the option to bolt on the cpu heatsink and form an active heatsink, in case I need to dissipate more power.

To remain fully passive, you will need a large and expensive heatsink to scale power up with this approach.

So another approach you may try is to get those LM2596 modules from ebay, and use it as the regulator.

[Alex30]

Haha not bad for less than a days work. I guess this is one of those supplies you'd beat with a brick and wouldn't even care about which is sometimes useful

EDIT
Now to design the coffee warmer on my PSU project...

[dannyf]

That was one of the primary goals of this projects. The power supply projects you see here are in my view to super-duper, with more focus on the show factor rather than the utilitarian factor of a power supply.

I hope this shows people, particularly someone starting out, what is possible with minimum and widely available parts and a few hours.

[dannyf]

With that behind me, I am going to build another el cheapo power supply, now geared more towards analog work -> meaning variable output, tracking dual rail power supplies.

More to come on that later.

[Kalvin]

Hint for the next project: Adjustable current limit is one of the most useful and most important properties of the power supply targeted for electronics design and testing. 

[hugo]

 .dc v2 1.25 15 10m for an LM317... how much current it could supply ?

[ludzinc]

Nice work!

Here's my el cheapo power supply panel with current limiting... and it too a lot longer than a few hours....

http://ludzinc.blogspot.com.au/2015/01/psu-console.html

[dannyf]

Your is a lot more full featured than mine.

What are those metal switches / buttons?