Yet another DIY Power Supply, Almost done

[Bassblaster]

Greetings. i recently got my hands on some 24V 300VA NO CENTER TAP transformers and i was trying to figure out what to do with them. i did some google searching for a 10A (capable) power supply. and after a while i found a schematic using 3x 2SD1047 outputs and an LM317 as the driver. i drew up this and i want a 2nd opinion before i order parts. The diode bridge will be externally heatsink mounted. Anyone see anything wrong with this setup? should i change anything? if all good i can start on a custom PCB.

 

[mv213]

Greetings. i recently got my hands on some 24V 300VA NO CENTER TAP transformers and i was trying to figure out what to do with them. i did some google searching for a 10A (capable) power supply. and after a while i found a schematic using 3x 2SD1047 outputs and an LM317 as the driver. i drew up this and i want a 2nd opinion before i order parts. The diode bridge will be externally heatsink mounted. Anyone see anything wrong with this setup? should i change anything? if all good i can start on a custom PCB.

View attachment 1037806

That will be a pretty beefy power supply! Do you have something planned for it?

 

[Bassblaster]

ALL THE BEEFINESS! But plans for it would be mainly playing with power amps or other big loads.
Since im getting into car audio foolishness, maybe i should invest in one of those server supplies that can do 150A at 13-14V

 

[restorer-john]

If you are primarily intending to use this for car audio 'foolishness' bear in mind you will be wasting a phenomenal amount of power as heat, running the gear at 14.4V (you'll be dropping around 18v over the output stage. At 10A there will be upwards of 180 watts or more going up as heat.

If it's just a variable 0-30V+ supply for general use, it's OK.

 

[petehall347]

some car audio will stand a bit over its rated voltage .. well i know some will and it will up the power output somewhat ..you need to study whatever it is first though . i used to use it for parties . 16v ish seemed a good voltage

 

[Bassblaster]

i know at lower voltages and higher current it will dissipate a fair bit of heat. but THIS supply won't be used for automotive stuff but maybe testing a head unit. the line i said about the foolishness would be car subwoofer amps. where i should i find a used server supply. i have a 100W amp in my truck and i measured 30A of current draw!!
ive ran car head units on the full 18V the TDA chips are rated for. good way to get almost 30WPC out of them. heatsinks dont like it to much. but anyway, the schematic i posted above, will that work good or should i change somithing? this supply will mostly be used for general messing around and other amps.

 

[petehall347]

seems there are many variations to do same thing . here is one example . http://www.reuk.co.uk/wordpress/electric-circuit/lm317-high-current-voltage-regulator/

 

[restorer-john]

Yes it will work. You can add another 100uF on the output in parallel with the 0.1uF and maybe a bleed resistor on the cap bank for switch off discharge.

Maybe an LED indicator and dropper resistor too.

 

[Bassblaster]

i thought about adding an indicator LED, but didn't for some reason. Updated the schematic What value should the bleeder resistor be? 4.7K?

 

[I LIKE MUSIC]

If you know the current requirement for the led and the voltage at R5, our friend Ohm's law will tell you the value of the resistor.

That voltage will sag a little under high current operation of the power supply, but that will not be an issue, you just want to limit the current to the led to a value safe for the led and enough to make the led visible at your required brightness.

BTW, technically R7 would be called a current limiting resistor, not a bleeder resistor.

 

[Bassblaster]

i meant R5, R7 2K is a little higher than what i would use for a 3.2V 20mA LED on a little over 30V, but i dont want it blinding me.
EDIT: Well im dumb i dont NEED R5 just use the LED as the bleeder

 

[merlynski]

Put a ~250 ohm resistor parallel to C7 as an output load to bias Q1-3 into conduction when there is no other load on the output.
Also a 1N4004 with the cathode to + and the anode to ground to protect the output from being accidentally reverse biased (-)

 

[Bassblaster]

This is the latest version of the schematic


I Also made up the first version of the PCB. i kept in mind that the transformer i have isnt going to be able to do 10A, i made sure the PCB was able to handle 10A. Since the PCB uses DC in and not AC, and therfor does not have a diode bridge on it, i could always run a beefy 24V 10A switch mode supply into it.


Im not too sure C6 and C7 are in the best possible place. red is top layer, blue is bottom. C4 may be moved away from the U1 due to heat. but i dont think U1 will get too hot

PCB and schematic can be viewed in more detail on the site i made them. no affiliation of course

https://easyeda.com/bassblaster505/30V_10A_supply-b4869512ca1343fb86e352bf99d527ec

 

[merlynski]

Make sure you have plenty of heatsinking for Q1-3. At max output current and minimum output voltage you may have over 75 watts dissipation on each transistor (worst case).

 

[merlynski]

The LM317 is not normally used as a driver for a power supply regulator. The way you have it configured eliminates the internal current limiting and thermal overload shutdown capability.as well as (probably) decreasing the output voltage regulation.

 

[Bassblaster]

I know the 317 isnt used as a driver much but iv seen it done before. Maybe i should add some low value resistors to the bases of the outputs? like they do in audio amps. or at least the ones ive been inside of

 

[merlynski]

I know the 317 isnt used as a driver much but iv seen it done before. Maybe i should add some low value resistors to the bases of the outputs? like they do in audio amps. or at least the ones ive been inside of

I have seen that also. Try it the way it is now. If you get current hogging by one of the output transistors you could increase the emitter resistors, or match the transistors Vbe. For a casual hobby bench supply you should be fine.

 

[sregor]

Most power supply and amps use a "flyback" diode - reverse diode across the output transistors C-E. This is to prevent energy storage on the output (capacitive or inductive) damaging the output devices. Nothing fancy. my 2 cents

 

[Bassblaster]

Most power supply and amps use a "flyback" diode - reverse diode across the output transistors C-E. This is to prevent energy storage on the output (capacitive or inductive) damaging the output devices. Nothing fancy. my 2 cents

i beleive that would be D2 in the latest schematic drawing

What do you guys think about swapping the 2SD1047's for TIP3055's? the TIP's are a little cheaper ($1.61 vs $3.71) and have similar specs.

 

[merlynski]

Looking at the Datasheets I have the 2SD1047 has a bit better DC current gain (hfe), but the Safe Operating Areas are about the same in this application, as long as they are kept from overheating. Both are on the edge of SOA at 30v/3A @ 25deg C. I would be tempted to use a 2N3055 in a TO-3 case and get better thermal characteristics. That would be more work, but will get the heat away from the PCB. (my OCD is showing )