power supply circuit

How about this one:

How much current do you want?

Lets just assume you make DC from the AC and then regulate down.

Doing a couple of figures for making 4A into lets say zero volts:

24V * sqrt(2) = 34V

34V * 4A = 136W

The the regulator circuit needs to get rid of 136W.

To do this, you need something like 10 TO-220 packages transistors screwed down to a quite large chunk of fins.

If you have enough fin area, about 5 to 7 TIP-35 or TIP-36 transistors would work for you.

Another question is how fast the current limiting or regulation needs to act. A lot of very practical designs can make a large spike of current when suddenly shorted. The power supply will easily survive this but if the short was you connecting it to a diode, the diode may be damaged.

Do you need instant (or nearly so) action on the current limiting?

Take a look at the LT3080:

Notice how it lets you set the output voltage all the way down to zero. This sort of design is a good place to start your thinking.

You can also think of the circuit as just a non-inverting op-amp amplifying the output of the pot you use to set the voltage. This is another good direction to go.

Regulated current or regulated voltage? Or do you want a lab-type supply with regulated voltage and smooth current limiting?

It sounds like you don't have a tremendous amount of experience with electronics -- you may find it helpful to get a copy of the ARRL handbook, which has some nice power supply designs as well as a _lot_ of good practical basic knowledge.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Do you need to implement control loops in software?
"Applied Control Theory for Embedded Systems" gives you just what it says.
See details at http://www.wescottdesign.com/actfes/actfes.html

Snip ">

My guess is it's for a glow plug in an RC model engine.... so regulated voltage, current limited.

Oops, forget I said that. :(

One thing to watch for: RMS current being drawn from the transformer exceeding the DC output current. The bigger the filter capacitor, the worse this gets.

Assuming that you will be using an adjustable voltage regulator, you should be OK with a few volts of filter capacitor ripple, so don't make the filter capacitors too big.

I would not try for more than 3-4 amps of DC from a 5 amp transformer unless the rating is for achievable DC with a "capacitor input" rectifier/filter circuit (capacitor immediately following the rectifier, as opposed to having a choke between the rectifier and the capacitor).

24 volts times sqr(2) is 33.9 volts. Allow loss of a couple volts in a bridge rectifier, a couple volts minimum drop in a voltage regulator that is doing its job well, a couple more for a current regulator, and that is about 28 volts. Allow 4 volts ripple on a moderate filter capacitor, and you get 24 volts.

As for a filter capacitor value: 4 volts ripple at 4 amps over the approx. 6 milliseconds or so that the capacitor has to supply current works out to 6,000 microfarads. I would not use more. Greater filter capacitance makes the current waveform drawn from the transformer spikier, which makes the RMS current drawn from the transformer greater. I might go for 4700 microfarads and only count on drawing 3-3.5 amps DC without overheating the transformer.

BEWARE - with mild-moderate overloads, transformers this size can take a couple hours to reach excessive temperature.

- Don Klipstein ( snipped-for-privacy@misty.com)

"Don Klipstein"

** That last claim is ABSOLUTE BULLSHIT !!!!

Nothing new for this ignorant, septic asshole to make them.

** Make it as big as you like and the rms ripple current will not increase. ** No it does not !!

The whole idea is an idiotic MYTH put about by complete fools like Klipstein.

** 20,000 uF would be a good value to use - cos increasing the filter capacitance above the minimum usable value has no effect on the rms current in the transformer windings.

...... Phil

Minor points - the transformer isn't rated to deliver 136W, and that 34V is unloaded peak.

When working out the power in the regulator, it is best to assume that the peak voltage is what you get. The rated voltage on the transformer is for a normal line voltage. In the high line case, you can get a little more.

I appreciate that, but my original comments stand. There is no valid reason to state "the regulator circuit needs to get rid of 136W" when the transformer is not rated to deliver that power.

Look up the LM338 three terminal regulator. It's rated at 5 amps and

35 volts from output to input. All you need is a couple resistors or a pot to set the output voltage. Use a 10 amp bridge rectifier to convert the transformer secondary AC to DC and a 20,000uF, or thereabouts filter cap. Connect the capacitor to the (+) input of the regulator and ground and adjust the resistors for the desired output votage. The regulator also has current limiting, but not sure what it is. You can get 1.2 volts to maybe 24 volts DC out, but you need a large heat sink when the output is set low and drawing 5 amps.

-Bill

But certainly not at the same time:

Go read the datasheet before you believe this last bit.

Earlier, I posted the simple math that suggests that you want about 10 TO-220 packaged transistors to get that much power.

Here's a bit of ascii art for a way to share the power around:

NPN R4 ----------------- ----/\\/\\---- ! Q2 \\ /e ! ! --- ! ! R2 PNP ! R3 ! +-/\\/\\--- -------+-----/\\/\\----+ ! e\\ / Q1 ! ! --- U1 ! ! R1 ! ------- ! --+--\\/\\/---+-----! LM317 !----+---+--- ------- ! ! ! +-/\\/\\--- ! \\ / \\ ! GND

The capacitors have been left out because I am too lazy to do them.

When U1 draws current R1 develops a voltage due to Ohms law. Once Q1's e-b draw is overcome, R2 sets how much voltage is needed on R1 per mA flowing in Q1.

R3 and R4 repeat the trick. If you set the resistor values correctly, you can spread the full load power over the three devices.

To get 10 or so devices, the Q2 R4 circuit becomes several transistors each with its own emitter resistor.

You seem to have a poor understand of what the rating of a transformer means.

to

On the contrary. A 24V 5A secondary is rated for 120VA. Your education suggests something different? Please inform us all as to YOUR transformer theory.

with

is

reason to

is

Yes that is the *rating* in VA and yes that works out to 120/1.5=80W of power from the rectifier. It however is far below the limit of what you can actually get out of the transformer. It is what the maker says the transformer can do forever on a normal line voltage and usually with an air temperature of some 50C. The transformer does not contain any significant current limiting. The output impedance is fairly low. The heat sinking on the regulator needs to be designed to withstand the high line case with a short circuit on the output. This is a lot more than the rated power of the transformer.

If the OP goes with less than the amount of heat sinking I suggested, I predict he will be replacing the transistors very early in the power supplies life.