Anyone built a variable power supply?

billybob

Member
Im wanting to build a variable 1 to 12 volt 30 amp DC power supply (more voltage the better) Im wondering if anyone has made one before

AnalogKid

Well-Known Member
Yes. If you are talking about a power supply with a regulated output, those started showing up in the 1930's - ish. For that kind of low voltage and moderately high current, probably the mid 1950's. IBM started producing transistorized computers in 1958, so they had to have low voltage/high current regulators.

ak

AnalogKid

Well-Known Member
Search for:

30 amp power supply schematic

10 seconds, 100 schematics.

ak

Pommie

I've got a 12V 30A SMPS here that cost me about US$25 off of ebay. Depends on what you intend to use it for. Mike. Last edited: gophert Well-Known Member Most Helpful Member Im wanting to build a variable 1 to 12 volt 30 amp DC power supply (more voltage the better) Im wondering if anyone has made one before Linear or smps? If smps is ok, Are you building from scratch (back to 120vAC source) or is starting with an xVDC supply and stepping the voltage down (and current up)? What quality of regulation are you looking for? What is the lowest voltage and lowest current you want to source from this supply? What is your budget? What do you plan to use it for? billybob Member I have a variable power supply that teaches 30 volts 5 Amps which is good for most things, but I need a power supply that steps down the 120 volts from the wall and gives a good amount of current for various experiments mostly with inverters and amplifiers. My budget is probably around$15 to $20 so ya pretty tight Nigel Goodwin Super Moderator Most Helpful Member I have a variable power supply that teaches 30 volts 5 Amps which is good for most things, but I need a power supply that steps down the 120 volts from the wall and gives a good amount of current for various experiments mostly with inverters and amplifiers. My budget is probably around$15 to $20 so ya pretty tight Well if you're talking linear PSU's, then your 'budget' isn't even enough for a mains transformer, the heatsinks, or even the reservoir capacitors - 30A is a pretty serious PSU and not cheap. billybob Member well I have a good selection of capacitors already, got some power transistors mostly npn. A lot of heats sinks. How much would a good 30 amp transformer be? Disregard the budget I said earlier billybob Member Would this work? What is the V with the arrow through it? Is it a volt meter? Visitor Well-Known Member If a lot of current at 12 volts would be useful, look for an Xbox power supply supply in a thrift store. These can usually be brought for less than$10, and can provide up to 14 – 16 amps depending on the version.

These supplies have a 5 volt supply at over an amp always available. The 12 volt supply is turned on by connecting one of the wires to the +5 volt line. Cut off the output connector to use. If I remember correctly, the connections are:

Black – ground. Connect all the black wires together.

Yellow – +12v. Connect all the yellow wires together.

Orange – +5v.

Blue or green – 12v enable. Connect to orange to switch on 12v output.

billybob

Member
Hey good to know, thank you. Goodwill is my friend

Visitor

Well-Known Member
The enable line takes almost no current to control, so if you get to the point of wanting to control a 12v load with a microcontroller, it's really simple.

I use one of these supplies to remote-control a column of 16 muffin fans* in a sliding glass door to draw air through the house to cool it off quickly.

* The muffin fans were 50 cents each at a computer recycle place. How could I pass them up???

Mike.

AnalogKid

Well-Known Member
Tubeaxial or tube-axial. These days the DC models use BLDC motors - Brush-Less DC.

ak

rjenkinsgb

Well-Known Member
If you want an off-the-shelf one, look at an EP-925, sold under various brand names; Manson, Voltcraft etc.

Or this is something similar to what you want, as a DIY one; I remembered seeing it in my garage a few weeks ago while looking for some other items.

The construction is somewhat primitive and cringeworthy by my presents standards, but it is around 40 years old; I remember building it when I still has a shed as a workshop at my parents house.

It also has 40 year accumulation of dust inside, which I cannot remove from the deeper recesses easily; but it's a realistic DIY high current adjustable supply, if a slightly odd one.

The only transformer I could obtain at the time that had an adequate power rating was a 24V one, far too high to regulate conventionally to 12V or less without dissipating crazy amounts of heat - so it's a "secondary switcher" design, what would now be called a buck regulator.

The outline regulator was in the applications notes of a data book, for some make of 78xx or 317 voltage regulator, using one of those as voltage sense and a "high current" driver for external transistors with some positive feedback for hysteresis.
I remember the principle but have no idea of the actual circuit, so I cannot provide that, sorry..

The bridge rec is the part with four upright terminals bolted to the base of the case in front of the big capacitor & the big silver cap the initial smoothing from that.

The blue capacitor is the output smoothing and the round aluminium item at the side of the two caps in the top view is a large "pot core", with the winding for the inductor. The diode that goes with that is attached directly to the top of the larger cap and insulated with duct tape

Most of the other small components are on the tagboard.

Well, it still works, so it can't be all that bad!

The socket at the top of the back panel with white wires is a plug-in 12V output for equipment, the bottom corner socket is main input.
The aluminium can across in front of the transformer is a mains RFI filter.

The left hand front panel switch is the on/off one and the right is a spring loaded one that switches the meter to voltage rather than current, while pressed. The voltage adjustment is the slotted pot on the rear panel.

[I've made others since this, but for other people; some much! larger, but I don't have any details or photos].

Edit - just searched for 78_ based switching regulator circuits & found several using pretty much an identical and familiar-looking configuration, though my version was adapted for much higher current & variable voltage:

Last edited:

JimB

Super Moderator
I like it.

All it needs is:
1 strip down
2 a good scrub to clean it up
3 a few updates, like an IEC mains inlet with filter, instead of that old Bulgin socket.
5 re wire it
You then have a sweet bench PSU.

JimB

billybob

Member
That’s awesome! I love the look of bare-bones electronic parts. You said you used LM317? How much current can that regulate?

Visitor

Well-Known Member
Let's talk about a trap when using linear regulators....

Say an LM317 is rated to provide 1.5 amps maximum. It's also rated for a maximum input voltage of 35 volts. So it's simple. Provide 35vdc to the input, and you can draw an amp and a half from 1.5 volts to 30-something volts, right?

Nope. You have to consider the voltage drop across the voltage regulator and the power dissipated by the regulator. In a linear regulator, the magic of regulating the voltage results in generating heat. The power dissipated by the regulator is:

P = (Vin – Vout) × I

So for this example,

P = (35 – 1.5) × 1.5 = 50 watts.

Yikes. Think of the heat generated by an (old) incandescent 60 watt light bulb. Touch it with your finger and you'll be burned. Imagine that much power dissipated by the small voltage regulator package. To make this work, you would need a large heat sink and possibly a fan too.

Does this mean we can't get an amp and a half at 1.5 volts? Are they trying to pull a fast one on us?

Not at all. If we want to reduce the power dissipated by the voltage regulator, we can reduce the voltage drop across the regulator. What if we drop the input voltage to 5vdc?

P = (5 – 1.5) × 1.5 = 5.25 watts.

Almost a 10× drop in power dissipation. With a reasonable heat sink, this could work.

One more piece of advice from lessons learned the hard way: the pinouts of negative regulators (like the LM7905) are not the same as for positive regulators. They will get extremely hot if you overlook this detail. I had a TO220 brand on the tip of my index finger for months from checking the regulator temperature when I overlooked this detail.

Nigel Goodwin

Super Moderator
I had a TO220 brand on the tip of my index finger for months from checking the regulator temperature when I overlooked this detail.
You should ALWAYS lick your finger first

However, I must admit to still having a(fully working) PIC16C84 with my fingerprint burnt in to the top of it - it was only powered from a PP3, so I never imagined that it could get very hot, so didn't use the lick trick.

rjenkinsgb

Well-Known Member
You said you used LM317? How much current can that regulate
On its own, it would only be one amp at most, but limited to rather less by power dissipation and heating as it's dropping so much voltage.

However in that circuit it's used as a driver for a complementary darlington (Sziklai pair) acting as a 30A power transistor - the three devices on the heatsink on the back; the MJ2955 PNP first device then the two 2N3055 in parallel as the second stage.

And as it's a switching regulator, the power dissipation is actually quite low; the transistors are either on or off, with the inductor and diode supplying the output current while the transistors are off.